Dumbarton Oaks Papers 54

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Nicolas Oikonomides 1934–2000

umbarton Oaks has lost a good friend. On 31 May 2000, Nicolas Oikonomides, Dumbarton Oaks’s advisor for Byzantine sigillography, died in Athens after a brief illness. The beginning of Professor Oikonomides’ long association with Dumbarton Oaks dates to 1973, the year he commenced a project to read and publish the 17,000 Byzantine lead seals in the collections of Dumbarton Oaks and the Fogg Museum. In many ways Professor Oikonomides was uniquely qualified to head the project and to edit the first four volumes of the Catalogue of Byzantine Seals at Dumbarton Oaks and in the Fogg Museum of Art, a series devoted to seals of the Byzantine provincial administration. Although as a youth he had originally intended to study the history of modern Greece at the University of Athens, Oikonomides soon turned to Byzantine history under the tutelage of Professor Dionysios Zakythinos, and his first writings in this discipline began to appear in the early 1950s. In 1958 he enrolled in a doctoral program in Paris, where he studied sigillography with one of the great specialists in the discipline, Pe`re Vitalien Laurent. His dissertation, an edition and commentary of the Taktikon Escorial, was the basis of a later volume, published in 1972, Les listes de pre´se´ance byzantines des IXe et Xe sie`cles. This book was then and remains a classic study of the evolution of titles and developments in the Byzantine system of administration. For almost two decades Professor Oikonomides maintained an association with French byzantinology, collaborating in the publication of seven volumes of the Archives de l’Athos series, beginning with the Actes de Dionysiou (1968) and continuing through the appearance of the fourth volume, Actes d’Iviron (1995). The colonels’ seizure of power in Greece in 1967 forced Professor Oikonomides to leave the country and to continue his career elsewhere. Eventually, in 1969, he settled in Canada, where he took a position in the history department at the University of Montreal. For an expatriate Greek with extensive academic training in France, francophone Montreal, with its large Hellenic community, was the perfect place to make the transition from Europe to North America. As he established himself at the university where he would remain for twenty years and twice serve as department chairman, Oikonomides continued to stay abreast of events in Greece and to maintain his contacts with Greeks abroad, not least through his broadcasting on Greek-language radio. Yet devoted as he was to his native land and culture, Oikonomides came to feel very much at home in Montreal and became a Canadian citizen, an allegiance of which he was also very proud.

D

x

NICOLAS OIKONOMIDES

At a reception commemorating his retirement from the University of Montreal in 1989, his colleagues and students paid homage to his contribution in teaching and research. He taught undergraduate courses and seminars in both Byzantine and modern Greek history while training and supervising a succession of graduate students. He was a practical, approachable teacher who believed that one learned by doing; he would lay out photographs of manuscripts or seals and have the student begin reading them to acquire the necessary skills and knowledge. While he insisted on high standards in the quantity and quality of research, above all in the primary sources, his guidance was always instructive and unobtrusive. It was, however, the wealth of his scholarly research and publications that truly set him apart, an achievement attributable not only to his extraordinary range of expertise, but also to his industry and use of time. When his successor as chairman expressed astonishment that Oikonomides had been so productive even when burdened with all manner of administrative responsibilities, it brought to mind Oikonomides once mentioning how he had transcribed a number of Athonite documents during plane trips between Montreal and Toronto (where he had been invited to give a seminar on Byzantine diplomatics). Professor Oikonomides was well versed in palaeography, and a good portion of his scholarly output reflects an abiding interest in using this tool to exploit unedited documents for information on legal precepts, the duties of bureaucrats, and the evolution of bureaucratic structures. Dumbarton Oaks drew on his expertise in these areas in the 1980s, when he became a key member of the team that produced the Oxford Dictionary of Byzantium, serving on the Advisory Board and writing all the entries on Byzantine diplomacy, diplomatics, and the fiscal system. Under Laurent’s early guidance, he had developed a further research skill: sigillography. It is possible that he would never have developed such mastery of this auxiliary discipline except for the conjunction of several unrelated circumstances. In 1969 Professor Herbert Hunger, while a visiting scholar at Dumbarton Oaks, had the opportunity to view the collection of some 12,000 Byzantine lead seals that Dumbarton Oaks had assembled and recommended that the institution undertake publishing them. Such a venture required a competent editor. With a trained sigillographer on the North American continent, the stage was set for Professor Oikonomides and Dumbarton Oaks to enter into a long-term arrangement whereby he served (from 1972 until his death) as advisor for Byzantine sigillography. The first phase of the project was an arduous one. It consisted of reading the 12,000 seals of the Dumbarton Oaks collection, in addition to the 5,000 more from the collection of the Fogg Museum of Art, and entering the transcriptions on note cards.1 By late 1979 the reading of the seals had been completed, and there was now lodged in the Dumbarton Oaks basement a vast “cartoteca” of seals. After an eight-year hiatus, the seals project was reborn in 1987 as a publications project. Professor Oikonomides rose to the challenge 1 One memory of this time that shall always remain with me (John Nesbitt) is a query of Professor Oikonomides as to where and how I should begin my portion of the work on the seals. I had, after all, never seen a lead seal in my life and, like most people, had no idea how to read one. He said that I should begin with the monogrammatic seals; if I could read those, I could read any of them. In the end, this was not true, for there are still many seals that elude my comprehension. Indeed, the fact that I can read and interpret any of them is really due to his patient guidance and the vast knowledge that he brought to our task.

NICOLAS OIKONOMIDES

xi

of bridging the gap between archive and dissemination in a way that few others would have. He realized that to publish them would require the use of new technology—the computer. If there was one area of expertise in which Professor Oikonomides openly delighted, it was his mastery of the intricacies of “MacWorld,” and one of his proudest achievements was the creation (with the help of Harvard technicians) of Sealshort, an inscriptional font. With this new font in hand it was possible to publish camera-ready copy, that is, catalogues that could be produced rapidly, accurately, and inexpensively. The question then became what to publish. The answer was relatively easy. For two decades Professor Oikonomides had been researching and publishing articles on changes in the provincial administration and the history of the Byzantine frontiers. The time had come to build upon such articles as “L’e´volution de l’organisation administrative de l’Empire byzantin au XIe sie`cle (1025–1118)” 2 and to publish seals relevant to historical geography and the functions and personnel of provincial bureaucracies. The energy that Professor Oikonomides brought to the accomplishment of this project (as well as others) is reflected in the fact that volume one appeared in 1991, and volume four will appear in 2001—a publication rate of almost one volume every two and a half years. This figure is of course exclusive of the many other seals publications in which he and Dumbarton Oaks were involved and of which he was either author or editor. To minimize discussion in the catalogues of how and why seals are assigned certain dates, Professor Oikonomides had the foresight to publish a booklet on these matters in 1986, Dated Lead Seals. To encourage the study of seals and their publication, in 1987 he initiated Studies in Byzantine Sigillography, a journal that soon became a means of overcoming one of the major stumbling blocks to the publication of seals, namely, bibliography. The proper cataloguing of seals requires information from many publications, including auction sales, on the existence of similar or parallel specimens. Addressing this issue, volume three (1993), volume five (1998), and volume six (1999) include lists of scholarly articles and auction catalogues appearing between the years 1931 and 1996. All of this Professor Oikonomides undertook while handling the competing needs of his home institutions in Greece. For shortly after the seals project was restarted (in 1987), he left North America to take the chair of Byzantine History at the University of Athens. Later he also took over direction of the Byzantine Studies Center at the National Hellenic Research Foundation, and in 1997 he became vice-president of the Executive Committee of the Foundation for Hellenic Culture. His scholarly production remained as rich as ever amidst these duties, and in 1996 he realized his longstanding intention to publish the definitive study of the middle Byzantine fiscal system, Fiscalite´ et exemption `a Byzance (IXe–Xes.). Nicolas Oikonomides’ scholarly reputation is assured, for his work will long be indispensable to research in every aspect of Byzantine history, but his friends and colleagues will remember him most of all for the pleasure of his company. He was a convivial, generous host with a ready sense of humor who loved to speak of his travels, experiences, and observations. Perhaps our most affectionate memory of Nikos will be the story of his rendezvous with U.S. customs officials. En route from Montreal to Washington, he pre2

TM 6 (1976).

xii

NICOLAS OIKONOMIDES

sented his visa for admission to the United States. When asked to explain his role with the “seals project,” Nikos responded that Dumbarton Oaks, a major research center, had come into possession of 17,000 seals, of which he was preparing a catalogue. As customs officials gathered round in amazement, Nikos warmed to the subject, explaining that the seals had come from various lands, principally Turkey and Greece; each one had been photographed, and now it was time to discuss the provenance and owners of the seals in a multivolume catalogue. It finally dawned on Nikos, as he beheld the incredulous faces before him, that the officials thought that he was talking about 17,000 marine mammals stored on an estate in the national capital. He then hastened to explain the meaning of “seal” in this particular context, but the lesson Nikos derived from this incident was that the United States truly was the greatest country on earth, for only Americans had the breadth of mind not just to envision but to support such a “seals catalogue.” We thought it appropriate to close with the following lines from Cavafy, which we offer in tribute to a great scholar who did so much to illuminate the history, institutions, and civilization of the medieval Greek world: ÔUph'rxe di´kaio", so␸o´ ", gennai'o" ÔUph'rxen e“ti to` a“riston ejkei'no, ÔEllhniko´ "— ijdio´ thta de` n e“c∆ hJ ajnqrwpo´ th" timiote´ ran⭈ eij" tou` " qeou` " euJri´skontai ta` pe´ ran. Il ´etait juste, sage, et courageux Et en fin Grec, plus que toute autre chose; L’humanite´ ne s’honore pas de plus belle qualite´ De meilleures ne se recontrent que parmi les dieux. He was just, wise, courageous And he was moreover that best of all, a Greek— Humanity has no more honorable quality; The ones beyond are found among the gods.

John Nesbitt Eric McGeer

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Church of Sts. Sergius and Bacchus in Constantinople and the Monophysite Refugees JONATHAN BARDILL

hen Cyril Mango first attempted to elucidate the circumstances surrounding the construction of the church dedicated to Sts. Sergius and Bacchus in the palace of Hormisdas in Constantinople, his proposed scenario raised objections from Richard Krautheimer and Thomas F. Mathews, to which Mango felt compelled to reply.1 I would not have ventured to address this controversial topic yet again, were it not for the emergence of some new evidence. The church (now Ku ¨¸cu ¨ k Ayasofya Camii) is located on the south side of the city, beneath the towering substructures of the curved end of the Hippodrome, just inside the sea walls along the shores of the Marmara.2 Today, the railway from Sirkeci Station runs in the narrow gap between the church’s south wall and the fortifications. Procopius indicates that the church stood within the palace of Hormisdas and that this palace had been the residence of Justinian and Theodora in the years before they ascended the throne in 527.3 Letters were written to Pope Hormisdas on 29 June 519, stating that Justinian had built (or perhaps, in truth, begun to build) a church of Sts. Peter and Paul in the palace of Hormisdas, and requesting relics of the two apostles and of St. Lawrence in order that they might be deposited there.4 Justinian supported Chalcedonian orthodoxy and had

W

This article was written while I held a British Academy Postdoctoral Fellowship at the Institute of Archaeology, Oxford. I am grateful to Sebastian Brock and an anonymous referee for references, and also to James Howard-Johnston and Cyril Mango for their comments following the presentation of a version of this paper to the Byzantine Studies Seminar in Oxford on 27 January 1995. 1 See C. Mango, “The Church of Saints Sergius and Bacchus at Constantinople and the Alleged Tradition ¨ B 21 (1972): 189–93; R. Krautheimer, “Again Saints Sergius and Bacchus of Octagonal Palatine Churches,” JO ¨ B 23 (1974): 251–53; T. F. Mathews, “Architecture et liturgie dans les premie`res ´eglises at Constantinople,” JO palatiales de Constantinople,” Revue de l’art 24 (1974): 22–29; C. Mango, “The Church of Sts. Sergius and Bacchus Once Again,” BZ 68 (1975): 385–92. Mango’s “Tradition” and “Once Again” are reprinted in C. Mango, Studies on Constantinople (Aldershot, 1993). 2 On the church, see W. Mu ¨ ller-Wiener, Bildlexikon zur Topographie Istanbuls (Tu ¨ bingen, 1977), 177–83, fig. 358; T. F. Mathews, The Byzantine Churches of Istanbul: A Photographic Survey (University Park, Pa.–London 1976), 242–59; P. Grossmann, “Beobachtungen zum urspru ¨ nglichen Grundriß der Sergios- und Bakchoskirche in Konstantinopel,” IstMitt 39 (1989): 153–59. 3 Procopius, Buildings, 1.4.1–8, ed. O. Veh, Prokop: Werke, vol. 5 (Munich, 1977); trans. H. B. Dewing and G. Downey, Procopius, vol. 7 (London, 1940). 4 For details of Sts. Peter and Paul, see Mango, “Tradition,” 189–90; and A. A. Vasiliev, Justin the First: An Introduction to the Epoch of Justinian the Great (Cambridge, Mass., 1950), 377–78.

2

CHURCH OF STS. SERGIUS AND BACCHUS

written earlier letters to the pope helping negotiate an end to the Acacian schism.5 It is, therefore, not unlikely that it was Justin I’s reconciliation with the papacy (achieved on 31 March 519) that motivated his nephew to build the new church.6 Justinian was made co-emperor with his uncle on 1 April 527; when the latter died exactly four months later, he became sole ruler, his wife was proclaimed augusta, and the imperial couple moved from the palace of Hormisdas into the Great Palace.7 Procopius explains that Sts. Sergius and Bacchus was built by Justinian alongside the earlier church of Sts. Peter and Paul, and he says that the two edifices were of similar size and shared the same atrium.8 The dedicatory inscription carved around the nave of Sts. Sergius and Bacchus is worth quoting in full: Other sovereigns have honoured dead men whose labour was unprofitable, but our sceptered Justinian, fostering piety, honours with a splendid abode the Servant of Christ, Begetter of all things, Sergius; whom not the burning breath of fire, nor the sword, nor any other constraint of torments disturbed; but who endured to be slain for the sake of Christ, the God, gaining by his blood heaven as his home. May he in all things guard the rule of the sleepless sovereign and increase the power of the God-crowned Theodora whose mind is adorned with piety, whose constant toil lies in unsparing efforts to nourish the destitute.9

The references to “our sceptered Justinian” and “God-crowned Theodora” indicate that the text was composed and inscribed in or after 527, the year in which the couple ascended the throne. Even if the capitals that support the inscribed entablature had been stockpiled or salvaged from an earlier structure (unlikely suggestions), the monograms on them, which are to be resolved as IOUCTINIANOU, BACILEVC, and QEODVRAC, provide the same terminus post quem.10 The monograms might also be adduced as evidence in favor of a terminus ante quem in 532. Those of Justinian and Theodora are of the box type (otherwise known as the block, square, or double-bar type), being constructed around a letter with two uprights (N in IOUCTINIANOU) or around two letters, each with a single upright (E and P in QEODVRAC). The cruciform style of monogram was a later development, which had not been introduced by 518.11 The absence of any cruciform monograms of Theodora on the Vasiliev, Justin the First, 179–83, 199. Vasiliev, Justin the First, 378; W. Dynes, “The First Christian Palace-Church Type,” Marsyas 11 (1964): 6. 7 See PLRE, II, 648–51 (“Iustinus 4”), 645–48 (“Iustinianus 7”); PLRE, III, 1240–41 (“Theodore 1”). 8 Procopius, Buildings, 1.4.1–8. 9 Translation from Mango, “Tradition,” 190. For the Greek, see, e.g., A. van Millingen, Byzantine Churches in Constantinople: Their History and Architecture (London, 1912), fig. 20 (by A. E. Henderson); S. G. Mercati, “Epigraphica,” Atti della Pontificia Accademia Romana di Archeologia, ser. 3, Rendiconti 3 (1925): 197–205. 10 For the monograms, see H. Swainson, “Monograms of the Capitals of S. Sergius at Constantinople,” BZ 4 (1895): 106–8. 11 The sculpture and brickstamps from St. Polyeuktos lack cruciform monograms. For the monograms, see R. M. Harrison, Excavations at Sarac¸hane in Istanbul, vol. 1 (Princeton, 1986), 162, fig. L; M. J. Vickers, “A ‘New’ Capital from St. Polyeuktos (Sarac¸hane) in Venice,” Oxford Journal of Archaeology 8 (1989): 228, fig. 2; and S. J. Hill, “The Brickstamps,” in Harrison, Excavations, 1:213, fig. B (B11, B28); 216, fig. C (C11, C13– C16). I will demonstrate elsewhere that the superstructures of the church of St. Polyeuktos were built using stamped bricks manufactured in and after 518, which is, therefore, the terminus post quem for the construction of the upper parts. E. Weigand, “Ein bisher verkanntes Diptychon Symmachorum,” JDAI 52 (1937): 130, placed the introduction of the cruciform monogram “zwischen 530 und 540” but lacked firm evidence for a terminus post quem. Such evidence now comes from the St. Polyeuktos excavations. 5 6

JONATHAN BARDILL

3

group of capitals used in Sts. Sergius and Bacchus may indicate that the type (or at least Theodora’s monogram of that type) had not appeared even by 527. In the churches of Hagia Eirene, Hagia Sophia (both ca. 532–537), and St. John at Ephesus (535/6–541),12 cruciform monograms of the empress are employed, either alone or in combination with her box monograms.13 The evidence might, therefore, be taken to indicate that the capitals in Sts. Sergius and Bacchus had been carved after Theodora’s accession in 527 but somewhat before the designs for the capitals in Hagia Eirene and Hagia Sophia had been prepared, ca. 533.14 Furthermore, no monogram of AUGOUCTA occurs on the capitals, although it does appear in both Hagia Eirene and Hagia Sophia, and this is perhaps another indication that the capitals were carved before ca. 533.15 These are, of course, arguments from silence, and there may be other explanations: the sculptors were unfamiliar with the most up-to-date imperial monograms, or the capitals were old stock or spoils taken from elsewhere. A terminus ante quem in 536 is firmer, since the acts of the council held in May and June of that year were signed by one Paul, “the presbyter and abbot of the holy apostles Sts. Peter and Paul and of Sts. Sergius and Bacchus in the palace of Hormisdas.” 16 Unfortunately, the five brickstamps that have so far been published are not helpful in clarifying the church’s date. The two types of rosette that Ernest Mamboury and Cyril Mango ascribed to the sixth century belong, in fact, to the Turkish period, being stamped on hexagonal tiles used to relay the floor in the gallery.17 The three other exFor the date, see C. Foss, Ephesus after Antiquity: A Late Antique, Byzantine and Turkish City (Cambridge, 1979), 88 n. 88. 13 See W. S. George, The Church of St. Eirene at Constantinople (London, 1912), fig. 7, pl. 16 (Hagia Eirene); C. G. Curtis and S. Aristarches, “ Ane j ´ kdotoi jEpigra␸ai´ Buzanti´ou,” JO ejn Kwnstantinoupo´ lei JEllhniko´ " Filologiko` " Su´ llogo", archaeological suppl. to vol. 16 (1885): 13–15 (Hagia Sophia); E. M. Antoniades, “Ek␸rasi" th'" Agi J ´a" So␸i´a", vol. 2 (Athens, 1908), 23, 29, 31, 34, 35, 68, 69, 205, 209, 213, 225, 231, 232, 236, 297, 309, 326, 335, 343, 356 (Hagia Sophia); H. Vetters, ed., Die Inschriften von Ephesos: Inschriften griechischer Sta¨dte aus Kleinasien 17.2, VII.2 (Bonn, 1981), no. 4363 (St. John at Ephesus). 14 Curtis and Aristarches, “ Ane j ´ kdotoi jEpigra␸ai´ Buzanti´ou,” 15, pl. 3, no. 8, suggest that the monogram on a capital in the southwest exedra at ground-floor level in Hagia Sophia gives the date A.M. 6042. The interpretation is doubtful, but if correct it would suggest that construction had reached the first storey by 533/4. See also C. G. Curtis, Broken Bits of Byzantium, pt. 2 (n.p., 1891), no. 15; W. R. Lethaby and H. Swainson, The Church of Sancta Sophia Constantinople (London–New York, 1894), 296; Antoniades, “Ek␸rasi" th'" Agi J ´a" So␸i´a", 2:29; H. Ka¨hler, Hagia Sophia, trans. E. Childs (New York–Washington, D.C., 1967), 14 with pl. 75; R. J. Mainstone, Hagia Sophia: Architecture, Structure and Liturgy of Justinian’s Great Church (London, 1988), 185. C. Foss, “Three Apparent Early Examples of the Era of Creation,” ZPapEpig 31 (1978): 241–46, does not discuss this capital. 15 Swainson, “Monograms,” 107, took the absence of the AUGOUCTA monogram to indicate that the church was built before Theodora became augusta. But this is impossible, since the entablature inscription and monograms of BACILEVC clearly indicate that the church was built after Justinian had become emperor, and Theodora is known to have become augusta on the same day that she and her husband were crowned. T. Mathews, The Early Churches of Constantinople: Architecture and Liturgy (University Park, Pa., 1971), 47, misrepresents Swainson’s argument, stating; “Theodora is not named in the monograms.” 16 Mango, “Tradition,” 191; Krautheimer, “Again,” 253; Mango, “Once Again,” 385–86. For Paul’s signatures, see ACO, ed. Schwartz, vol. 3 (Berlin, 1940), 46 no. 34, 129 no. 54, 144 no. 56, 158 no. 54, 164 no. 54, 173 no. 59. In no. 34, Paul styles himself presbu´ tero" kai` hJgou´ meno" tw'n aJgi´wn Pe´ trou kai` Pau´ lou tw'n ajposto´ lwn, Sergi´ou kai` Ba´ kcou tw'n martu´ rwn. 17 See E. Mamboury, “Les briques byzantines marque´es du chrisme,” Annuaire de l’institut de philologie et d’histoire orientales 9 (1949): 449–62 (where nos. 16 and 17, in fact, represent the same stamp); C. Mango, “Byzantine Brick Stamps,” AJA 54 (1950): 19–27, fig. 1 (where the two different types are shown). The 12

4

CHURCH OF STS. SERGIUS AND BACCHUS

amples, although certainly sixth century, cannot at present be dated with greater precision.18 We will require a larger sample of material that can be said with certainty to come from the Justinianic fabric of the church, if brickstamps are to clarify matters.19 Thus, the capitals were designed, and the inscription composed, in the period 527– 536, possibly between 527 and 533. Mango suggested that the epithet “sleepless,” which is used to describe Justinian in the entablature inscription, reflected the emperor’s reputation, known from Procopius and John Lydus, for working late into the night.20 He deduced that “the inscription was composed at a time when Justinian’s nocturnal habits had achieved some notoriety,” 21 therefore probably several years after his accession in 527. The same might also be argued on the basis of the description of Theodora’s “unsparing efforts to nourish the destitute.” It is, however, unclear whether these claims have some historical basis or are merely conventional topoi. Certainly, the inscription contains standard themes. The descriptions of Theodora’s mind as “adorned with piety” (eujsebi´h ␸aidru´ netai) and of Justinian as “fostering piety” (eujsebi´hn . . . aje´ xwn) recall references in the dedicatory inscription of St. Polyeuktos to Anicia Juliana’s mind “full of piety” (eujsebi´h" plh´ qousan)22 and to her “heeding piety” (eujsebi´h" ajle´ gousa).23 The invocation of St. Sergius to guard Justinian’s rule and to increase Theodora’s power is reminiscent of Juliana’s invocation of all the saints to whom she had dedicated churches to protect her, her son, and his daughters.24 However, it is less clear whether the more specific references to Justinian’s sleeplessness and to Theodora’s care for the destitute are equally traditional themes. Since the emperor and empress would have moved from the palace of Hormisdas into the Great Palace upon their accession in 527, it is not certain why they chose to locate a magnificent new church in their former residence, where Justinian had built a basilica in honor of the Apostles Peter and Paul only a few years earlier (ca. 519). Krautheimer proposed that the location of the church might be explained by the fact Turkish date of the hexagonal tiles is noted, e.g., by George, The Church of St. Eirene, 62. See also S. Eyice, ˘lalar,” Sanat Tarihi Yıllıg˘ı 9–10 (1979–80): 155–62. “Osmanlı devri Tu ¨ rk yapı sanatında damgalı tug 18 See I. P. Meliopoulos, “ jExakri´bwsi" ajrcaiw'n topoqesiw'n,” JO ejn Kwnstantinoupo´ lei JEllhniko` " Filologiko` " Su´ llogo" 29 (1907): 222–31, 231, section G⬘, nos. 1–2; van Millingen, Byzantine Churches in Constantinople, 79, no. 1 (mentioned earlier by E. M. Antoniades, “Ek␸rasi" th'" Agi J ´a" So␸i´a", vol. 1 [Athens, 1907], 83 n. 36). The 6th-century stamp bearing the twelfth indiction published by Meliopoulos could conceivably refer to the period from 1 September 533 to 31 August 534, but we cannot be sure that the brick was found in the fabric of the church, nor that it was new when it was used. 19 The only other evidence that might be adduced to date the church is a confused passage in the unreliable Patria, 3.39, Scriptores originum Constantinopolitanarum, ed. T. Preger, vol. 2 (Leipzig, 1907), 231–32, with G. Dagron, Constantinople imaginaire (Paris, 1984), 320–21. 20 See Procopius, Secret History, 12.20, 12.27, 13.28, ed. Veh, vol. 1 (Munich, 1981); trans. H. B. Dewing, Procopius, vol. 6 (London, 1935); Procopius, Buildings, 1.7.8–9; John Lydus, De Magistratibus, 3.55, ed. R. Wu ¨ nsch, Ioannis Lydi De Magistratibus (Leipzig, 1903), trans. A. C. Bandy, On the Powers or the Magistracies of the Roman State (Philadelphia, 1983). 21 Mango, “Tradition,” 190; idem, “Once Again,” 388–89. 22 Palatine Anthology, 1.10.25–26, ed. H. Stadtmueller, Anthologia Graeca, vol. 1 (Leipzig, 1894), trans. W. R. Paton, The Greek Anthology, vol. 1 (London–New York, 1927). 23 Palatine Anthology, 1.10.16. 24 Palatine Anthology, 1.10.37–39. For similar comparisons between these epigrams, see C. L. Connor, “The Epigram in the Church of Hagios Polyeuktos in Constantinople and Its Byzantine Response,” Byzantion 69 (1999): 511–12.

JONATHAN BARDILL

5

that the foundations had been laid shortly before Justinian and Theodora ascended the throne. However that may be, the two churches in the palace of Hormisdas would not have been easily accessible to the imperial couple until a connection had been established between the palace of Hormisdas and the Great Palace, and this work does not appear to have been undertaken until after 532.25 Mango, besides being concerned about the location chosen for the church, wondered why the imperial couple had decided to dedicate it to Sts. Sergius and Bacchus.26 He proposed that explanations for both the location and dedication were to be found in John of Ephesus’ Lives of the Eastern Saints. It is recorded there that the empress Theodora accommodated a community of over five hundred Monophysite refugees within the great halls (triklinoi) of the palace of Hormisdas,27 and Mango suggested that the church of Sts. Sergius and Bacchus had been built for their use. First, he observed that the entablature inscription refers to Theodora, “whose constant toil lies in unsparing efforts to nourish the destitute,” and suggested that this could be taken to refer to the shelter afforded the persecuted Monophysites.28 Second, John of Ephesus states that in one of the great halls in which the Monophysites were lodged was a “martyrium,” 29 which Mango asserted must refer to the church of Sts. Sergius and Bacchus itself.30 The inscription’s reference to Theodora’s charitable works need not refer specifically to the Monophysite refugees, and Krautheimer dismissed it as a mere topos.31 Furthermore, there is no conclusive proof that John of Ephesus’ martyrium is the church of Sts. Sergius and Bacchus—and Krautheimer even made the implausible suggestion that the Syriac word rendered “martyrium” in Brooks’s translation (be¯th sa¯hde¯) was simply a “reliquary” or “relic.” 32 Despite the lack of corroborative evidence, however, it must be admitted that Mango’s theory provides a particularly good explanation for the location of the church in the palace of Hormisdas. Additionally, since John of Ephesus tells us that the refugees came from Syria, Armenia, Cappadocia and Cilicia, Isauria and Lycaonia, Asia, Alexandria, and Constantinople,33 the hypothesis also provides a possible explanation for the dedication of the church to St. Sergius, to whom only one other church in Constantinople was dedicated34 but who was greatly revered in Syria.35 Procopius, Buildings, 1.4.1–2 (probably written ca. 554) states that the connection was made when Justinian became emperor. However, a Syriac account of the conversations held in the palace of Hormisdas in 532 suggests that the connection was established shortly after the discussions: see S. Brock, “The Conversations with the Syrian Orthodox under Justinian (532),” OCP 47 (1981): 92 with n. 17. 26 Mango, “Tradition,” 191; idem, “Once Again,” 388. It may be noted that only Sergius is mentioned in the dedicatory inscription. 27 The relevant part of the Syriac text is edited with an English translation in “John of Ephesus: Lives of the Eastern Saints,” trans. E. W. Brooks, PO 18 (1924): 600, 676–84. 28 Mango, “Tradition,” 191. 29 Brooks, “Ephesus,” 682. 30 Mango, “Tradition,” 192. 31 Krautheimer, “Again,” 253. 32 Ibid., 252. 33 Brooks, “Ephesus,” 677. 34 For the church, which was near the cistern of Aetius, see R. Janin, La ge´ographie eccle´siastique de l’Empire byzantin, vol. 1, Le sie`ge de Constantinople et le patriarcat oecume´nique, pt. 3, Les ´eglises et les monaste`res, 2nd ed. (Paris, 1969), 454. 35 On the cult of St. Sergius in Syria and Mesopotamia, see E. K. Fowden, The Barbarian Plain: Saint Sergius Between Rome and Iran (Berkeley–Los Angeles–London, 1999). 25

6

CHURCH OF STS. SERGIUS AND BACCHUS

Krautheimer raised a number of more or less serious objections to Mango’s position. In Krautheimer’s opinion, the refugees would have arrived in Constantinople only after the council of 536 at which Monophysitism was condemned, hence after the church of Sts. Sergius and Bacchus had been built. Furthermore, he argued that it was unlikely that such a fine church would have been constructed for refugees.36 The first objection is not serious, since persecution began in the East as early as 521, ceasing only during the peaceful negotiations of 530–536.37 Thus, Theodora might have begun to provide shelter for persecuted monks from the time of her accession in 527.38 Krautheimer’s second criticism, however, was tacitly accepted by Mango, who shifted his argument somewhat, asserting in a later article: “The fact that the Monophysites were lodged at such a smart address shows that this [construction] happened when they were being courted rather than when they were being persecuted.” 39 It must indeed be admitted that the de´tente between 530 and 536 would have been the most likely period for the imperial couple to have built a church for the use of a Monophysite community. According to Elias’ Life of John, Bishop of Tella, a delegation of Monophysite monks arrived in Constantinople in 531,40 and we first hear of the Monophysites’ association with the palace of Hormisdas in 532, when discussions between Monophysite and Orthodox bishops were held in the heptaconch triclinium of the palace.41 Mango suggested that the church of Sts. Sergius and Bacchus was built for these visiting Monophysite dignitaries and that construction would have been started around 531, making the church a contemporary of Hagia Sophia.42 The Monophysite refugees—who would have established themselves in the palace of Hormisdas between 527 and 530—must also have had access to the church of Sts. Sergius and Bacchus, Mango argued, since John of Ephesus refers to their martyrium in that palace and since Theodora’s care for the persecuted monks is proclaimed in the entablature inscription. Another of Krautheimer’s objections to Mango’s thesis rested upon the signature of the abbot Paul on the acts of the anti-Monophysite council of 536. From this evidence, Krautheimer deduced that Paul was the abbot of an Orthodox monastery that had nothing whatsoever to do with the community of Monophysites assembled elsewhere in the palace of Hormisdas. Krautheimer suggested that Paul would have ensured that the Monophysites were kept well away from his churches of Sts. Peter and Paul and Sts. Sergius and Bacchus.43 In response, Mango plausibly countered that Paul might have been a Monophysite abbot who changed sides in 536, either willingly or under pres36 Krautheimer, “Again,” 252: “it seems to me unlikely that a structure that permanent and splendid was built for a refugee camp.” 37 See W. H. C. Frend, The Rise of the Monophysite Movement (Cambridge, 1972), 247, 260–63, 273–75; J. J. van Ginkel, “John of Ephesus: A Monophysite Historian in Sixth-Century Byzantium” (Ph.D. diss., Groningen, 1995), 27–29, 168–69. 38 Mango, “Tradition,” 192 n. 10, considered that the refugees may have begun to arrive in Constantinople as early as 527. In “Once Again,” 386, Mango was prepared to push the arrival of the refugees back only as far as 531. 39 Mango, “Once Again,” 392. 40 Trans. E. W. Brooks, CSCO, Scriptores Syri, ser. 3, vol. 25 versio (Leipzig, 1907), 39. 41 ACO, ed. Schwartz, vol. 4.2 (Berlin, 1914), 169. That the discussions took place in the palace of Hormisdas is confirmed by an independent Syriac account: see Brock, “Conversations,” 92. 42 Mango, “Once Again,” 392. 43 Krautheimer, “Again,” 252–53.

JONATHAN BARDILL

7

sure.44 It must be admitted, however, that his alternative suggestion that an Orthodox abbot might have been appointed to lead the refugee community when persecution resumed in 536 seems much less likely.45 It is certainly possible that the abbot Paul was the head of the Monophysite community: John of Ephesus states that the refugees in one great hall of the palace of Hormisdas “had as it were the organization of one convent and one service, and an archimandrite and a steward, and servitors and a table together, and were continually governed according to the whole perfect organization of a convent.” 46 In fact, the same author provides further information that may best be explained by linking the abbot Paul with the Monophysite monastery in the palace of Hormisdas. John records that sometime after Theodora’s death in 548, Orthodox opponents of the Monophysites attempted to sully the community’s reputation: And while a few of the holy men were in these cells, the adversaries introduced some women with their husbands, and others who were not chaste . . . and whereas these men thought to defile the saints’ dwelling [i.e., the Monophysite accommodation], God purified it by a sudden fire, in that fire fell and burnt the whole of that place, only a small portion of it escaping. . . . And so at last it was given to the martyrs’ chapel of the holy Mar Sergius [i.e., the church of Sts. Sergius and Bacchus], and a monastery was built on that spot and it remains to the present time to the glory of God.47

Justinian, respecting the request of his late wife that the Monophysite community should be preserved,48 then moved it to another imperial property, the house of Urbicius.49 Thus, according to John of Ephesus’ account, many of the halls of the palace of Hormisdas that had been home to the Monophysite refugees were destroyed in a fire sometime after 548 and before 565/6 (when John was writing).50 The church of Sts. Sergius and Bacchus survived the conflagration and was given the devastated ground nearby, on which a monastery was founded.51 This testimony, therefore, suggests that the church of Sts. Sergius and Bacchus had no monastery proper before 548. How, then, do we explain that already in 536 one Paul could claim to be abbot of the monastery of Sts. Peter and Paul and Sts. Sergius and Bacchus? Mathews simply suggested that John of Ephesus’ information concerning the foundation of the monastery is unreliable,52 but this seems most unlikely given John’s first-hand knowledge of the refugee community.53 It would seem reasonable to conclude that Paul was the head of the Monophysite monastic comMango, “Once Again,” 389. Ibid., 392. 46 Brooks, “Ephesus,” 678. 47 Ibid., 683–84. 48 Ibid., 680. 49 Ibid., 683. 50 For the date of composition, see ibid., 680 n. 2, 681 n. 2, and E. W. Brooks, “Introduction,” in PO 17 (1923): iii–xv. 51 This is perhaps the monastery to which Cedrenus, Historiarum Compendium, ed. I. Bekker, vol. 1 (Bonn, 1838), 642–43, refers as being for “famous men.” 52 Mathews, “Architecture,” 24. 53 On the date and location of John’s residence in Constantinople, see Brooks, “Introduction,” iv–vi. He became head of Constantinople’s Monophysite community after the death of Theodosius of Alexandria in 566, by which time it had been moved to the house of Urbicius. See van Ginkel, “John of Ephesus,” 32–33, 214. 44 45

8

CHURCH OF STS. SERGIUS AND BACCHUS

munity housed in the halls of the palace of Hormisdas. If so, by 536, this community, which was not housed in a monastery proper, had apparently been granted access to both the churches of Sts. Peter and Paul and Sts. Sergius and Bacchus. So far, so good. Unfortunately, however, there are two strong objections to Mango’s suggestion that the church of Sts. Sergius and Bacchus is to be identified with the martyrium described by John of Ephesus. Krautheimer rightly observed that the Monophysite accommodation appears to have been arranged in haste to cope with an emergency situation and that the monks, who are said to have erected altars everywhere in the halls of the palace of Hormisdas, do not appear to have had access to any church.54 It may be added that John of Ephesus states explicitly that the martyrium was inside a triklinos and was clearly not a church in its own right but rather a chapel set up within a hall of the palace.55 The second objection is decisive. John of Ephesus describes at some length the occasion on which the great hall containing the martyrium collapsed under the weight of the congregation.56 The disaster was terrible, and the cries of hundreds of refugees were heard in the Great Palace, bringing the imperial couple to the scene. Mathews noted this passage and rightly concluded that since the church of Sts. Sergius and Bacchus still stands today, it cannot be identified as the martyrium that collapsed.57 Mango, although mentioning the disaster, did not address this serious objection in his reply.58 On the basis of these two major objections, it would appear that we must conclude, along with Krautheimer and Mathews, that there is no demonstrable link between the church of Sts. Sergius and Bacchus and the community of Monophysite refugees based in the halls of the palace of Hormisdas. One additional possibility, however, might answer these objections: a closer reading of John of Ephesus suggests that the surviving church of Sts. Sergius and Bacchus may have been built to replace the hall whose collapse he describes. Crucially, John records that, after the disaster, Justinian “made the hall itself that fell an
azga¯.” 59 Brooks translates
azga¯ as “portico,” but it can also mean “arch,” “vault,” “vaulted roof,” or “vaulted building.” 60 That John could here be referring to the construction of a new building with a domed roof is demonstrated by the use of the same Syriac term in discussions of the vaulting of two sixth-century domed churches.61 Michael the Syrian describes the earthquake of 7 May 558, which caused a portion of the vaulting of Hagia Sophia in Constantinople to collapse. From other accounts of the disaster, it is clear that the “vault” to which Michael refers must be either the eastern part of the dome or the eastern semidome.62 Verse 8 of the inauguration anthem of Justinian’s domed 54 Krautheimer, “Again,” 252; Brooks, “Ephesus,” 676–79, with descriptions of temporary cells and booths on 678 and reference to makeshift altars on 679. 55 Brooks, “Ephesus,” 682. 56 Ibid., 681–82. 57 Mathews, “Architecture,” 24: “visiblement, les moines officiaient ailleurs.” 58 Mango, “Once Again,” 386. 59 Brooks, “Ephesus,” 683. 60 See R. Payne Smith, Thesaurus Syriacus, vol. 1 (Oxford, 1879), col. 104. 61 I am extremely grateful to Sebastian Brock for providing, at my request, the parallel references referred to here, and for giving translations where unavailable or debatable (letter to the author of 2 December 1994). 62 Michael the Syrian, Chronicle 9.31, ed. J.-B. Chabot, Chronique de Michel le Syrien, vol. 2 (Paris, 1901), 262, 320. Regarding this collapse, Paul the Silentiary reports: “but the curve of the eastern arch slipped off and a portion of the dome was mingled with dust: part of it lay on the floor, and part—a wonder to behold— hung in mid-air as if unsupported” (trans. C. Mango, The Art of the Byzantine Empire, 312–1453 [Englewood

JONATHAN BARDILL

9

church of Hagia Sophia in Edessa states that “its entire ceiling (i.e., dome) is fastened on the vaults (
azge¯),” but the description is unfortunately too vague to determine exactly which architectural elements (semidomes or squinches) are meant.63 Given that the collapsed triklinos had been used as a martyrium, it is not unlikely that the vaulted structure built to replace it was a domed church.64 We cannot, of course, be certain that this new vaulted building was the church of Sts. Sergius and Bacchus that survives today; but, if we grant that it was, then we have a satisfactory explanation for the church’s location and dedication.65 There is a further indication that the new, vaulted edifice was, in fact, a church, since John claims that it was constructed in thanks to God for the miraculous survival of all those involved in the disastrous collapse of the martyrium: “this story filled the king and queen and the magnates also, and the rest of the people throughout the city with astonishment, and called forth praise and thanksgiving to God, and rendered the community of the blessed men itself honourable in the eyes of everyone, so that at last the king sent, and made the hall itself that fell a vaulted building.” 66 There is undoubtedly some exaggeration here, and John’s claim that the community was “honourable in the eyes of everyone” might simply be taken to indicate that the events occurred during the de´tente of 530–536. But if we give credence to the story of the collapse itself and identify the vaulted building with Sts. Sergius and Bacchus, then it was primarily this disaster that motivated the construction of the church for the refugees’ use—although we may be sure that Justinian’s desire to court the Monophysite dignitaries who were then present in Constantinople explains the grandeur of the new edifice. I conclude by summarizing the possible sequence of events. A community of Monophysite refugees was established by Theodora in the halls of the palace of Hormisdas after she and her husband had moved into the Great Palace in 527. According to John of Ephesus, this community was organized in the manner of a convent, with an archimandrite and stewards, and it employed one of the palace halls as a martyrium. His account makes no suggestion that the monks were granted access to the church of Sts. Peter and Cliffs, N.J., 1972], 80–81). For other sources concerning this collapse, see C. Mango, “Byzantine Writers on the Fabric of Hagia Sophia,” in Hagia Sophia from the Age of Justinian to the Present, ed. R. Mark and A. H. C ¸ akmak (Cambridge, 1992), 51–53. 63 See primarily K. E. McVey, “The Domed Church as Microcosm: Literary Roots of an Architectural Symbol,” DOP 37 (1983): 95; A. Palmer (with L. Rodley), “The Inauguration Anthem of Hagia Sophia in Edessa: A New Edition and Translation with Historical and Architectural Notes and a Comparison with a Contemporary Constantinopolitan Kontakion,” BMGS 12 (1988): 132, 160; K. E. McVey, “The Sogitha on the Church of Edessa in the Context of Other Early Greek and Syriac Hymns for the Consecration of Church Buildings,” Aram 5 (1993): 333, 357. The form of church at Edessa is unclear. I wonder whether it may have been an aisled tetraconch, similar to those at Seleucia-Pieria, Apamea, Bosra, or Rusafa. Verse 7 of the hymn would then refer to the four main arches that supported the dome, and verse 8 would speak of the semidomes that sprang backward from these arches, thus surrounding the dome “like crags jutting from a mountain.” 64 The Life of Behnam and Sarah mentions a martyrium made “in the form of small vaults (
azgo¯ne¯).” Syriac text in P. Bedjan, ed., Acta martyrum et sanctorum, vol. 2 (Paris, 1891), 432. 65 It remains to be determined whether the irregularities in the design of the church can be explained by the suggestion that it occupies the site of an earlier palatial hall, and whether anything of that palatial hall (of uncertain date) survives. It may also be worthwhile considering the possibility that the double-shell design was chosen by those for whom Sts. Sergius and Bacchus was built, since the plan of the church is closely related to those of the aisled tetraconchs of Syria, Jordan, and Palestine. 66 Brooks, “Ephesus,” 682–83, corrected.

10

CHURCH OF STS. SERGIUS AND BACCHUS

Paul.67 On one occasion, the hall that served as a martyrium collapsed, and the imperial couple determined to build a vaulted structure to replace it. This was presumably the church of Sts. Sergius and Bacchus. The construction of such a grand building for a Monophysite community, and at imperial command, is most likely to have occurred during the peaceful negotiations of 530–536; the evidence of the monograms on the capitals might be used to argue for a date no later than ca. 533. Theodora’s care for the refugee community, which had grown up during the preceding years of persecution, was proclaimed in the entablature inscription of the new church. The dedication of the church to Sergius was intended to appeal in particular to the refugees from the frontier regions of Syria and Mesopotamia, where the saint was most popular. By means of this and other gestures, Justinian and Theodora hoped to encourage them both to come to terms with their Orthodox opponents and to remain firm in their opposition to Persia.68 By 536 one abbot is known to have overseen both Sts. Sergius and Bacchus and Sts. Peter and Paul, so the latter, too, was apparently available for use by the Monophysite monks. Perhaps it was during the construction of Sts. Sergius and Bacchus that the refugees had first been granted access to the adjoining church. The abbot in 536, one Paul, was forced to sign the acts of the anti-Monophysite council. At some point after 548, the halls containing the temporary monastic cells of the Monophysites were destroyed by fire, and the community was transferred to the house of Urbicius. Only then was a proper monastery founded for the churches of Sts. Peter and Paul and Sts. Sergius and Bacchus. According to this interpretation of John of Ephesus’ account, the church of Sts. Sergius and Bacchus is not to be identified with the martyrium he describes but with the vaulted building that replaced it on the same site, and the church was built for a community of Monophysite refugees, not, as Krautheimer and Mathews held, for imperial use.69 The alternative to this proposal is to conclude that the Monophysites who were lodged 67 Some monasteries in Mesopotamia and northern Syria consisted of a large, rectangular hall, to the side of which was attached a small room serving as an oratory. See, e.g., G. Tchalenko, Villages antiques de la Syrie du nord, vol. 1 (Paris, 1953), 178–81; G. Bell, The Churches and Monasteries of the Tur
Abdin (London, 1982), ´ glises de Me´sopotamie du Nord: Ambar et 6–10, 10–13, 31–35, 35–38, 38–39; M. Mundell Mango, “Deux E Mar Abraham de Kashkar,” CahArch 30 (1982): 47–70. I thank Marlia Mundell Mango for this observation. 68 See Fowden, Barbarian Plain, 130–33, who observes that Justinian and Theodora’s alliance with St. Sergius was proclaimed not only by the construction of the church in Constantinople (possibly to house the relic of the saint’s thumb, translated to Constantinople by Anastasius between 514 and 518) but also by sending a gift of a gem-encrusted gold cross to the saint’s shrine in Rusafa between 527 and 540. I. Shahıˆd, “The Church of Sts. Sergius and Bacchus in Constantinople: Who Built It and Why?” BSCAbstr 22 (1996): 84, observes that St. Sergius was a Roman, a soldier, had probably fought the Persians, and was buried near the Persian border. He postulates that Justinian would have adopted the saint as a palladium during the Persian War of 527–532. 69 Mango’s “Tradition” and “Once Again” were attempts not only to link the Monophysites with Sts. Sergius and Bacchus but also to discredit thereby the alleged tradition of building imperial palace-churches (i.e., churches built within a palace and used for the celebration of a specifically imperial ritual) on a double-shell plan (i.e., with a domed core and ambulatory). I preferred not to conflate the two issues here and have addressed only the former. Krautheimer was not convinced by Mango’s arguments against the existence of the tradition (R. Krautheimer, Early Christian and Byzantine Architecture, 4th ed., rev. R. Krautheimer and ´ urcˇic´ [Harmondsworth, 1986], 224 n. 22, and, in general on the postulated tradition, 76–78, esp. 76 n. S. C 22). There can be no doubt that Sts. Sergius and Bacchus was built within the palace of Hormisdas and that this palace was connected (probably after 532) to the Great Palace by Justinian (see above, note 25). But if my hypothesis is accepted, the church was certainly not built for imperial use and, therefore, cannot be used to support whatever case there may be for a tradition of double-shell imperial palace-churches.

JONATHAN BARDILL

11

in the halls of the palace of Hormisdas had no link with the monastery of Sts. Peter and Paul and Sts. Sergius and Bacchus. But in that case, an explanation would have to be found for the existence of an Orthodox monastery of Sts. Peter and Paul and Sts. Sergius and Bacchus not far from a community of refugee Monophysite monks, and in a vacated imperial residence. Oxford University

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Organizational Setup and Functioning of the Fish Market in Tenth-Century Constantinople GEORGE C. MANIATIS

ish, being an important source of animal protein, played a very important role in the diet of Constantinople’s large population in the tenth century. Yet, our understanding of the way the fish industry actually worked within the prevailing economic and institutional milieu is incomplete. Research heretofore has focused on the guild of the fish merchants (ijcquopra´ tai) as part of the state regulatory mechanism rather than on the functioning of the fish market per se. The distinction between wholesale and retail fresh fish markets has been overlooked and its implications remain unascertained. The connection between the fish markets and the processing segment of the industry has been missed, and the latter’s effect on fresh fish pricing has not been explored. The impact of the organizational setup and key institutional parameters on the conduct of the individual players, the operation of the downstream markets, and the nature of the competitive process remains elusive. In addressing these important issues, in-depth analysis of key operational aspects of the industry is crucial as it provides the requisite supporting structure.1 It offers a perspective for probing the rationale for the guild system as applied to the fish industry, for ascertaining the behavioral patterns of guild members and enforcing state officials within the existing legislative framework, and for detecting the character of interactions between fisherman and wholesaler, wholesaler and retailer, and fisherman or fish merchant and processor. It also establishes the dynamics of inter-seller and inter-buyer relations, gauges the extent of state intervention, and puts in perspective the role of the market mechanism.

F

1 The Book of the Eparch ( jEparciko` n Bibli´on) lays out the basic provisions that governed the guild of the fish merchants (17.1–4). The Greek text was published with emendations by J. Nicole, Le Livre du Pre´fet ou ´ dit de l’Empereur Le´on le Sage sur les Corporations de Constantinople (Geneva, 1893), and was reprinted in l’E J. Zepos and P. Zepos, Jus Graecoromanum (Athens, 1931), 2:371–92. English translations are by A. E. R. Boak, “The Book of the Prefect,” Journal of Economic and Business History 1 (1929): 597–619; and E. H. Freshfield Roman Law in the Later Roman Empire—Byzantine Guilds Professional and Commercial (Cambridge, 1938). A recent critical edition of the Greek text as well as a German translation is by J. Koder, Das Eparchenbuch Leons des Weisen (Vienna, 1991). Literary sources provide virtually no evidence on the organization and operation of the fish market in the capital or on the nature of the interaction of the players involved. On the other hand, legal texts and the regulatory regime, judiciously interpreted and supplemented with sound economic analysis, supply the requisite building blocks for the construction of a workable analytical framework helpful in ascertaining the likely character of the fish market and the behavior of the agents.

14

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

THE LEGAL REGIME OF THE FISHING GROUNDS The long-standing legal framework applicable to fishing activities established the unimpeded right to fish anywhere in the maritime space: O J kwlu´ wn me ejn qala´ ssh aJlieu¯ sai uJpo´ keitai th¯ peri` u”brew" ajgwgh' (Those barring me from fishing in the sea are subject to the actio injuriarum) (Basilics 53.6.5); Eij kai` sunhqe´ " ejsti parano´ mw" tina´ " kwlu´ ein eJte´ roi" aJlieu´ ein pro` tw'n ijdi´wn oi“kwn, o”mw" oJ kwlu´ sa" ejne´ cetai th' peri` u”brew" ajgwgh' (Though illegal, it is common for people to block others from fishing in front of their homesteads; nevertheless, those so doing are subject to the actio injuriarum) (Basilics 60.21.13; Synopsis Basilicorum, A.XXXVI.1). This situation was viewed as unconducive to efficient fixed-net fishing,2 apparently because (a) it resulted in missed opportunities to exploit available grounds suitable for this type of fishing; and (b) it threatened depletion of existing fishing grounds through overfishing or the harvesting of undersized fish, since the absence of enforceable property rights tends to foster shortsighted resource management.3 Moreover, this policy infringed upon the legitimate right of owners to exploit waterfront property. Since fixed-net fishing was a major source of the capital’s high-value fish supply,4 steps were taken to promote the utilization of all suitable shorelines and their sustainable development, to encourage proper management, and to ensure the availability of requisite investments in fixed and working capital. These goals were met through legislative action: (a) to combine littoral properties whose frontages were short of the prescribed length for the installation of fixed-net fisheries; and (b) to prevent third parties from preempting the equal right of waterfront property owners to set up and operate a fixed-net fishery in front of their land (pro´ qura qala´ ttia). In essence, the enacted legislation aimed to set operational priorities, modify prevailing business practices, and end the haphazard status quo ante of open access, which had hindered the full exploitation of the natural resource and resulted in significant loss of economic, social, and fiscal benefits. Potentially, these measures, aside from rectifying a perceived inequity vis-a`-vis tax-paying waterfront property owners and reducing the tension between property owners and free riders, would have streamlined fishing operations, encouraged more responsible behavior, increased the daily catch (and hence supply), lowered the price of species otherwise in limited supply, and increased employment opportunities—all while adding to the revenue collected from the tax levied on the market value of the catch at the waterfront (aJliei´a, aJlieutikh´ ). Furthermore, the extension of property rights to the sea floor would have raised littoral property values, broadened the tax base, and increased property tax revenues. Emperor Leo VI (886–912) sought to effect these remedies through several novels.5 Novel 56 extended the rights of waterfront property owners to the sea floor, in effect restricting the legal notion of res communis and the attendant right of open access and 2 These are fisheries established in shallow sea waters, lagoons, or river estuaries employing a technique whereby nets were anchored on poles driven into the sea floor (ejpocai´). 3 On the impact of such practices on the stock of fish, see H. S. Gordon, “The Economic Theory of a Common-Property Resource: The Fishery,” Journal of Political Economy 62.2 (1954): 129, 135, 139. 4 Most of the world’s highest valued fisheries are located in coastal waters. E. A. Loayza and L. M. Sprague, A Strategy for Fisheries Development, World Bank Discussion Paper no. 135 (Washington, D.C., 1992), xii. 5 For the text of Novels 56, 57, 102, 103, and 104, see Jus Graecoromanum, 1:126–27, 170–73; H. Monnier, Les Novelles de Le´on le Sage (Paris, 1923), 120–22, 145–46; C. A. Spulber, Les Novelles de Le´on le Sage (Cernauti, 1934), 234–37, 302–6; P. Noailles and A. Dain, Les Novelles de Le´on VI le Sage (Paris, 1944), 212–17, 334–43.

GEORGE C. MANIATIS

15

establishing a legally protected private property right where none existed before. Unlike other fishing methods, the lasting character of fixed-net installations gave the preempting occupant an enduring user right that precluded access to all others, notably the long-standing owners of shoreline properties. This pre-emptive right came to be viewed as unfair, because it vitiated the equal rights of use of these fishing grounds by waterfront property owners and, based on past experience, was judged to be unsuitable for the promotion of this type of fishing. The practical significance of Novel 56 was that, henceforth, the owner had the inalienable and enforceable right to prevent third parties from occupying the extension of his shoreline property into the sea. He alone was entitled to install and operate (or lease) a fixed-net fishery and could transfer (e.g., sell) his right at will. Previously, he had not had these rights and, in fact, was subject to an actio injuriarum if he attempted to stop prospective operators from setting up such a fishery. Potentially, Leo’s measure would have reduced disputes and led to better husbandry of the fixed-net fisheries by curbing excessive harvesting (which, by reducing the fish population, increases the costs to other fishermen),6 and by holding in check the use of detrimental fishing techniques (e.g., use of a too-small mesh size), that affect the size and weight, as well as the reproductive rates of the fish stock, with similar adverse effects.7 Such practices are difficult to resist under an open access system, which provides virtually no incentive to prudent resource management. With Novel 57, Leo codified as law a long-standing custom that established the minimum distance between fisheries (182.5 ojrguiai´ ⫽ about 350 m), thereby ensuring that fisheries did not encroach on one another. Novel 104 directed how disputes likely to arise from the violation of the prescribed distances would be resolved. Finally, Novel 102 provided a remedy for an owner whose waterfront property (qala´ ttio" nomh´ ) was short of the prescribed length for the establishment of a fixed-net fishery by mandating participation of owners of contiguous properties: if they did not agree voluntarily to form a joint operation—lacking a conventional communio (koinwni´a), a mandatory partnership (communio rerum) was to be formed. The rationale for this ruling was that, even though the obligatory communio rerum (hJ tw'n pragma´ twn koinw´ nhsi") might infringe individual rights, such a coercive union put adjoining properties into productive use, something that was advantageous to all participants and the common good. The profits from the forced joint undertaking were to be divided equally rather than in proportion to the extent of the shoreline property owned (Novel 103), on grounds that each partner’s share of the catch could not be determined with precision and, more important, that the major part would be useless without the lesser. The complement rendered operational something that was previously valueless. 6 These costs, referred to as “externalities,” are generated within the fishery and are imposed by fishermen on other fishermen. The offenders do not consider them as such in making their decisions, because these costs are not borne by them—they are the consequence of their actions but external to them, and they may be unaware that their decisions are imposing costs on others, or they may be aware but have no incentive to refrain from such actions. Real externalities affect harvest levels of other fishermen, whereas pecuniary externalities affect their income levels. For details, see R. S. Johnston, Fisheries Development, Fisheries Management, and Externalities, World Bank Discussion Paper no. 165 (Washington, D.C., 1992), 1–5. 7 R. Turvey, “Optimization and Suboptimization in Fishery Regulation,” American Economic Review 54 (1964): 64–76; Johnston, Fisheries, 4.

16

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

Gilbert Dagron is critical of this rationale, the absence of legal reasoning, and the internal inconsistency within the novels, because he felt they led to unnecessary privatization of the public domain, weakening of authority, proliferation of privileges, multiplication of law suits, compromise of individual rights, and institution of preferential property rights.8 Because of these alleged shortcomings, Dagron concludes that these measures “met with partial success,” and that Novels 56 (peri` proqu´ rwn qala´ ssh") and 103 (division of profit in communio rerum) fell into desuetude, as is evidenced by the fact that they were not retained in Attaleiates’ compilation (1072–73).9 He suggests further that only Novels 57, 102, and 104 were implemented.10 Yet, inclusion of the crucial Novel 56 in the Synopsis Minor (mid-12th or early 13th century),11 the Prochiron Auctum (end of the 13th century),12 and the Hexabiblos (14th century)13 suggests that it was likely being implemented well beyond the tenth century and apparently satisfactorily so, since it was not abrogated. Moreover, implementation of Novel 102 (communio rerum) presupposed that Novel 56 was in full force. Though the method and solution adopted by the state to advance its objectives may be criticized on narrow legal, political, or sociological grounds, the thrust of the novels had tangible economic merit. The measures induced smaller property holders, apparently numerous along the waterfront of the Sea of Marmara and the Black Sea, “who preferred to go against their own interests rather than cooperate with their neighbor” (Novel 102), to work together productively. A balance had to be struck between the advantage of securing property rights and sensible limits to the absolute freedom in their exercise within the framework of an organized society.14 To be sure, high stakes might 8 This critical review of the legislation enacted by the novels from a legal standpoint is presented in G. Dagron, “Lawful Society and Legitimate Power: “Ennomo" politei´a, e“nnomo" ajrch´ ,” in Law and Society in Byzantium: Ninth–Twelfth Centuries, ed. A. E. Laiou and D. Simon (Washington, D.C., 1994), 43–51; idem, “Poissons, peˆcheurs et poissonniers de Constantinople,” in Constantinople and Its Hinterland, ed. C. Mango and G. Dagron (Aldershot, 1995), 61–67. 9 Po´ nhma, Jus Graecoromanum, 7:491. 10 Dagron, “Lawful Society,” 47–48 and n. 90; see also ibid., n. 49. 11 Neara` l∆ (n"∆)tou' kurou' leo´ nto´ " fhsin, w”ste despo´ zein e”kaston tou' parakeime´ nou aujtv' qalatti´ou proqu´ rou kai` ku´ rion ei«nai tou´ ", ei“ tine" boulhqei'en cwri`" th'" aujtou' ejpitroph'" th'" tw'n proqu´ rwn wjfelei´a" ajpolau´ ein, ejkdiw´ kein (Leo’s Novel 56 stipulates that ownership over the extension of a property line into the nearby sea floor implies that the owner can evict those who attempt to profit from the use of these grounds without his permission). Synopsis Minor, A.32. 12 Qespi´zomen toigarou'n, tw'n ijdi´wn proqu´ rwn e”kaston despo´ zein kai` ku´ rion ei«nai tou´ twn, w”ste tou` " mh` pro´ qura e”conta" dhladh` mh` despo´ ta" gi´nesqai´ pote (1). (1)⫽Nov. Leonis 56 (We therefore decree that waterfront property owners shall have ownership and sovereign authority over the extension of their property into the sea floor [pro´ qura], and those who have no such property can never acquire property rights and exercise dominion). Prochiron Auctum, XXXVIII.78. 13 Peri` proqu´ rwn qalatti´wn hJ triakosth` neara` 3 tou' kai´saro" Le´ onto" fhsin w”ste despo´ zein e”kaston tou' parakeime´ nou aujtv'⭈ kai` ku´ rion ei«nai tou', ei“ tine" boulhqei'en cwri`" th'" aujtou' ejpitroph'" th'" tw'n proqu´ rwn ajpolau´ ein wjfelei´a", ejkdiw´ kein. Note 3 reads: “Neara´ 56 (Emperor Leo’s Novel 56 stipulates that owners of waterfront properties have rights on their extension into the sea floor [proqu´ rwn qalatti´wn], and as owners they have the right to turn away anyone who attempts to profit from the occupancy of such property without their permission)”: Hexabiblos, 2.1.54. On the dates of Synopsis Minor, Prochiron Auctum, and Hexabiblos, see P. Collinet, “Byzantine Legislation from the Death of Justinian (565) to 1453,” Cambridge Medieval History (Cambridge, 1923), 4:722–23. 14 From his perspective, the emperor acted as the impartial, rational, and ideal legislator—a forerunner of classical utilitarians: “that action is best, which procures the greatest happiness for the greatest numbers”; F. Hutcheson, An Inquiry Concerning Moral Good and Evil (1725), quoted in J. Rawls, A Theory of Justice (Cam-

GEORGE C. MANIATIS

17

give rise to disagreements and litigation, but the courts could mediate disputes to resolve contested rights and normalize relations. Indeed, the very existence of property rights may be explained in terms of the necessity for the orderly exploitation and conservation of a resource.15 The impetus for legislative action evidently was a change in economic conditions, since the definition of new private property rights16 is largely an economic rather than legal calculus. As the value of a common property resource increases because of prospective higher net return, individuals are more prone to try to secure private property rights over it. People tend to exercise rights when they believe that gains from exploiting an underutilized resource will exceed the costs, and they neglect to do so when the likely gains from pursuing such an economic activity are deemed insufficient. In a sense, properties in the public domain are those that people have refrained from laying claim to. Whenever conditions or perceptions change, new gainful opportunities are likely to emerge and activities previously not worth undertaking are perceived as valuable enough to repay the effort spent.17 The process is reinforced and facilitated if, concurrently, state and private interests can be advanced as well: the convergence of public and private interests propels legislative initiatives to sanction transfer of rights from public to private ownership. The littoral property owners’ self-interest in opening fixed-net fishing grounds and securing exclusive exploitation rights to them conjoined with the state’s bridge, 1971), 22 n. 9. In the utilitarians’ view, society’s institutions should be arranged so as to achieve the greatest collective satisfaction. It is immaterial how this sum of satisfactions is distributed among individuals. The violation of the right of a few is vacated by the greater good achieved and shared by many—by social expediency. On the utilitarian doctrine, see Rawls, Theory, 22–26, 30–31; R. Nozick, Anarchy, State, and Utopia (New York, 1974), 28–30; B. Russell, A History of Western Philosophy (New York, 1945), 183, 627–29, 773–82; J. A. Schumpeter, History of Economic Analysis (New York, 1954), 130–34; R. A. Posner, The Economics of Justice (Cambridge, 1983), 48–87. Depending on their position along the ideological spectrum, critics have faulted utilitarianism as having failed to take seriously the distinction between persons and as having mistaken impersonality for impartiality (Rawls, Theory, 183–92); as violating the neutrality of the state between its citizens (Nozick, Anarchy, 30–33); as being unable to explain concrete government actions (Schumpeter, History, 428–29); or as lacking a method for calculating the effect of a decision or policy (Posner, Economics of Justice, 54–55). For a confutation of the criticisms leveled against classical utilitarianism as they apply to economic analysis and a theory of “constrained utilitarianism,” see Posner, Economics of Justice, 60–115. Utilitarian hypotheses are basic to the prescriptive part of economic theory referred to as welfare economics, often described as “applied utilitarianism.” A convincing argument can be made that there are circumstances under which state action may prove beneficial to those governed, in the sense that it helps them achieve their own ends while safeguarding other persons’ rights. When the welfare of the members of the economy is in part dependent on each others’ activity and persons pursuing their own immediate interests act in a manner contrary to the interests of the others, if voluntary arrangements to rectify adverse consequences cannot be relied upon, then it becomes advantageous to all members of the society to have the activities of every individual restricted by coercive measures. In such instances, state intervention and the extension of government power, although it involves restriction of choice and decision, will not necessarily affect the citizenry disadvantageously. Cf. W. J. Baumol, Welfare Economics and the Theory of the State (Cambridge, 1952), 140–42. 15 Gordon, “Fishery,” 134. 16 Property rights are understood as rights individuals appropriate over their own labor and the goods they possess, and are predicated on legal rules and their enforcement, organizational forms, and behavioral norms—in short, on the institutional framework. D. C. North, Institutions, Institutional Change and Economic Performance (New York, 1990), 33. For the concept, emergence, and role of property rights, see H. Demsetz, “Toward a Theory of Property Rights,” American Economic Review 57.2 (1967): 347–59. 17 Y. Barzel, Economic Analysis of Property Rights (New York, 1989), 65, 74–75; North, Institutions, 51.

18

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

larger interest to extend land ownership into the sea and remedy the perceived suboptimal utilization of this vital natural resource.18 ORGANIZATION OF THE FISH INDUSTRY Fishermen as Suppliers Fishermen resided in the capital and the nearby littoral communities dotting the shores of the Bosporos, the Sea of Marmara, and the Black Sea. They brought their catches to the city, which was by far the largest consumption center. Fishermen were not allowed to sell their catches at the fishing grounds or directly to consumers (Book of the Eparch, 17.3), which suggests that their operations were distinct from those of the fish merchants.19 If they managed to own a boat and the requisite gear, fishermen worked as independents: either solo, or by forming partnerships. Otherwise they worked as hired hands for larger fishermen who might own one or more fishing boats or for operators of fixed-net fishing establishments. Because of the importance of the fish industry, and despite its labor intensive nature, a substantial amount of capital must have been invested in fishing vessels and tackle. This suggests that the size distribution of fishing operations was rather wide. Fishermen very likely had difficulty in banding together voluntarily because of their very large number and wide dispersion. This, coupled with unimpeded entry and exit, probably contributed to a competitive organizational structure. Unlike the fish merchants, fishermen were not organized into an obligatory guild, probably because the nature of their activities and their spatial distribution, which extended beyond city limits, made it impractical. The Fish Merchants’ Guild The fish trade was organized by law (Book of the Eparch) into a guild, with its own chiefs appointed by the eparch, one for each of the designated retail fish markets in the 18 Under a regime of free and open access, a change in the perceived value of a common-property resource has important ramifications for fishery management and development: greater risk for stock depletion due to intensified fishing effort (i.e., increased inflow of capital and labor into the fishery), overfishing, or capture methods (e.g., smaller mesh size); greater likelihood of over-investment and hence waste of resources; or higher incidence of conflicts between different groups of fishermen and between existing fishermen and new entrants, e.g., as a result of competition for the same stock of fish in a given location (World Bank, A Study of International Fisheries Research [Washington, D.C., 1992], 14–15). To prevent this from occurring, the property regime would inevitably have to move away from free property access, and a more rational system of property ownership and resource use would have to be devised. Effective fishing management cannot be achieved without assigning some form of property rights to fishing grounds (World Bank, Study, 27; D. W. Bromley and M. M. Cernea, The Management of Common Property Natural Resources [Washington, D.C., 1989], 22). However, adoption of access controls leads to administrative allocation of exclusive use rights that benefit one set of users at the expense of another—a wealth redistribution decision that can only be made at the political level (World Bank, Study, 27). But any action with redistributive effects (e.g., taxation) unavoidably encroaches on an individual’s rights. Nevertheless, this is a small price to pay, a trade-off, for the achievement of larger state and public ends, all the more so, since continuation of a free access policy without privatization of property rights over fishing grounds is hardly a workable alternative. 19 ¨ckle, Spa¨tro¨mische und byzantinische Zu ¨nfte (Leipzig, 1911), 45. Cf. A. Sto

GEORGE C. MANIATIS

19

capital (ejcou´ sh" mia'" eJka´ sth" kama´ ra" kai` to` n prostateu´ onta 17.1).20 In those economic activities governed by a guild structure, guild membership was a condition sine qua non for the practice of the trade,21 albeit only for the proprietor of the business establishment. Indeed, workers were denied membership, as the guild was strictly an association of business owners and operators.22 The law does not set a numerus clausus for the fish merchants’ guild. Nor does it stipulate conditions of entry, which is not unusual (compare, e.g., sections 5, 7, 9, 10, 11, 13, 14, 15, 18, 19, 21). Apparently, entry conditions were similar to those applicable to guilds for which the law was explicit (e.g., 2.10, 3.1, 4.5, 6.6, 7.3, 8.13, 12.2, 16.1): the prospective candidate had to produce several respected and honest persons,23 who vouched for him before the eparch that he was worthy to exercise the trade (ejp∆ ajgaqh' uJpolh´ yei), that he would not contravene regulations in force (mhde` n para` ta` diatetagme´ na pra´ ttesqai), and that he was a person of means (eu“poro"). As in most other guilds, slaves were admitted into the fish merchants’ guild since they were not explicitly excluded, provided they were sponsored by their masters. The fish were retailed in vaulted chambers (ejn kama´ rai"), which kept them cool and fresher longer. ¨ckle, Byzantinische Zu ¨nfte, 45. Sto 21 This view is also held by A. P. Christophilopoulos, To` jEparciko` n Bibli´on Le´ onto" tou' Sofou' kai` aiJ Suntecni´ai ejn Buzanti´v (Athens, 1935), 4, 36, 50; G. Mickwitz, “Die Organisationsformen zweier byzantinischer Gewerbe im X. Jahrhundert,” BZ 36 (1936): 71 and n. 3; A. D. Sideris, JIstori´a tou' Oijkonomikou' Bi´ou (Athens, 1950), 264; B. Mendl, “Les Corporations Byzantines,” BSl 22 (1961): 302, 304, 312–18. 22 Many authors maintain that guild membership was not compulsory and, hence, industrial and trade ¨ckle, Byzantinische Zu ¨nfte, activities could be conducted outside the guild system: Nicole, Livre du Pre´fet, 80; Sto `n 8 ff and n. 5; Boak, “Book of the Prefect,” 608; G. Zoras, Le Corporazioni Bizantine: Studio sull’ Eparciko j Bibli´on dell’Imperatore Leone VI (Rome, 1931), 172; R. S. Lopez, “Silk Industry in the Byzantine Empire,” Speculum 20 (1945): 16 and n. 3; S. Runciman, “Byzantine Trade and Industry,” in Cambridge Economic History of Europe (Cambridge, 1987), 2:154 n. 12, 156; G. Ostrogorsky, History of the Byzantine State (Oxford, 1968), 254; D. Simon, “Die byzantinischen Seidenzu ¨ nfte,” BZ 68 (1975): 36–39; A. Muthesius, “The Byzantine Silk Industry: Lopez and Beyond,” Journal of Medieval History 19 (1993): 32. Yet, it would be simplistic to entertain the notion that while a state-mandated guild system was in force, the parallel conduct of commercial activities by the same crafts would be allowed to go on outside the purview and control of the authorities. It would serve no purpose to put in place an organizational structure, detail operational functions, and enact elaborate regulations concerning admission, obligations, and conduct, only to let businesses operate outside the system. Such practice would have caused the guild system to founder. In fact, the practice of the chandler and soapmaker crafts by non-guild members was explicitly prohibited (11.1; 12.6). 23 ¨r Sozial und WirtschaftsL. M. Hartmann, “Zur Geschichte der Zu ¨ nfte im fru ¨ hen Mittelalter,” Zeitschrift fu ¨ckle, Byzantinische Zu ¨nfte, 58; Zoras, Corporazioni Bizantine, 82; and Christophilogeschichte 3 (1894): 24; Sto poulos, jEparciko` n Bibli´on, 50 and n. 5, 51, hypothesize that five persons had to vouch for new entrants even when the law is silent, as stipulated in provisions 2.10, 4.5, 8.13. However, in the case of the raw silk dealers, 6.6 stipulates “para´ tinwn ejnti´mwn kai` crhsi´mwn ajndrw'n,” i.e., by certain (tinwn) men, suggesting that ¨ckle, Byzantinische Zu ¨nfte, 56; Zoras, the number of vouchers could vary. Similarly, it has been argued (Sto Corporazioni Bizantine, 82; Christophilopoulos, jEparciko` n Bibli´on, 50) that the vouchers had to be members of the same guild. This requirement appears only in the case of the merchants of domestically produced silks (4.5), and there is no reason to believe that the law intended it to apply to other guilds as well; if it did, it would have so stated, given the importance of the vouchers’ testimony in making admission decisions. Though guild members probably were not excluded, relying exclusively on member-vouchers would have resulted in the creation of “closed shops” through attrition and forestallment of entry under false pretext. In turn, this would have led to concentration of economic power and privileged entry by relatives of existing members, a situation that would hardly have been tolerated by authorities imbued with an anti-monopoly mind-set and competitive spirit, as reflected in Basilics, 19.18.1 and the Book of the Eparch, which not only set the official industrial and commercial policy but was also an extension of the legislation that sought to restrain monopoly and promote competition (e.g., 18.5, 20.3). 20

20

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

On the other hand, aliens were not allowed to join, not because they were viewed with suspicion by the authorities, as has been suggested,24 but rather to preserve the field for the locals. As in many other instances, the law does not indicate whether a fee had to be paid by new members entering the guild of the fish merchants. In the law’s silence, scholarly opinion on this issue is divided. Some think that, unless stipulated, there was no such obligation;25 however, their arguments do not seem convincing. The assertion that member fees were specified only for the more affluent or prestigious guilds, for example, those involved in silk manufacturing and trade or soap-making, and that guilds involved in the alimentary trades were exempted to avoid burdening their members is hardly plausible, since the law is also silent in the case of the guilds of goldsmiths (2.10) and bankers (3.1), presumably among the more affluent and respected professions. Moreover, the one-time fee was nominal and affordable by anyone intent on setting up shop. More likely, the law’s silence indicates that a customary standard fee of two nomismata generally applied (6.6, 7.3). In the few cases where the fee was explicitly set, it introduced higher charges, usually for manufacturing establishments, of three (8.13), six (4.5), and twelve (12.2) nomismata. Like the members of all other guilds, fish merchants were not allowed to practice any other trade concurrently; they had to choose one and give up the other, informing the eparch of their choice (18.5). Several hypotheses have been advanced concerning the rationale for this ruling. In the context of the fish merchants’ guild, the aim was to foster labor specialization and to facilitate state supervision of trade activities,26 to keep the guilds from banding together and becoming a threat to the state,27 and to ensure public order and political stability.28 Fundamentally, the purpose was to prevent the concentration of economic power in a few hands. By prohibiting firms from expanding into activities conducted by other guilds, whether individually or through partnerships, and thereby effecting a union of trades and horizontal or vertical integration, potential threats to the regime were forestalled. Insistence on strict division of labor among guilds prevented enterprise growth unrelated to market demand, control of the market through formation of a few large enterprises, and weakening or elimination of competition based on sheer market power—in short, monopolization or lessening of competition. The Processing Segment of the Industry The organizational structure of the fishing industry included a post-harvest sector for the processing of unsold fish and marketing of preserved fish. Some of the processing was probably done by fishermen and fish merchants, mainly when the quantities involved 24 ¨ckle, Byzantinische Zu ¨nfte, 59, 119; Zoras, Corporazioni Bizantine, 79–80; Christophilopoulos, jEparciSto ko` n Bibli´on, 52. 25 ¨ckle, Byzantinische Zu ¨nfte, 60–61; Zoras, Corporazioni Bizantine, 86–87; Christophilopoulos, jEparciko` n Sto Bibli´on, 52 and n. 5. Contra, Hartmann, “Geschichte,” 24. 26 ¨ckle, Byzantinische Zu ¨nfte, 98; Zoras, Corporazioni Bizantine, 78; C. M. Macri, L’organisation de l’e´conomie Sto urbaine dans Byzance sous la dynastie mace´doine (Paris, 1925), 46, 58. 27 ´ tude historique sur les corporations professionnelles chez les romains (Louvain, 1895), 1:50, 354 J. P. Waltzing, E ´ conomie urbaine, 41. ¨ckle, Byzantinische Zu ¨nfte, 98; Zoras, Corporazioni Bizantine, 78; Macri, E and n. 3; Sto 28 ´ conomie urbaine, 41. Zoras, Corporazioni Bizantine, 78; Macri, E

GEORGE C. MANIATIS

21

were small and the activity was carried out in their households. At peak harvests, however, when the retail market could not absorb the entire daily catch, large quantities of fish must have been diverted to salting, pickling (in brine or vinegar), or drying. This activity required basic technical, procurement and marketing skills, processing capacity in the form of sheds, equipment and tools, storage space, and working capital for operating expenses. An allusion to the craft of curers (tariceutai´, taricopra´ tissai, taricopw'lai)29 and the export of preserved fish to Italy during Roman times30 strongly suggest the existence of an organized processing activity completely independent of fishing and fish trading. Conceivably, these independent craftsmen might also have been involved in meat preservation, allowing them to utilize the excess capacity that resulted from the seasonality of fish processing. Preserved fish (and meat) were staples, consumed mainly by the masses,31 and the quantities marketed must have been substantial. Ship crews departing from the capital also bought and stored quantities of preserved fish for their voyages.32 But in keeping with the principle of “one man—one trade” (18.5), processors could sell only to grocers (saldama´ rioi), who had exclusive retail distribution rights and were conveniently located in the city (13.1). As was the case in other manufacturing and trade activities, the fact that the processors were not organized into a guild may be attributed to their relatively small number, which did not justify the formation of an independent guild. MARKET ORGANIZATION The Wholesale Market The law directed wholesale fish merchants to make their purchases from fishing boats that put in at designated wharfs and nearby beaches; they were not permitted to buy fish from fishermen at the fishing grounds (oiJ ijcquopra´ tai th` n ejxw´ nhsin poiei´twsan ejn tai'" aijgialoi'" kai` ejn tai'" ska´ lai" ajpo` tw'n katairo´ ntwn ploi´wn, mh` ajnerco´ menoi aujtoi` ejn tai'" ejpocai'" kai` ejn tai'" loipai'" a“grai", 17.3). The reason for the ruling was to prevent the sale of fish in small quantities, that is, retail (wJ" a‘n mh` katakermati´zoito hJ dia´ prasi"). This suggests that not only were retail fish merchants enjoined from purchasing directly from fishermen, but that fishermen themselves were not permitted to retail their catch. In the same vein, provision 53.6.7 of the Basilics directed that fishermen could not be forced to carry their wares to towns and sell them there themselves, but had the legal right to dispose of them through third parties (oiJ aJliei'" oujk ajnagka´ zontai ta` ei“dh fe´ rein eij" ta` " po´ lei" kai` di∆ eJautw'n pipra´ skein, ajll∆ a“deian e“cousin eJte´ roi" aujta` metadido´ nai). The provision prohibits, in effect, local authorities from compelling fishermen to alter their long29 ¨ckle, Byzantinische Zu ¨nfte, 46; J. M. Maspero, Papyrus grecs d’e´poque byzantine (Milan, 1973), 1:53, no. Sto 67023, v. 8; Dagron, “Poissons,” 69. 30 J. Andre´, L’Alimentation et la cuisine `a Rome (Paris, 1961), 113, and the sources cited therein. 31 ¨ckle, Byzantinische Zu ¨nfte, 46; J. W. Nesbitt, “Mechanisms of Agricultural Production on Estates of the Sto Byzantine Praktika” (Ph.D. diss., University of Wisconsin, 1972), 48–49, 57; Dagron, “Poissons,” 69. 32 Nesbitt, “Mechanisms,” 49–50. Given the high level of commercial traffic in the capital, the demand for preserved fish by departing ships and individuals must have been considerable. Freshfield, Byzantine Guilds, ´ conomie urbaine, 42–43; and Runciman, “Byzantine Trade and Industry,” 158, misinterpreting 41; Macri, E provision 17.2, maintain that preserved fish might not be sold to persons leaving the city. Provision 17.2 did not prohibit the export of preserved fish as long as it was processed from unsold surpluses.

22

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

standing practice of selling their catch to intermediaries.33 Clearly, both provisions allude to the existence of a two-tier fish trade and of two distinct markets: a wholesale market at the waterfront34 and a retail one in the urban center of the capital, where the retail fish merchants had stalls in the designated fish markets (oiJ ijcquopra´ tai e“stwsan ejn tai'" legome´ nai" megi´stai" kama´ rai" th'" po´ lew" ajpempolou'nte" tou` " ajgreuome´ nou" ijcqu´ a", 17.1).35 The existence of a separate wholesale market and independently functioning wholesalers is further supported by the fact that this activity required substantial financial resources in the form of working capital for the spot purchase of the fish and, possibly, for the extension of credit to retailers, a form of non-price competition to promote sales and increase market share. This also suggests that a rather limited number of fish merchants had the requisite financial resources to engage in wholesale fish trade, as well as the potential to exercise monopsonistic/monopolistic pricing power. On the other hand, although there were no explicit restrictions, it is unlikely that wholesalers were involved in the retail trade. The fact that a distinct profit margin is stipulated for this particular activity seems to support this view. Indeed, there would be no reason to differentiate between wholesale (17.3) and retail (17.1) profit margins if these activities were undertaken by the same persons—the two margins would have been lumped together. In any case, this natural division of labor, such as was found between sheep merchants and butchers (15.2–5), was also consonant with official policy. Moreover, the low profit margin set for waterfront transactions (0.7%) is suggestive of a large turnover, which could only be handled by entrenched wholesalers. The statutory differentiation of profit margins between operations conducted at the waterfront and those carried on in the fish markets (8.33%) is further indication that retailers did not purchase directly from fishermen but only through wholesalers. If individual purchases were made directly from fishermen by the totality of the guild membership, an internal mechanism for organizing collective purchases and subsequent distribution of the purchased lot among the members would have been required. Yet, the law does not enjoin the fish merchants, as it does in other instances (5.3, 6.8, 9.3), to pool their funds to make collective purchases directly from Middlemen, usually well-known to fishermen, provided an outlet for their catch and eliminated their marketing risks, thereby satisfying their quest for security. 34 Zoras, Corporazioni Bizantine, 179, and Dagron, “Poissons,” 70–71, also interpret provision 17.3 as implying that the purchases from the fishermen were made wholesale. However, they do not distinguish an independently functioning wholesale market per se, and they do not recognize the existence of wholesalers, qua wholesalers, who did all the purchasing as virtual intermediaries and, in turn, sold to retailers adding a stipulated profit margin. 35 R. Janin, Constantinople byzantine: De´veloppement urbain et re´pertoire topographique (Paris, 1964), 98, stretching the evidence (ibid., n. 6), maintains that the fish was sold by shopkeepers located in the various quarters of the capital or by peddlers (“marchands ambulants”) roaming the city. N. Oikonomides, in Hommes d’affaires grec et latins `a Constantinople (XIIIe–XVe sie`cles) (Montreal, 1979), 99 n. 178, and “Entrepreneurs,” in The Byzantines, ed. G. Cavallo (Chicago, 1997), 153, and Dagron, “Poissons,” 69, also subscribe to the presence of “poissonniers ambulants” in the city. K. P. Matschke, “Situation, Organisation und Aktion der Fischer von Konstantinopel und Umgebung in der Byzantinischen Spaetzeit,” BBulg 6 (1980): 286, drawing on 14thcentury sources (ibid., nn. 36, 37) goes even further, asserting that fishermen sold their catch either to fish merchants or they marketed it themselves directly to consumers. In light of the unambiguous formulation of provision 17.1, the mandatory guild organization of the fish trade, and the strict division of labor among crafts, the hypotheses advanced are hardly convincing—at least with the knowledge of the conditions prevalent in the 10th century. 33

GEORGE C. MANIATIS

23

the fishermen. Finally, wholesale operations involving a rather small number of dealers facilitated the chiefs’ task of ascertaining purchase prices of wholesalers (17.3) and retailers (17.1). Indeed, this task would have been almost impossible had the chiefs been required to contact dozens of retailers roaming the waterfront and deal directly and individually with as many fishermen, each striking a deal at a different price for each species of fish.36 Designation of a rather narrowly bounded area along the water’s edge for the operation of the wholesale market37 facilitated the discharge of supervisory and other tasks by state officials (17.1, 4), including collection of the tax levied on the fishermen once they had sold their catch.38 A more subtle purpose was to confine sellers and buyers in a particular locality to form a unified wholesale market and prevent market fragmentation—proximity being the operative medium. This ensured the unfettered function of the market mechanism by forestalling diversion of the catch, stealthy transactions, and the emergence of a black market for scarcer superior quality fish. It instilled competitive pricing behavior, led to the establishment of uniform prices for fishes of similar grades, and curbed attempts to evade tax by concealing part of the catch. In addition, localization of transactions, development of specialized trading, and transaction of business on specified market days increased market efficiency and reduced transaction costs by expediting the information flow regarding the quality of the fish, the size of the catch, and the range of prices.39 Market localization, repeated dealings, common values, and enforceable legal rules promoted personalized exchange between buyers and sellers, which reduced uncertainty40 and kept transaction costs very low.41 Finally, the directive was also prompted by concern to uphold fundamental principles of the guild organization, namely, to ensure equal access to commercial opportunities for all members and curb potential monopsonistic proclivities of entrenched members, who might otherwise be prone to buying up the bulk of the catch before other guild members had a chance to make purchases (forestalling). The Retail Market The law required that each fish market have a chief (prostateu´ onta), whose duty was to ensure that the profit margin of the retail fish merchants not exceed the legal maximum of 8.33 percent, a figure to be ascertained by comparing the wholesale purchase price at the waterfront and the retail price of the fish in the capital’s markets (ejcou´ sh" See also below, p. 28. The location of the wholesale fish market is placed along the Golden Horn and near the Small Gate (Mikra` Pu´ lh): Oikonomide`s, Hommes d’affaires, 97–99 and map on p. 105; Dagron, “Poissons,” 69–70. 38 Apparently, this tax was levied on fishing activities and fishing installations (e.g., biba´ ria). It could take the form of rent on state concessions. See P. Lemerle et al., Actes de Lavra (Paris, 1982), 4:163–64, and (Paris, 1977), 2:162 and passim; R. J. Loenertz, De´me´trius Cydone`s: Correspondance (Vatican City, 1960), 165–66; Matschke, “Situation,” 282. 39 Cf. G. J. Stigler, “The Economics of Information,” Journal of Political Economy 69 (1961): 213, 216; North, Institutions, 27–30, 34. 40 North, Institutions, 22, 25. 41 For instance, costs associated with ascertaining the condition or other attributes of the fish, checking on the background of the agents involved, protecting property rights, or enforcing agreements in the fish trade, financial arrangements, etc. 36 37

24

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

mia'" eJka´ sth" kama´ ra" kai` to` n prostateu´ onta ejpithrou'nta, o”pw" te` kai` hJ ejxw´ nhsi" ejn th' qala´ ssh ge´ gone kai` o”pw" hJ ajpempo´ lhsi" gi´netai, 17.1). To be able to discharge this function, the chiefs had to be present at the consummation of wholesale deals, so that they could be fully informed of the prevailing prices. Also, they had to oversee the orderly conduct of the transactions, ensuring that no part of the catch was withheld from the market and preventing collusive behavior by the parties to the exchange. In addition, every day at daybreak the chiefs had to apprise the eparch42 of the quantity of the white fish caught during the night, so that the sale in the retail fish markets could be effected in accordance with his instructions (i”na kata` to` n tou´ tou prosdiorismo` n hJ ajpempo´ lhsi" pro` " tou` " ejn th' po´ lei gi´nhtai, 17.4). Finally, the chiefs saw to it that no fish was salted unless it remained unsold and ran the risk of spoilage (17.2). On the other hand, the task of ensuring that the fishing boats moored at the piers or put in at the designated beaches and that fish merchants (wholesalers and retailers) did not make purchases directly from the fishermen by going out to the fishing grounds (17.3) was delegated to the coast guard under the paraqalassi´th" and his staff. THE CONTROVERSY OVER PROFIT MARGINS AND PERQUISITES The ambiguous formulation of provisions 17.1 and 17.3 of the law has given rise to various interpretations and controversy. Jules Nicole differentiates between the one chief of the guild (chef de la corporation) and the several chiefs overseeing the fish markets (chefs des halles). In his view, the profit (be´ne´fice) of the chief of the guild amounted to four follis per nomisma (1.4%), an arbitrary increase over the two follis stipulated in the law (17.3), whereas the profit of the chiefs of the fish markets came to one miliaresion per nomisma, or 8.33 percent (17.1). The profit margin of the fish merchants was set at two follis per nomisma, or 0.7 percent (17.3).43 These percentages were based on the prices paid to fishermen.44 Edwin H. Freshfield also distinguishes between the guild master who received (following Nicole) 1.4 percent and the “superintendents” of the fish markets who received 8.33 percent for their services, while the fish merchants earned 0.7 percent.45 Dagron too distinguishes between the chief representing the eparch (prosta´ th") and the overseers posted in each fish market in the city (prostateu´ onte"). The overseers stationed at the fish markets were not representatives of the eparch or remunerated by him, but were head fish merchants (chefs-poissoniers) retained from among guild members by the prefectural authority (l’authorite´ pre´fectorale) to ensure the functioning of each fish market in conformity with established regulations. The discharge of this task entitled the overseers to a stipulated reward that they could not exceed and were to share with their Most likely the assessor (su´ mpono"), a high-ranking official on the eparch’s staff whose duties, inter alia, included adjusting weights and measures to conform with changes (positive or negative) in purchase prices of wheat or wine (18.4, 19.1). 43 The higher denominations of the Byzantine currency and their equivalency were 1 nomisma ⫽ 12 miliaresia ⫽ 24 keratia ⫽ 288 follis. 44 Nicole, Livre du Pre´fet, 67, 68 and n. 1. 45 Freshfield, Byzantine Guilds, xix, xx, 40, 41. Attention to the distinction between two sets of chiefs made ` sto` 12o Aijw'na. Parby Nicole and Freshfield was first drawn by E. Papagianni, “Monacoi` kai` Mau´ rh Agora j athrh´ sei" se` Problh´ mata tou' ∆Eparcikou' Bibli´ou,” Buzantiaka´ 8 (1988): 67–69. 42

GEORGE C. MANIATIS

25

fellow members (associe´s) or workers. The total profit (be´ne´fice) amounted to 9.7 percent and was fixed at two stages: 0.7 percent at the time of the purchase of the fish from the fishermen and applied to all fish merchants in each fish market,46 plus another 0.7 percent for their prosta´ th"… and 8.33 percent for the prostateu´ wn of each fish market at the time of the resale on the retail market. The 8.33 percent made up the commercial profit proper, accruing to the head merchants in each fish market and distributed to the other members, whereas their 1.4 percent supplementary “benefit” defrayed transport costs from the waterfront to the fish markets and the payment of gratuities (sunh´ qeiai) to numerous government officials,47 an expense that was shared equally by the fish merchants and their patron.48 Finally, it has been argued that the chiefs of the fish merchants were entitled to compensation both as members of the guilds and as appointees and representatives of the eparch. When the chiefs took part in the fish trade, they gained 0.7 percent as did all other fish merchants; when they acted as overseers representing the eparch, they collected the special reward of 8.33 percent.49 Though imaginative, these interpretations of the law about the entitlement and allocation of profit margins and perquisites appear strained. Indeed, the formulation of provision 17.1 seems to suggest that the chiefs of the guild of the fish merchants received, as a perquisite for performing their supervisory duties, one miliaresion per nomisma, that is, 8.33 percent, on the realized sales in the capital’s fish markets (w”ste ejn tv' nomi´smati aujto` n ejnapokerdai´nein miliari´sion e”n). Provision 17.3 meanwhile suggests that both fish merchants and their chiefs each earn two follis apiece on the nomisma, that is, 0.7 Dagron, “Poissons,” 71 n. 66, correctly disputes Koder’s amendment of the fish merchants’ profit margin from 2 follis to 2 keratia (from 0.7% to 8.33%) in provision 17.3 in his edition of the law (Eparchenbuch, 128). More precisely, though, 0.7% was the profit margin set for the wholesalers. See below, p. 26. 47 Dagron, “Poissons,” 70, 72, appears to accept uncritically the inflated assertions of a 12th-century letter writer: I. Tzetzes, Epistulae, ed. P. A. M. Leone (Leipzig, 1972), 79–84, no. 57, a source of dubious validity. Tzetzes maintains that the fish merchants were coerced by the eparch’s lieutenants to pay them part of their earnings (meri´da doqh'naiÚ ibid., 81–82). The hyperbolic tone and sweeping pronouncements of the letter, suggesting, inter alia, widespread extortion and implicating high-ranking officials among the eparch’s staff, including the chiefs of the fish merchants’ guild (referred to as ejpithrhtai´) and others not even remotely connected with the fish trade (e.g., prwtokagkela´ rioi [chiefs of bureaus], dome´ stikoi [private secretaries], manda´ tore" [messengers], ejpo´ ptai [inspectors], kai` pa´ sh ta´ xei peri` e“parcon [and everybody in the eparch’s office]), casts serious doubt on the author’s accuracy and fairness. Tzetzes’ disparaging remarks actually refer to gratuities paid to collectors of various taxes and their retinue and to ad hoc inspectors. See G. Rouillard, “Les taxes maritimes et commerciales d’apre`s des actes de Patmos et de Lavra,” Me´langes Charles Diehl (Paris, 1930), 1:284–85, 288–89; N. Oikonomide`s, Fiscalite´ et exemption fiscale `a Byzance (IXe–XIe s.) (Athens, 1996), ¨hbyzantinischen Staates (Munich, 1958), 168, 174. In 77–80, 88–89; J. Karayannopoulos, Das Finanzwesen des fru this context, A. Harvey, Economic Expansion in the Byzantine Empire, 900–1200 (Cambridge, 1989), 106, aptly comments that tax collectors “were always maligned by Byzantine letter writers and reality is clouded by a mass of rhetoric.” In the case of the fish trade, apparently no ad hoc tax was levied on fish merchants, and hence no tax collectors were involved. So, gratuities were limited to customary gifts in kind (fish) on holidays and special occasions, and to officials with direct supervisory functions over the fish markets (probably city inspectors). It would seem therefore that the extent and frequency of the gratuities is exaggerated. On the officials serving in the eparch’s office, see J. B. Bury, The Imperial Administrative System in the Ninth Century (London, 1911), 69–73, 139; R. Guilland, “Etudes sur l’histoire administrative de l’Empire byzantin– L’Eparque, I. L’Eparque de la ville— JO E“parco" th'" Po´ lew",” BSl 41 (1980): 22–26. 48 Dagron, “Poissons,” 71–72. 49 Papagianni, “Parathrh´ sei",” 69–70, 73. Consistent with his interpretation of the stipulated profit, Dagron contests this view: “Poissons,” 71 n. 67. 46

26

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

percent (ajpokerdai´nonte" ªoiJ ijcquopra´ taiº kaq∆ e’n no´ misma ajna` fo´ lei" du´ o kai` oiJ tou´ twn prosta´ tai ajna` fo´ lei" du´ o), based on waterfront prices. Anastasios P. Christophilopoulos,50 concurring with Panayotes Angeletopoulos,51 maintains that the stipulated 8.33 percent award in provision 17.1 pertains both to fish merchants and the chiefs of the guild. In his view, the notion that provision 17.1 establishes the remuneration of the chiefs and provision 17.3 that of the fish merchants is “unacceptable,” on grounds that the latter’s 0.7 percent profit margin is paltry and unrealistic compared to the 8.33 percent received by their chiefs. Besides, he notes, the chiefs’ rate is explicitly set at 0.7 percent in provision 17.3. Christophilopoulos, therefore, correctly suggests that aujto´ n (him) in provision 17.1 should be read aujtou´ " (them), that is, the fish merchants. With this emendation, the gross profit margin of the retail fish merchants becomes 8.33 percent, which makes sense, since they had to cover transport costs and other operating expenses. In fact, if the respective turnover rates are taken into account, it is comparable with those of other retailers, for example, the 4.2 percent (plus the 16.66% allowed for operating costs) earned by bakers (18.1) or the 16.6 percent by grocers (13.5). Yet, it can hardly be accepted that provision 17.3, as emended, refers to perquisites of the chiefs as well. In fact, provision 17.3 sets the profit margin for the wholesale fish merchants, who bought large quantities of fish from the fishermen, and this explains the prescribed small profit margin of 0.7 percent, which is not insignificant when applied to the volume handled. Contrary to common belief,52 the chiefs of the guilds were not, and realistically could not have been, members of their respective guilds, as they could hardly be trusted to be impartial in the discharge of their functions because they had a vested interest. The law nowhere indicates that the chiefs of the professional guilds ought to be members, nor was the eparch directed to select them from among guild members.53 Since all other officials involved in the guilds’ activities were public functionaries, it would be odd if the chiefs of the guilds were the exception. Given the nature of their responsibilities, they had, a fortiori, to be outsiders (or at least former guild members who had given up their trade permanently and could be trusted) and their salaries had to be paid directly by the treasury. In implementing the law, the state employed disinterested civil servants paid by the fiscus rather than self-interested guild members, because only in this way could it ensure that they would fulfil their duties competently and impartially. Therefore, contrary to what has been asserted,54 the chiefs of the fish merchants were Christophilopoulos, jEparciko` n Bibli´on, 62 n. 2. P. Angeletopoulos, “Shmei'a tina` tou' Buzantinou' Dikai´ou,” Dikaiosu´ nh 2 (1924): 72. 52 ¨ckle, Byzantinische Zu ¨nfte, 84; Boak, “Book of the Prefect,” 599; Christophilopoulos, jEparciko` n Sto Bibli´on, 47–48, 49 and n. 1; Oikonomides, “Entrepreneurs,” 155; Papagianni, “Parathrh´ sei",” 69; Matschke, “Situation,” 282; J. Koder, “∆Epagge´ lmata scetika´ me´ to´ n ejpisitismo´ sto´ ∆Eparciko´ Bibli`o,” in ÔH kaqhmerinh´ ´ dieqne´ " sumposio, 1988 (Athzwh´ sto´ Buza´ ntio. Tome´ " kai´ sune´ ceia sth´ n eJllhnistikh´ kai´ rwmai¨kh´ para´ dosh. A ens, 1989), 370–71, maintains that the chiefs of the guilds, though officials of the state, were chosen from among their members. 53 The only exception was the association (su´ llogo") of notaries, who performed quasi-judiciary functions and had to meet special requirements, in particular, knowledge of the law, in order to qualify (1.1 and 1.2). Besides, the association of notaries was not a guild. 54 Nicole, Livre du Pre´fet, 67, 68 and n. 1; Freshfield, Byzantine Guilds, xix, xx; Zoras, Corporazioni Bizantine, 197; Christophilopoulos, jEparciko` n Bibli´on, 62 n. 2; Runciman, “Byzantine Trade and Industry,” 159; Papagianni, “Parathrh´ sei",” 67, 69; Dagron, “Poissons,” 71–72. 50 51

GEORGE C. MANIATIS

27

not remunerated for discharging their assigned supervisory duties, because they were salaried employees of the state.55 However, provision 17.3 allowed the chiefs to accept and share an extra reward totaling 0.7 percent of the market value of the catch as compensation for the unpleasantness, hardship, and inconvenience inherent in the exercise of their duties—they had to be on the waterfront by daybreak every day, regardless of weather conditions, and function in uninviting surroundings. Extra payment for services viewed as being to the advantage of all parties and rendered under taxing conditions or beyond the call of duty were not unusual. The chief of the cattle market inspectors, for example, received four miliaresia for each stolen animal he recovered and, when disputants submitted to his arbitration, he was entitled to a fee of six follis for every nomisma of the disputed amount (21.10). Similarly, the supposed distinction between chiefs of the guild and chief overseers of the fish markets is specious and unwarranted on a number of grounds.56 In the first place, it is not supported by the text of section 17 of the Book of the Eparch, which, if anything, suggests equality among the chiefs. In provision 17.3, the expression kai` oiJ tou´ twn prosta´ tai ªajpokerdai´nonte"º ajna` fo´ lei" du´ o does not imply that the chiefs of the fish merchants were involved in the fish trade, as has been argued,57 since the discharge of the dual function of trader and overseer would not only be impractical but present a conflict of interest. Nor does the provision suggest any division of labor among the chiefs: after being assigned to their posts, the chiefs oversaw both the wholesale and retail fish trade, and also called on the eparch to report on the size of the catch of white fish. Provision 17.1 clearly states that the chiefs oversaw how the purchase in the waterfront and the selling in the markets proceeded (o”pw" te` kai` hJ ejxw´ nhsi" ejn th' qala´ ssh ge´ gone kai` o”pw" hJ ajpempo´ lhsi" gi´netai).58 Such differentiation introduces an unnecessary and unaccounted for new stratum of public officials at a level below the chiefs, something not attested in the law, or in any other source for that matter.59 Interposition of such “superintendents” or “head fish merchants” appointed by the “prefectural authority” is arbitrary and unfounded. There is no compelling reason to make this distinction for the guild of the fish merchants when it is not made for any other guilds that also had more than one chief and operated under similar conditions, including dealers in raw silk (6.4), pork dealers (16.3), and tavernkeepers (19.1), and probably, as has been suggested, for 55 In fact, administrators in Byzantium were paid directly from the treasury: A. Kazhdan and G. Constable, People and Power in Byzantium: An Introduction to Modern Byzantine Studies (Washington, D.C., 1982), 151. 56 ¨ckle, Byzantinische Zu ¨nfte, 79; G. Mickwitz, Die Kartellfunktionen der Zu ¨nfte und ihre Bedeutung bei der Sto ´ conomie urbaine, Entstehung des Zunftwesen (Helsinki, 1936), 220; Zoras, Corporazioni Bizantine, 66–67; Macri, E 72–75; A. Vogt, Basile Ier et la civilisation byzantine `a la fin du IXe`me sie`cle (Paris, 1908), 143, make no such distinction. A contributing factor in making this dichotomy may have been the awkward formulation of provision 17.1: “ejcou´ sh" mia'" eJka´ sth" kama´ ra" kai` to` n prostateu´ onta ejpithrou'nta, o”pw" . . .”; probably it was read as “each fish market having a chief-supervisor.” If the comma is placed after prostateu´ onta, as it should be, the passage reads “each fish market has a chief, who sees to it how . . .” Koder, apparently believing that the comma was misplaced, dropped it altogether from the Greek text: Eparchenbuch, 126. 57 ¨nfte, 84; Papagianni, “Parathrh´ sei",” 69–70; Dagron, “Poissons,” 71–72; Koder, ¨ckle, Byzantinische Zu Sto “∆Epagge´ lmata,” 370–71. 58 Cf. Zoras, Corporazioni Bizantine, 197. 59 E.g., in the Kletorologion of Philotheos (Bury, Imperial Administrative System, 139), which lists the eparch’s supervisory staff.

28

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

the guilds of the perfume dealers (10), candlemakers (11), butchers (15), and bakers (17) on whom the law is silent.60 The existence of lower-level superintendents would obviate the need for more than one chief. Yet, provision 17.4 clearly stipulates that the chiefs (plural) of the fish merchants (oiJ tw'n ijcquopratw'n prostateu´ onte") should report collectively to the eparch on the size of the white fish catch. Because of their low rank in the administrative hierarchy, it is unlikely that such “superintendents” or “head fish merchants” would be authorized to communicate directly with the eparch, in effect supplanting the chiefs of the guild who truly represented the eparch. In Dagron’s profit allocation scheme, wholesale and retail profit margins are combined, although they may well not have accrued cumulatively to the same individuals, since apparently wholesalers did not retail fish. Further, transport costs were borne primarily by the retailers and differed for each of them, whereas gratuities were shared by both wholesalers and retailers unequally, a fact that calls for discrete and individual treatment of these costs.61 Also, it is very unlikely that workers participated in any profit-sharing scheme. Finally, as fish was purchased in designated locations, it would be necessary that all chiefs be familiar with the size and composition of the catch and prevailing wholesale prices in order to be in a position to advise the eparch and make sure that fish merchants did not exceed the prescribed profit margins (17.1, 17.3). A related issue that deserves comment pertains to the punishment meted out for infractions arising from the conduct of the fish trade. The last sentence of provision 17.4 stipulates that whoever contravened the ordinance should be flogged, shorn, and expelled from the guild. It has been argued that the punishment was applied both to the chiefs of the guild and to the fish merchants.62 Apparently, the proponents of this view were misled because the relevant passage was inappropriately included at the end of provision 17.4, which directs the chiefs to report the size of the catch of white fish to the eparch, instead of forming a separate provision 17.5 dealing with all conceivable transgressions alluded to in provisions 17.1, 17.2, and 17.3. There can be no doubt that the tenor of the passage, oiJ de` para` tau'ta tolmw'nte" diapra´ ttesqai, clearly refers to breaches by fish merchants and pertains to all infractions; they were not, and could not be, applicable to the chiefs. Not being members of the guild, the chiefs could not possibly be expelled, while as state officials they were subject to the disciplinary actions provided ¨ckle, Byzantinische Zu ¨nfte, 79. Sto It is not clear from provision 17.1 whether the retailers’ markup was based solely on the wholesalers’ purchase price or if it also included their profit margin, thereby raising the base on which the retailers’ own markup was computed. Judging from the directive that the chiefs ascertain the price paid for the fish at the waterfront (17.1) and the bakers’ rather low profit margin (4.2%), probably the retailers’ markup was estimated on the wholesalers’ purchase price (⫽waterfront price)—an unambiguous formula, suggesting that the retailers had to absorb the wholesalers’ markup. Also, the chiefs’ reward for common service to all fish merchants was shared by wholesalers and retailers, probably on the basis of their financial means. Very likely, the chiefs were paid in cash by the wholesalers from amounts withheld from the retailers at the time of the purchase. Finally, the stipulated profit margins were: (a) maximum, implying that they might not always be attainable due to market conditions; and (b) gross, suggesting that operating expenses, transport costs, contributions to the chiefs, ad hoc gratuities, etc., would also have to be absorbed. According to him, the chiefs were of lowly status and members of the fish merchants’ guild (p. 84). 62 ¨ckle, Byzantinische Zu ¨nfte, 84, 128–29, 133; Papagianni, “Parathrh´ sei",” 66 n. 23. Papagianni mainSto ¨ckle the penalties applied only to chiefs of the guild. Yet, Sto ¨ckle (p. 84) is quite tains that according to Sto emphatic that guild members were also subject to the same penalties. 60 61

GEORGE C. MANIATIS

29

for civil servants in cases of dereliction of duty. Such provisions certainly could not be included in the Book of the Prefect, whose rulings applied strictly to members of the guilds.63 PRICE FORMATION Composition of the Catch The fish markets of Constantinople were supplied by migratory (notably of the tunny and mackerel families),64 semi-migratory, and sedentary species. The Sea of Marmara and the Black Sea were major fishing grounds, but they experienced seasonal migratory patterns, good and bad brood years,65 changeable climatic conditions, and other inherent uncertainties66 that could result in considerable volatility of the annual, seasonal, and daily catch.67 The composition of the catch can be broadly distinguished into superior quality white fish (leukoi` ijcqu´ e"), which included half a dozen or so highly prized species in limited supply and commanding of relatively high prices,68 and common, mostly inferior fish that accounted for the bulk of the catch and fetched considerably lower prices.69 The latter’s relative abundance provided a low-cost diet for the low-income inhabitants of the capital, while the much lower priced preserved fish remained the mainstay of the 63 ´ conomie urbaine, 91–97; Zoras, Corporazioni Cf. Christophilopoulos, jEparciko` n Bibli´on, 63–65; Macri, E Bizantine, 143–47. 64 For the seasonal migratory pattern of such fishes, see A. Koraes, ed., Xenokra´ tou" kai` Galhnou'Ú Peri` th'" ajpo` tw'n ejnu´ drwn trofh'" (Paris, 1814), 65; Dagron, “Poissons,” 57–58. The back and forth migration of fish between the Black Sea and the Sea of Marmara resulted in two intensive harvesting peaks each year, each lasting one to four weeks. Dagron, “Poissons,” 58. 65 Gordon, “Fishery,” 126. 66 As the adage goes, aJliou' [aJlie´ w"] kai` poulopia´ stou a“dhlo" bi´o" (the life of the fisherman and of the bird catcher [hunter] is uncertain), cited by P. Koukoules, Buzantinw'n Bi´o" kai` Politismo´ " (Athens, 1952), 5: 339. In the same vein, . . . ouj pare´ stin ojya´ rion, tw'n aJlie´ wn pantelw'" ajpotuco´ ntwn a“gra" (There was no fish because the fishermen put to sea but returned empty-handed) in D. Papachryssanthou, ed., “Un confesseur du second Iconoclasme: la vie du patrice Nice´tas (836),” TM 3 (1968): 329, chap. 2.2. Cf. also H. Delehaye, ed., Les saints stylites (Brussels, 1923), 212.7–11, 16–18; 213.19–20. Sometimes, fishermen not only failed to catch any fish but also lost their fishing gear in stormy seas: vita of Thomais of Lesbos, AASS, Nov. 4:240, chap. 21.1–8. 67 For the various fishing modes and devices employed in Byzantium, see Koukoules, Buzantinw'n Bi´o", 331–41; Dagron, “Poissons,” 60–64. 68 They included red mullet (sunagri´"), red snapper (ejruqri'no", leqri´ni), grey snapper (sfuri´"), surmullet (fagkri´, sunagri`" hJ mono´ fqalmo"), bass (la´ brax), grey mullet (kestreu´ ", ke´ falo"), mullet (tri´glh, mparmpou'ni) and porgy (tsipou'ra, cruso´ fru" oJ ejpi´cruso"). As these species are found only in the eastern Mediterranean, the English equivalents are proximate. 69 On the several dozens of species found in the area, their qualities and popular regard see Koukoules, Buzantinw'n Bi´o", 79–86; K. Krumbacher, “ÔO Yarolo´ go",” Das Mittelgriechische Fischbuch (Munich, 1903), 361–62; Koraes, Peri` tw'n ejnu´ drwn trofh'", passim; D. C. Hesseling and H. Pernot, Poe`mes prodromiques en grec vulgaire (Amsterdam, 1910), 3:52, vv. 94–96, 99; 55, vv. 152–54, 159–63; 56, vv. 179–80, 204–5; 57, vv. 216 j–l, y; 59, vv. 258–59; 63, vv. 325a, 325c, 325k; 4: 74, vv. 27–28; J. Darrouze`s, ed., Epistoliers byzantins du Xe sie`cle (Paris, 1960), 180–81, letter 19; 199–200, letter 45; 322–23, letter 6; 364–65, letter 29; 366–67, letter 32; R. Romano, ed., Pseudo-Luciano, Timarione (Naples, 1974), 69, vv. 538–42. It is noteworthy, that the scombroid family of tunnies (qu´ nno", qunni´", koli´a", o“rkunqo", phlamu´ ") and mackerel (sko´ mbro") were viewed as the lowest quality fish and were described as stinking (qu´ nna hJ brwmiare´ a): Hesseling and Pernot, Poe`mes, 3:52, v. 99; crummy, lean, and foul (palami´de" potape´ ", sacne` " kai` brwmisme´ ne",): ibid., 3: 57, 2161; heavy (barei'"), constipating (oujk eujsto´ maco"), indigestible (du´ spepto"), flatulent (fusw´ dh"), impure (yafaroi´), and unwholesome (kako´ cumoi): Koraes, Peri` tw'n ejnu´ drwn trofh'", 4, 65–66.

30

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

poorest people all year round. It is noteworthy in this context that most contemporary scholars suppose that the white fish referred to in provision 17.4 were tunnies.70 The notion that white fish ⫽ tunnies is inaccurate, since the description of white fish (⫽white flesh) hardly fits the darkish, low-quality, migratory tunny fish varieties and mackerel available locally and in large quantities.71 Transaction Patterns Given the scores of fish species included in the daily catch, a practical and fair way had to be devised for pricing, computation of profit margins, and exercise of the chiefs’ supervisory task. To this end, and to observe the letter and spirit of the law, it is likely that fishermen divided their catch into white fish and common fish. The former included a rather small number of similar species of superior quality, some belonging to the same family and being close substitutes for each other;72 they were in limited supply and constant demand, fetching similar high prices. The latter accounted for the bulk of the catch and consisted of diverse species, including the highly seasonal migratory varieties (e.g., tunnies). This diverse group was subdivided into species or clusters of species of similar quality that were close substitutes of each other and exhibited minimal, if any, price differential. In this way, the range over which prices had to be negotiated at the waterfront was narrowed significantly, a reasonable basis for calculating the wholesalers’ and retailers’ markups could be conveniently established, and chiefs were able to keep records, compare buying and selling prices, and ensure that profit margins did not exceed the prescribed levels.73 The law did not stipulate that the fish merchants’ guild was required to purchase the entire daily catch; but, to ensure that all landed fish were traded (i.e., actually offered for sale), it provided that only the surplus fish (eij mh` tou` " peritteu´ onta" ªijcqu´ a"º) could be salted to prevent spoilage (17.2). This frustrated efforts to withhold stealthily part of the catch for the purpose of marketing it locally or exporting in processed form in order to maintain or raise prices. The pressure of excess supply, giving rise to surpluses of fish fit for salting and pickling,74 must have been felt particularly during seasonal peaks, Nicole, Livre du Pre´fet, 68 and n. 2; Boak, “Book of the Prefect,” 616; Freshfield, Byzantine Guilds, 41 and ´ conomie urbaine, 74; Dagron, “Poissons,” 70. n. 1; Mickwitz, Kartellfunktionen, 220; Macri, E 71 See notes 68 and 69. Harvey also holds that it was the daily catch of the less common high-quality fish (⫽white fish) that was reported to the eparch: Economic Expansion, 170–71. 72 Products are considered close substitutes if a change in the price of any of them will affect noticeably the quantity purchased of the others. The more close substitutes a family or cluster of species has, the greater the potential for substitution in the face of price changes and the greater the tendency for price uniformity, whether the stimulus comes from the demand or the supply side. 73 There are indications that fish was sold by weight or number for a fixed price: Dagron, “Poissons,” 72–73. However, differences in size, and hence in weight, would hamper sale by number, probably confining this practice to cheaper varieties of fish. Sale by weight was common for bread (18.1.4) and wine (19.1), where price fluctuations were reflected by keeping the price constant and varying the quantity sold. The regulation is founded on the belief that consumers would find it less difficult to accept a smaller portion for ¨ckle, Byzantinische Zu ¨nfte, 100–101. a certain price than to be made to pay more for the same portion: Sto This psychological gimmick introduced a false aura of price stability but had no practical effect different from varying prices for a fixed quantity. 74 Fishes that were cured included sardines (sa´ rdai, qri´ssai), tunnies (phlamu´ ", qu´ nno"), mackerel (sko´ mbo"), swordfish (xifi´a"), and grey mullet (kestreu´ "). See Koukoules, Buzantinw'n Bi´o", 5: 81–2, 84–85 and the sources cited therein; Hesseling and Pernot, Poe`mes, 2: v. 42b; 3: vv. 94, 179. 70

GEORGE C. MANIATIS

31

when the absorptive capacity of the fresh fish market was strained. In these situations fishermen might be forced to make difficult decisions: either reduce the size of the catch or, at least temporarily, accept prices that minimally covered average variable costs (see Fig. 1).75 To be operational, therefore, the notion of surplus must reflect market realities and the decision-making process of the players, as the day market set a price for each fish species (or family of fish) independently of the fishermen’s costs. This price depended on the extent of the processible surplus, the prevailing market (demand and supply) conditions in the fresh and preserved fish markets, and the bargaining power and skills of the buyers and sellers. At times, these prices could be so depressed as to result in negative profits. Indeed, one might expect that processors, even if they operated competitively, would be inclined to offer very low prices, in the knowledge that their refusal to buy would mean total loss for the sellers, thereby driving prices to unprofitable levels. Also, the fact that preserved fish was consumed by the poorest segments of the population and was marketed through the intermediation of the grocers, who added a sizable markup, would tend to depress further the processors’ offer price. To embody these economic circumstances, the notion of “surplus” as applied in provision 17.2 should be understood as meaning unsold fish about to spoil that could not be disposed of at unprofitable prices for any stretch of time. Put differently, fishermen could absorb infrequent but not persistent losses, since in the end they had to fetch prices that enabled them to cover their full costs and earn a normal profit in order to stay in business. Processors procured their fish from both fishermen and fish merchants,76 although their primary source was very likely to have been the fishermen. Wholesalers, being in a better position to assess the proximate quantities the market could absorb each day, would limit their purchases to these estimates, since any additional amounts would remain unsold and would have to be turned over to processors at low prices, thereby cutting into their profits. On the other hand, if the price the wholesalers offered was less than the average variable costs of the most efficient fishermen, say in an attempt to take advantage of larger catches during the migration seasons, while the price the fishermen ex75 In Figure 1, P⫽price⫽average revenue (AR)⫽marginal revenue (MR); total revenue⫽TR; TC⫽total cost; MC⫽marginal cost; ATC⫽average total cost; AVC⫽average variable cost; AFC⫽average fixed cost. The general principle under perfect competitive conditions is that the individual fisherman (like any enterprise owner) as a price-taker will adjust his output so as to maximize his net profit, i.e., the excess of total revenue over total cost. To do so, in the short run he will equate P⫽MC. Thus, if the going price is Pa, the fisherman will supply the quantity OQa of species x and earn a net profit AB ⫻ OQa, which is the difference between TR⫽AD ⫻ OQa and TC⫽BD ⫻ OQa. By the same token, unit profit⫽AB⫽AD(⫽AR)–BD(⫽ATC), where BD(⫽ATC)⫽BC(⫽AFC) ⫹ CD(⫽AVC). If the price of fish fell to Pb, in the short run the fisherman might decide to stay on temporarily supplying OQb, despite the fact that he does not cover his full cost, as long as Pb is higher than AVC. The reason is that at Pb the fisherman minimizes his loss as he covers AVC plus part of AFC. However, at any price below Pc which just covers AVC, say Pd, he cannot cover his current expenses, let alone his fixed costs, and hence he cannot continue in business by making negative profits. It follows that in the short run the fishermen’s supply curve is the portion of the MC curve that lies above the AVC curve (SS’). It is also clear that, in order to remain in business in the long run, the fisherman must equate P⫽MC⫽ATC, i.e., the going price must cover his full cost. (By assumption, ATC includes normal profit.) For details see C. E. Ferguson Microeconomic Theory (Homewood, Ill., 1969), 227–49; L. G. Anderson, The Economics of Fisheries Management (Baltimore, 1986), 56–89; A. W. Stonier and D. C. Hague, A Textbook of Economic Theory (London, 1957), 157–61. 76 Provision 17.2 advisedly specifies those who sell fish (toi'" ajpempolou'si) to indicate that either group could dispose of the surplus (⫽unsold) fish.

Price & Cost S´ MC

P = AR = MR

Pa

A

ATC

B AVC Pb C S

Pc Pd

0

1

Qc

Qb

D Qa

Quantity

Pricing strategies under perfect competitive conditions

Price

S

Db

Da Pb

______________________

Dc Pa

Pc

0

2

______________________

______________________

Qa

Quantity

Pricing and demand in a competitive day market (f ixed supply) environment

Price MSC SRS

Pb SRS Pa

________________________

0

3

________________________

D

Qb

D´ Qa

Quantity

Effect of persistent price levels below average variable costs

Price

S

Dc

Dp Pa

______________________

Pc

______________________

Pb

______________________

0

4

Dc´

Qa

Quantity

Effect of an alternate f loor price (sales to processors) in face of lowered demand

32

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

pected to obtain by selling to processors was at least equal to their average variable costs, the fishermen, understandably, would balk and prefer to sell to processors, rationalizing their action as disposing of a bona fide surplus. Certainly, surpluses must have often emerged at the distribution level as well, as a result of miscalculation of demand by wholesalers and retailers or from unanticipated shifts in demand. In such instances, merchants would find themselves at the end of the day with unsold fish becoming stale and no alternative but to sell to processors at whatever price they could fetch. Thus, aside from preventing the waste of an essential nutritional resource, an important effect of the existence of a processing stage as a separate activity undertaken by independent businessmen was the development of an active market for preserved fish and an attendant price structure for certain varieties of processed fish, which, by siphoning off an important part of the catch that could not be absorbed, alleviated the pressure of excess supply. Finally, intriguing questions of operational significance can be raised concerning the manner in which business transactions were organized and conducted among the parties to the exchange in the wholesale fish market. Being fewer in number, did wholesalers act and bargain collectively, or did they deal on an individual basis with the fishermen? Did they negotiate with major fishermen or a designated spokesman for the fishermen? Did fishermen develop ties with particular wholesalers to whom they tendered their catch and whom over time they came to trust, accepting the price they offered unquestioningly? To what extent did the vulnerability of the fishermen because of the perishable nature of their product affect their bargaining power? Could the numerous and unorganized fishermen forge a common counteractive stance through some cooperative scheme in their dealings with the wholesalers, who were organized and fewer in number? Did the fishermen tacitly recognize a price leader, who would take the initiative to set a price and whose subsequent price changes would be matched by the others? Who took the first step to post a price, the wholesalers or the fishermen? Were prices set on a take-it-orleave-it basis, or was there opportunity for bargaining? Was the fish auctioned at the waterfront? Could wholesalers also take advantage of the numerous and competitive retailers? Evidently, transaction patterns and the ensuing give-and-take affected decision making and price determination. Wholesale Price Determination The fact that fish is a perishable commodity, more so under the conditions prevalent in the tenth century, gave rise to a day market in the sense that all sellers (fishermen, wholesalers, retailers) were compelled to offer their product on the same day and accept any price, since unsold quantities could not be stored. This had far-reaching implications for the functioning of the fresh fish market and price determination. Since the market period lasted a single day and sellers could not adjust the quantities offered, for example, by withdrawing part of the catch, the supply of fish in the aggregate and for each species was fixed and limited to the size of the daily catch. In technical terms, this fundamental feature translated into a perfectly inelastic (vertical straight line) supply curve for each seller and each variety (or family) of fish, meaning that the position (⫽strength) of the demand curve alone determined the price at which the market would be cleared. And, since the fixed supply in a day market ceases to be related to the seller’s cost structure,

GEORGE C. MANIATIS

33

price can fall below average variable costs and sales can result in subnormal and even negative profits (see Fig. 2).77 The only way fishermen could react to persistent price levels that did not cover their average variable costs in the short run78 and their average total (fixed plus variable) costs in the longer run was either by reducing the size of their catch or for some of them to withdraw from the market. Thus, the adjustment ultimately would take the form of a fall-off of the daily quantity of fish offered for sale until prices were raised to levels that would again make it possible for fishermen to operate profitably and stay in business (see Fig. 3).79 As already discussed, the fresh and processed fish markets were connected for certain fish varieties, a link that added a new dimension to and had important implications for pricing in this portion of the catch. When faced with excess supply, fishermen had no alternative but to dump any unsold quantities on the fish-processing market. The price they obtained in this market might vary, depending on demand and supply conditions for processed fish, the quantity of the surplus fish offered for processing, and the bargaining power and negotiating skills of the parties to the exchange. The fact that unsold fresh fish would ultimately go to waste could have weakened the negotiating position of the fishermen (or the fish merchants, for that matter), possibly pushing prices below even average variable costs and setting in motion the fishermen’s defense mechanism—cutting back on the fish supply. Thus, the demand for low-value fish about to spoil and that could be utilized only in processed form effectively set a minimum below which the price of fresh fish in the day market could not fall (see Fig. 4).80 77 In the day market period and under competitive conditions, fishermen cannot adjust their output and hence the supply of fish is fixed at OQa as seen in Figure 2. The price at which the market will be cleared is determined, therefore, by demand alone, while supply alone determines the market equilibrium quantity. Thus, if the day market demand is given by Da, the market equilibrium price will be Pa, where demand equals supply. If demand is greater, say Db, the price would also be greater, Pb, but the market equilibrium quantity would be the same as before because the supply is fixed. Similarly, if demand was less, Dc, the price would be lower, Pc, while the same quantity would be marketed. For details see Ferguson, Microeconomic Theory, 226–27; Stonier and Hague, Textbook of Economic Theory, 153, 156–57. 78 Depending on individual circumstances, a price below average variable cost could mean that a fisherman might not be able to earn even a subsistence wage. This can also happen to inefficient (i.e., high-cost) fishermen at prices even above average variable cost. 79 In Figure 3, the original day market equilibrium is achieved at price Pa and quantity OQa, shown by the intersection of the day market supply curve MSC and the demand curve DD⬘. However, if this price is unprofitable, fishermen will adjust to the new conditions by restricting their catch. The short-run supply curve SRS reflects the downward adjustment in supply as fishermen react to prevailing unprofitable prices. At any price below Pb no fish will be supplied, whereas at progressively higher prices more will be offered, and vice versa. In this example, quantity OQb is supplied at price Pb as indicated by the intersection of the demand curve DD⬘ and the short-run supply curve SRS, which is higher than the original day market price Pa. Longer term adjustments will be affected by the variability of the factors of production, the presence or absence of external economies and diseconomies, and the imperative that fishermen cover their full costs. Hence, the long-run supply curve may be flat, rising, or even falling. It is worth noting that, strictly speaking, the terms “short run” and “long run” are not definite periods of time but sets of operative conditions and, in consequence, there can be no sharp or exact distinction between the two. Whether conceived as sets of conditions or as time periods, the two merge. For details on this last point, see D. S. Watson, Price Theory and Its Uses (New York, 1968), 190. 80 In Figure 4, the day market equilibrium price is Pa and the quantity of fish supplied OQa, corresponding to the intersection of consumer demand Dc and supply curve S. If consumer demand drops to Dc’, exhausting the absorptive capacity of the market and lowering price to Pb, while the processors’ demand Dp was higher resulting in price Pc greater than Pb, it is evident that the fishermen would rather sell to proces-

34

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

Gunnar Mickwitz argues that the fish merchants (he does not differentiate between wholesalers and retailers) had to abide by regulations obliging them to make purchases collectively by forming a cartel. Prices, therefore, were set cooperatively and competition among guild members was avoided; this enabled them to exercise monopsonistic purchasing power over the fishermen and acquire fish at lower than competitive prices. By suggesting this pricing pattern—and in a sweeping generalization—Mickwitz groups the fish merchants with the importers of raw silk, Syrian silks or linen.81 Yet, the law does not even hint that the fish merchants had a statutory obligation to act in concert in their purchases, whereas it does so stipulate in the case of the textile importers (6.8, 5.3, 9.3). Moreover, while squeezing external textile suppliers may have been an acceptable trade policy, taking advantage of local fishermen would have run counter to anti-monopoly legislation and the preamble of the Book of the Eparch, which stresses that the parties to commercial transactions should not brazenly exploit one another, that the financially stronger should not do injustice to the weaker, that no person should oppress his fellow man, and that all actions should be governed by the rule of law. It follows that actions on the part of fish merchants that were damaging to fishermen would certainly have prompted official intervention. To be sure, protection of consumers’ interests was a very high priority of the eparch. Nonetheless, it is unlikely that he would decline to protect the interests of toiling suppliers of a basic staple in the face of unfair competitive practices. Provision 17.3 leaves no doubt that the state did not intervene in the wholesale phase of fish trade and that fish prices at the waterfront were determined by market forces. Yet, the fact remains that wholesalers were a rather small group, were organized into a guild, and presumably had the tacit support of the retailers, while the fishermen were many and unorganized. Prima facie, this market structure suggests that wholesalers could, potentially, negotiate from a position of strength and exercise monopsonistic buying power by acting in concert and driving all-or-nothing deals. Their position may have been further strengthened because the perishability of the fish diminished the fishermen’s maneuverability and bargaining power. Nevertheless, the potency of institutional arrangements, the instability of collusive schemes, and the threat of new entry suggest that attempts by wholesalers to influence purchasing prices were probably intermittent and short-lived at best, as underlying market fundamentals tended to foster competitive behavior among wholesalers. Fishermen and wholesalers had antipodal interests: the former had every incentive to aim for the highest possible prices, the latter at the lowest. Wholesalers would counteract the fishermen’s demands, since their profitability, given their fixed and narrow markup, depended on sales volume. Ultimately, volume was a function of the quantity marketed and the price prevailing at the retail resale market, a variable that, in turn, depended on the strength and elasticity of consumer demand. Disaggregated demand for fish was elastic, even for common fish varieties despite their low prices, because such purchases accounted for a significant proportion of family incomes and many very close sors. Therefore, the price offered by processors, Pc, in this example, establishes a floor below which the price of fresh fish in the day market cannot fall. 81 Mickwitz, Kartellfunktionen, 216, 223.

GEORGE C. MANIATIS

35

substitutes were available. This meant that lower prices resulted in greater quantities sold and greater wholesaler profit. An initial price might be announced at the waterfront by either party to the exchange, but most likely it was set by the group that could more readily predict price, either because they were better informed or ostensibly wielded greater bargaining power—wholesalers. In quoting prices, wholesalers might have acted in concert, an action open to them because their relatively small number enabled them to behave as oligopsonists. Fishermen might accept the posted price, particularly if their marketing skills were inadequate, or if they were weary of bargaining or had developed good and longstanding relations with individual wholesalers. Varying cost structures and different perceptions about operative price levels would have made cooperation among fishermen difficult. However, they might counter-offer, possibly through an entrenched fisherman or a tacitly acknowledged price leader, and through bargaining ultimately arrive at a price accepted by all. Or, fishermen might feel that auctioning their fish in designated locations would yield better results by inviting the wholesalers to bid openly— which might or might not be a truly competitive process, if the wholesalers acted in collusion. Legally, however, price declines had to be the outcome of the free play of market forces, the interplay of demand and supply. Concerted action by the wholesalers to depress prices could, therefore, be construed by the authorities as a conspiracy to monopolize the market in violation of anti-monopoly legislation. Nevertheless, an oligopsony enjoying stable membership and experience with price changes over a long period can result in such predictable patterns of response that it leads effectively to an unspoken understanding that all rivals observe and accept. Such a pattern could be one in which all wholesalers tacitly recognize a price leader who assumes the responsibility to set a lower price than would otherwise obtain and whose price changes are promptly matched by the others. In this instance, the oligopsony implicitly becomes a quasi-unified group of wholesalers appearing to take advantage of the many small fishermen who supply them competitively. Yet, as long as no collusive agreement is made, such practice is not unlawful. Parallel business behavior based on recognized interdependence of rival price policies that lead to price uniformity cannot be construed as actual or tacit agreement.82 Furthermore, sometimes the authorities might have found it difficult to ascertain to what extent a price decline was due to excess supply and in what measure to monopsonistic buying pressure. To the extent that a small number of entrenched fishermen controlling a substantial portion of the daily catch engaged in a contest with the wholesalers, a bilateral oligopoly would emerge. As the contesting groups can exercise control over the price and thereby render the pricing mechanism inoperative, price becomes indeterminate within a wide range, and bargaining power and negotiating skills come to be determining factors. The price, therefore, can fall in either extreme, or uncertainly between the two extremes. If neither group dominates, there is a good chance that prices will gravitate toward competM. Singer, Antitrust Economics: Legal Cases and Economic Models (Englewood Cliffs, N.J., 1968), 111–13; J. S. Bain, Industrial Organization (New York, 1968), 309–10, 312. 82

36

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

itive levels. It is noteworthy that the emergence of a bilateral oligopoly may not necessarily be the product of collusive behavior; as long as there was rivalry or imperfect collusion on either side of the wholesale fish market, the behavior was not culpable and the outcome cannot be viewed suspiciously.83 More often than not, however, effective exercise of monopsony (monopoly) power in concentrated market structures requires collusive action. In this setting, the wholesaler guild members must develop enforceable schemes to restrict purchases, have the ability to set prices and discipline fractious members, and buy the acquiescence of the overseeing chiefs—conditions difficult to put in place. What is more, collusive arrangements typically tend to be imperfect and fragile. Wholesalers may not be able to concur on the price to be quoted or, eager to increase their own market share, may not adhere to the set buying price, bidding it up in order to secure additional supplies at the expense of their competitors. Clearly, such action undermines the cohesiveness of the group, as it forces competitors to follow suit, matching bids and raising prices. Also, an aggressive pricing policy, forcing weaker fishermen out of the market or to reduce the size of their catch, could have adverse effects on the wholesalers’ prospective profitability. A price squeeze on fishermen would be in violation of anti-monopoly legislation and bound to attract the attention of authorities. Furthermore, the very short trading period at the waterfront and the need for spot decisions narrowed considerably the window of opportunity and imposed a severe limit on the wholesalers’ ability to posture and reach consensus, rendering concerted action less likely. Finally, since there were no legal barriers to entry and since economic barriers were not unsurmountably high, wholesalers were under the constant threat of new entrants—a concern that tended to dilute the effectiveness of their potential monopsony power. Under such circumstances, the force of monopsonistic buying power is attenuated and prices tend to move toward competitive levels. A further potent deterrent of collusive business behavior, by wholesalers, retailers, and fishermen alike, was anti-monopoly legislation. Monopolization of a commodity was strictly prohibited. Individual actions as well as express or tacit agreements among groups aiming to drive prices down through predatory tactics or maintain prices at levels above those dictated by market forces were illegal and explicitly forbidden in fish transactions. Violators were subject to steep fines and so were members of the eparch’s staff if they failed to enforce the law (Book of the Eparch, 18.5; Basilics, 19.18.1; Synopsis Basilicorum, 24.1). Detection of conspiratory actions would have prompted the authorities to intervene either ipso jure or at the request of the aggrieved party. It seems, therefore, that it would have been difficult to put into practice, let alone enforce, unfair methods of competition under the watchful eye of the state appointed chiefs and their informers. Finally, there is a lingering perception among contemporary scholars that guilds effectively limited competition among their members.84 However, in the fish trade the fact that the wholesalers had a statutory buying prerogative as a group did not confer ipso J. S. Bain, Pricing, Distribution, and Employment (New York, 1953), 394–96, 432–36. Mickwitz, Kartellfunktionen, 216, 223, 228, 229, 234; Lopez, “Silk Industry,” 18; Oikonomides, “Entrepreneurs,” 156. 83 84

GEORGE C. MANIATIS

37

facto exercisable monopsonistic pricing power on individual guild members, since the wholesalers did not act on command, in unison, or in compliance with internal regulations entailing disciplinary action for non-compliance. The law did not impose price discipline on guild members to thwart intra-guild competition and protect an individual member’s share in the total business or to ensure equality of economic results. Rather, the emphasis was on fostering equality of opportunity—equal access to the daily catch. In reality, wholesalers were competing among themselves for market share, which meant acquiring large quantities of fish, as this was the only way to enhance profitability in the face of fixed and small profit margins. In sum, the unified and closely supervised market85 fostered a give-and-take environment that helped forge divergent quotations into uniform prices for the various families of fish and for fish of similar grades. A confluence of factors was conducive to an unencumbered interplay of market forces, and monopsonistic or monopolistic pricing conduct, to the extent that it was practiced and went unnoticed, probably was the exception and not the rule. Institutional arrangements, the unworkability of collusive arrangements, unimpeded entry, short trading periods, and self-interest attenuated the exercisable monopsony or monopoly power of the parties to the exchange, fostered a competitive rather than collusive attitude in inter-buyer and inter-seller relations, and allowed the free play of market forces to steer wholesale fish prices toward competitive levels. In a highly localized market, price differences become known almost instantaneously, prompting swift adjustments that result in price uniformity. On the other hand, while the preserved fish market set a rock-bottom price for a significant part of the daily catch by providing an outlet for unsold fish, the fishermen’s most effective way to stem falling prices, when market conditions or unfair competitive action compelled them to sell at unprofitable price levels, was to reduce the size of the catch on subsequent market days. Retail Price Determination Retail prices, reflecting patterns of consumer demand,86 influenced wholesale prices;87 but actual daily spot prices at the waterfront were also affected by the quantities supplied. Given that wholesalers were few and encountered a large number of competitively buying retailers, potentially, at least, they were in a position to exercise monopolistic pricing power over the latter. By acting in concert, they could raise the base price on which their profit margins were calculated. Practically, however, this was not possible, as the chiefs were directed to ensure that the retailers’ markup did not exceed the legally prescribed maximum, and to do this they had to compare wholesale purchase prices with See above, p. 23. Consumer demand for fish was a function of fish prices, buyer income and its distribution, preferences, prices of other commodities, particularly of close substitutes, and religious custom. Consumption of blooded fish was forbidden to the religious during Lent and on Lenten days (some 150 days of the calendar year in all), when the laity substituted bloodless fish (shellfish): Nesbitt, “Mechanisms,” 58–59. Koukoules argues that the Byzantines’ “love” for fish was so great that they did not abstain from fish consumption even during fast days despite the Church’s prohibition: Buzantinw'n Bi´o", 79. Low income families may have had no alternative for the relatively inexpensive staple. 87 See above, pp. 29–30, 30–31, 34–35. 85 86

38

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

asking retail prices. Transparency in the wholesalers’ transactions was no less a matter of concern. Such determinations, however, presupposed full knowledge of prevailing waterfront prices and hence the physical presence of the chiefs at the closing of wholesale deals,88 a fact that clearly precluded falsification of prices by the wholesalers. Attempts by individual wholesalers to reduce prices and accept lower profit margins to expand their sales were bound to be frustrated as well. Given their small number and the high degree of market localization, the rest of the group would immediately match the new price, eliminating inter-seller price differences. Wholesalers had no incentive to take such action, since they were aware in advance that their competitors would respond without delay to their price moves. Similarly, if the retailers’ demand turned out to be less than what the wholesalers anticipated and the latter were compelled to reduce prices, recognized price interdependence among rivals would prompt adjustments that would wipe out any price differential and establish price uniformity. Market localization, the small number of wholesalers, recognized price interdependence, and the institutional structure left little margin for price differences in the marketplace, and this effectively limited the search for competitive prices on the part of the retailers. It is therefore likely, at least until they had cause to believe otherwise, that retailers usually accepted the wholesalers’ uniform quotations as representing truly competitive waterfront prices plus their fixed markup, particularly when wholesalers had a reputation for fair dealing and long-standing relationships had been established. Correspondingly, retail prices were also uniform for each species or grade of fish, reflecting waterfront prices plus the retailers’ profit margin. Whether posted retail prices were always actually attained depended on the strength of the day-market demand, which might at times force retailers to reduce prices and accept lower profit margins, the more so since at the end of the day unsold fish would have to be disposed of to processors at much lower prices. The risk of unprofitable sales would be minimized in subsequent market days by adjusting purchases in light of the most recent experience. White Fish: Markup versus Price Setting The eparch’s highly unusual daily intervention to determine the terms of the retail sale of white fish only (17.4) raises several intriguing issues: why was the eparch’s intervention limited only to white fish, when for all other fish a uniform markup was stipulated statutorily? Did he fix price(s) or profit margin(s)? If the latter, why did the law not set the same (8.33%) profit margin for all species? Did the eparch fix a uniform markup for all white fish species, or did he set individual rates? Dagron maintains that, because of the importance of the white fish in the diet of the capital’s residents, the authorities sought to fix their price to mitigate seasonal variations in supply,89 a view held by others as well.90 Against this view, one should note first that, See above, pp. 21–24. Dagron, “Poissons,” 70 and n. 65, 71. On the misunderstanding of what species are referred to as “white fish,” see above, p. 30 and notes 68, 69, 71. 90 ¨ckle, Byzantinische Zu ¨nfte, 80–81, 102; Nicole, Livre du Pre´fet, 68; Freshfield, Byzantine Guilds, 41; Zoras, Sto Corporazioni Bizantine, 71, 197; Harvey, Economic Expansion, 170–71; G. I. Bratianu, “La question de l’approvisionnment de Constantinople `a l’e´poque byzantine et ottomane,” Byzantion 5 (1929–30): 91–92; Koder, “∆Epagge´ lmata,” 370. 88 89

GEORGE C. MANIATIS

39

in conformity with official policy applied to staple articles, the eparch was bound to fix profit margins on white fish retail sales and not their price.91 If the intention was to fix prices, the eparch would have intervened and set wholesale prices, which were the bedrock for establishing retail prices; yet, he was not involved in this crucial stage of price formation. Second, to ensure a degree of fairness, prices must be set in consideration of the position and shape of shifting demand curves for different species, as well as differing cost structures for individual fish merchants—information central authorities would have difficulty in obtaining. Third, proximate retail prices were already reflected in the prevailing waterfront wholesale prices, and this obviated the need for outright pricefixing; establishing markups was far more convenient. Fourth, establishing administratively the price of a perishable commodity in limited and volatile supply entailed the risk of pricing the white fish out of the market and causing artificial shortages and waste, or of cutting unnecessarily and unfairly into retailers’ earnings. On the other hand, fixing maximum profit margins effectively allowed market forces to determine the selling price, while it introduced a degree of flexibility into the system that met the authorities’ main objectives: to ensure fair prices, avoid waste, and prevent profiteering. While a ceiling was set on the selling price when the prevailing market price tended to exceed waterfront prices plus the retailers’ stipulated markup, a weaker market demand would force retailers to accept lower markups, break even, or even suffer a loss on that market day in order to dispose of their perishable wares. The profit reduction suffered in such instances was probably marginal and infrequent, as experience and the daily nature of the trade allowed for price adjustments on subsequent days. Fifth, though intrusive, profit capping is much less disruptive of the functioning of the market mechanism than outright pricefixing. Finally, fixed profit margins would not have translated into equal economic results for members of the fish merchants’ guild, as individual net profits varied depending on the volume of sales, running costs, and overall efficiency of operations. Rather, the practice fostered a competitive spirit, as it provided an incentive to capture a larger share of the market to enhance profitability, because an increase in sales not only broadened the basis on which markups were calculated but also reduced the unit cost of sales by spreading overhead costs over more units. The fixing of ad hoc profit margins by the eparch in the case of the white fish, in contrast to the uniform markups fixed by law for all other species, was probably prompted by the high price differential between white fish and other fish categories. White fish’s superior quality, limited and volatile supply, and high price elasticity of demand92 resulted in wide price fluctuations, a circumstance that called for varying markThe law fixed profit margins on bread (18.1), fish (17.1), foodstuffs such as preserved meat, salted fish, legumes, olive oil, cheese, etc., and on earthenware, glassware, and hardware (13.1). 92 Although the demand for fish as a generic commodity can be fairly inelastic, the demand for a particular species can be very elastic, implying that a given fall in price causes a relatively larger increase in the amount bought, and vice versa. This is because the more narrowly and specifically fish is defined the more substitute varieties it has, and the more elastic is the demand for each variety. White fish species had a high price elasticity of demand because at high prices demand is generally elastic; most superior quality, high priced white fish species were very close substitutes; and these fish represented a relatively high percentage in the buyer’s budget. And the more close substitutes a species has and the greater its importance in the buyer’s budget the more elastic is the demand for it. Hence, even small changes in the available quantity could cause substantial price volatility. 91

40

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

ups and explains the frequency of reporting. Applying the going markup (8.33%) would have yielded much higher profit margins for white fish sales, which would have been unfair. The guiding principle in fixing markups for white fish species was to set, each time, a figure that would make profit margins commensurate with that earned on common fish varieties. Also, the fact that the law was preoccupied with the size of the white fish catch indicates that few species were involved, compared to the host of ordinary fish varieties, whose sheer numbers made setting variable markups impractical. The relative abundance throughout the year of tunnies, mackerel, and other ordinary fish,93 measures ensuring that the entire catch be marketed, and competition from close substitutes and preserved fish, tended to check appreciable upward price movements for these species at times of shortage,94 to reduce the amplitude of seasonal price fluctuations, and to assure the lowest possible prices. Any attempt, as suggested above, to fix lower prices in low-yield seasons would have backfired and resulted in even lower catches and attendant tendency for prices to rise further. On the other hand, other factors would have tended to stem significant price declines and contributed to price stability. An appreciable fall in price would enable some families to buy more, while lower income groups who could not previously afford fresh fish would enter the market, thereby raising the demand for and the prices of these species. At the same time, fishermen, at least the more entrenched, might try to reduce their catch to prop up prices. These processes would have been reinforced, since the demand for preserved fish absorbed part of the migratory fish varieties, setting minimum prices and narrowing significantly the amplitude of seasonal price oscillations. Understandably, the authorities’ focus was on establishing profit margins for white fish (⫽white flesh, high quality), which was subject to wider price fluctuations all year round. This explains why the chiefs were directed to report to the eparch daily, not seasonally, and only on the size of the white fish catch. CONCLUDING OBSERVATIONS Commercial fishing was a critical segment of the fish industry, and measures were taken to ensure full utilization, sustainable development, and efficient management of important fishing grounds, such as those used for fixed-net fishing. The industry itself 93 The lack of concern for the size of the common fish catch and the statutory establishment of the retailers’ markup seem to imply that fish were caught in fairly large quantities, even in otherwise lean times, and that their prices therefore tended to be relatively low. 94 Inferior quality fish (fresh or salted) was a chief staple for very low income families in the capital. When prices were high, they spent a large part of their budget on this article and were forced to curtail their consumption of other more expensive foodstuffs. Since fish was the cheapest food, they would buy more of it, not less, even though now its price was higher. This paradox (“Giffen effect”) results in an unconventional upward sloping demand curve, and fish in this instance is an “inferior good.” The implication is that in such circumstances a strong market demand involving too many households will reinforce the upward price spiral. However, such occurrences have been rare, e.g., bread in England, potatoes in Ireland during the 19th century, and rice in less developed Asian countries in recent times. The reason is that, even if some people view fish as an inferior good, others may not, and when less is purchased, or even when buyers leave the market altogether, the upward pressure on price is relieved, and so the market demand curve (the horizontal summation of all individual demand curves) will be normal (downward sloping)—at a higher price less fish is bought. On the notion of the Giffen effect, see Stonier and Hague, Textbook of Economic Theory, 63–65, 68–70; Watson, Price Theory, 93.

GEORGE C. MANIATIS

41

was organized into several independent sectors with distinct markets and structural characteristics: a large competitive harvesting sector composed mainly of small unorganized fishermen; a quasi-oligopsonistic/oligopolistic wholesale trade sector involving a limited number of entrenched fish merchants; a competitive retail sector; and a rather concentrated surplus fish processing sector. The fact that only the fish merchants, wholesalers and retailers, were organized into a guild, and that wholesalers and processors were small and potentially influential groups, while the fishermen formed a large unorganized cluster, heightened the importance of competition and the dynamics of inter-buyer, interseller, and seller-buyer interaction. The perishability of fish gave rise to a day market, where demand alone became the price determining factor, while the existence of a processing sector alleviated the pressure of excess supply for certain quantitatively important fish varieties by absorbing unsold surpluses. Realistic interpretation of the legal provision suggests that unsold fish were disposed of at prices that, though at times might result in negative profits, in the long run would not distress fishermen and would enable them to stay in business. Thus, the link between the fresh and processed fish markets had significant implications for price determination, as the sizable demand for preserved fish and the alternative it offered for diverting unsold fish effectively set a floor price in the fresh fish day market for a considerable part of the daily catch, particularly of inferior quality varieties at peak harvests. The law leaves no doubt that prices at the waterfront for all fish species were determined by market forces. These wholesale prices formed the basis for the calculation of wholesaler and retail profit margins, which were fixed statutorily on a permanent basis. State officials interfered with the terms of retail sales only in the case of the high-quality white fish, which was in limited supply, commanded high prices, and experienced price volatility; but even then they only set comparable profit margins, not prices, in line with established policy for necessities. The maximum retail price was essentially the waterfront price plus the retailers’ markup. Whether posted wholesale and retail prices were actually attained depended on the strength of the day-market demand and the extent, apparently real, of inter-seller and inter-buyer competition. Monopsonistic pricing at the wholesale level, to the extent that it was practiced collusively and remained unnoticed was probably not sustainable. The fragility of collaborative schemes, unimpeded new entry, self-interest, short trading periods, and antimonopoly legislation, all helped shape an antagonistic rather than cooperative attitude in inter-buyer relations, as wholesalers vied to increase their turnover and expand their market share. Such market conduct was conducive to a rather unrestrained interplay of market forces, driving prices toward competitive levels. Recognized interdependence of price policies created a strong tendency toward price uniformity among rivals. Furthermore, the wholesalers’ buying prerogative did not ipso facto confer exercisable monopsonistic pricing power on individual guild members, or on retailers for that matter, as the wholesalers did not act on command, in unison, or under threat of disciplinary action for non-compliance. The law did not impose price discipline on the guild members to stifle intra-guild competition and protect an individual member’s market share or ensure equality of economic results. Instead, the emphasis was on fostering equality of opportunity. The same arguments can be adduced to substantiate the inefficacy of the potential monopsonistic buying power of the processors as well. All in all, the players at every

42

FISH MARKET IN TENTH-CENTURY CONSTANTINOPLE

market level were competing against each other, and such competitive spirit within the unified and closely supervised wholesale and retail markets fostered a give-and-take that moved divergent quotes toward uniform prices for the various families and grades of fish. Bethesda, Md.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Naissance d’une vocation: aux sources de la carrie`re byzantine de Thomas Whittemore RE´MI LABRUSSE AND NADIA PODZEMSKAIA

INTRODUCTION: PORTRAIT DE THOMAS WHITTEMORE EN BYZANTINISTE VOYAGEUR homas Whittemore (1871–1950) est toujours demeure´ une personnalite´ ´etrange: admire´ pour l’inlassable activite´ qu’il a de´ploye´e, `a partir de 1930, en faveur des ´etudes byzantines, il n’en a pas moins suscite´ les re´serves de la plupart des byzantinistes qui l’ont ˆtoye´ pour le caracte`re approximatif de son activite´ scientifique. Avant tout homme co d’action, il s’est pourtant passionne´ment engage´ dans un domaine auste`re, la byzantino` ´etait de mise une rigoureuse ´erudition. Et, dans son activite´, s’il a toujours su logie, ou ˆ t certain du secret mettre en valeur avec brio les re´sultats qu’il obtenait, il a cultive´ un gou sur les moyens pour y parvenir. De meˆme pour son personnage: il s’est prote´ge´, en quelque sorte, derrie`re une brillante fac¸ade d’homme du monde et s’est tre`s peu exprime´, en de´finitive, sur lui-meˆme, sur ses motivations et sur les ´ele´ments fondateurs de son action. Aujourd’hui, comment aborder sa vie? Peut-on lever un coin du voile et trouver des cle´s ´ tats-Unis, l’Europe et l’Orient, en faveur ˆle qu’il a joue´, entre les E pour interpre´ter le ro du de´veloppement des ´etudes byzantines? Faut-il, pour y parvenir, le conside´rer avant ˆle de byzantiniste officiel, fondateur du Byzantine Institute of America tout dans son ro dont Dumbarton Oaks est actuellement l’he´ritier, et, surtout, responsable de la remise au jour et de la restauration des mosaı¨ques de Hagia Sophia `a Istanbul? Ou faut-il au contraire s’attacher `a son personnage de voyageur et d’infatigable mondain, meˆle´ `a la fois `a l’art et `a la politique? Du premier aspect, il suffit de rappeler les ´ele´ments essentiels, bien connus de tous

T

La re´daction de cet article a ne´cessite´ la consultation de plusieurs fonds d’archives, dont les principaux sont indique´s, en notes, de la manie`re suivante: AM: Archives Matisse, Paris. AD: Archives Duthuit, Paris. BAR/CERYE: Bakhmetieff Archives, Committee for the Education of Russian Youth in Exile, Rare Book and Manuscript Library, Columbia University, New York. BB: Bibliothe`que byzantine (fonds Thomas Whittemore, Institut byzantin), Paris. DO: Dumbarton Oaks, Washington, D.C. FK: Fonds prive´ Kondakov, Muse´e litte´raire, Prague. ISGM: Isabella Stewart Gardner Museum, Boston. Nous remercions, pour leur aide et leurs conseils, Mme Ekaterina Basargina, le Pr. Gilbert Dagron, Mlle Anna-Claude Delay, M. Claude Duthuit, Mrs. Vivian Endicott Barnett, le Pr. Andre´ Guillou, Mme Kopecka, Mme le Pr. Angeliki Laı¨ou, le Pr. William L. MacDonald, le Pr. Cyril Mango, Mme Catherine Piganiol, Mrs. Ellen Scaruffi, Mrs. Susan Sinclair, le Pr. Jean-Pierre Sodini, le Pr. Jean-Michel Spieser, Mme Natalia Teteriatnikov, le Pr. Vladimir Vavrinek, et Mme Zahradnikova.

44

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

les byzantinistes:1 Whittemore a su persuader Mustafa Kemal d’autoriser la transformation d’Ayasofya Camii en muse´e, par de´cret pre´sidentiel de mars 1934, apre`s quatre cent quatre-vingt un ans de conse´cration au culte musulman, et d’y rendre `a nouveau visibles les mosaı¨ques restaure´es par les soins du Byzantine Institute. En tant que directeur de cet institut, il a appose´ sa signature sur la couverture de ces jalons importants des ´etudes byzantines que sont les quatre volumes de rapports de campagne consacre´s `a la restaura` partir de 1934, il a participe´ `a tous les congre`s internationaux ´ glise.2 A tion de la Grande E d’e´tudes byzantines. Et c’est encore lui qui, en 1948, peu avant de mourir, a permis `a Paul Underwood de lancer les premiers travaux `a l’e´glise du Christ de la Chora, alias Kariye Camii, destine´e aussi `a ˆetre transforme´e en muse´e; lui, enfin, qui est `a l’origine de la constitution des fonds documentaires et bibliographiques de la Bibliothe`que byzantine, `a Paris, et de Dumbarton Oaks, `a Washington. ` propos du second aspect, il n’est pas inutile de mentionner brie`vement quelques A te´moignages, tant a ´ete´ remarquable la passion de Whittemore `a fre´quenter, dans le ˆ ts personnels, tout ce que le monde international comptait de brillant. cadre de ses gou ` Boston, de`s le de´but du sie`cle, il se lie avec Isabella Stewart Gardner et Charles Eliot A ` Berlin, en 1908, il rend hommage `a Henri Matisse et le rencontre peu apre`s Norton. A 3 ` `a Paris. A Oxford, ou ˆte´s de Paul Morand, qui ` il ne fait que passer, on le trouve aux co 4 ` lui de´die son premier roman, ´ecrit en 1910. A Paris encore, il est l’ami de Gertrude ` Vienne, en 1913, Stein, qui le met en sce`ne dans une pie`ce ´ecrite en 1913, IIIIIIIIII.5 A ` Moscou, il fre´quente Josef Strzygowski, qui le conduit dans l’atelier de Gustav Klimt.6 A 7 ` Sofia, en 1918, il est introduit aupre`s du patriarche de l’e´glise russe, Tikhon (Fig. 1). A en 1924, il s’entretient de la philosophie de Henri Bergson avec le roi de Bulgarie.8 En Voir notamment N. B. Teteriatnikov, Mosaics of Hagia Sophia, Istanbul: The Fossati Restoration and the Work of the Byzantine Institute (Washington, D.C., 1998), 26 sq.; G. Constable, “Dumbarton Oaks and Byzantine Field Work,” DOP 37 (1983), 171–76; W. L. MacDonald, “Whittemore, Thomas,” Dictionary of American Biography, suppl. 4 (1946–50) (New York, 1974), 890–91. 2 The Mosaics of St. Sophia at Istanbul: Preliminary Report on the First Year’s Work, 1931–1932. The Mosaics of the Narthex (Paris, 1933); Second Preliminary Report: Work Done in 1933 and 1934. The Mosaics of the Southern Vestibule (Paris, 1936); Third Preliminary Report: Work Done in 1935–1938. The Imperial Portraits of the South Gallery (Boston, 1942); Fourth Preliminary Report: Work Done in 1934–1938. The Deesis Panel of the South Gallery (Boston, 1952). 3 Sur ces deux points, voir ci-apre`s. 4 Paul Morand, Les Extravagants: sce`nes de la vie de bohe`me cosmopolite, pre´sentation et notes de Vincent Giroud (Paris, 1986). Ce texte de jeunesse ´etait demeure´ ine´dit; il est de´die´ “a` Thomas Whittemore,” et porte cette citation de Gilles de Caen (XIIe s.) en exergue: “On les nomme `a juste titre extravagants ceux-la` qui vagabondent hors des chemins fraye´s et des coutumes ´etablies” (p. 17). 5 Il y apparaıˆt sous les initiales “T——S WH——.” Voir James R. Mellow, Charmed Circle: Gertrude Stein & Company (New York–Washington, D.C., 1974), 186. Voir ´egalement L’Autobiographie d’Alice B. Toklas (Paris, 1934), 158. 6 “I have been all day with Strzygowski and tomorrow we go together to visit Klimt,” Whittemore `a Isabella Stewart Gardner, de Vienne, le 17 octobre 1913, ISGM. 7 En 1912, l’e´veˆque de Londres lui remet une lettre de recommandation pour le me´tropolite de Moscou (BB). 8 “[Whittemore] told me that his friend, the Emperor of Bulgaria, continually reads Bergson; that one of his books lies always at his side; that he finds in his thought constant aid. Anyhow we find Bergson in Bulgaria.” Matthew Prichard `a Isabella Stewart Gardner, le 5 juillet 1924 (ISGM). Le 21 juin 1924, `a 16 h. 30, Thomas Whittemore a ´ete´ rec¸u par le roi de Bulgarie en audience. Lettre de la meˆme date d’un conseiller aulique, DO. 1

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

45

´ thiopie, le 2 novembre 1930, il assiste, en compagnie d’Evelyn Waugh, au couronneE ment de l’empereur et est rec¸u par lui, pour lui pre´senter ses recherches sur le christianisme.9 En Roumanie, en 1935, la princesse Bibesco lui offre l’hospitalite´ et tire de ses ` Londres, en 1946, il flatte conversations avec lui un roman ne´o-byzantin extravagant.10 A les inclinations artistiques d’une princesse irakienne en lui faisant remarquer que sa pein` New York, en 1944, il ravive l’inte´reˆt de Fernand Le´ger ture “approuve Matisse.” 11 A 12 ` pour l’art byzantin. A Istanbul, en 1948, il fait de meˆme avec Le Corbusier, auquel il entrouvre la porte de Hagia Sophia pour lui re´ve´ler l’e´tat d’avancement de ses travaux.13 ´ tat, `a WashEt lorsqu’il meurt, le 8 juin 1950, c’est dans les couloirs du De´partement d’E 14 ington, alors qu’il allait rendre visite `a John Foster Dulles. Comment ces deux versants ont-ils pu, de fait, s’allier au cours de la longue carrie`re de Whittemore? Comment cette personnalite´ `a l’e´vidence multiple a-t-elle ´ete´ conduite `a consacrer tant d’efforts `a un domaine traditionnellement re´serve´ `a l’e´rudition la plus discre`te? Nous proposons, pour tenter d’y re´pondre, d’emprunter deux voies qui, chaDu 16 octobre au 29 novembre 1930, Whittemore se rend de Londres `a Djibouti et `a Addis-Abbeba pour le couronnement de l’empereur d’Ethiopie, Haı¨le´ Se´lassie´, qui a lieu le 2 novembre. Il est muni d’une lettre d’introduction (signature illisible, `a en-teˆte du muse´e d’histoire naturelle de Chicago): “Oserai-je m’adresser `a votre Majeste´ dans le but de pre´senter mon ami Mr. Thomas Whittemore, savant ame´ricain distingue´, s’occupant d’e´tudes sur l’arche´ologie et l’ancienne ´epoque de l’histoire chre´tienne. Peut-eˆtre vous plaira-t-il de le recevoir et d’eˆtre entretenu par lui sur la nature et l’importance de ses recherches” (BB). Voir ´egalement ` Evelyn Waugh, en 1931, raconte sa participation `a ces meˆmes ´eve´nements et ´evoque Remote People, le texte ou longuement sa rencontre avec le “professeur W.” (paru en franc¸ais sous le titre Hiver africain [Paris, 1992], notamment 67–111). 10 Voir la lettre de la princesse Marthe Bibesco `a Whittemore, le 2 octobre 1935: “Be well assured that I will write nothing, or better publish nothing—as I have written much about you in my ‘memento’—until you tell me to do so. I hope above all hopes that you will be able to fulfill your promise and come to stay with me at Mogosoe¨a end of October or beginning of November (even better). All is ready for you to visit the painted convents of Bucovina, with every possibility of serious investigations” (BB). Son roman Theodora, le cadeau de l’Orient, finalement publie´ en 1953 `a Paris, est de´die´ au “professeur Thomas Whittemore.” L’avant-propos du livre est constitue´ d’une conversation fictive avec ce dernier et se conclut ainsi: “C’est ainsi que l’inspiration m’est venue de composer ce re´cit sur place, pendant mes se´jours `a Constantinople, `a la suite de mes conversations avec le professeur Whittemore que j’allais voir re´gulie`rement, chaque anne´e, au cours des travaux re´ve´lateurs qu’il poursuivait `a Sainte-Sophie” (p. XIV). 11 La lettre que celle-ci lui adresse en retour ne manque pas de piquant: “Cher Professeur, . . . J’avais tellement aime´ cette phrase `a propos de ma peinture: ‘Elle approuve Matisse.’ . . . Personne n’aurait dit excepte´ vous—une imminence artistique [sic]—la seule chose qui aurait ´ete´ dite dans ce cas. Merci de l’avoir compris car j’ai justement aime´ chez lui ce qu’il est venu prendre chez nous.” Lettre de la princesse “Ze´id el ˆtel Ritz, le 10 de´cembre 1946, BB. Hussein of Irak” de Londres, ho 12 En 1944, Whittemore organise au Metropolitan Museum of Art de New York une exposition de ses ` cette occasion, travaux `a Hagia Sophia (sous forme de photographies, de moulages et de releve´s sur toile). A Fernand Le´ger, qui vivait `a l’e´poque `a New York, re´dige un article enthousiaste (“Byzantine Mosaics and Modern Art,” Magazine of Art [avril 1944], 144–45) qui est `a la fois un ´eloge de l’art non-mime´tique des Byzantins et un plaidoyer pour l’usage moderne de la mosaı¨que. 13 Le 4 octobre 1948, Whittemore fait visiter Hagia Sophia et Kariye Camii `a Le Corbusier, qui s’inte´resse aux mesures de ces ´edifices. Voir Le Modulor: essai sur une mesure harmonique `a l’e´chelle humaine applicable universellement `a l’architecture et `a la me´canique (Paris, 1950), 196: “L’avion me de´posait `a Istamboul le 3 octobre ` son ´equipe dernier. Le lendemain le Professeur Wittemore [sic] me faisait les honneurs de Sainte-Sophie ou de jeunes arche´ologues recherche les mosaı¨ques enfouies depuis des sie`cles sous le badigeon.” L’exemplaire du livre conserve´ `a la Bibliothe`que byzantine est de´dicace´ “pour le Prof. Wittemore [sic] en souvenir d’une si bre`ve et fructueuse entrevue `a Sainte-Sophie de Stamboul, avec ma plus vive sympathie, avril 1950.” 14 Article ne´crologique du Boston Globe, le 8 juin 1950. 9

46

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

cune, apportent un ´eclairage diffe´rent sur le cheminement de sa carrie`re et contribuent, ce faisant, `a expliquer la singulie`re gene`se de sa vocation byzantine. I. MATTHEW STEWART PRICHARD, UN MAITRE SOCRATIQUE Premie`res anne´es bostoniennes Pour comprendre, `a ses origines, le lien de Thomas Whittemore avec les ´etudes byzantines, il faut remonter aux toutes premie`res anne´es de sa formation, dans les cercles d’esthe`tes bostoniens du tournant du sie`cle. Certes, c’est la litte´rature anglaise, et non ˆme´ de Tufts l’arche´ologie de la Me´diterrane´e orientale, que le jeune homme, sorti diplo College en 1894, enseigne dans la meˆme universite´, `a partir de l’anne´e universitaire 1895–96. Ce dandy de´licat, fe´ru de poe´sie, appre´cie´ pour la qualite´ de ses lectures publiques de Rudyard Kipling, noue alors des relations d’e´le´gante mondanite´ avec tout ce que Boston comptait `a l’e´poque de personnalite´s riches et intellectuelles. De´ja`, cependant, son activite´ est tourne´e vers les arts plastiques.15 Plus pre´cise´ment, ses pre´occupations le conduisent `a s’inte´resser, dans les premie`res anne´es du sie`cle, `a la question de la diffe´rence entre arts mineurs et arts majeurs, ainsi qu’au proble`me des rapports entre art et liturgie, qu’il ´evoque avec les milieux catholiques franciscains de Boston, dont il est proche.16 C’est aussi par ce biais qu’il de´veloppe avec Isabella Stewart Gardner (1840–1924) des liens qui vont rapidement se re´ve´ler de´cisifs, en donnant `a son existence une orientation intellectuelle et esthe´tique dont le fil conducteur se retrouve tout au long de sa ` Fenway Court, le palais ve´nitien reconstitue´ de la ce´le`bre collectionneuse, carrie`re. A transforme´ en muse´e prive´ `a compter du 1er janvier 1903, Whittemore a pu entrer en contact avec un cercle qui, entoure´ d’œuvres rares et prestigieuses, re´cemment acquises ˆt ou en cours d’acquisition, s’efforc¸ait d’allier une re´flexion esthe´tique originale `a un gou ` la meˆme ´epoque, le Museum of Fine Arts, pour les images d’un ´eclectisme raffine´. A situe´ `a quelques dizaines de me`tres de Fenway Court, ´etait en pleine expansion, engage´ lui aussi dans une re´flexion complexe—et conflictuelle—sur la pre´sentation des œuvres d’art, la finalite´ des collections, l’organisation et les missions d’un muse´e des beaux-arts.17 Les deux institutions rivalisaient, donnant ainsi un tour d’autant plus aigu `a leurs efforts pour de´finir leurs de´marches respectives. C’est dans ce contexte, chez Isabella Stewart Gardner dont ils ´etaient les intimes, que Whittemore a rencontre´ Okakura Kakuzo ¯ (1862–1913) (Fig. 2) et, surtout, Matthew Stewart Prichard (1865–1936) (Fig. 3), tous deux membres de l’e´quipe du Museum of Fine Arts. En 1902–3, il donne une se´rie de quatre confe´rences sur les sujets suivants: “Greek Sculpture,” “The Development of Gothic Architecture,” “English Cathedrals,” et “Canterbury Cathedral” (DO). 16 En 1906, Whittemore donne une confe´rence intitule´e “‘Minor Arts’ Lecture,” dont le texte dactylogra` , contestant l’appellation d’art mineur, il rephie´ est conserve´ au Isabella Stewart Gardner Museum, et ou grette l’e´poque me´die´vale, en Occident, lorsque cette distinction n’avait pas encore cours: “Strictly speaking, perhaps, we could say that there are no minor arts. There is art turned, let us say, into less important spaces, into minor channels. . . . It would seem impossible for our very best men to equal in interest and artistic grasp most of the work done in mediaeval cathedrals, merely to fill in spaces with something ornamental” (p. 1, ISGM). Le texte se poursuit par une longue ´evocation des conceptions japonaises de l’art. 17 Voir W. M. Whitehill, Museum of Fine Arts, Boston: A Centennial History, 2 vols. (Cambridge, Mass., 1970), notamment “The Battle of the Casts,” 1:172–217. 15

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

47

´ le`ve d’Ernest Fenollosa, grand de´fenseur des valeurs esthe´tiques du Japon tradiE tionnel, face au mouvement de modernisation occidentale qui caracte´risait le Japon de l’e`re Meiji, Okakura avait ´ete´ appele´ en 1904 par le muse´e de Boston pour en organiser la collection d’objets d’Extreˆme-Orient.18 Entre avril 1904 et sa mort, en septembre 1913, il effectua ainsi de fre´quentes alle´es et venues entre Boston, le Japon, la Chine et l’Inde,19 afin d’acque´rir des œuvres pour le compte du muse´e, qui lui doit encore aujourd’hui ` une ´epoque ou ` l’art asiaune part essentielle de ses collections extreˆme-orientales.20 A tique ´etait encore, tre`s souvent, aborde´ comme un exemple parmi d’autres d’exotisme, pour des raisons de curiosite´ ethnographique plus que pour une re´elle appre´ciation esthe´tique, le muse´e de Boston, en partie sous l’influence d’Okakura, s’est place´ `a l’avantgarde d’un nouveau rapport avec ces formes d’art, rapport fonde´ sur la volonte´ d’engager un ve´ritable dialogue entre traditions: “In conclusion,” ´ecrit par exemple Okakura, en 1905, dans le Bulletin du muse´e, “I wish to allude to the importance of giving the public greater opportunities for approaching the real meaning of Asiatic art. Universal as the spirit of art is, its forms must differ as so many expressions of different ideals of race and philosophies of life. Japanese and Chinese arts require to be interpreted from within like European art, and their productions are to be treated neither as curiosities nor phantasies, except from the inattentive.” 21 Par lui, qu’il fre´quentait `a Fenway Court et avec lequel il ´etait en correspondance,22 Whittemore a donc pu avoir la re´ve´lation de l’importance et de la cohe´rence interne d’esthe´tiques non-occidentales qui, en tournant leur ´energie vers les arts dits mineurs, avaient instaure´ une articulation entre l’art et la vie tout `a fait ´etrange`re `a celle qui avait pre´valu en Europe depuis la Renaissance. Whittemore disciple de Prichard? Mais, davantage encore que d’Okakura, c’est d’un de ses amis, Anglais temporaire´ tats-Unis, que Whittemore a tire´ le meilleur de ses intuitions esthe´tiques. ment exile´ aux E Matthew Stewart Prichard avait commence´ `a travailler dans le domaine artistique en rejoignant en 1892 le cercle des amis d’Edward Perry Warren, `a Lewes House, dans le Sussex23 (Fig. 4). Pendant dix ans, il avait contribue´ `a constituer la collection d’antiques 18 Voir P. Chalfin, “Japanese Art in Boston,” The Burlington Magazine ( janvier 1906), 220–22. Sur Okakura Kakuzo ¯, voir Yasuko Horioka, The Life of Kakuzo¯, Author of The Book of Tea ([Tokyo], 1963); E. Grilli, “Okakura Kakuzo ¯: A Biographical Sketch,” en Okakura Kakuzo ¯, The Book of Tea (Rutland, Vt.–Tokyo, 1995), 119–33; Christopher Benfey, “Tea with Okakura,” New York Review of Books, 25 mai 2000, 43–47; et Okakura Kakuzo ¯, Collected English Writings, 3 vols. (Tokyo, 1984). 19 ` il a rendu Okakura a se´journe´ en Inde entre novembre 1901 et octobre 1902, notamment `a Calcutta ou visite `a Rabindranath Tagore. 20 Voir Saeko Yamawaki et al., ´edd., Okakura Tenshin and the Museum of Fine Arts, Boston (Boston-Nagoya, 1999). 21 “Japanese and Chinese Paintings in the Museum,” Bulletin of the Museum of Fine Arts [Boston] ( janvier 1905), 6, repris in Okakura Kakuzo ¯, Collected English Writings, 2:85. Okakura, a surtout publie´ en 1906 le livre qui l’a rendu mondialement ce´le`bre, ve´ritable manuel d’esthe´tique orientale applique´e, `a l’usage de ` l’on retrouve exprime´es les meˆmes ide´es. l’Occident moderne, The Book of Tea (New York, 1906), ou 22 ˆ t 1911, de France (Provins), en route pour l’Egypte, Whittemore ´ecrit `a Isabella Stewart Le 25 aou Gardner: “Two letters, one from you and one from Okakura, his written in Seattle the day before he sailed for Japan” (ISGM). La correspondance entre les deux hommes n’a malheureusement pas ´ete´ conserve´e. 23 Voir D. Sox, Bachelors of Art: Edward Perry Warren and the Lewes House Brotherhood (Londres, 1991).

48

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

de ce riche amateur ame´ricain, et `a organiser la vie de la communaute´ de Lewes House, fonde´e sur une sorte d’ultra-esthe´tisme `a connotations nietzsche´ennes. En 1902, Pri´ tats-Unis, `a Boston, sur l’invitation du fre`re d’Edward chard quitte l’Angleterre pour les E Perry Warren, Samuel Dennis Warren, qui venait d’eˆtre ´elu pre´sident du muse´e des beaux-arts. Il va travailler dans la nouvelle ´equipe du muse´e entre 1902 et 1906, date `a ` partir de 1908, il s’installe `a Paris, ou ` il re´side jusqu’en 1914 laquelle il de´missionne.24 A avant de passer la guerre comme prisonnier anglais en Allemagne et de se re´installer de´finitivement `a Londres, en 1918.25 Pendant ses cinq anne´es de se´jour `a Boston, en lien avec Okakura et avec Isabella Stewart Gardner, Prichard a formule´ pour la premie`re fois la grande conviction qui a oriente´ toute sa pense´e: que les arts orientaux sont intrinse`quement supe´rieurs `a la tradi` ils font de l’art un moyen et pas une fin. tion europe´enne classique, dans la mesure ou ` A partir de la`, `a l’occasion d’un voyage en Italie en 1906, il de´couvre dans l’art byzantin le paradigme ide´al, `a la charnie`re entre Orient et Occident, susceptible selon lui de redonner vitalite´ et ´elan `a l’art europe´en: de meˆme, selon Prichard, que la civilisation byzantine avait su se libe´rer de l’emprise de l’art mime´tique gre´co-romain, de meˆme l’art moderne europe´en devrait pouvoir s’e´manciper des principes the´oriques mis en place `a la Renaissance, en restaurant un lien authentiquement ve´cu entre l’art et l’existence quotidienne, `a l’oppose´ de l’attitude de contemplation gratuite propose´e par les chefs d’œuvre des maıˆtres classiques, et impose´e par les muse´es qui les recueillent, les exposent et en organisent le culte, `a l’abri de la vie re´elle. Prichard se de´finit d’emble´e comme un philosophe et non comme un historien d’art: en ce´le´brant l’art oriental, en de´nonc¸ant avec violence la Renaissance occidentale et la pratique des muse´es, il veut donner `a sa pense´e une dimension prospective, et non pas seulement arche´ologique. Et c’est `a Paris qu’il trouve finalement la pense´e philosophique et l’œuvre artistique susceptibles de re´ˆte´, et l’œuvre de Matisse de pondre `a ses aspirations: la philosophie de Bergson d’un co l’autre (Fig. 5). Byzance, Bergson et Matisse vont donc constituer a` partir de 1909 la triade sacre´e de ce personnage singulier, presque totalement de´pourvu d’œuvre publie´e,26 mais doue´ d’un intense pouvoir de se´duction intellectuelle, qui s’est exerce´ sur le petit cercle de disciples qu’il s’est toujours attache´ `a rassembler autour de lui. La pense´e de Prichard a exerce´ sur Whittemore une influence profonde et durable: la volonte´, en particulier, de restaurer un lien fort et dynamique entre l’art et la vie, et l’approche des œuvres artistiques en tant qu’e´lan global de cre´ation collective, en lien avec la socie´te´ dans son ensemble, ont imme´diatement se´duit le jeune Ame´ricain. Ces positions lui permettaient en effet d’articuler l’un avec l’autre son esthe´tisme exacerbe´, d’une part, et son pragmatisme fondamental, son puissant besoin d’action, d’autre part. Pour autant, Whittemore faisait-il partie du cercle des “prichardiens” stricto sensu? Pas exactement. Il est certain que les deux hommes ont ´ete´ en contact `a Boston puisque, de`s Voir W. M. Whitehill, “Some Correspondence of Matthew Stewart Prichard and Isabella Gardner,” Fenway Court 1974 (Boston, 1975), 14–29. 25 Pour des indications biographiques plus de´taille´es concernant Prichard, voir R. Labrusse, “Matthew Stewart Prichard,” New Dictionary of National Biography (Oxford, `a paraıˆtre). 26 Sa seule publication importante est, `a compte d’auteur, le texte d’une confe´rence prononce´e `a la Taylorian Institution d’Oxford en de´cembre 1919: Matthew Stewart Prichard, Greek and Byzantine Art (Londres, 1921). En revanche, Prichard ´etait un ´epistolier extreˆmement prolixe. 24

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

49

1910, de passage `a Paris, l’Ame´ricain va rendre visite `a son ami anglais, hospitalise´ apre`s avoir contracte´ une fie`vre scarlatine: “Mr. Whittemore whom I think you have met,” ´ecrit Prichard `a Isabella Stewart Gardner, le 15 juin 1910, “had come to Paris the very day a doctor had examined me, and in spite of the risks of infection he stayed by me with great courage to the last moment.”27 Un an plus tard, lors d’un nouveau passage, Whittemore se charge de donner des nouvelles d’Isabella Stewart Gardner `a Prichard: “I have heard wonderful things about you lately, for Professor Whittemore was in Paris for thirty hours and saw me.” 28 Pourtant, Prichard ne reconnaıˆtra jamais vraiment `a Whittemore une valeur intellectuelle digne d’estime. Sa mondanite´ l’agace, de meˆme que la superficialite´ de ses jugements, et plus encore sa tendance `a reprendre `a son propre compte des ide´es ´emises par les autres. De`s juillet 1911, apre`s une rencontre `a Londres, il s’en plaint de´ja` `a Isabella Stewart Gardner: “In London, . . . I saw Whittemore, too, who feels always a little as if the seed had fallen among thorns and that he is moving it to good soil.” 29 Pre`s de quinze ans plus tard, en 1925, son opinion n’a gue`re change´ puisqu’il ´ecrit `a l’un de ses plus proches disciples, Georges Duthuit (1891–1974) (lequel venait d’e´pouser la fille de Matisse et se pre´parait `a mener une double carrie`re de critique d’art contemporain et de byzantiniste): “Whittemore was on his way to America where he went last week. He knows many persons of wealth in America and has a shrewd knowledge of things, but his first thought is not to compromise himself by sharing unorthodox views or acknowledging heterodox friends. You can see him when he is again in Paris but must not count on much help from him. . . . Had he lived at the time, he would have dined constantly with Dives, have called regularly on Pilate, have had conversations with John the Baptist and been in the front row at the Crucifixion. As far as I know, Americans are unable to understand or absorb such a notion as dynamic art; they have not the gift of the necessary yielding, attention or passion.” 30 En de´finitive, dans les jugements de Prichard `a l’e´gard de Whittemore, affleure presque toujours cette irritation—d’autant plus aigue¨ qu’elle concerne un homme dont ˆ ts et les ide´es sont proches des siens. Son pragmatisme “ame´ricain” l’exaspe`re;31 les gou l’universalite´ de ses fre´quentations lui semble coupable; enfin, son enthousiasme de “by` il lui paraıˆt fonde´ sur une pense´e apzantiniste” ne´ophyte l’inquie`te dans la mesure ou 32 proximative. ISGM. ISGM. 29 ISGM. 30 ˆ t 1925, AD. Lettre du 5 aou 31 Prichard ´etait revenu de Boston avec un vif sentiment anti-ame´ricain. Le 10 de´cembre 1913, il conclut le re´cit, `a Georges Duthuit, de sa rencontre avec un jeune peintre ame´ricain `a Paris par cette remarque: “All Americans go in the direction of matter which with painters often takes the road of talking” (AD). 32 ` un long passage est consacre´ `a Voir la lettre de Prichard `a Georges Duthuit du 14 mai 1914, ou Whittemore: 27 28

Whittemore has turned up, more like a sucking pig than ever, endowed with perpetual youth. He is as ever inclined to listen, at first, afterwards he mounts his high horse. He has tried ambassadors on me—what Briand said to him—but I am adamant, I tell him that they are wicked people. He has brought some Byzantine coins, and has showed me the copper ones of which many are quite wonderful. He took me to a new restaurant: I have quite a list of new restaurants to which to take you. . . . He has just left a yacht and gives a few days to Paris before travelling on an archaeological

50

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

Le travail `a Hagia Sophia Pourtant, apre`s la fondation du Byzantine Institute of America, si la personnalite´ de Whittemore semble toujours insignifiante `a Prichard, le sens historique de sa destine´e, en revanche, a pris `a ses yeux une importance croissante. D’abord, il se fe´licite de son engagement officiel en faveur des ´etudes byzantines, ainsi que du lien ´etabli entre ces travaux arche´ologiques et l’art contemporain, tel qu’incarne´ par Matisse: “As a lecturer,” e´crit-il `a Duthuit le 29 fe´vrier 1928, “in N.Y. Univ. on Byzantine Art, [Whittemore] introduces the work of Matisse; and in the Near East he is in touch with Byzantine things and authorities. In the U.S., Maclagan of the V.&A. declared that without knowing him it would be little use for an authority on art to visit that country, for he knows all that world.” 33 Plus ge´ne´ralement, Prichard voit dans la passion byzantine de quelques-uns de ses disciples une occasion pre´cieuse de re´volutionner les valeurs esthe´tiques et, `a partir de ˆte´s de la`, les relations humaines en Occident. Pour ce faire, il place Whittemore aux co Duthuit, qui organise en 1931 la premie`re exposition internationale d’art byzantin, dans les salles du palais du Louvre `a Paris,34 et de Royall Tyler (1883–1953), qui ´etait depuis les anne´es dix un collectionneur enthousiaste d’art byzantin et un ami de Matisse:35 “Is it not strange that three men whom you know as well as Royall Tyler, Whittemore and myself should all from different angles be concerned with a type of artistic expression which offends all the academic canons? That means revolt. But there are more massive signs than that. The churches are empty—of men, at least. The universities are finding that modern languages are supplanting the classics; I think marriage as an institution is tottering, and the home is threatened. The war gave the knock-out blow, but all was ready to overthrow nineteenth century institutions.” 36 ˆ t-il marginal—et une position re´volutionnaire conCe lien entre le byzantinisme—fu journey in Bulgaria with a great savant. But in spite of my best efforts his archaeology is very weak. His coming to me pleases me of course, and I am entirely at his order, to read his ms., to tell him of his subjects, and to give him moral dicta. I was very frank to him about America, his listening to these tirades is the price he pays for things which are useful to him, but when I told him that I doubted if [Richard] Strauss’ music was worth a straw he promptly turned the conversation, for has he not already sworn allegiance to Strauss personally? (AD). 33 ` partir du ler septembre 1928, Whittemore est ´elu assistant professor of fine arts `a la New York AD. A University. Cette ´election avait ´ete´ envisage´e de`s 1926, mais n’avait pas abouti pour des raisons budge´taires. Le 31 janvier 1927, le directeur du de´partement, le ge´ne´ral Charles S. Sherril, ´ecrivait au candidat: “We are hoping that you will be able to give us a course in Byzantine Art which Dr. Shapley discussed with you last year. It should be given at the Metropolitan Museum of Art” (BB). Eric Maclagan ´etait le directeur du Victoria and Albert Museum et a travaille´ `a l’e´poque pour quelques mois `a Harvard. Il ´etait l’un des proches de Prichard `a Londres. 34 Exposition internationale d’art byzantin (Paris, muse´e des Arts de´coratifs, Palais du Louvre, pavillon de Marsan, 28 mai–9 juillet 1931). Duthuit ´etait secre´taire ge´ne´ral de l’exposition. 35 Tyler, riche Bostonien ´eduque´ `a Oxford, avait commence´ `a former sa collection d’art byzantin avant 1914, `a Paris, notamment chez Joseph Brummer. Ami de Prichard et de Matisse, il a mene´ paralle`lement une carrie`re de diplomate international, en particulier en Hongrie, et de spe´cialiste de l’art byzantin. Il figure parmi les plus importants preˆteurs de l’exposition de 1931, et est l’auteur, avec Hayford Peirce, de L’Art byzantin, 2 vols. (Paris, 1932–34) (1e`re ´edition en un volume: Byzantine Art [Londres, 1926]). Il a conseille´ Mildred et Robert Woods Bliss dans la constitution de leur collection, le´gue´e `a Harvard University en 1940 (voir D. Sutton, “An Oasis of Scholarship,” Apollo 119 [avril 1984], 2–6). Son fils, William Tyler, a ´ete´ directeur de Dumbarton Oaks entre 1969 et 1977. 36 Prichard `a Mrs. Montgomery Sears, le 20 janvier 1933, BB. Voir ´egalement cette lettre de Prichard `a Una Pope-Hennessy, de Londres, le 27 janvier 1932: “Whittemore, Royall Tyler, Duthuit have all been in town, and all my faculties have been engaged” (ISGM).

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

51

corde avec la tournure de plus en plus utopique—voire ´esote´rique—prise par la pense´e de Prichard qui, dans les dernie`res anne´es de sa vie, relie avec une obstination presque fascinante les changements radicaux subis par les socie´te´s occidentales, d’une part, et le de´veloppement des ´etudes byzantines, d’autre part. Une telle reˆverie, Whittemore ne l’a jamais formule´e que de manie`re prudente, nuance´e et, au fond, superficielle, dans la ` cela lui permettait de persuader des donateurs potentiels de l’importance mesure ou poe´tique des taˆches auxquelles allaient servir leurs fonds. Prichard savait cela, mais il reconnaissait aussi `a Whittemore ses qualite´s propres: une capacite´ de´concertante `a se de´placer incessamment `a travers le monde—“extraordinary traveller!” ne peut-il s’empeˆcher de s’exclamer, peu apre`s la fondation du Byzantine Institute37 —et une extreˆme habilete´ `a nouer, en toute occasion, des relations au plus haut niveau, tant intellectuelles que politiques—“he knows everybody everywhere.”38 Or, pour l’avance´e concre`te des connaissances sur l’art byzantin, ces qualite´s sont pre´cieuses, comme le prouve la mise en œuvre—pour le moins remarquable, d’un simple point de vue diplomatique39 —du grand chantier de Whittemore, la restauration de Hagia Sophia de Constantinople, en 1931 (Figs. 6 et 7): “I told Helen [Sears] of the great importance that I attach to Whittemore’s work in Sta-Sophia’s, where he tells me he had made a good beginning. Byzantine work and all the early work of the Near East is going to be the gospel of the next few centuries, and everything saved from destruction belonging to that movement will become a sign post to enfranchisement.”40 C’est pourquoi, `a la faveur du de´veloppement de l’activite´ du Byzantine Institute, dans les dernie`res anne´es de la vie de Prichard, sa relation avec Whittemore est davantage devenue celle d’un maıˆtre `a un disciple. Le directeur du nouvel institut va re´gulie`rement chercher la bonne parole `a Londres, pour l’inte´grer `a la description de ses travaux de restauration et la diffuser, de cette fac¸on, aupre`s de ses nombreux interlocuteurs: “Day in, day out, [Prichard] advised upon [the work’s] processes to the last, minutest detail, remaining all the time a prophetic figure in the background, to offer wisdom when the time approached to tell officially of the work’s advance.” 41 Whittemore est meˆme alle´ plus loin: il a demande´ `a Prichard d’e´crire pour lui une partie essentielle du texte final des deux premiers rapports de fouilles publie´s sous sa signature, en 1933 et 1936, et traitant respectivement des mosaı¨ques de´couvertes dans le narthex et dans le vestibule sud de la ´ glise.42 C’est ce qui ressort des ´souvenirs confie´s par Henry Roman Walker, l’un Grande E 37 ˆ t 1930, AD. Voir ´egalement une des dernie`res lettres adresse´es par Prichard `a Georges Duthuit, le 3 aou Prichard `a Isabella Stewart Gardner, le 5 juillet 1924: “Whittemore is here. He illustrates Bergson’s indication that the present is not a reality. . . . Whittemore is never in a place; he was, he will be; he comes from and is going to; but he is never ‘here’” (ISGM). 38 Prichard `a Georges Duthuit, le 5 octobre 1925, AD. 39 En juin 1931, un de´cret pre´sidentiel de Mustafa Kemal autorise le Byzantine Institute `a mettre au jour, restaurer et conserver les mosaı¨ques de la Ayasofya Camii, sous la direction de Thomas Whittemore. Les travaux commencent en de´cembre de la meˆme anne´e par le narthex. En mars 1934, un nouveau de´cret pre´sidentiel transforme la mosque´e en muse´e. Mustafa Kemal visite personnellement l’e´difice en fe´vrier 1935, c’est-a`-dire apre`s sa se´cularisation, geste symbolique de grande importance dans le cadre de la politique laı¨que activement mene´e par la Re´publique turque. 40 Prichard `a Mrs. Montgomery Sears, le 1er mai 1932, BB. 41 H. R. Walker, Prichard’s Biography, texte dactylographie´, chap. 8, p. 1, ISGM. 42 The Mosaics of St. Sophia at Istanbul: Preliminary Report, et Second Preliminary Report. Dans les deux rapports, le nom de Prichard apparaıˆt simplement dans la liste finale des remerciements (respectivement p. 24 et p. 35).

52

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

des proches de Prichard, `a John Pope-Hennessy, en 1970: “The St-Sophia work, you will be aware, owes as much, in absentia, to MSP, as to his pupil (and extremely unphilosophical) Thomas Whittemore who used to send him material and trust to Matt to edit it, which he would do often enough to the extent of re-writing, entirely.” 43 En ce qui concerne le premier rapport, un cahier manuscrit, ´ecrit de la main de Prichard, confirme cette hypothe`se: le texte, qui porte sur la description de la mosaı¨que du tympan central et fait une synthe`se finale sur les mosaı¨ques du narthex, est, `a de tre`s mineures variations pre`s, celui des dix dernie`res pages du livre44 (Fig. 8). En ce qui concerne le second rapport, Prichard ´ecrit le 12 mars 1935 `a Una PopeHennessy: “Whittemore approaches, to find how little I have done for him. But I have thought a little, and am pretty clear that we are touching a new dispensation.” 45 Et, un an avant sa mort, le 13 septembre 1935, il confie `a Henry Roman Walker: “What work I have done has been about Byzantine things. Like yourself, I hunger and thirst for our studies, and as soon as I can I count on returning to continue them. Whittemore’s Report will soon be in the proof stage, and until that is passed I want to remain near Oxford.” 46 De sorte que, lorsque le lecteur contemporain lit, sous la signature de Thomas Whittemore, des analyses opposant le caracte`re “classique,” descriptif et fige´ de la mosaı¨que de la lunette du narthex de Hagia Sophia, d’une part, et le caracte`re purement “byzantin,” dynamique et ´emotionnel des croix justiniennes qui rythment toute la de´coration de l’e´difice (Figs. 9 et 10), c’est la prose et la pense´e de Matthew Stewart Prichard qu’il rencontre, et non celles de Whittemore: “The great mosaic is classical, if we mean thereby a descriptive composition, which, based on the intelligent translation of perception into symbols, idealises the result by emphasising the truth, beauty, goodness and geometrical construction of its content. . . . The crosses reflect, in their immediate appeal to our emotional life, an approach to action that is more characteristic to Asia than of the mystical transfiguration of matter which have found favor in Europe.” 47 Et de meˆme lorsqu’on peut lire, en conclusion du meˆme volume: “The aim of Byzantine art was essentially not form but function.” 48 De Prichard `a Matisse Quelques anne´es plus tard, en 1942, Henri Matisse ´ecrit en marge d’un article que lui consacre Louis Aragon: “la croix byzantine, l’imagination et la raison”; et Aragon pre´cise que cette note a ´ete´ ´ecrite “en relation avec la mention faite des mosaı¨ques de ISGM. ´ tant donne´ les corrections au sein du manuscrit de Prichard et sa nume´rotation DO (box A, book 35). E des planches, diffe´rente de celle qui figure dans le texte final, il ne peut pas s’agir d’une copie poste´rieure du texte publie´ (The Mosaics of St. Sophia at Istanbul: Preliminary Report, 14–24), mais bien d’un premier projet. La date de 1933–34 inscrite, d’une autre main, en teˆte du cahier, est fausse; la re´daction du cahier est ante´rieure `a 1933. 45 ISGM. 46 ISGM. 47 The Mosaics of St. Sophia at Istanbul: Preliminary Report, 24. On retrouve mot pour mot, dans la communication de Whittemore au IVe Congre`s international des ´etudes byzantines, ces passages ´ecrits par Prichard: T. Whittemore, “Mosaics of Aya Sophia,” Actes du IVe Congre`s international des ´etudes byzantines (Sofia, 1934), publie´ sous la re´daction (sic) de B. D. Filov, Bulletin de l’Institut Arche´ologique Bulgare 10 (Sofia, 1936), 206. 48 The Mosaics of St. Sophia at Istanbul: Preliminary Report, 33. 43 44

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

53

Sainte-Sophie.” 49 Entre-temps, alors que Prichard est mort en 1936, Whittemore a renoue´ des relations re´gulie`res avec Matisse, qui a dessine´ une se´rie de portraits de lui en 1937. Il a envoye´ `a l’artiste des photographies des mosaı¨ques de Hagia Sophia50 et des exemplaires des rapports publie´s51 (Fig. 11); il a ´evoque´ le sens de ces de´couvertes et, naturellement, s’est fait la` encore l’e´cho de son maıˆtre anglais, dont Matisse connaissait du reste bien la pense´e, pour avoir assiste´ `a son de´veloppement de´cisif, `a Paris, entre 1909 et 1914.52 Les relations de Whittemore avec Matisse fournissent ainsi un te´moignage exemplaire de l’influence exerce´e par Prichard, aussi bien sur le peintre franc¸ais, d’ailleurs, que sur le byzantiniste ame´ricain. Au de´part, Whittemore a rencontre´ Matisse par l’interme´diaire des Stein, inde´pendamment de Prichard qui n’e´tait pas encore installe´ `a Paris. De`s 1908, sur la recommandation personnelle de Matisse, il a visite´ `a Moscou la collection du grand collec` la fin du tionneur et principal patron de Matisse `a l’e´poque, Sergueı¨ Chtchoukine.53 A 54 mois de de´cembre de la meˆme anne´e, de retour de Moscou, il visite avec enthousiasme, `a Berlin, la premie`re grande exposition personnelle consacre´e `a Matisse en Allemagne, ´ la cole`re de la plupart des critiques et avait laisse´ le peintre exposition qui avait provoque 55 passablement effondre´. C’est pourquoi, plus de trente ans apre`s, en 1942, dans ses conversations avec Pierre Courthion, celui-ci garde un vif souvenir d’une des seules manifestations de sympathie qui lui soient alors parvenues, celle de Whittemore: “Rentrant `a ˆtel, je trouve des lauriers, avec un grand ruban rouge. Un de mes admirateurs, mon ho un Ame´ricain, qui se trouvait `a Berlin—je n’en savais rien—m’envoyait cette grande couronne. Je veux dire en passant que c’e´tait Thomas Whittemore qui, charge´ par l’Institut byzantin d’Ame´rique, vient de faire apparaıˆtre des splendeurs de mosaı¨ques, en dirigeant l’enle`vement de´licat de la couche de chaux qui les ensevelissait depuis des sie`cles, dans les coupoles de Sainte-Sophie de Constantinople.” 56 “Matisse-en-France,” 1942, en L. Aragon, Henri Matisse, roman, 2 vols. (Paris, 1971), 1:73. “Il y a, dans un coin de la pie`ce aux dessins des photos des mosaı¨ques byzantines de Sainte-Sophie auxquelles il [Matisse] revient souvent, parlant de ce qui l’occupe” (ibid.). 51 ` Henri Matisse. Hommages Le premier volume, conserve´ dans les Archives Matisse, porte la de´dicace: “A respectueux. Thomas Whittemore” (AM). 52 Voir R. Labrusse, Matisse: la condition de l’image (Paris, 1999), chap. 5, et Henri Matisse: “la re´ve´lation m’est venue de l’Orient” (Florence, 1997), 354–55. 53 “Il me semble qu’un sie`cle s’est ´ecoule´ depuis 1908 lorsque vous m’avez donne´ une lettre pour M. Stchoukine `a Moscou” Whittemore `a Matisse, le 20 de´cembre 1935, AM. 54 Le 23 de´cembre 1908, Whittemore ´etait encore `a Moscou, comme en te´moigne une note de te´le´phone inse´re´e dans un catalogue en franc¸ais de la galerie Tre´tiakov, ´edition de 1905, BB. 55 ¨ ber Henri Matisse,” Werk, no 6 Sur cette exposition de Matisse chez Paul Cassirer, voir H. Purrmann, “U ( juin 1946), 185–92. 56 Conversations avec Pierre Courthion, 1941, exemplaire dactylographie´, AM. Voir ´egalement les souvenirs de Gertrude Stein: “Puis Matisse eut une grande exposition `a Berlin. Je me rappelle si bien un jour de ` nous devions aller de´jeuner `a Clamart avec les Matisse. Quand nous arrivaˆmes, ils printemps charmant, ou ´etaient tous assemble´s autour d’une ´enorme caisse, dont on venait d’o ˆter le couvercle. Nous nous joignıˆmes `a eux; dans la caisse, il y avait la plus grande couronne de laurier que l’on eu ˆ t jamais tresse´e, et elle ´etait orne´e d’un beau ruban rouge. Matisse montra `a Gertrude Stein une carte qui se trouvait dans le paquet. On y voyait inscrit: ‘a` Henri Matisse, triomphant sur le champ de bataille de Berlin’ et c’e´tait signe´: ‘Thomas Whittemore.’ Thomas Whittemore ´etait un arche´ologue bostonien et un professeur `a Tufts College, un grand admirateur de Matisse, et c’e´tait un tribut qu’il envoyait au maıˆtre. Mais Matisse de´clara sur un ton encore plus lugubre que d’ordinaire: ‘Enfin, pourtant, je ne suis pas encore mort!’” L’Autobiographie d’Alice B. Toklas, 122–23. 49 50

54

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

En 1910, lorsque Whittemore visite (quelques semaines avant Matisse lui-meˆme) la grande exposition internationale d’art musulman organise´e `a Munich, il rele`ve la similitude de certains tissus persans avec l’art de Matisse.57 Et peut-eˆtre est-ce `a la meˆme ´epoque qu’il devient proprie´taire d’une toile de 1904, La Terrasse `a Saint-Tropez, et l’offre `a Isabella Stewart Gardner, pour son muse´e de Boston (Fig. 12)—un geste de la pre´cocite´ duquel il s’est toujours enorgueilli: “Je crois ˆetre le premier `a avoir apporte´ un de vos tableaux sur l’autre rive de l’Atlantique,” ´ecrit-il `a Matisse en 1935. “Depuis ce jour et jusqu’ `a l’heure pre´sente, les lec¸ons trace´es par votre pinceau m’ont guide´ dans mes ´etudes inlassables de l’art byzantin et copte.” 58 Pourtant, ces relations n’ont jamais pris l’ampleur ni l’intensite´ de celles que Prichard a, lui, ´etablies avec le peintre entre 1909 et 1914. Whittemore, comme `a l’accoutume´e, demeure aupre`s de Matisse une apparition occasionnelle, fugitive et mondaine. Les relations entre les deux hommes ne se resserrent, en fait, qu’a` partir de 1935, lorsque le byzantiniste demande au peintre de le portraiturer: “Je serai responsable,” lui ´ecrit-il alors, “dans une certaine mesure, de cette re´ve´lation des mosaı¨ques de Sainte-Sophie que j’apporte au monde. Il serait peut-eˆtre d’une certaine importance de conserver mes traits pour l’histoire de ces travaux qui sera ´ecrite un jour, mais je sens qu’il n’y a qu’une main ˆtre. Vous souvenez-vous qu’au de´but de notre connaissance, qui puisse le faire,—la vo vous vouliez prendre un croquis de moi, mais je ne suis jamais reste´ assez longtemps `a Paris pour que cet espoir puisse se re´aliser. Serait-il possible que vous fassiez maintenant, au cours de cet hiver ou au printemps, soit `a Paris soit `a Nice, si je viens vers vous, ce croquis qui me situera sur un plan byzantin, ainsi que seul, dans le monde entier, vous pouvez le faire?” 59 Finalement, les se´ances de pose ont lieu en de´cembre 1936 et au de´but du mois de janvier 1937, aboutissant `a deux grands portraits au fusain (Figs. 13 et 14) et deux dessins ` le byzantiniste, coiffe´ du chapeau brode´ au trait `a l’encre de Chine (Figs. 15 et 16), ou ` ˆt-il, ouzbek qui, paraı ne le quittait jamais, apparaıˆt “flatte´,” selon Pierre Matisse.60 A partir de cette date, les ´echanges se sont fait plus fre´quents, plus intimes: celui que Whit`s temore appelait encore en 1935 “Cher maıˆtre,” est devenu en octobre 1937 “Mon tre Deux exemplaires du catalogue de l’exposition de Munich (Ausstellung von Meisterwerken muhammedanischer Kunst. Musikfeste. Muster-Ausstellung von Musik-Instrumenten. Amtlicher Katalog [Munich, 1910]) sont conserve´s dans les Archives Whittemore de la Bibliothe`que byzantine. L’un (3e ´edition) est annote´ de la main de Whittemore et on y lit sur la page de garde: “In room no 3 the garment—with black background—[2 mots illisibles] is done in Matisse’s manner” (BB). L’œuvre en question est reproduite in F. Sarre et F. R. ¨nchen 1910 (Munich, 1912), 3: pl. 199, Martin, Die Ausstellung von Meisterwerken muhammedanischer Kunst in Mu cat. 2,362 (fragment de soie `a figures multicolores sur fond noir, Perse, 1550–1650, coll. Kelekian, Paris). Matisse, pour sa part, a visite´ l’exposition d’art islamique entre le 7 et le 14 octobre 1910, et il y a retrouve´ Prichard, qui avait passe´ tout l’e´te´ `a Munich. Voir Labrusse, Henri Matisse, 355 et 358. 58 Lettre du 20 de´cembre 1935, AM. La date du don de Whittemore au Gardner Museum demeure toutefois hypothe´tique. Alfred Barr s’en est enquis, mais sans succe`s, lorsqu’il pre´parait sa monographie sur Matisse (lettre `a Martin Mower, Fogg Art Museum, Cambridge, Mass., du 9 mars 1951, archives of the Museum of Modern Art, New York). De plus, Whittemore a achete´ pour le compte d’Isabella Stewart Gardner au moins trois des cinq dessins de Matisse pre´sents dans sa collection, et peut-eˆtre quatre. Voir K. E. Haas, “Henri Matisse: ‘A Magnificent Draughtsman,’” Fenway Court 1985 (Boston, 1986), 37 et 48 note. 59 Whittemore `a Matisse, le 20 de´cembre 1935, AM. 60 “Whittemore est venu me montrer le portrait que Papa a fait de lui, qui est ressemblant mais flatte´. Je veux dire qu’il n’est pas aussi bien que le portrait” (Pierre Matisse `a Henri Matisse, 9 fe´vrier 1937, Pierre Matisse Archives, The Pierpont Morgan Library, New York). 57

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

55

ˆt cellecher et grand ami.” Et cette entente, interrompue par la guerre, reprend aussito ci acheve´e, comme en te´moigne le ton de leur correspondance entre 1945 et 1949. On peut constater, `a partir de cette correspondance, que Whittemore se fonde toujours sur l’he´ritage de Prichard pour nourrir ses conversations avec le peintre. Pour le remercier de ses portraits, par exemple, il remarque, dans une tonalite´ typiquement ˆtre et nul autre que prichardienne: “Plusieurs traits rapprochent l’art byzantin avec le vo vous ne saurait mieux expliquer les relations qui existent entre ces deux grands arts, se´pare´s dans le temps, mais unis dans une meˆme tendance vers le dynamisme inte´gral.” 61 Et pre`s de dix ans plus tard, apre`s la guerre, le ton n’a gue`re change´: Maintenant, je suis en route pour l’Orient. J’ai vu votre Exposition `a Paris,62 c’est un stimulant qui m’a donne´ beaucoup d’e´ner` travers de longs de´dales, vous avez abouti `a la vision comple`te d’inte´gration. J’ai gie. A beaucoup appre´cie´ la plaquette pour l’Exposition. ˆt je serai `a Istanbul. Les travaux de de´gagement des mosaı¨ques se poursuivent Biento et maintenant, nous sommes en possession d’un catalogue complet de peintures qui s’e´chelonnent presque sur mille ans. Les Byzantins comprenaient l’espace, mais n’ont pas commis la faute d’essayer de le reproduire et de cre´er ainsi un obstacle `a leur vision.63

Pour Whittemore, `a la fin de sa vie, l’art byzantin se re´sumait, dans le sillage des ide´es de Prichard, `a une architecture incluant tous les arts, voue´e `a la cre´ation d’un milieu favorable `a l’accomplissement des ce´re´monies. C’est ce qu’il allait chercher, par exemple, ` il a fait dans les monaste`res du Mont Athos (Figs. 17 et 18) ou de Patmos (Fig. 19), ou plusieurs se´jours au cours des anne´es vingt et trente. Il s’agissait pour lui d’un art performatif, en quelque sorte, dont la mobilite´, l’adaptabilite´ structurelles ´etaient principaleˆle de la lumie`re: le baˆtiment de l’e´glise byzantine devenait “une ment fonde´es sur le ro composition en termes de lumie`re et de couleurs construites” ou, “d’un point de vue fonctionnel, un re´ceptacle pour la lumie`re” 64 (Fig. 20). C’est `a partir de cette orientation qu’il ce´le`bre le succe`s d’un syste`me de´coratif respectant la de´finition de l’espace propose´e par les murs, n’utilisant pas cet espace inte´rieur comme un mur d’exposition, mais fusionnant avec lui et l’organisant; un syste`me qui ne cherche pas `a appesantir la conscience du spectateur avec des informations d’ordre intellectuel ou `a reque´rir de lui la contemplation d’un spectacle fictif, mais qui vise a` le diriger vers la re´alite´ de la ce´re´monie, en dynamisant sa perception; qui n’est pas, autrement dit, anime´ par l’ide´e de beaute´ en soi, mais par celle de rapport `a autrui. Ces convictions, Whittemore les a re´affirme´es inlassablement dans ses confe´rences. C’est meˆme ce qui l’a rendu ce´le`bre dans le grand public, surpris de voir soudain l’art byzantin de´crit en termes de “dynamisme inte´gral,” de participation de l’individu au groupe, et ˆ t pas de contemd’expansion de´corative “de´sinte´resse´e,” tout cela dans un but qui ne fu 61 ` titre de comparaison, voir cette lettre de Prichard `a Georges Duthuit, Lettre du 16 janvier 1937, AM. A du 27 septembre 1926: “We turn now to the fiery dynamism of life’s creative or integral consciousness” (AD). 62 Il s’agit de la re´trospective d’œuvres de Matisse, contenant trente-sept toiles, au Salon d’automne 1945 (28 septembre–29 octobre). 63 Lettre du 29 janvier 1946, AM. 64 “Haghia Sophia is a composition in light and constructed color” (“Notes on Light in Haghia Sophia, Istanbul,” exemplaire dactylographie´, 26 janvier 1945, p. 3, DO); “Haghia Sophia is functionally a receptacle of light” (ibid., p. 11). La` encore, Prichard est tout proche: “Architecture as an art is composition in terms of light” (The Mosaics of St. Sophia at Istanbul: Second Preliminary Report, 33)

56

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

plation esthe´tique mais d’incitation a` agir.65 Ce seront d’ailleurs parmi ses derniers mots, publie´s apre`s sa mort, en 1952, `a propos de la grande De´isis de la tribune sud de Hagia ` se meˆlent, `a la manie`re caracSophia, qui avait ´ete´ restaure´e en 1938—des remarques ou te´ristique de Whittemore, une approche esthe´tique dont on connaıˆt de´sormais les origines, et une approche technique, ou, pour mieux dire, pragmatique, reposant sur les travaux de son ´equipe: Only the Byzantine solved the problems of the architectural use of painting in mosaic. They never destroyed the integrity of the wall by space composition. The construction of the panel is in a field of small irregular wavy areas, produced by three superimposed layers of plaster of different depths and mixtures. . . . The Church was the workshop of the mosaicist who, in the light of the architectural organism of building, created, in life’s time, the manifestation which started the gaze on a new course of action.66

II. “SAUVER LA CULTURE RUSSE”: UNE PASSION FONDATRICE La contribution russe au Byzantine Institute of America Entre sa fondation en 1930 et la mort de Thomas Whittemore, en 1950, le Byzantine Institute of America67 a consacre´ son activite´ `a la restauration et `a l’e´tude de trois monuments byzantins d’Istanbul: Hagia Sophia au cours de la premie`re de´cennie, les ´eglises du Christ de la Chora (Kariye Camii) et Sainte-Marie Pammakaristos (Fethiye Camii) au cours de la seconde de´cennie. Auparavant, ces baˆtiments avaient traverse´ une pe´riode d’abandon due notamment `a l’e´clatement de la Premie`re Guerre mondiale et `a la fermeture pre´cipite´e de l’Institut arche´ologique russe de Constantinople (RAIK) en 1914. Actif depuis 1894, cet Institut, dont les nombreuses activite´s sont illustre´es dans les seize volumes monumentaux des Izvestija RAIK, avait incarne´ pendant vingt ans la pre´sence dominante `a Istanbul de l’e´cole arche´ologique russe, qui, malgre´ plusieurs tentatives, n’a pas ´ete´ re´tablie apre`s 1914.68 Avec la cre´ation du Byzantine Institute of America, on assiste donc `a un important changement dans l’arche´ologie byzantine d’Istanbul, marque´ par l’affirmation de la primaute´ ame´ricaine dans ce domaine. ´ tats-Unis, en mars 1937, Les comptes rendus de presse `a l’occasion de ses confe´rences donne´es aux E sont `a cet ´egard re´ve´lateurs. Le titre choisi par le Boston Evening Transcript est: “The Criterion of Byzantine Art Is Not Beauty, Pr. Whittemore Said. . . . He described light as one of the building materials in Byzantine architecture, as in the Byzantine conception of architecture organized light becomes constructive color” (3 mars 1937). Et The Arizona Daily Star: “Byzantine Religious Theme Is Outlined in Whittemore Talk. . . . We admire the objects of Greek art, but Byzantine art is one of the spiritual principles with which we have been out of touch since the Renaissance. We only begin now to perceive those spiritual principles and our need of them in our social order” (31 mars 1937). 66 T. Whittemore, The Mosaics of St. Sophia at Istanbul: Fourth Preliminary Report, 21–22. 67 Cre´´e en 1930 comme´ “voluntary association,” l’Institut byzantin d’Ame´rique fut “incorporated” en 1934: T. Whittemore, “Report on the Byzantine Institute,” Renaissance News, no 3 (Autumn 1948), 50. La mort de Whittemore provoqua de fait sa liquidation, et les travaux sur le terrain furent repris par Dumbarton Oaks: voir Constable, “Dumbarton Oaks and Byzantine Field Work,” 172. La bibliothe`que parisienne de l’institut ´ cole nationale des langues orientales vivantes et prit le nom de la Bibliothe`que byzantine. fut donne´e `a l’E 68 Sur les diffe´rentes tentatives du renouvellement du RAIK jusqu’en 1945, voir E. Basargina, “Russkij archeologicˇeskij institut v Konstantinopole: Archivnye fondy” (L’institut arche´ologique russe `a Constantinople: Les fonds d’archives), Archivy russkich vizantinistov v Sankt-Peterburge, re´d. I. Medvedev (Saint` propos des copies de peinture byzantine `a Istanbul: Les Pe´tersbourg, 1995), 88 sqq., et N. Podzemkaia, “A artistes ´emigre´s et la naissance de l’Institut byzantin d’Ame´rique,” Histoire de l’art, no 44 ( juin 1999), 123–40. 65

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

57

Si, apre`s 1914, l’arche´ologie russe a perdu son statut officiel `a Istanbul, la “pre´sence russe” n’y a pas pour autant ´ete´ ´elimine´e. Au contraire: au de´but des anne´es vingt, Constantinople a fourni un refuge aux restes de l’arme´e blanche en de´route et est ainsi devenue, pour quelques anne´es, un centre important, un point de passage pour les ´emigre´s. C’est graˆce `a ces milieux russes d’Istanbul, notamment aux savants et aux artistes e´mi´ tats-Unis, que la me´moire de la tradition gre´s, disperse´s par la suite en Europe et aux E arche´ologique russe a pu ˆetre pre´serve´e, et que Whittemore a pu l’utiliser au profit de son institut. Celui-ci a en effet accompli un long travail au sein de l’e´migration russe, au fil des anne´es vingt—travail souterrain, reste´ dans l’ombre, qui apparaıˆt pourtant comme un ´ele´ment fondamental dans l’histoire du Byzantine Institute of America: ce faisant, Whittemore a pu assumer l’he´ritage d’une tradition d’e´tudes byzantines et atteindre son but `a Istanbul, graˆce `a l’argent rassemble´ aupre`s de riches familles de Boston, ainsi qu’aux pourparlers mene´s avec virtuosite´ aupre`s des autorite´s turques. Pre´fac¸ant le premier volume du Bulletin of the Byzantine Institute, l’organe officiel de la nouvelle institution,69 il rendait hommage aux nombreux savants slaves qui avaient “inspire´ la fondation du Byzantine Institute”: “This initial number of the Bulletin serves to recall a long collaboration with numerous Slavic archaeologists and scholars, whose creative and active association helped to inspire the foundation of the Byzantine Institute of America. In acknowledgement of this debt we offer at the outset as a part of the contents of this present issue a record of some of our efforts in Bulgaria. . . . ”70 Cette collaboration remonte, selon Whittemore, aux travaux mene´s par les membres de l’Institut en ´ gypte, en Russie, en Turquie et `a travers les Balkans avant 1930,71 ce qui situe l’histoire E du Byzantine Institute of America bien avant sa naissance “officielle.” Dans le meˆme texte, il indique qu’une des premie`res fouilles “entreprises par l’Institut” [sic] avait concerne´ la basilique de Belovo en Bulgarie, repe´re´e (“investigated”) en 1922 et fouille´e (“excavated”) en 192472 (Fig. 21). Dans d’autres contributions au Bulletin, la “ge´ne´reuse initiative” ou l’“aide ge´ne´reuse” de Thomas Whittemore sont ´egalement mentionne´es `a propos des fouilles mene´es en 1920 par Ivan Velkov, directeur du Muse´e national de Sofia, sur le site d’une basilique pale´ochre´tienne `a Me´sembria73 et `a propos des fouilles ´ glise Rouge de Perusentreprises en 1921 par l’Institut arche´ologique bulgare sur l’E ´ glise Rouge furent exe´cute´s par S. N. Pokrovskij, alors tica.74 Les plans et les releve´s de l’E que les copies des fresques, publie´es dans le premier volume du Bulletin, sont de la main d’Olga Wivdenko-Belokopytov, ´ele`ve de Nikodim P. Kondakov. C’est Kondakov, en effet, qui a incarne´ la principale re´fe´rence intellectuelle de Whittemore, pendant la pe´riode de son activite´ bulgare, et c’est `a lui que le premier Bulletin rend un “hommage particulier,” en le pre´sentant comme “l’un des promoteurs les plus importants d’un ´equipement de haute qualite´ technique dans le champ de la recherche byzantine.” 75 D’apre`s les journaux de Kondakov conserve´s `a Prague, WhitteLa publication du premier volume, retarde´e par la guerre, eut lieu en 1946. T. Whittemore, avant-propos `a BByzI 1 (1946), 1. 71 Ibid. 72 Ibid. 73 I. Velkov, “An Early Christian Basilica at Mesembria,” BByzI 1 (1946), 61–70. 74 ´ glise Rouge de Perustica,” BByzI 1 (1946), 15–42. A. Frolov, “L’E 75 Whittemore, avant-propos `a BByzI 1 (1946), 1 (“one of the chief promoters of a high technical equipment in Byzantine research”). 69 70

58

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

` il lui a transmis des livres more lui a rendu visite `a plusieurs reprises, `a Sofia en 1921 (ou de la part de Gabriel Millet et s’est entretenu avec lui des travaux `a Me´sembria), et `a Prague en 1923.76 Signe de l’importance de ce contact, le premier Bulletin of the Byzantine Institute contient un texte du ce´le`bre savant sur le tre´sor de Nagy-Szent-Miklos: ce texte, ´ecrit Whittemore, fut donne´ par Kondakov, peu avant sa mort en janvier 1925, “comme une marque d’amitie´” au Byzantine Institute (sic).77 Outre cette publication et les re´fe´rences directes, dans le premier Bulletin of the Byzantine Institute, au “Patriarche des byzantinistes,” 78 l’he´ritage de Kondakov est sensible dans les contributions de savants ou d’artistes russes qui ont travaille´ sous sa direction intellectuelle, notamment Andre´ Grabar dont les “premiers essais sur l’art bulgare, en 1920 et 1921, furent guide´s par le regrette´ N. P. Kondakov, alors professeur `a Sofia,” 79 ou Olga Wivdenko-Belokopytov, qui suivit les conseils de son maıˆtre pour copier les fresques de ´ glise Rouge.80 Ce sont ces Russes, ´ele`ves directs ou indirects de Kondakov, qui ont l’E fourni un apport de´cisif `a l’e´laboration des mate´riaux rassemble´s en Bulgarie au cours des anne´es vingt, et publie´s dans le premier volume du Bulletin of the Byzantine Institute. Andre´ Grabar a ´ete´ charge´ de l’e´dition de l’ouvrage entier; la pre´sentation des fouilles ´ glise Rouge a ´ete´ confie´e `a Andre´ Frolov, qui s’est appuye´ sur les releve´s exe´cute´s de l’E ´ glise Rouge par Olga Wivdenko-Belokopytov, ainsi que sur le texte de celle-ci intitule´ L’E (Perouchtitza): Fresques, date´ “Paris, fe´vrier 1933,” dont une dactylographie est conserve´e `a la Bibliothe`que byzantine `a Paris, accompagne´e d’une carte de visite avec l’adresse parisienne de l’artiste et d’une de´dicace `a Whittemore. Dans d’autres publications et travaux dirige´s par Whittemore, l’apport des collaborateurs russes apparaıˆt aussi fondamental. Parmi eux, on peut mentionner Michel Malinine, ´editeur du second et dernier volume du Bulletin (1956), consacre´ aux ´etudes coptes et de´die´ `a la me´moire de Walter Ewig Crum, ou bien Nikolaj Kluge, ancien membre du RAIK, qui exe´cuta au cours des anne´es trente les releve´s des mosaı¨ques de Hagia Sophia pour le Byzantine Institute of America. Souvent, en outre, la contribution des Russes est demeure´e invisible, anonyme: la Bibliothe`que byzantine, par exemple, conserve un grand nombre de textes manuscrits ou dactylographie´s, semble-t-il, sur la commande de Whittemore, afin de pre´parer ses propres publications ou des recherches entreprises par l’institut. Il s’agit souvent de traductions en anglais ou en franc¸ais d’ouvrages russes, ou bien de re´sume´s et comptes rendus faits `a partir de ces ouvrages. On trouve par exemple `a la Bibliothe`que byzantine deux traductions dactylographie´es en anglais de fragments du XIe chapitre de l’ouvrage 76 N. Kondakov, Journal (FK). Je remercie Mesdames Zahradnikova et Kopecka de m’avoir automise´e `a consulter le Journal de Kondakov lors du colloque Symposium Kondakovianum qui eut lieu `a Prague du 15 au 17 fe´vrier 1995. 77 Whittemore, avant-propos `a BByzI 1 (1946), 1 (“in friendly gesture”). Kondakov fut le premier `a attribuer le tre´sor de Nagy-Szent-Miklos `a l’art protobulgare. Il avait vraisemblablement un projet plus vaste sur ce sujet: dans une lettre manuscrite non date´e, conserve´e `a la Bibliothe`que byzantine `a Paris, il faisait part `a Whittemore de son de´sir de publier `a ce sujet un texte assez long (50–75 pages in-4o, environ 30 dessins), en s’appuyant sur les nombreux mate´riaux qu’il avait rassemble´s au cours de longues anne´es d’e´tudes. 78 A. Grabar, “Kondakov and the Treasure of Nagy-Szent-Miklos,” BByzI 1 (1946), 3 (“this Patriarch of Byzantinists”). 79 A. Grabar, La peinture religieuse en Bulgarie (Paris, 1928), V. 80 Dans son Journal, Kondakov ´evoque les visites au cours desquelles O. Wivdenko-Belokopytov lui montrait ses aquarelles (FK).

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

59

du pe`re Pavel Florenskij, Colonne et fondement de la ve´rite´ (Moscou, 1914), contenant notamment la dixie`me lettre intitule´e “Sophia,” avec une description des diffe´rents types ˆnes de Hagia Sophia, ce qui montre que, en 1919 au plus tard,81 Whittemore avait d’ico cherche´ `a inte´grer la pense´e the´ologique russe `a sa formation de byzantiniste. Et ce n’est sans doute pas un hasard si le nom du grand philosophe russe Vladimir Soloviev est ´evoque´ dans le texte de Whittemore intitule´ Notes on Light in Hagia Sophia, Istanbul.82 La majeure partie de ce travail souterrain de traduction et de recherche fut assure´ par les deux plus proches collaborateurs de Whittemore `a Paris, Boris Ermolov (Fig. 22) ˆle majeur revenait `a Ermolov, secre´taire de Whitteet Vladimir Rayevsky (Fig. 23). Le ro more en Europe, recrute´ au de´but des anne´es vingt et fide`le, sa vie durant, `a la personne de Whittemore, qui le mettra au nombre des quatre trustees du Byzantine Institute of America.83 C’est `a travers lui que passait toute la correspondance parisienne du Byzantine Institute; c’est souvent par son interme´diaire que Whittemore communiquait avec les ` titre d’exemple, c’est Ermolov qui, en 1931, lui milieux russes `a Paris (Figs. 24 and 25). A ´ cole des Hautes E ´ tudes `a pre´senta le jeune Andre´ Frolov, ´ele`ve de Gabriel Millet `a l’E ´ glise Rouge, et Ermolov Paris; Whittemore le chargea de travailler sur le mate´riel de l’E 84 ˆ t verse´. fit en sorte qu’un salaire mensuel lui fu Le Byzantine Institute of America et le Committee for the Education of Russian Youth in Exile L’efficacite´ de la collaboration de Whittemore avec les Russes, sur plus de trois de´cennies, s’explique par le fait que, de`s l’origine, cette collaboration s’est de´veloppe´e au sein d’un cadre institutionnel dont la principale incarnation fut, de`s 1922, le Committee for the Education of Russian Youth in Exile: cette organisation ´etait destine´e `a aider de jeunes ressortissants russes, exile´s `a la suite de la Re´volution d’octobre et de la guerre civile, `a poursuivre leurs ´etudes en Europe. C’est sur elle qu’en 1930 s’est greffe´ le Byzantine Institute of America. La continuite´ entre les deux institutions est frappante: toutes deux inspire´es et dirige´es par Whittemore, elles avaient leur sie`ge principal `a Boston, ´etaient ge´re´es par le meˆme personnel, notamment Seth Gano comme tre´sorier ou adjoint-tre´sorier `a Boston et Boris Ermolov comme secre´taire en Europe,85 et ´etaient finance´es par les meˆmes personnes issues de la haute socie´te´ bostonienne. Seul le domaine d’activite´ a change´, entraıˆnant le changement de de´nomination. ` l’origine, des sections du Committee for the Education of Russian Youth in Exile A ` leurs teˆtes furent place´s des Russes furent ouvertes `a Sofia, Belgrade, Berlin et Paris. A ´emigre´s: en Bulgarie, Alexandre et Dmitrij Ermolov; en Yougoslavie, Alexandre Brandt; La premie`re dactylographie, de vingt-cinq pages, est date´e de la fin de 1919, avec des corrections et rajouts manuscrits (BB; M SS/III 2). La seconde dactylographie, de dix-sept pages, n’est pas date´e, et contient des corrections et des suppressions manuscrites (BB; M SS/III 1). 82 Texte dactylographie´ de treize pages, date´ du 26 janvier 1945, p. 4, BB; M SS/I 14. 83 D’apre`s le §16 de son testament, Whittemore avait confie´ `a quatre personnes—John Nicholas Brown, Francis H. Taylor (directeur du Metropolitan Museum of Art, New York), Boris N. Ermolov (bibliothe´caire du Byzantine Institute, Paris) et Seth Gano—le soin d’administrer son legs (Testament of Thomas Whittemore, §16, BB). 84 B. Ermolov `a S. Gano, le 27 de´cembre 1931, BAR/CERYE, box 12. 85 ˆle de secre´taire en Dans la premie`re moitie´ des anne´es vingt, c’est Nadine Somov qui a rempli le ro Europe. 81

60

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

`a Berlin, Nadine Somov et ses aides Boris Ermolov et Vladimir Rayevsky.86 En 1924, la section berlinoise fut transfe´re´e `a Paris avec ses boursiers, et le comite´ prit possession d’une partie des locaux de l’Institut slave, rue Michelet: la section parisienne devint alors ` son apoge´e, vers le milieu des anne´es vingt, le sie`ge principal du comite´ en Europe. A l’activite´ du comite´ prit une ampleur conside´rable: `a titre d’exemple, en 1924–25, il financ¸ait environ quatre cents ´etudiants, pour la plupart en France, et cinq cents enfants dans les ´ecoles en Bulgarie. Parmi les boursiers en France, on comptait entre autres le litte´raire et slavisant Dimitrij Stremouchov et l’e´gyptologue Michel Malinine. Vers la fin des anne´es vingt, le comite´ a fortement re´duit son activite´; en 1934, il y avait encore quelques boursiers, mais en 1934–35, la distribution de bourses cesse de´finitivement. Il s’agissait moins, selon Boris Ermolov, secre´taire exe´cutif du comite´ pour l’Europe occidentale, d’un de´clin que d’une nouvelle orientation ne´cessite´e par les traˆt inte´reˆt, de´sormais, `a subventionner vaux de Whittemore `a Istanbul: le comite´ avait pluto les ´etudiants en art byzantin ou en langue grecque, en relation directe avec le Byzantine Institute of America qui avait `a l’e´poque trois ans.87 Mais la ve´ritable raison expliquant la cessation des activite´s du comite´, c’est que Whittemore cherchait alors `a collaborer avec ` il a repris les savants sovie´tiques: `a la fin des anne´es vingt, il a voyage´ en U.R.S.S. ou contact avec des arche´ologues, des historiens de l’art et des restaurateurs. Il en a rapporte´ plusieurs ouvrages de´dicace´s, actuellement conserve´s `a la Bibliothe`que byzantine `a Paris. ` une feuille colle´e porte Ainsi d’un exemplaire de Orient oder Rom de Josef Strzygowski, ou un texte dactylographie´ signe´ de la main de Whittemore: “This copy of Orient oder Rom was sent by Strzygowsky to D. Ainalov. Ainalov gave it to me in 1925 when I was seeing him in Leningrad daily. He was living in two rooms near the porte coche`re of the Cheremetev Palace on the Fontanka. The pencil notes are in the hand of Ainalov.” 88 Du nouvel inte´reˆt de Whittemore pour le de´veloppement de la science sovie´tique en matie`re de conservation et de restauration des monuments de l’ancien art russe et de l’art byzantin, te´moignent son texte intitule´ “The Place of Four Icons in the History of Russian Art” dans le Deuxie`me recueil de´die´ `a la me´moire de The´odore Uspenskij,89 aussi bien que des documents conserve´s `a la Bibliothe`que byzantine, notamment une traduction anglaise du texte d’Alexandre Anissimov sur la pe´riode pre´mongolienne de la peinture russe, faite par Vladimir Rayevsky en 1928.90 Paralle`lement, Whittemore s’est naturelle86 Le bref re´sume´ ici propose´ de l’activite´ du Comite´ s’appuie sur une premie`re ´etude du fonds du Committee for the Education of Russian Youth in Exile, conserve´ actuellement dans les Bakhmetieff Archives, Rare Book and Manuscript Library, Columbia University, New York (BAR/CERYE). Il contient 35 000 unite´s de conservation et n’a encore jamais ´ete´ ´etudie´. 87 “Dear Mr. Gano, I acknowledge with thanks the receipt of your letter of April 16th with a cheque for $134—to be distributed among our stipendiates in May. In sending them this subvention I have warned them that the June subvention will be the last and that under no circumstance the Committee can take the responsibility of a further prolongation of its help. Personally, I believe that if such a help could be possible next school year, we should direct it in a different way and give subventions only to students having a certain connection with the work of the Byzantine Institute: studying Byzantine Art, Greek language, and who could usefully assist the B. I., if necessary, in future” (B. Ermolov `a S. Gano, le 1 mai 1934, BAR/CERYE, box 12, folder Correspondence 10). 88 BB (D. VI. 11). 89 L’Art byzantin chez les slaves: l’ancienne Russie, les slaves catholiques (Paris, 1932), 212–18. 90 A. Anissimov, The Pre-Mongolian Period of the Old Russian Painting (Voprosy restavracii, vol. 2), trad. V. N. Rayevsky (1928), soixante-douze pages dactylographie´e (BB).

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

61

ment efforce´ de faire disparaıˆtre toute mention de ses relations avec les ´emigre´s russes.91 Cette volonte´ contribue `a expliquer que son activite´ de soutien aux ´emigre´s russes soit reste´e relativement me´connue des chercheurs travaillant sur l’histoire de l’e´migration.92 Whittemore et le soutien aux re´fugie´s de guerre Le Byzantine Institute of America trouve donc son origine dans le Committee for the Education of Russian Youth in Exile; mais, si l’on recule dans le temps, on s’aperc¸oit que celui-ci, `a son tour, ne fut que la continuation d’autres institutions inspire´es par Whittemore, qui se succe´de`rent `a partir de 1916. On en trouve un aperc¸u dans un texte de Seth Gano date´ du 22 de´cembre 1924 et intitule´ Committee for the Education of Russian ` l’on apprend l’existence de deux institutions ante´rieures: Refugees in Youth in Exile,93 ou Russia, et le Committee for the Rescue and Education of Russian Children. Cette succession a correspondu aux trois phases de l’activite´ de Whittemore: actions ge´ne´rales de soutien entre 1916 et 1919 (Refugees in Russia); coexistence entre ces actions et des actions ´educatives entre 1919 et 1921 (Committee for the Rescue and Education of Russian Children);94 actions uniquement ´educatives `a partir de 1922 (Committee for the Education of Russian Youth in Exile). Les documents conserve´s dans les archives de Columbia University `a New York permettent de retracer l’activite´ de Whittemore en Russie avant la constitution en 1922 du Committee for the Education of Russian Youth in Exile. Le de´but de cette activite´ remonte `a 1915 lorsque, apre`s la de´faite militaire russe, de grandes masses de re´fugie´s se pre´cipite`rent vers l’est de la Russie. Ce sont eux dont Whittemore, `a l’origine, se pre´occupa: “Thomas Whittemore’s work for Russia began in 1915, when during the retreat from Galicia of the Russian Army under Grand Duke Nicholas, the country was devastated, the cities burned, and the people driven from their homes, to prevent provisioning of the advancing enemy. A tide of millions of destitute peasants, mostly women, old men and children, swept into Russia in a great migratory wave, driven before the army, and rivers of human misery poured over the land. Typhus, cholera and typhoid thinned their ranks. Starvation and untold suffering were their daily lot. Mr. Whittemore, who was in the Balkans when the disaster came, hastened, and did what he could single91 Boris Ermolov avanc¸ait cette raison pour expliquer le refus de Whittemore, en 1933, de communiquer le nombre de boursiers soutenus par le Committee for the Education of Russian Youth in Exile en Europe Occidentale entre 1921 et 1932, et la somme totale de´pense´e par son comite´, `a Michail M. Fedorov, pre´sident de l’Association d’aide aux ´etudiants ´emigre´s russes (Paris), qui pre´parait un ouvrage consacre´ au travail ´educatif parmi les ´emigre´s russes au cours des dix dernie`res anne´es: “He does not want now his name to be published in connection with anything dealing with Russian emigrants abroad, which is quite natural. . . . I think that the actual archaeological work of Mr. Whittemore and his relations with Soviet scientists obliging him to avoid conspicuous connections with Russian emigrants would be a sufficient explanation” (B. Ermolov `a S. Gano, le 8 fe´vrier 1933, BAR/CERYE, box 12, folder Correspondence 10). 92 Le nom de Whittemore est absent dans l’ouvrage de N. Struve, Soixante-dix ans d’e´migration russe, 1919– 1989 (Paris, 1996). M. Raeff ne lui consacre que quelques lignes dans Russia Abroad: A Cultural History of the Russian Emigration, 1919–1939 (New York–Oxford, 1990; ´edition russe [Moscou, 1994], 43). 93 BAR/CERYE (box 17, file 6). 94 Selon Gano, le comite´ s’intitulait entre 1919 et 1922 “Committee for the Rescue and Education of Russian Youth in Exile,” tandis que, dans la plupart des documents de l’e´poque, il est mentionne´ comme “Committee for the Rescue and Education of Russian Children.”

62

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

handed, with some help from the Grand Duchess Tatiana Committee, to relieve the suffering.” 95 Auparavant, Whittemore avait de´ja` eu l’expe´rience du proble`me des re´fugie´s de ` il ´etait alle´ dans le cadre de ses inte´reˆts arche´oguerre, d’abord en Bulgarie en 1912, ou 96 ` il se trouva au de´but de la Grande Guerre logiques, et ensuite au nord de la France, ou en “service me´dical actif ”: “In 1914, before Canteen or Red Cross service was established, Mr. Whittemore found hundreds of wounded soldiers on the quays at Calais, waiting for transportation to England. They were hungry and cold; nothing was being done for them. He bought tea and biscuits and induced the people to bring hot water and help him to feed these men. The people of the place kept up their work, while Mr. Whittemore went to Boulogne and other ports to institute the same service. Going farther into France, he found ‘acres of wounded’ undergoing operations without anaesthetics. He cabled to America for money and then travelled between London, where supplies could be bought, and the improvised French Hospitals, carrying chloroform and other necessities.” 97 Ce texte, tire´ d’un document au titre ´eloquent, Facts to be used wholly or in part in introˆ ˆetre re´dige´ selon les consignes de Whittemore luiducing Mr. Thomas Whittemore, a du meˆme, qui ´etait extreˆmement attentif `a la construction de sa propre image et `a la fortune posthume de son œuvre: plus tard, il laissera des dispositions pre´cises concernant son legs litte´raire afin qu’il soit examine´ et ´eventuellement publie´ sous forme de “historical record.” 98 Dans ce document, l’œuvre de Whittemore en Russie apparaıˆt donc comme celle d’un sauveur, d’un ange gardien, “fre`re parmi le peuple”: “He went like a brother among the people, distributing food and medicine with his own hand, consoling the sick and dying, trying to encourage the able bodied to work and improve conditions.”99 C’est en ces termes qu’est de´crit son travail pour les re´fugie´s de guerre au sein du comite´ de soutien pre´side´ par l’une des filles du tsar, Tatjana Romanov. L’ouest de la Russie ´etait de´vaste´ par l’arme´e russe qui reculait en pratiquant la politique de la terre ˆ le´e. Les populations de´place´es, affame´es, se fixe`rent dans la province de Samara, bru ` ou le gouvernement arrangea des baraques dans lesquelles on dormait sur des bancs `a quatre range´es, propices `a la propagation rapide de maladies comme le typhus. Whittemore chercha `a agir `a la fois individuellement et `a travers le comite´ de Tatjana. En tant qu’agent de ce comite´, il obtint l’autorisation de circuler librement dans toute la Russie pour organiser et surveiller l’aide aux re´fugie´s. Au cours de l’e´te´ 1916, il revint en Ame´rique et fonda son premier comite´, Refugees in Russia, avec des sections dans plusieurs villes, dont la seule permanente fut la section bostonienne. Parmi les membres de cette section figurent Mr. et Mrs. Ralph Adams Cram, Katharine P. Loring, Mary C. Wheelwright, Mrs. Roger Wolcott, et d’autres, qui tous ˆle actif dans les diverses institutions dirige´es par Whittemore, toujours joueront un ro ´epaule´ par Seth Gano comme tre´sorier.100 Whittemore devint l’agent du comite´ en Russie S. Gano, Committee for the Education of Russian Youth in Exile (22 de´cembre 1924), p. 1. “In 1912 my archaeological studies took me to Bulgaria, where after the Balkan Wars I worked to some extent with the Bulgarian refugees”: Relatia [sic] of My Journey in Russia: November–December 1915 and January & February 1916, p. 1, BAR/CERYE, box 20, folder Constantinople, 1919–22. 97 Texte dactylographie´ date´ “1917, 1918–1919,” p. 1, BAR/CERYE, box 17, file 2. 98 Testament of Thomas Whittemore, §16. Cette taˆche n’a pas ´ete´ assure´e par ses trustees. 99 Facts to be used wholly or in part in introducing Mr. Thomas Whittemore, p. 1. 100 BAR/CERYE, box 22, folder “Refugees in Russia—Boston Committee (1916).” 95 96

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

63

et utilisa l’argent rassemble´ pour les besoins des re´fugie´s.101 Il devait poursuivre cette activite´ sous le gouvernement provisoire et sous les Bolcheviks. Chaque anne´e, il faisait de brefs se´jours en Ame´rique afin de trouver de nouveaux soutiens politiques et financiers; il y donnait des confe´rences sur l’e´tat de la Russie et en rapportait de l’argent et des provisions. L’un de ces voyages est souvent mentionne´ dans les documents: Whittemore aurait quitte´ la Russie au cours de l’e´te´ 1917 et y serait revenu en novembre en passant par le Japon et la Sibe´rie, avec, comme bagage personnel, soixante caisses de provisions et des milliers de dollars.102 ˆpitaux, les L’aide apporte´e par Whittemore fut distribue´e dans les orphelinats, les ho camps de re´fugie´s et les foyers qui abritaient leurs enfants, `a Nizhnij-Novgorod, Kazan, Samara, Irkutsk, Vjatka, Vologda. Travaillant en collaboration avec les Zemstvos russes et avec des personnes inde´pendantes, Whittemore apporta son concours `a un grand nombre d’institutions et d’organisations charitables. Apre`s la Re´volution d’octobre, l’aide aux re´fugie´s devint de plus en plus difficile: des groupes d’enfants furent transfe´re´s des villes vers les campagnes, dans les manoirs de la noblesse; en juin 1918, deux mille dollars furent de´pense´s pour semer des pommes de terre; des ´ecoles furent ´egalement organise´es pour apprendre aux enfants `a survivre par eux-meˆmes: les filles apprenaient `a coudre, les garc¸ons `a fabriquer des chaussures. Whittemore et le soutien aux ´emigre´s russes Revenant en Russie au printemps 1919, Whittemore ne put prendre la route de Pe´trograd et dut passer par Constantinople pour rejoindre le sud de la Russie. Un an plus tard, le 19 fe´vrier 1920, il relatait cet ´episode dans une confe´rence au Boston Committee for Russian Relief, dans le salon du Copley Plaza Hotel `a Boston: “Mr. Chairman, ladies and gentlemen, after speaking to you here in Boston about a year ago, I went to Constantinople, expecting to find my way immediately to Russia; but it was just after the fall of Odessa, and I found that Constantinople had become a Russian city. The streets were flowing with Russians. There were grand dukes and Cossacks and peasants and princes, and I saw that there was work to be done for the Russians there, for the Russian refugees.” 103 ´ tats-Unis `a Constantinople, l’amiral Marc Sur les conseils du haut commissaire des E L. Bristol, Whittemore re´quisitionna l’e´glise allemande `a Constantinople afin de la transformer en dortoir pour les convalescents. L’objet du comite´, pre´cisait-il, ´etait de secourir en priorite´ les re´fugie´s russes qui retournaient en Russie, `a Ekaterinograd, “where the Copie d’une lettre de S. Gano `a K. Loring, le 15 juillet 1919, BAR/CERYE, box 15. “He returned to Russia after a short trip to America in November 1917, by way of Japan and Siberia, taking shoes and surgical supplies. In Japan, where he had a committee organized and working for him, he secured drugs, hospital supplies and clothing, although an embargo had been placed by the allies on all goods destined for Russia, he was able to enter and cross Siberia with sixty cases of supplies as luggage, jealously guarded by him in the baggage car. On the way he met the American Embassy staff coming out of Russia, but found the American Ambassador, Mr. Francis, at Vologda and continued his work with the sympathetic assistance of Mr. Francis and of the late Mr. Madden Summers, Consul General.” Gano, Committee for the Education of Russian Youth in Exile, pp. 1–2. 103 Address of Charles Eliot & Thomas Whittemore, le 19 fe´vrier 1920, stenographic record, p. 2, BAR/CERYE, box 20. Sur Constantinople, “ville russe,” voir P. Dumont, “Les anne´es blanches,” en S. Yerasimos, ´ed., Istanbul, 1914–1923: capitale d’un monde illusoire, ou l’agonie des vieux empires (Paris, 1992), 184–203; et J. Deleon, The White Russians in Istanbul (Istanbul, 1995). 101 102

64

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

` partir de ce moment, l’actigreat General Denikin had his headquarters” 104 (Fig. 26). A vite´ de Whittemore change d’orientation: de´sormais, elle est lie´e aux Russes “blancs,” aux ´emigre´s, et se recentre sur le sud de la Russie et sur Constantinople, d’abord, puis seulement sur Constantinople, et, `a partir de 1922, sur les pays d’accueil de la diaspora russe en Europe. Toutes les couches de la socie´te´ et tous les niveaux culturels ´etaient repre´sente´s parmi les Russes rassemble´s `a Constantinople, ce qui, pour les organisations de soutien, constituait une situation comple`tement nouvelle. Whittemore fut, semble-t-il, un des premiers `a s’en rendre compte et `a re´orienter, en conse´quence, son activite´ vers les taˆches culturelles et ´educatives, dont la premie`re fut la construction d’e´coles.105 Cette nouvelle strate´gie, apre`s avoir rencontre´ quelques re´ticences `a Boston,106 finit par s’imposer; dore´navant, le soutien offert par Whittemore aux re´fugie´s russes se situera exclusivement dans le domaine de l’e´ducation, d’abord pour les enfants des re´fugie´s, ensuite pour les ´etudiants. Cette orientation n’e´tait pas de´nue´e d’intentions politiques: l’intervention ame´ricaine pour la sauvegarde de la culture russe, selon Whittemore, devait empeˆcher que, conforme´ment `a la tradition, l’Allemagne s’assuraˆt la primaute´ sur ce terrain. Ainsi expliquait-il `a Katharine Loring le besoin urgent d’organiser l’impression de livres et de manuels en langue russe: “Leipzig printed great quantities of Russian books before the war. Leipzig today will print Russian books cheaper than they can be printed elsewhere. Shall we let Russia go back to Leipzig for her books, or shall we help her establish new printing relations in America and England? This is an indication of how all Russian life will flow back to Germany unless new canals are out for it.” 107 Aider la Russie pour faire obstacle `a la supre´matie de l’Allemagne et pour accroıˆtre ´ tats-Unis: telles ´etaient les motivations qui ont pousse´ Whitteˆle international des E le ro more en Russie pendant la Grande Guerre et pendant les premie`res anne´es qui ont suivi la Re´volution d’octobre; tel ´etait, du moins, le sens de ses confe´rences sur l’e´tat de la Russie, devant le public ame´ricain, destine´es `a mettre en garde l’opinion publique contre l’e´ventualite´ d’une coalition russo-allemande et `a encourager l’engagement ame´ricain pour la “Russie ve´ritable.” Pourquoi ce souci? Parce que, au-dela` du soutien aux re´fugie´s, avec sa dimension politique, la Russie, les Russes, la vie russe constituaient un des inte´reˆts fondamentaux de Whittemore, et ceci depuis une date assez ´eloigne´e: selon Igor Grabar, Whittemore ˆnes en Russie de`s 1911.108 Cette connaissance de la Russie, nourrie par avait ´etudie´ les ico une curiosite´ et une sensibilite´ aigue¨s, a fait de Whittemore un te´moin privile´gie´ des anne´es cruciales de la guerre et de la Re´volution. Voici comment Charles W. Eliot, par exemple, le pre´sentait lors de sa confe´rence de fe´vrier 1920 `a Boston: Address of Charles Eliot & Thomas Whittemore, p. 3. Le 13 mai 1920, il demande `a son tre´sorier de l’argent pour acheter des livres et d’autres objets scolaires et pour organiser l’impression de manuels en cyrillique (caˆblogramme de T. Whittemore, de Constantinople, `a S. Gano, c/o Lee Higginson & Co., Boston, le 13 mai 1920; BAR/CERYE, box 15, folder Correspondence Whittemore II, 1918–21). 106 Voir une copie de la lettre de S. Gano `a Whittemore, le 21 juillet 1920, BAR/CERYE, box 15, folders Gano–Whittemore. 107 Brouillon de lettre de Whittemore `a K. Loring, American Embassy, Constantinople, le 16 mai 1920, p. 2, BAR/CERYE, box 15. 108 I. Grabar `a V. M. Grabar, le 30 avril 1929, en I. Grabar, Lettres, 1917–1941 (Moscou, 1977), 214. 104 105

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

65

Mr. Whittemore is much the most interesting and convincing witness on Russia that I have heard during the last six years. He is the only witness, coming from Russia, who has been able for four years past to understand and speak Russian, and to find his way about the country without any guide whatever, and to deal with all sorts and conditions of men directly, personally, without the intervention of any agent. That alone makes him supremely interesting as a witness to the condition of things in Russia. He began to go to Russia in 1908 and had made many studies of the Russian people, their art, their church, before the war began. And he has been steadily in Russia ever since 1916.” 109

Meˆme tonalite´ dans une lettre re´dige´e par l’amiral Bristol, `a la demande de Seth Gano, et date´e du 13 novembre 1923: “I have known Mr. Whittemore for over four years and a half. During that time he has visited Constantinople frequently while at the same time spending much time in Russia.110 From the very beginning I found him most interested in everything pertaining to Russia and also that he had been quite familiar with things in that country before the war, especially with the personal and artistic side of life, and has continued that interest ever since I have known him.” 111 Te´moin direct et par la` meˆme irremplac¸able, Whittemore pouvait proposer `a la socie´te´ bostonienne des re´cits ˆ lante actualite´, sur la situation en Russie. Lors de sa de premie`re main, et d’une bru confe´rence de fe´vrier 1920, il fit une description de´taille´e de l’e´tat dans lequel il avait ` il ´etait arrive´ six heures environ apre`s le de´part des Bolchetrouve´ la ville d’Odessa, ou ` ´etaient inviks, en venant de Constantinople: compte rendu de sa visite sur les lieux ou stalle´s la TchK et l’Osvag (bureau de propagande), sur la politique des Bolcheviks envers les enfants, enfin sur la puissance de l’Arme´e rouge et la de´faite du ge´ne´ral De´nikine, qu’il connaissait personnellement. Il conclut en ´evoquant la question de la terre et le proble`me du blocus ´economique impose´ par les Allie´s `a la Russie—strate´gie redoutable, selon lui, puisqu’elle servait surtout les Bolcheviks. Les lettres Whittemore, en forme de journal, constituent un autre te´moignage: il y de´crit, jour apre`s jour, des ´eve´nements historiques de premie`re importance, `a commencer par la re´volution de fe´vrier 1917. Il fait alterner des descriptions de sce`nes de rue, des observations d’ordre ge´ne´ral et des synthe`ses sur le “caracte`re russe,” en se fondant `a la fois sur son inte´reˆt pour les gens simples et sur ses relations avec des personnalite´s, puisqu’il connaissait aussi bien l’une des filles du tsar, le patriarche Tikhon et les ministres du gouvernement provisoire `a Pe´tersbourg, que les dirigeants bolcheviks `a Moscou et les commandants des arme´es blanches dans le sud de la Russie.112 Dans un texte publie´ en 1919 sous le titre “The Bolshevist and the Cubist” dans The Touchstone and the American Art Student Magazine, il parle de sa visite de la galerie Chtchoukine peu de temps apre`s la Re´volution d’octobre, guide´ par le ce´le`bre collectionneur. Sa description ´evoque la re´ception du cubisme, notamment de l’œuvre de Pablo Picasso, par les visiteurs russes, qui voyaient, selon lui, un reflet de l’actualite´ en Russie dans cette “destruction formelle.” Et Whittemore renche´rissait: Address of Charles Eliot & Thomas Whittemore, p. 1. ` , sans doute par une erreur de dactylographie, le mot “Russia” a On a corrige´ le texte de la lettre ou ´ete´ remplace´ par une re´pe´tition du mot “Constantinople.” 111 M. Bristol `a S. Gano, United States High Commission, American Embassy, Constantinople, le 13 novembre 1923, BAR/CERYE. 112 Voir les re´ponses au questionnaire de l’American Red Cross: American Committee for the Relief of Russian Refugees, Lee Higginson & Co., BAR/CERYE. 109 110

66

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE The creations of Picasso are preternatural, they are permeated with the anguish and terror of existence—the fear of birth and death. All the violence and change and growth of Russia since the Revolution began are mirrored here. It is not I alone who say so. Left to myself I might not have caught the analogy. It is Russians themselves who find here their own “madded” land. “Don’t you see,” Mr. Schoukine said to me one day, “It is Russia.” And surely as one stands in this room with the silent, suffering people, one discerns there in this misty mirror the dim outlines of their own enigmatic visage.113

Ce jugement sur l’œuvre “destructive” du Picasso cubiste place Whittemore en continuite´ avec une certaine tradition russe, dont te´moignent les textes de Nikolaj Berdjaev, ` titre d’exemple, voici ˇ ulkov, publie´s en 1913–14.114 A Jakov Tugendhold ou Georgij C comment Berdjaev commenc¸ait son article sur Picasso: “Quand on entre dans la salle Picasso de la Galerie S. I. Chtchoukine, on est envahi par un horrible sentiment de terreur; cette sensation est lie´e non seulement `a la peinture et au destin de l’art, mais aussi au destin de la vie cosmique.” 115 Dix ans de pre´sence en Russie, et une telle sympathie avec la sensibilite´ russe, ont ainsi fait de Whittemore un interlocuteur privile´gie´ aussi bien pour les Russes de l’inte´rieur que pour les exile´s `a Constantinople puis en Europe: avec lui ils pouvaient collaborer et ´etablir une relation de confiance, ce `a quoi rend hommage une de´dicace du pe`re Pavel Florenskij sur un exemplaire de Colonne et fondement de la ve´rite´ (Moscou, 1914) ` l’auteur re´sume l’enjeu spirituel conserve´ actuellement `a la Bibliothe`que byzantine, ou de l’activite´ de Whittemore, en ´echo `a leurs discussions sur le destin de la Russie et sur la mission de l’Ame´rique (Fig. 27): To Thomas Whittemore from the author. Emplis de tristesse et de de´sespoir, nous avons coutume de penser: l’Ouest est `a notre Ouest. Mais, tourne´e vers l’Est par la volonte´ de Dieu, notre Patrie biface voit de nouveau sa voisine, maintenant—a` son Est. L’Est de notre Est,—ne serait-ce pas ceci? J’interroge l’aˆme de l’Ame´rique avec un timide espoir. Qu’il en soit ainsi—et une nouvelle lumie`re s’ouvrira au monde. La rencontre, pour moi flatteuse, avec le fils de l’Ame´rique m’a permis de lui poser cette question. 1918 V, 9 (22), Moscou, Jour de Saint Nicolas Thaumaturge. Preˆtre Pavel Florenskij116

Cette conception de la Russie en tant que pays biface, tourne´ `a la fois vers l’Est et vers l’Ouest, a profonde´ment impre´gne´ la psychologie du peuple russe, sa culture, sa litte´rature et son art depuis les origines. Le lien traditionnel ´etabli en Russie entre identite´ culturelle et identite´ nationale explique que les milieux russes proches de Whittemore aient vu dans la “destruction syste´matique de la culture” par la re´volution et la guerre civile un danger pour la vie nationale russe toute entie`re. La destruction de la culture, T. Whittemore, “The Bolchevist and the Cubist,” The Touchstone and the American Art Student Magazine 4 ( janvier 1919), 318. 114 ˇ ulkov, “Demony i sovremennost’ (Mysli o francuzskoj ˇzivopisi)” (De´mons et modernite´: Pense´es sur G. C la peinture franc¸aise), Apollon, no 1–2 ( janvier–fe´vrier 1914), 64–75; J. Tugendhold, “Francuzskoe sobranie ˇˇcukina” (La collection franc¸aise de S. I. Chtchoukine), Apollon, no 1–2 ( janvier–fe´vrier 1914), 5–46; S. I. S N. Berdjaev, “Picasso,” Sofia, no 3 (mars 1914), 57–62. Pour les textes en langue franc¸aise, voir le dossier La Russie et Picasso, Les Cahiers du Mnam, no 4 (avril–juin 1980), 299–323. 115 Berdjaev, “Picasso,” 57; traduit du russe. 116 Stolp i utverzˇdenie istiny (BB; S VIII. 20); traduit du russe. 113

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

67

par la perse´cution des individus et par l’ane´antissement du syste`me d’e´ducation, ´etait conside´re´e comme “le plus grand crime des Bolcheviks.” 117 Afin de reconstruire cette culture, il fallait donc re´instaurer un syste`me d’e´coles `a tous les niveaux et faire rede´marrer l’activite´ culturelle en dehors des ´ecoles, en faisant appel `a l’Ame´rique et `a son messager, Thomas Whittemore: “America is the only country that seems to be in ordinary conditions of life and to her that falls the magnificent and noble task of preserving the world’s culture,” ´ecrivait une correspondante russe de Whittemore. “So it is to the Americans that I address myself with this letter and ask them to take in consideration what I have honour to present to them and to give me an opportunity of developing in detail my plan of reconstruction of Russian culture.” 118 Te´moin de l’exode des Russes `a Constantinople, l’arche´ologue ame´ricain a ´ete´ l’un des premiers `a preˆter l’oreille `a cet appel et `a vouloir “pre´server la vie nationale russe en exil” 119 pour re´pondre aux besoins profonds de la Russie: “Tout le reste est temporaire et provisoire. Ceci seulement influencera le futur de la Russie.” 120 En premier lieu, il a cherche´ `a cre´er les conditions d’une ´education en langue russe, dans des ´ecoles russes, avec des manuels et des professeurs russes, pour les enfants: “Some one may say,” ´ecrivait Whittemore `a Katherine Loring, de Constantinople, le 16 mai 1920 “why not put these children into American, French, English, and Greek schools? It is just not to do this that we work.” 121 Dans un deuxie`me temps, il a pense´ `a mettre sur pied un programme permettant aux ´etudiants russes de poursuivre leurs ´etudes dans les universite´s europe´ennes, re´pondant ainsi `a un vœu exprime´ parmi les exile´s: “The only hope for Russia is to have a fraction of the sane and cultured class return, so that the regeneration of Russia shall not be set back for another generation.” 122 Voici comment l’amiral Bristol a de´crit ce tournant dans l’activite´ de Whittemore: “When the refugees began to pour out of Russia as a result of the Bolshevik revolution, Mr. Whittemore had in mind the saving to the world and to Russia those minds which had developed art, literature and the sciences in Russia. In carrying out this idea he desired to select from the refugees those young men who could continue their education or be educated in the arts and sciences, and, where possible, under the old professors from Russia, so that some day, if it were possible, these young men thus educated could go back to assist Russia in her reconstruction. In this education it was desired that these young men should retain the best of Russian psychology and of Russian sentiment and thus have at heart the true interests of Russia.” 123 Dans son ensemble, l’activite´ de Whittemore au sein du Committee for the Education of Russian Youth in Exile s’inscrivait dans un projet vaste et ambitieux, qu’on pourrait qualifier aujourd’hui d’utopique: au lendemain de la de´baˆcle des arme´es blanches, il s’agissait de sauvegarder hors de Russie une “ve´ritable culture russe,” pre´serve´e de “l’esMme Bobrinskoy `a Whittemore, Constantinople, le 14 mai 1922, 1, BAR/CERYE, box 20, file “On the Bolcheviks.” 118 Ibid., p. 2. 119 Whittemore `a Miss Loring, le 16 mai 1920, 1 (“to preserve Russian National life in exile”). 120 Whittemore `a S. Gano, le 8 juin 1921, BAR/CERYE, box 15. 121 Whittemore `a Miss Loring, le 16 mai 1920, p. 1. 122 Texte signe´ par Euge`ne Nebolsine et date´ du 10 novembre 1922, p. 4 (BAR/CERYE, box 20, file “On the Bolcheviks”). 123 M. Bristol `a S. Gano, United States High Commission, American Embassy, Constantinople, le 13 novembre 1923, BAR/CERYE. 117

68

` RE BYZANTINE DE THOMAS WHITTEMORE LA CARRIE

prit de de´cadence” de certaines cultures europe´ennes. Les lettres envoye´es `a Whittemore entre 1922 et 1924 par un de ses boursiers, Vladimir Rayevsky, ´etudiant en litte´rature `a Berlin avant de devenir un des principaux collaborateurs du Committee for the Education of Russian Youth in Exile puis du Byzantine Institute of America a` Paris, te´moignent de cet ´etat d’esprit.124 Rayevsky y raconte ses ´etudes en philologie romane et germanique, faites “pour la Russie, uniquement pour la Russie . . . uniquement pour cette ´etrange, myste´rieuse Russie.” Apre`s avoir beaucoup voyage´ pendant trois ans, avoir fait de nombreuses rencontres et avoir risque´ sa vie sur les champs de batailles de la guerre civile, il de´clare que son choix d’origine, pour de telles ´etudes, s’est ave´re´ juste.125 Convaincu de la haute importance de ces ´etudes pour une future reconstruction de la culture russe, Rayevsky vouait un grand respect au travail de Whittemore contre la “mort spirituelle” de la Russie: “Russia never ceased to be great, even in its distress. But in the latter case, it could see better and clearer its very friends. And as the work of Mr. Hoover has saved the Russian population from the hunger, death, just as your work, dear Mr. Whittemore, shall certainly save Russia from the most horrid kind of death, not from physical, but from spiritual one, to which it is now approaching (as there are no persons in the country younger than 25, who could read or write, the popular education does not exist at all now), from the extinction of cultural striving.” 126 C’est dans ce contexte qu’a pu se de´velopper l’image de Whittemore comme messager de Dieu. En concluant sa confe´rence `a Boston le 19 fe´vrier 1920, Whittemore citait trois lettres qui lui avaient e´te´ adresse´es par des Russes, dont une provenant de Kharkov: “The School for Blind Children in Kharkoff has great need of everything and, kept up only by gifts, is in a critical position actually because of the civil war and unrest in Russia. Unhappy blind children starve and suffer from the cold and lack of clothes. But in this time God himself thus sent a messenger from the great American nation, who provided the children with material for clothes at his first coming, and now as winter approaches he comes again, probably through the prayers of the children, and has brought with him money for the vegetables which are needed through the winter. As the curator of the school, and in the name of all the blind children I express my hearty and deep gratitude of the highly munificent institutions and its messenger.” 127 De`s le de´but de son engagement dans la Grande Guerre, Whittemore s’est pre´sente´ en messager de la grande nation ame´ricaine, sauveur des re´fugie´s. Dans les anne´es 1916– ˆte´ des re´fugie´s russes; enfin, `a partir de 1919, il l’a 19, il a pre´cise´ cette image du co ´etendue en apparaissant comme le sauveur de la nation russe toute entie`re. C’est ce ro ˆle charismatique qu’il endosse sur une photographie prise lors d’un repas de Noe¨l en 1926, 124 Dans la lettre `a Whittemore de Berlin le 15 juin 1922, V. Rayevsky parle de la de´cadence de la culture allemande, impre´gne´e d’un esprit de revendication, en la comparant avec la culture russe qui, tout en ne´cessitant un “spiritual revival,” est pourtant davantage digne d’espoir (BAR/CERYE, box 15). 125 “One can never know one’s self without comparing with others similar ones, that is, why here in exile we feel ourselves more Russians than when having been at home; then it was not remarked, then it seemed a self-evident truth. To bring there the knowledge of Western soul expressed in language, art, literature, to expose it in the whole beauty of its historical development, to show the ways of foreign thought by wakening conscience of our own ways, to serve as a little piece of gravel in the mighty process of building Russian culture, is that really quite useless task, is that complete aimlessness, as the noisiest authorities of our time ‘try to prove sometimes’” V. Rayevsky `a Whittemore, Berlin, le 14 mars 1924, BAR/CERYE, box 15. 126 Ibid. 127 Address of Charles Eliot & Thomas Whittemore, p. 22.

´ MI LABRUSSE AND NADIA PODZEMSKAIA RE

69

avec Nadine Somov et les ´etudiants boursiers du Committee for the Education of Russian Youth in Exile (Fig. 28). Lors de re´unions avec ses boursiers, Whittemore prononc¸ait des discours sur la “ve´ritable Russie,” dans lesquels il les persuadait que leurs ´etudes ´etaient destine´es `a servir plus tard la reconstruction de leur patrie en souffrance et les confirmaient ainsi dans la certitude de l’utilite´ de leur travail pour la nation. En te´moignent les lettres de Rayevsky: “You have spoken once to us about this process of the formation of the Russian State and now the persons come from there, relate us about the existence of an other national face of Christian Russia, which is developing little by little under the present unnational and unnatural [sic] Government and will appear soon out of the present chaos and deformation. The important role of every educated Russian in the period of the Russian Revival is quite clear.” 128 L’e´nergie de´ploye´e par Thomas Whittemore dans son action pour la Russie et pour les Russes entre 1916 et 1924 s’appuyait sur la conviction que, ce faisant, il influenc¸ait directement le processus historique et contribuait meˆme `a le cre´er. Cre´ation, vie, myste`re—voila` les mots-cle´s qu’on retrouve dans tous les documents de l’e´poque pour expliquer le contenu et l’enjeu de son activite´: Starvation and cold will continue to support anarchy. Organized outside relief is now barred from Russia; but this relief, carried by a well known and trusted individual and dispensed mainly through the Russian people themselves offers a safe and sure means of helping the real Russians to LIVE and to make a beginning of self government.129 Now, what is happening in Russia? . . . Russia is changing from a government to a state. It was a government. It has been nothing but a government since Peter the Great started. . . . Now, Russia is passing out of that paternal despotism and Czardom, through this destructive despotism of Bolshevism into freedom in the form of a state. Russia will certainly reappear among the nations in some definite and powerful form of federal union, and it may have, as some think it will have, the casting vote in Europe. . . . I think it is all in the hands of life, you may say. Life is going to solve the question, and life will be equal to the situation. Life is taking it in hand. The form of a state is a mystery, and that is why we cannot understand what is going on in Russia today.130

L’espoir de changements imminents en Russie nourrissait cette conviction, qui a donc ˆ progressivement s’amenuiser dans la seconde moitie´ des anne´es vingt. Avec elle, l’acdu tivite´ du Committee for the Education of Russian Youth in Exile a perdu sa raison d’eˆtre et, paralle`lement, au de´but des anne´es trente, la restitution `a l’humanite´ des mosaı¨ques de Hagia Sophia et la rede´couverte de la civilisation byzantine est devenue le moteur intellectuel et spirituel de l’activite´ de Whittemore. Mais il n’y a pas eu de solution de continuite´ avec ce qui avait ´ete´ conc¸u, pre´pare´ et ´elabore´ au cours des deux de´cennies pre´ce´dentes, en lien avec les ´emigre´s russes: la preuve en est qu’ils ont continue´, de fac¸on certes moins visible et souvent anonyme, `a collaborer de manie`re active au sein du Byzantine Institute of America (Fig. 29). Paris V. Rayevsky `a Whittemore, le 18 novembre 1923, BAR/CERYE, box 15. Facts to be used wholly or in part in introducing Mr. Thomas Whittemore, p. 2. 130 Address of Charles Eliot & Thomas Whittemore, pp. 18–19, 21. 128 129

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Le patriarche de Moscou Tikhon, Moscou, photographie dédicacée à Thomas Whittemore, 1918 (photo: Bibliothèque byzantine, Paris)

2

Isabella Stewart Gardner (en haut à droite) et Okakura Kakuzo (en bas au milieu) avec trois autres personnes, Gloucester, Massachusetts, 1910 ( photo: Isabella Stewart Gardner Museum, Boston)

3

John Briggs Potter, Matthew Stewart Prichard, Boston, 1905, plume et aquarelle sur papier, 45 × 31,3 cm ( photo: Isabella Stewart Gardner Museum, Boston)

4

Matthew Stewart Prichard en blouse d’artiste, Lewes House, Sussex, vers 1895 (photo: Isabella Stewart Gardner Museum, Boston)

5

6

Henri Matisse, Matthew Stewart Prichard, Issy-les-Moulineaux, avril-mai 1914, eauforte sur vélin, 20 × 15 cm, épreuve numérotée et signée, avec une dédicace autographe “à Mr. Pritchard [sic] en témoignage de ma profonde sympathie. Henri Matisse” (collection particulière)

Thomas Whittemore assistant au déplacement d’un fragment architectural à Hagia Sophia, Istanbul, 1934 ( photo: Bibliothèque byzantine, Paris)

7

Thomas Whittemore sur les échafaudages, avec deux restaurateurs, devant la mosaïque du tympan de la porte royale du narthex de Hagia Sophia, Istanbul, 1935

8

Matthew Stewart Prichard, première page d’un cahier manuscrit avec une partie du texte de The Mosaics of St. Sophia at Istanbul: Preliminary Report, 1933 (Dumbarton Oaks)

9

Tympan de la porte royale du narthex de Hagia Sophia, tiré de Thomas Whittemore, The Mosaics of St. Sophia at Istanbul: Preliminary Report, 1933 (photo: Archives Henri Matisse, Paris)

10

Croix de la voûte du narthex de Hagia Sophia, tiré de Thomas Whittemore, The Mosaics of St. Sophia at Istanbul: Preliminary Report, 1933 (photo: Archives Henri Matisse, Paris)

11

Thomas Whittemore, page de garde de The Mosaics of St. Sophia at Istanbul: Preliminary Report, 1933, avec une dédicace autographe: “À Henri Matisse, Hommages respectueux, Thomas Whittemore” (photo: Archives Henri Matisse, Paris)

12

Henri Matisse, La Terrasse à Saint-Tropez, 1904, huile sur toile, 72 × 58 cm ( photo: Isabella Stewart Gardner Museum, Boston)

13

Henri Matisse, Thomas Whittemore, Paris, janvier 1937, fusain sur papier, 51,2 × 39 cm (collection particulière)

14

Henri Matisse, Thomas Whittemore, Paris, janvier 1937, fusain sur papier, 50 × 38,5 cm (collection particulière)

15

Henri Matisse, Thomas Whittemore, Paris, janvier 1937, plume et encre noire sur papier, 38 × 28,5 cm (photo: musée Matisse, Nice)

16

Henri Matisse, Thomas Whittemore, Paris, janvier 1937, plume et encre noire sur papier, 38 × 28,4 cm ( photo: Fogg Art Museum, Harvard University Art Museums, Cambridge, Mass., bequest of Thomas Whittemore)

17

18

Thomas Whittemore et trois autres personnes avec un moine du monastère russe Saint-André, Mont Athos, vers 1925 (photo: Bibliothèque byzantine, Paris)

Thomas Whittemore entouré de moines, Mont Athos, vers 1925 (photo: Bibliothèque byzantine, Paris)

19

Thomas Whittemore lors d’une procession de reliques, Patmos, mardi de Pâques 1938 ( photo: Bibliothèque byzantine, Paris)

20

Étude d’effets de lumière à Hagia Sophia, Istanbul, 1944, photographie prise sur les instructions de Thomas Whittemore (photo: Archives Henri Matisse, Paris)

21

22

Thomas Whittemore (à droite) sur un chantier des fouilles de l’église de Belovo, Bulgarie, été 1924 (photo: Bibliothèque byzantine, Paris)

Boris et Dmitrij Ermolov, Sofia, Bulgarie, janvier 1923, dédicacé “to our dear friend Prof. Th. Whittemore from his very thankful and affectionate Boris and Dmitri Ermoloff ” (photo: Columbia University, Rare Book and Manuscript Library, New York)

23

Vladimir Rayevsky, vers 1920 ( photo: Columbia University, Rare Book and Manuscript Library, New York)

24

Thomas Whittemore, Boris Ermolov, Vladimir Rayevsky, et Nicolas de Witte, Paris, église russe Saint-Serge, janvier 1945 (photo: Bibliothèque byzantine, Paris)

25

Henri Grégoire, André Frolov, Boris Ermolov (secrétaire général) et Vladimir Rayevsky dans la salle du Byzantine Institute à l’École des langues orientales, Paris, vers 1948 (repris de: Cent-cinquantenaire de l’École des langues orientales. Histoire, organisation et enseignements de l’École nationale des langues orientales vivantes, Paris, 1948, 411, pl. 3)

26

27

Le général Denikine et ses troupes, Essentouki, vers 1920, collection Thomas Whittemore ( photo: Bibliothèque byzantine, Paris)

Père Pavel Florenskij, dédicace à Thomas Whittemore sur la page de garde de Colonne et fondement de la vérité ( Moscou, 1914) (photo: Bibliothèque byzantine, Paris)

28

Thomas Whittemore et Nadine Somov entourés des boursiers du Committe for the Education of Russian Youth in Exile, Paris [?], dîner de Noël 1926 ( photo: Columbia University, Rare Book and Manuscript Library, New York)

29

Alexandre Iakovlev (1887–1938), Thomas Whittemore, dessin dédicacé à Thomas Whittemore, 1937, crayon sur papier, localisation inconnue (repris de: Alexis Gritchenko, My Encounters with French Artists: Memoirs of an Artist, New York, Alexis Gritchenko Foundation, 1968)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Invention of the Counterweight Trebuchet: A Study in Cultural Diffusion PAUL E. CHEVEDDEN

he counterweight trebuchet represents the first significant mechanical utilization of gravitational energy. In the military realm, this artillery weapon played a significant role in warfare across Eurasia and North Africa. It unleashed a revolution in siegecraft and provided the impulse for dramatic changes in military architecture to counter the greater destructive force of gravity-powered artillery. In the political realm, the emergence of the centralized state owes something to this machine, according to Joseph Needham and Robin Yates, due to the increased resource mobilization by the state that the new technology necessitated. In the field of technology, it influenced the development of such practical devices as clockwork, as Lynn White has demonstrated. According to White, this weapon may even have affected the evolution of pure science during the Middle Ages. This subject has been taken up by Vernard Foley, who has argued that the counterweight trebuchet played a role in the greatest single advance in physical science of the medieval period, the innovations in theoretical mechanics associated with Jordanus of Nemore.1 The counterweight trebuchet was the product of a technological tradition that began in ancient China, was further advanced in the technologically sophisticated civilizations of Islam and Byzantium, and was brought to its fullest development in Western Europe. This machine was a collective achievement of four civilizations and stands as one of the greatest products of multiculturalism in the field of technology. The development of the counterweight trebuchet dramatically illustrates technological adaptation

T

1 P. E. Chevedden, “Fortifications and the Development of Defensive Planning during the Crusader Period,” in The Circle of War in the Middle Ages, ed. D. J. Kagay and L. J. A. Villalon (Woodbridge, Suffolk, 1999), 33–43; J. Needham and R. D. S. Yates, Science and Civilisation in China, vol. 5, Chemistry and Chemical Technology, pt. 6, Military Technology: Missiles and Sieges (Cambridge, 1994), 239–41; L. White, Jr., Medieval Technology and Social Change (Oxford, 1962), 103; idem, Medieval Religion and Technology (Berkeley, 1978), 14, 54, 88–89, 238, 268, 269, 283–85, 308; P. E. Chevedden, L. Eigenbrod, V. Foley, and W. Soedel, “The Trebuchet: Recent Reconstructions and Computer Simulations Reveal the Operating Principles of the Most Powerful Weapon of Its Time,” Scientific American (July 1995): 66–71. Research on this article has been supported by a grant from the National Endowment for the Humanities. I would like to thank Robert I. Burns, S. J., George T. Dennis, S. J., Les Eigenbrod, Vernard Foley, Donald J. Kagay, Werner Soedel, Sarolta A. Takacs, and Theresa M. Vann for their discussion on the topics considered in this paper. I also thank the Thesaurus Linguae Graecae at the University of California, Irvine, for running searches for Greek terms pertaining to artillery.

72

INVENTION OF THE COUNTERWEIGHT TREBUCHET

spurred by the dynamics of conflict and contact over the wide expanse of Eurasia and North Africa. The counterweight trebuchet left its mark on warfare, political institutions, technology, and on pure science, yet its origins and early development remain obscure. Current scholarship has advanced little beyond the conclusions reached more than a century ago. It has been assumed that the machine came into use around 1200 and shortly afterwards developed amazing capabilities.2 In the sketchbook of Villard de Honnecourt produced between 1220 and 1240, a gravity-powered trebuchet is depicted that utilized a counterweight box with a volume of about eighteen cubic meters. This box, according to a recent study, could have carried a mass weighing up to thirty tons. It has been estimated that a trebuchet with a mass of this capacity could launch a 100-kilogram projectile more than 400 meters and a 250-kilogram projectile more than 160 meters. With a mass half this weight, a projectile of 100 kilograms could be flung 217 meters and one of 60 kilograms hurled 365 meters. Such a high level of performance would have been astounding if it were achieved only a few decades after the introduction of the counterweight trebuchet. Rapid development of this kind strains credulity.3 For a discussion of the historical development of the trebuchet, see G. Dufour, Me´moire sur l’artillerie des ´ tudes sur le passe´ et l’avenir de l’artillerie, anciens et sur celle du Moyen Age (Paris, 1840), 87–112; L.-N. Bonaparte, E 6 vols. (Paris, 1848–71), 2:26–61; E. E. Viollet-le-Duc, Dictionnaire raisonne´ de l’architecture du XIe au XVIe sie`cles, ¨hler, Die Entwickelung des Kriegwesens und der Kriegfu ¨hrung in der 10 vols. (Paris, 1854–68), 5:218–42; G. Ko Ritterzeit von Mitte des II. Jahrhunderts bis zu den Hussitenkriegen (Breslau, 1890), 3:139–211; R. Schneider, Die Artillerie des Mittelalters (Berlin, 1910); H. Yule, ed., The Book of Ser Marco Polo, the Venetian Concerning the ¨tz im MittelKingdoms and Marvels of the East, 3d ed., 3 vols. (London, 1926), 2:161–69; B. Rathgen, Das Geschu alter (Berlin, 1928; repr. Du ¨ sseldorf, 1987), 578–638; K. Huuri, “Zur Geschichte des mittelalterlichen Geschu ¨ tzwesens aus orientalischen Quellen,” StOr 9.3 (Helsinski, 1941); White, Medieval Technology and Social ´, “Machines de jet me´die´vales,” Gladius 10 (1972): 25–43; idem, Forteresses de la Change, 102–3, 165; J.-F. Fino France me´die´vale: construction, attaque, de´fense, 3d ed. (Paris, 1977), 149–63; D. R. Hill, “Trebuchets,” Viator 4 (1973): 99–115; J. Needham, “China’s Trebuchets, Manned and Counterweighted,” in On Pre-Modern Technology and Science. Studies in Honor of Lynn White, Jr., ed. B. S. Hall and D. C. West (Malibu, 1976), 107–45; S. A. ´ tudes Song, ser. 1, Histoire et institutions, pt. 2, ed. F. Aubin Shkoliar, “L’Artillerie de jet `a l’e´poque Sung,” E (Paris, 1971), 119–42; idem, Kitaiskaia doognestrel’naia artilleriia: materialy i issledovaniia (Moscow, 1980); C. M. Gillmor, “The Introduction of the Traction Trebuchet into the Latin West,” Viator 12 (1981): 1–8; D. J. C. King, “The Trebuchet and Other Siege-Engines,” Chateau Gaillard 9–10 (1982): 457–69; R. D. S. Yates, “Siege Engines and Late Zhou Military Technology,” in Explorations in the History of Science and Technology in China, ed. Li Guohao, Zhang Mehgwen, and Cao Tianqin (Shanghai, 1982), 414–19; R. Rogers, “The Problem of Artillery,” App. 3 of Latin Siege Warfare in the Twelfth Century (Oxford, 1992), 254–73; Needham and Yates, Science and Civilisation in China; Chevedden et al., “The Trebuchet”; P. E. Chevedden, “The Artillery of King James I the Conqueror,” in Iberia and the Mediterranean World of the Middle Ages. Essays in Honor of Robert I. Burns, S.J., ed. P. E. Chevedden, D. J. Kagay, and P. G. Padilla (Leiden, 1996), 179–222; idem, “The Hybrid Trebuchet: The Halfway Step to the Counterweight Trebuchet,” in On the Social Origins of Medieval Institutions. Essays in Honor of Joseph F. O’Callaghan, ed. D. J. Kagay and T. M. Vann (Leiden, 1998), 179–222; P. E. Chevedden, Z. Shiller, S. R. Gilbert, and D. J. Kagay, “The Traction Trebuchet: A Triumph of Four Civilizations,” Viator 31 (2000): 433–86. 3 Villard de Honnecourt, Album de Villard de Honnecourt, Architecte du XIIIe sie`cle, ed. J. B. A. Lassus and A. Darcel (Paris, 1858); idem, Facsimile of the Sketch-Book of Wilars de Honecourt, An Architect of the Thirteenth Century, with comments and descriptions by J. B. A. Lassus and J. Quicherat, trans. and ed. R. Willis (Lon¨ttenbu ¨ches ms. Fr 19093 der Pariser Nadon, 1859); idem, Villard de Honnecourt: Kritisch Gesamtausgabe des Bauhu tionalbibliotek, ed. H. R. Hahnloser (Vienna, 1935); idem, Carnet de Villard de Honnecourt d’apre`s le manuscrit conserve´ `a la Bibliothe`que nationale de Paris (no. 19093), ed. A. Erlande-Brandenburg et al. (Paris, 1986); Violletle-Duc, Dictionnaire raisonne´, 5:218–42; R. Bechmann, Villard de Honnecourt: La pense´e technique au XIIIe sie`cle et sa communication (Paris, 1991), 255–72. 2

PAUL E. CHEVEDDEN

73

It is now known that medieval fortifications and defensive planning began a process of revolutionary change shortly after 1200 in order to counter the greater destructive power of the counterweight trebuchet and to exploit this new artillery for use in the defense of strongpoints.4 It is unlikely that a transition in fortification design would have followed immediately after the first appearance of the counterweight trebuchet. After all, the bastion system of defensive planning used in the sixteenth century, which arose out of a need to withstand the devastating blows of even more effective gunpowder artillery, did not emerge until well after the introduction of efficient cannon.5 One would expect the counterweight trebuchet to have undergone a process of development before defensive planners were required to come up with positive countermeasures to thwart it. Hence, an earlier dating for the introduction of the counterweight trebuchet appears likely. Before examining the historical evidence on this question, a brief introduction to the trebuchet is in order. HURLING MOUNTAINS AND HILLS By the end of the sixth century, a new class of artillery had replaced the stoneprojectors of classical antiquity.6 This class of artillery, conventionally denoted by the On the changes in military architecture resulting from the introduction of the counterweight trebuchet, see P. E. Chevedden, “The Citadel of Damascus” (Ph.D. diss., University of California, Los Angeles, 1986); idem, “Fortifications and the Development of Defensive Planning.” Clive Foss and David Winfield suggest that Byzantine military architecture underwent changes during the reign of Manuel I Komnenos (1143–80) ¨tahya (Oxto counteract the counterweight trebuchet: C. Foss, Survey of Medieval Castles of Anatolia, vol. 1, Ku ford, 1985), 77, 83; C. Foss and D. Winfield, Byzantine Fortifications: An Introduction (Pretoria, 1986), 48. For alternate views regarding the dating and purpose of some of the changes in military architecture noted by Foss and Winfield, see R. W. Edwards, review of C. Foss, Survey of Medieval Castles of Anatolia, Speculum 62.3 (1987): 675–80; Chevedden, “Fortifications and the Development of Defensive Planning.” 5 On the new system of fortifications of the gunpowder era, see J. R. Hall, “The Early Development of the Bastion: An Italian Chronology, c. 1450–c. 1534,” in Europe in the Late Middle Ages, ed. J. R. Hale, L. Highfield, and B. Smalley (London, 1965), 466–94; C. Duffy, Siege Warfare: The Fortress in the Early Modern World, 1494– 1660 (London, 1979); S. Peper and N. Adams, Firearms and Fortifications: Military Architecture and Siege Warfare in Sixteenth-Century Siena (Chicago, 1986). 6 ¨tze de Saalburg (1918; repr. Bad On the artillery of classical antiquity, see E. Schramm, Die Antiken Geschu Homburg, 1980); E. W. Marsden, Greek and Roman Artillery: Historical Development (Oxford, 1969); idem, Greek ¨mischer Ballisten and Roman Artillery: Technical Treatises (Oxford, 1971); N. Gudea and D. Baatz, “Teile Spa¨tro aus Gornea und Orsova (Ruma¨nien),” Saalburg Jahrbuch 31 (1974): 50–72; D. Baatz, “The Hatra Ballista,” Sumer 33.1 (1977): 141–51; idem, “Das Torsionsgeschu ¨ tz von Hatra,” Antike Welt 9.4 (1978): 50–57; idem, “Recent Finds of Ancient Artillery,” Britannia 9 (1978): 1–17, pls. 1–5; idem, “Teile Hellenistischer Geschu ¨ tze aus Griechenland,” AA (1979): 68–75; idem, “Ein Katapult der Legio IV Macedonica aus Cremona,” Ro¨mische Mitteilungen 87 (1980): 283–99; idem, “Hellenistische Katapulte aus Ephyra (Epirus),” Athenische Mitteilungen 97 (1982): 211–33; idem, “Katapultteile aus dem Schiffswrack von Mahdia (Tunesien),” AA (1985): 677–91; idem, “Eine Katapult-Spannbuchse aus Pityus, Georgien (UDSSR),” Saalburg Jahrbuch 44 (1988): 63–64; ¨mische Jagdarmbrust,” Archa¨ologisches Korrespondenzblatt 21 (1991): 283–90; idem, Bauten und idem, “Die Ro ´ le´ments d’une catapulte romaine Katapulte des ro¨mischen Heeres (Stuttgart, 1994); D. Baatz and M. Feuge`re, “E trouve´e `a Lyon,” Gallia 39 (1981): 201–9; A. G. Drachmann, The Mechanical Technology of Greek and Roman Antiquity (Copenhagen, 1963), 186–91; idem, “Biton and the Development of the Catapult,” in Prismata, ¨r Willy Hartner, ed. Y. Maeyama and W. G. Saltzer (WiesbaNaturwissenschaftsgeschichtliche Studien. Festschrift fu den, 1977), 119–31; B. C. Hacker, “Greek Catapults and Catapult Technology: Science, Technology and War in the Ancient World,” Technology and Culture 9 (1968): 34–50; Y. Garlan, Recherches de poliorce´tique grecque (Athens, 1974), 212–25; J. G. Landels, Engineering in the Ancient World (Berkeley, 1978), 99–132; A. W. Lawrence, Greek Aims in Fortifications (Oxford, 1979), 43–49; W. Soedel and V. Foley, “Ancient Catapults,” Scientific American 240 (March 1979): 150–60; P. Fleury, “Vitruve et la nomenclature des machines de jet romaines,” 4

74

INVENTION OF THE COUNTERWEIGHT TREBUCHET

French term trebuchet, consisted of a beam that pivoted around an axle that divided the beam into a long and short arm. At the end of the longer arm was a sling for hurling the missile, and at the end of the shorter one pulling ropes were attached, or, in later versions, a counterweight. To launch a projectile, the short arm, positioned aloft, was pulled downward by traction or gravity or by a combination of both forces. The impetus applied to the beam propelled the throwing arm of the machine upward and caused the missile to be hurled from the sling (Figs. 1–5). Three distinct forms of this artillery developed: the traction trebuchet, powered by crews pulling on ropes; the hybrid trebuchet, powered by both a pulling-crew and gravity power; and the counterweight trebuchet, activated solely by the force of gravity obtained by the fall of a large pivoting mass. The traction trebuchet, invented by the Chinese sometime before the fourth century B.C., was partially superseded at the beginning of the eighth century by the hybrid trebuchet. This machine appears to have originated in the realms of Islam under the impetus of the Islamic conquest movements. By the ninth century, the hybrid trebuchet was being used in the Middle East and the Mediterranean world, as well as in northern Europe. The introduction of the counterweight trebuchet marked a breakthrough in the development of mechanical artillery. It was the first fully mechanized pivoting-beam artillery weapon powered exclusively by the force of gravity. The elaboration of the trebuchet in its three forms increased the destructive power of mechanical artillery considerably. The most powerful Chinese traction trebuchet with a 250-man pulling-crew was capable of throwing a stone-shot weighing between 57 and 63 kilograms a distance of more than 75 meters.7 The Miracula of St. Demetrius written by John I, archbishop of Thessalonike, is the first historical work to provide a detailed description of the traction trebuchet. The Avaro-Slavs employed a battery of fifty traction trebuchets ( petrobo´ loi, petroboloi) when they laid siege to Thessalonike in 597, and the archbishop of the city graphically details the devastation wrought by them: These trebuchets [ petroboloi] had quadrilateral [trusses] that were wider at the base and became progressively narrower toward the top. Attached to these machines were thick axles plated with iron at the ends, and there were nailed to them pieces of timber like beams of a large house. Hanging from the back side of these pieces of timber were slings and from the front strong ropes, by which, pulling down and releasing the sling, they propel the stones up high and with a loud noise. And on being discharged they sent up many great stones so that neither earth nor human constructions could withstand the impacts. They also covered those quadrilateral-shaped trebuchets [ petroboloi] with planks on three sides, so that those inside launching them might not be wounded by arrows [shot] by those on the city walls. And since one of these, with its planks, had been consumed by fire from an incendiary arrow, they returned, carrying off the machines. On the following day they again brought up these trebuchets [ petroboloi] covered with freshly skinned hides and planks, and placing them closer to the city walls, shooting, they hurled mountains and hills against us. For what else might one term these immensely large stones?8 REL 59 (1981): 216–34; P. E. Chevedden, “Artillery in Late Antiquity: Prelude to the Middle Ages,” in The Medieval City under Siege, ed. I. Corfis and M. Wolfe (Woodbridge, Suffolk, 1995), 131–73; A. Wilkins, “Reconstructing the Cheiroballistra,” Journal of Roman Military Equipment Studies 6 (1995): 5–59. 7 This quantitative data is taken from the Chinese military treatise, Wu jing zong yao (The essentials of the military classics), completed in 1044. For this important data, Shkoliar, Kitaiskaia doognestrel’naia artilleriia, 358–69; and Chevedden et al., “Traction Trebuchet,” table 3. 8 John I, archbishop of Thessalonike, Miracula S. Demetrii, ed. P. Lemerle, Les plus anciens recueils des miracles de saint De´me´trius et la pe´ne´tration des slaves dans les Balkans, 2 vols. (Paris, 1979), 1:154. The rendering of this

PAUL E. CHEVEDDEN

75

How and when did the Avaro-Slavs acquire the use of the traction trebuchet? These people did not possess the technology of mechanized siegecraft when they first appeared in the Balkans in the sixth century. Theophylactus Simocatta, the historian for the reign of Maurice (582–602), tells us that they first obtained this technology through a captured Byzantine soldier named Bousas. In exchange for his life, Bousas constructed for the Avaro-Slavs a missile-throwing siege machine referred to as a helepolis (eJle´ poli", citytaker), which was instrumental in capturing the fortress of Appiareia in Moesia Inferior (in northern Bulgaria) in 587. Theophylactus states that: Bousas taught the Avars to construct a certain siege machine, for they [the Avars] happened to be most ignorant of such machines, and he built the trebuchet [helepolis] to hurl missiles. Soon thereafter the fortress was leveled, and Bousas collected judgment for their inhumanity, having taught the barbarians something frightful, the technology of besieging. Thence the enemy captured effortlessly a great many of the Roman cities by making use of this original device.9

The machine that leveled the fortress of Appiareia in 587 is most certainly the same machine that “hurled mountains and hills” against Thessalonike in 597. The term helepolis, therefore, corresponds to petrobolos (petrobolo") and refers to the traction trebuchet, which the Avaro-Slavs first acquired from the Byzantines in 587. Hence, we can assume that the Byzantines were making full use of the traction trebuchet prior to this date and place the introduction of this machine into the Mediterranean region at least as far back as the sixth century. The stone-projector that was capable of leveling fortresses and hurling mountains and hills was partially superseded, in the eighth century, by an even mightier piece of artillery: the hybrid trebuchet. This ordnance was far more powerful than its progenitor, and it easily outstripped the artillery of the classical world. In 1218, crusaders besieging Damietta in the Nile delta utilized a hybrid trebuchet that launched stone-shot weighing 185 kilograms, six times as heavy as that of the most commonly used large ancient catapults.10 Counterweight trebuchets could do much more. Their projectiles probably reached a common maximum of 300 kilograms, and during the fourteenth century there passage on the artillery used by the Avaro-Slavs at Thessalonike is based upon the translation by Speros Vryonis, Jr., in “The Evolution of Slavic Society and the Slavic Invasions in Greece: The First Major Slavic Attack on Thessaloniki, A.D. 597,” Hesperia 50 (1981): 384, but a number of changes have been made to indicate that the text describes trestle-framed traction trebuchets, not artillery of the classical world. Vryonis used the Latin word ballistrae to translate the Greek term petroboloi, leaving the reader with the impression that these machines are tension or torsion catapults, rather than traction trebuchets. Lynn White, Jr., was the first scholar to draw attention to this passage as a description of a traction trebuchet (L. White, Jr., “Technology, Western,” DMA 11: 660). A scholarly debate still rages over the date of the Avaro-Slavic siege of Thessalonike during Maurice’s reign in the 6th century, with Lemerle, Yannopoulos, and Whitby arguing for 586 and Vryonis for 597. On this dating question, see Lemerle, S. De´me´trius, 2:50–61; P. A. Yannopoulos, “La pe´ne´tration slave en Argolide,” BCH, suppl. 6 (1980): 323–71; M. Whitby, The Emperor Maurice and His Historian: Theophylact Simocatta on Persian and Balkan Warfare (Oxford, 1988), 117–21; Vryonis, “Evolution.” 9 Theophylact Simocatta, Historiae, 2.16.10–11, ed. C. de Boor and P. Wirth (Stuttgart, 1972); translated in Vryonis, “Evolution,” 388, with the following change: the term helepolis (city-taker) is translated as “trebuchet,” not as “siege engine.” For another translation of this passage, less expertly rendered, see M. Whitby and M. Whitby, The History of Theophylact Simocatta: An English Translation with Introduction and Notes (Oxford, 1986), 66. 10 Sa¯wı¯rus Ibn al-Muqaffa
, Bishop of al-Ashmunayn, History of the Patriarchs of the Egyptian Church, ed. and trans. Y.
Abd al-Ması¯h. , A. Su ¯ ryal Atiya, A. Khater, and O. H. E. Khs.-Burmester, 4 vols. (Cairo, 1942–74), 3.2:129 (Arabic text), 218 (trans.).

76

INVENTION OF THE COUNTERWEIGHT TREBUCHET

are reports of counterweight trebuchets that launched stones weighing between 900 and 1,200 kilograms.11 The counterweight trebuchet was so far superior to any piece of artillery yet invented that its introduction brought about a revolution in siegecraft that rendered existing systems of defense obsolete. This gravity-powered siege-engine could discharge missiles of far greater weight than the traction or hybrid machines, and it could do so with remarkable accuracy. The machine was thus able to deliver devastating blows against the same spot of masonry time after time, and this made it potentially capable of demolishing the strongest fortified enclosures. The introduction of the counterweight trebuchet led to an increase in the scale of warfare and produced revolutionary changes in military architecture in order to counter the greater destructive power of this new artillery. A THING WHICH AMAZED EVERYONE The earliest reference to the counterweight trebuchet (trabuchus) that has been cited in a European historical source records its use at the siege of Castelnuovo Bocca d’Adda in northern Italy in 1199.12 The earliest extant illustration of a counterweight trebuchet is found in a military manual written for Saladin around 583/1187 by Murd.¯ı ibn
Alı¯ ibn Murd.¯ı al-T . arsu ¯ sı¯.13 Relying principally on these two sources, most historians of artillery have concluded that the counterweight trebuchet was first introduced at the end of the twelfth century.14 Newly adduced historical evidence, however, suggests that the counterweight trebuchet can be traced as far back as the early twelfth century or perhaps earlier. Information previously dismissed or not taken into account indicates that attempts to produce a fully mechanized trebuchet powered exclusively by gravitational energy may have been underway by the end of the eleventh century. The latter half of the eleventh century witnessed increased military activity in the eastern Mediterranean, as the Byzantine Empire, Western European crusaders, and the sultanate of the Great Seljuqs, joined by their dependents and local Atabeg dynasties, all contended for power. The upsurge in armed conflict in this region provided a great stimulus for the development of military technology, particularly artillery. At the first major military operation of the First Crusade—the siege of Nicaea (6 May to 19 June 1097)—there is conclusive evidence that new creative approaches were being attempted in the design of heavy artillery. Anna Komnene records that her father, the emperor Alexios I Komnenos (1081–1118), constructed large trebuchets, referred to as eJlepo´ lei" (helepoleis, city-takers), of several types, “and most of them were not fashioned according to conventional designs for such machines but followed ideas which he had devised himself and which amazed everyone.” Alexios supplied the crusaders with this artillery and with it they subjected Nicaea to a continuous bombardment of stone-shot.15 Huuri, “Geschu ¨ tzwesens,” 64. Ibid., 171. 13 See below, note 63. 14 In his study of Latin siege warfare in the 12th century, Rogers found no references to counterweight trebuchets. “Such weapons were in use in the early 1200s,” he concludes, “and thus it is likely that Latins developed or acquired the knowledge of this type of artillery during the later twelfth century” (R. Rogers, Latin Siege Warfare in the Twelfth Century [Oxford, 1992], 265). 15 Anna Komnene, Alexiade: Re`gne de l’empereur Alexis I Comne`ne (1081–1118), 11.2.1, 11.2.5, ed. and trans. B. Leib, 3 vols. (Paris, 1937–45) (this edition of Anna’s chronicle is hereafter cited as AK). Two English 11 12

PAUL E. CHEVEDDEN

77

Not long after the siege of Nicaea, decisive evidence for the appearance of the counterweight trebuchet begins to appear in a variety of historical sources. The Byzantine historian, Niketas Choniates, offers very strong evidence for the existence of the counterweight trebuchet in his accounts of two sieges—Zevgminon in 1165 and Nicaea in 1184— by providing descriptive details of the machine (see below). An eye-witness account of the Norman siege of Thessalonike in 1185 speaks of “newly invented heavy artillery” being used and describes one monstrous trebuchet affectionately called “The Daughter of the Earthquake” and “Their Mother” (see below). T . arsu ¯ sı¯’s military manual presents the earliest full-length description and illustration of a counterweight trebuchet. Although this treatise was probably produced around 583/1187, the information contained in it on the counterweight trebuchet unquestionably dates from an earlier period (see below). The system of artillery nomenclature also provides persuasive evidence for the earlier existence of the counterweight trebuchet. Beginning in the second decade of the twelfth century, new terms are applied to pieces of heavy artillery in Syriac, Arabic, and Greek chronicles. Trebuchets are identified in these sources as “big,” “great,” “huge” or “frightful” machines (see below), indicating a change in technology: the introduction of the counterweight trebuchet. The building, deployment, and use of new types of heavy artillery at the siege of Nicaea in 1097 now takes on increased significance precisely because the counterweight trebuchet was about to make its debut. But can we be sure that Emperor Alexios I was engaged in the development of the counterweight trebuchet? Historians examining Alexios’s inventive efforts during the siege of Nicaea in 1097 have been unable to ascertain what he was working on. Worse, his achievements in 1097 have been discredited. Anna’s account of her father’s construction of helepoleis and their role in the siege of Nicaea has been regarded as either highly misleading or false. In his military history of the First Crusade, John France misconstrues Anna’s account of Alexios’s helepoleis by claiming that the “machines” provided by the emperor to the crusaders were merely “designs for machines,” not actual weapons. He then rejects her account on the grounds that the crusader army “had considerable knowledge of siegecraft.” 16 Randall Rogers proposes two possibilities regarding the emperor’s helepoleis: that either they played no part in the siege because they were probably never constructed, or, if they were constructed, they probably never reached Nicaea and so had no effect on the outcome of the siege. Ironically, both scholars credit Alexios with providing the crusaders with substantial support. France acknowledges that the emperor furnished “enormous” assistance, while Rogers assesses this aid as “considerable.” Both are adamant, however, that Byzantine assistance included no siege machines. Rogers concludes, “that crusader operations against Nicaea were organized and conducted within the expedition, and it seems that the siege engines involved in these efforts were built and operated by crusaders.” According to Rogers, the Byzantine contribution to the crusader siege-engines was limited to the supply of fastenings.17 translations of Anna Komnene’s chronicle can be consulted on the question of artillery at Nicaea: The Alexiad of the Princess Anna Comnena, trans. E. A. S. Dawes (London, 1928), 271–72 (hereafter AK [Dawes]), and The Alexiad of Anna Comnena, trans. E. R. A. Sewter (Harmondsworth, 1969), 336–37 (hereafter AK [Sewter]). A. Serper, “La prise de Nice´e d’apre`s la ‘Chanson d’Antioche’ de Richard le Pe`lerin,” Byzantion 46 (1976): 411–21. 16 J. France, Victory in the East: A Military History of the First Crusade (Cambridge, 1994), 162, 165, 368. 17 Rogers, Latin Siege Warfare, 16–25.

78

INVENTION OF THE COUNTERWEIGHT TREBUCHET

France’s arguments for dismissing Anna’s account are fallacious. Anna’s account clearly indicates that Alexios furnished the crusaders with actual machines, not designs for machines. Whether or not the crusaders had considerable knowledge of siegecraft is irrelevant to the question of siege machines being constructed by Alexios and supplied to the crusaders. A conclusion that these machines were not built and provided to the crusaders does not follow from the fact that the crusaders had considerable knowledge of siegecraft. During World War II, Great Britain had considerable knowledge of weaponry and military technology, yet that nation still required the supply of an enormous amount of arms and war munitions from the United States. Rogers’s arguments are equally dubious. He falsely states that the evidence regarding the engineers and craftsmen responsible for siege machinery at Nicaea is contradictory. There is no conflicting evidence regarding Byzantine assistance to the crusaders at Nicaea with respect to helepoleis. Greek sources substantiate the episode; Latin sources are silent on the matter. No Latin source provides evidence that contradicts the Greek sources for this incident. Rogers attempts to undermine the credibility of this episode by arguing that “there is no reference in Greek or Latin sources to a unit of Byzantine engineers serving with a crusading leader or with Tatikios’ archers and infantrymen positioned near Raymond of St. Gilles’s camp.” He further objects that “no such devices appear in any of the Latin sources, some of which do mention the emperor’s role in arranging the blockade of Lake Ascanius.” 18 The fact that the Greek sources do not provide additional information on this episode is no reason for discounting them, and the fact that Latin sources make no mention of it at all is no reason for rejecting the Greek accounts that do. As a final argument, Rogers wrongly claims that Anna’s terminology for Alexios’s siege machines is vague and that no such machines figure subsequently in her narrative. Both France and Rogers fail to offer an assessment of the credibility of Anna’s account of Byzantine siege machines at Nicaea based upon substantive evidence. Before attempting such an assessment, it is essential to dispel any notion that Anna’s terminology for the siege machines is vague or that the machines described at Nicaea do not appear subsequently in her chronicle. Anna states that her father undertook the construction of machines called helepoleis. This term has been interpreted by Rogers as “siege machines” and rendered as “machines” by France.19 The term helepolis was not a general name designating any siegeengine, however. It referred to the capital weapon of mechanized siegecraft in the Byzantine arsenal. It had originally been used to identify a giant mobile siege-tower. As a number of large mobile siege-towers were designed to carry battering-rams, the term helepolis came to designate the machine that later became known as the cat-castle, a mobile siege-tower fitted with a ram-mantlet. Later the term was applied to a ram-carrying mantlet.20 As the capital weapon of siege warfare changed, with the introduction of new Ibid., 21. For helepolis as “siege machine,” see ibid., 21–22; as “machine,” see France, Victory in the East, 162, 165; and as a generic term for any siege-engine, see D. Sullivan, “Tenth-Century Byzantine Offensive Siege Warfare: Instructional Prescriptions and Historical Practice,” in Byzantium at War (9th–12th c.), National Hellenic Research Foundation, Institute for Byzantine Research, International Symposium 4 (Athens, 1997), 191. The ODB (1:195) identifies the eJlepo´ lei" as a wooden tower (s.v. “Artillery and Siege Machinery”). 20 For references to helepolis as a large mobile siege-tower, see E. W. Marsden, Greek and Roman Artillery: Technical Treatises (Oxford, 1971), 70–73, 84–90; Garlan, Poliorce´tique grecque, 225–34; O. Lendle, Text und 18 19

PAUL E. CHEVEDDEN

79

or improved siege machines, the term helepolis was applied to the foremost siege-engine of the day, to keep pace with actual practice. From the sixth century onward, the term helepolis was used by a number of authors to designate heavy artillery: the trestle-framed trebuchet in its traction, hybrid, or counterweight forms.21 Anna’s terms for artillery—“city-taker” (eJle´ poli", helepolis), “rock-thrower” (petrobo´ lo", petrobolos), and “stone-thrower” (liqobo´ lo", lithobolos)—all bear a Hellenistic pedigree. These terms, however, do not refer to the artillery of the Hellenistic era. By the sixth century, the heavy artillery of classical antiquity had been replaced by the trestleframed traction trebuchet, and terms for ancient siege-engines, helepolis and petrobolos, were applied to the new machine (see above). Obviously the artillery denoted by these terms had changed fundamentally, while the terminology did not. Although the Byzantines did coin new terms for various types of trebuchets—alakation (alaka´ tion), lambdarea (lambdare´ a), manganon (ma´ gganon), manganikon (magganiko´ n), petrarea (petrare´ a), tetrarea (tetrare´ a), and cheiromangana (ceiroma´ ggana)—old-fashioned terms associated with mechanized warfare of the ancient world were widely used—helepolis, lithobolos, petrobolos, and sphendone¯ (sfendo´ nh). By the fourteenth century, Latin loan words came into vogue: praikoula (prai´koula) and tri[m]poutseto (tri[m]poutse´ to). The terms manganon and manganikon were used by some authors specifically to denote the pole-framed trebuchet and by other authors to identify any model of trebuchet. The traction trebuchet with a lambda-shaped trestle frame was termed a lambdarea or labdarea (labdare´ a), while the trebuchet with a triangular-shaped trestle frame was named helepolis, petrobolos, petrarea (or tetrarea), and sphendone¯. As the trestle-framed trebuchet evolved from traction to hybrid to counterweight forms, the names for it (helepolis, petrobolos, and petrarea/tetrarea) remained the same. The pole-framed trebuchet was identified as an alakation and a lithobolos, while the pole-framed “hand-trebuchet,” which was operated by a single man, was designated a cheiromangana. Following the Latin domination of the Byzantine Empire, the two most widely used counterweight trebuchets were identified by foreign loan words. The European bricola, a gravity-powered machine using a single pole for its frame (Fig. 5), was referred to as a praikoula, and the trestle-framed machine was called by its French name trebuchet (tri[m]poutseto).22 Untersuchungen zum technischen Bereich der antiken Poliorketik (Wiesbaden, 1983), 36–70; as a cat-castle, see I Macc. 13:43–44; as a battering-ram, see Onasander, Strathgiko´ ", in Aeneas Tacticus, Asclepiodotus, Onasander, with an English Translation by Members of the Illinois Greek Club (Cambridge, Mass., 1986), 508–9; Josephus, The Jewish War, trans. H. St. J. Thackeray (Cambridge, Mass., 1997), 2.553, 3.121, 3.226, 3.230, 5.275, 5.279, 5.281, 5.299, 5.317, 5.329, 5.473, 5.479, 6.236.26, 6.221, 6.393; Ammiani Marcellini Rerum gestarum libri qui supersunt, ed. W. Seyfarth (Leipzig, 1978), 23.4.10. In the 10th century, Leo the Deacon noted that the Romans used the term helepolis to refer to a battering-ram (Leonis diaconi Caloensis historiae libri decem, ed. C. B. Hase [Bonn, 1828], 2.7; 25.13). 21 Theophylact Simocatta, Historiae, 2.16.11, 2.18.2, 3.11.2, 6.5.4, ed. de Boor and Wirth; Michael Attaliates, Historia, ed. I. Bekker, CSHB (Bonn, 1853), 151; Nicetae Choniatae Historia, ed. J. L. van Dieten, CFHB, 11 (Berlin–New York, 1975), 18, 26, 38, 163, 168, 180, 282, 287, 298, 605, 606, 619, 624, 636 (hereafter Choniates, ed. van Dieten). For additional references to trebuchets in Byzantine sources from the 6th century onward, see G. T. Dennis, S.J., “Byzantine Heavy Artillery: The Helepolis,” GRBS 39 (1998): 99–115. 22 For tri[m]poutseto, see The Chronicle of Morea: A History in Political Verse, Relating the Establishment of Feudalism in Greece by the Franks in the Thirteenth Century, ed. J. Schmitt (Groningen, 1967), 852, 1412, 1481, 1700, 8430, 9155. For praikoula, or prekoula, see Nicetae Choniatae Historia, ed. I. Bekker, CSHB (Bonn, 1835), 174.27; 175.29. I would like to thank George T. Dennis, S.J., for the references to praikoula.

80

INVENTION OF THE COUNTERWEIGHT TREBUCHET

Greek authors clearly had a number of terms to choose from to denote the trebuchet. Some, like Anna, were partial to the old-fashioned terminology, while others used new terms or settled for a nomenclature combining old names with new.23 It is not unusual that Anna would use Hellenistic terms for the artillery of the Komnenian period, since instances abound in her text of old words being used to refer to new things.24 Descriptive details provided in the Alexiad indicate that Anna used the term helepolis to denote heavy artillery, and not a mobile siege-tower. This machine is employed like the petrobolos to batter walls, and it is used aboard ships in order to assist in the siege of a maritime city (Durazzo on the Adriatic Sea).25 Since helepolis designates the capital weapon of its time, 23 On Byzantine terms for artillery, see Huuri, “Geschu ¨ tzwesens,” 79–93, 148, 212–16; Chevedden, “Artillery of King James I,” 64–67; idem, “Hybrid Trebuchet,” 199–200; Sullivan, “Byzantine Siege Warfare,” 189–90; Dennis, “Byzantine Heavy Artillery.” 24 G. Buckler, Anna Comnena: A Study (Oxford, 1929), 488–97. 25 At the siege of Kastoria, Alexios I Komnenos used helepoleis and petroboloi to breach the walls of the city (AK 6.1.2; AK [Dawes] 138; AK [Sewter] 181). At the siege of Dristra, Alexios made a breach in the walls with helepoleis (AK 7.3.2; AK [Dawes] 173; AK [Sewter] 222). At the siege of Chios, Constantine Dalassenos battered the walls “with a host of helepoleis and petroboloi, and the ramparts between the two towers were destroyed” (AK 7.8.3; AK [Dawes] 183; AK [Sewter] 234). At the siege of Durazzo in 1081, Robert Guiscard’s “plan was to surround [the city] the moment he got there with helepoleis by land and sea—for two reasons: first, he would terrify the inhabitants; secondly, having isolated them completely, he would take the city at the first assault” (AK 3.12.2; AK [Sewter] 131). Dawes’s translation of this passage distorts the literal meaning of the Greek text and implies that helepoleis were only sited on land, not aboard ships: “His plan was to surround [the city], when he reached it, with [helepoleis] both on the land- and sea-side so as to strike dismay into the hearts of the inhabitants and also by thus hemming them in completely, to take the town by assault” (AK [Dawes] 95). Trebuchets were commonly mounted on ships in order to assist in the siege of a coastal city. According to Marsden, ship-mounted artillery could assist in such a siege in three ways: “First, it helped the besieger to block the harbour and thus cut the city off completely from outside contacts. Secondly, the besieger could compel the defenders to expend considerable energy in protecting the maritime area of their city and thus prevent them concentrating their entire defensive efforts against his assaults by land. Finally, if conditions were favourable, the besieger might actually direct his principal attack, properly supported by artillery, against a selected point on the seaward side, having taken fuller advantage of the privilege, naturally enjoyed by all besiegers, of choosing which section of the circuit to attack” (Marsden, Greek and Roman Artillery: Historical Development, 169–70). Arabic sources record the use of ship-mounted trebuchets at the sieges of Baghdad in 198/813 and 251/865, and during the Zanj revolt in 269/883 (Abu ¯ Ja
far Muh. ammad ibn Jarı¯r al-T . abarı¯, Tarı¯kh al-rusul wa-al-mulu ¯k [Annales], ed. M. J de Goeje et al., ser. 3, 15 vols. [Leiden, 1879–1901], 3:936, 1626, 2025 [the references are to series and page number, not to volume and page number]). Arabic sources note that the Almohads used sea-borne trebuchets against Christian forces defending Mahdia in 553/1158 (
Alı¯ ibn
Abd al-La¯h ibn Abı¯ Zar
, al-Anı¯s al-mut.rib bi-rawd. al-qirt.¯as, ed. C. J. Tornberg [Uppsala, 1843], 129); and crusaders used a ship-mounted trebuchet against Muslims defending Acre in 1189–92 and against Muslims attacking Acre in 1291 (Ibn al-Muqaffa
, History of the Patriarchs, 3.2:89 [Arabic text], 151 [trans.]; al-Malik al-Muayyad Abu ¯ al-Fida¯, al-Mukhtas.ar fı¯ akhba¯r al-bashar [Cairo, 1907–8], 4:25). In 1142, Emperor John II Komnenos lashed fishing boats and light transports together to form a platform on which ¨lu helepoleis were placed in order to bombard the fortifications along Lake Pousgouse¯ (Beys¸ehir Go ¨ ) (Choniates, Historia, ed. van Dieten, 38; O City of Byzantium, Annals of Niketas Choniate¯s, trans. H. J. Magoulias [Detroit, 1984], 22). During the “Fourth” Crusade, the crusaders used sea-borne trebuchets to assault Constantinople in 1203 and 1204 (Geoffrey de Villehardouin, Chronicles of the Crusades, trans. M. R. B. Shaw [Harmondsworth, 1963], 68, 70, 89; Robert of Clari, The Conquest of Constantinople, trans. E. H. McNeal [New York, 1936], 70, 71, 73). In 1241 the Genoese launched an amphibious attack on Savona using ship-mounted trebuchets (Annali genovesi di Caffaro e de’ suoi continuatori, ed. L. T. Belgrano and C. Imperiale di Sant’Angelo, 4 vols. [Rome, 1890–1929], 3:120). In 1249 the crusaders launched an assault on Damietta using trebuchets mounted on ships (M. Paris, Additamenta (London, 1639), 157; F. W. Brooks, “Naval Armaments in the Thirteenth Century,” Mariner’s Mirror 14 [1928]: 120–21; J. Hewitt, Ancient Armour and Weapons in Europe, 3 vols. [Oxford, 1855–60], 1:352). The Hundred Years’ War also witnessed the use of trebuchets mounted on vessels

PAUL E. CHEVEDDEN

81

Anna almost certainly uses the term to refer to the most powerful piece of artillery in the Byzantine siege arsenal: the trestle-framed hybrid trebuchet.26 Hellenistic usage distinguished artillery by size. A large catapult was designated a “rock-thrower” ( petrobolos) and a small catapult a “stone-thrower” (lithobolos).27 Accordingly, Anna employs these two terms to identify the two basic types of traction-powered artillery of her day. She designates the larger trestle-framed trebuchet as a petrobolos28 and the smaller pole-framed trebuchet as a lithobolos.29 On one occasion, petrobolos is used interchangeably with helepolis. This is apparently due to the fact that both machines have a trestle frame.30 The linkage of lithobolos with the pole-framed machine seems certain because Anna records the placement of this artillery piece on walls and on mobile siegetowers, the very locations where this light artillery weapon was generally sited.31 Following the siege of Nicaea, Anna records that the Byzantines made subsequent use of helepoleis at the sieges of Laodicea in 1104 and Mylos.32 Anna cites the use of helepoleis prior to Nicaea at the sieges of Aretai,33 Durazzo,34 Kastoria,35 Apollonias,36 Dristra,37 Chios,38 and Abydos.39 The conclusion that Anna’s terminology is vague or that she makes no further mention of helepoleis after the siege of Nicaea is not supported by the evidence. When referring to artillery, Anna uses the term helepolis more than any other word. The main objection to Anna’s account of Alexios’s helepoleis seems to be that it comes from a biased source and, therefore, cannot be regarded as credible. According to Rog-

(J. Sumpton, The Hundred Years War: Trial by Battle [Philadelphia, 1991], 207). The Castilian fleet under Pedro the Cruel attacked Barcelona on 11 June 1359 with trebuchets mounted on the poops of their galleys (Peter IV of Aragon [III in Catalonia], Cro`nica de Pere e Cerimonio´s, in Les quatre grans cro`niques, Jaume I, Bernat Desclot, Ramon Muntaner, Pere III, ed. F. Soldevila [Barcelona, 1971], 6.24); and the Genoese assaulted Kyrenia on the island of Cyprus on 10 February 1374 with sea-borne trebuchets (Leontios Makhairas, Recital Concerning the Sweet Land of Cyprus Entitled “Chronicle,” ed. R. M. Dawkins, 2 vols. [Oxford, 1932], 1:463–65). Marino Sanudo describes the construction of a long-range trebuchet for use on a ship, and Mariano Taccola describes and illustrates a coastal fortress being assaulted by a sea-borne trebuchet (Marino Sanudo, called Torsello, Liber secretorum fidelium crucis super Terrae Sanctae recuperatione et conservatione [Hanover, 1611], 80; Mariano Taccola, De machinis: The Engineering Treatise of 1449, 2 vols. [Wiesbaden, 1971], 1:119; 2:200). 26 AK 2.8.5, 3.12.2, 4.1.2, 4.4.4, 4.5.1, 6.1.1, 6.1.2, 6.13.1, 7.3.2, 7.8.3, 7.8.10, 9.3.3, 10.9.3, 11.1.3, 11.1.6, 11.2.1, 11.11.7, 13.3.11, 13.5.4, 13.5.6, 14.1.6; AK (Dawes) 62, 95, 98, 105, 138, 163, 173, 183, 187, 219, 258, 270, 271, 297, 330–31, 336, 361; AK (Sewter) 94, 131, 135, 141, 143, 181, 210, 222, 234, 237, 275, 319, 334, 336, 366, 404, 409, 410, 438. 27 Lawrence, Greek Aims in Fortifications, 72. 28 AK 2.8.5, 6.1.2, 7.8.3, 9.3.3, 11.2.5, 13.2.2, 14.1.6, 14.2.11; AK (Dawes) 62, 138, 183, 219, 272, 326, 361, 366; AK (Sewter) 94, 181, 234, 275, 337, 399, 438, 443. 29 AK 4.1.1, 4.1.2, 4.4.6; AK (Dawes) 98, 104; AK (Sewter) 135, 142. 30 The term petrobolos is used interchangeably with helepolis in Anna’s account of the siege of Nicaea (AK 11.1.3, 11.2.1, 11.2.5; AK [Dawes] 270–72; AK [Sewter] 334, 336, 337). 31 AK 4.1.1, 4.1.2, 4.4.6; AK (Dawes) 98, 104; AK (Sewter) 135, 142. 32 AK 11.11.7, 13.5.4, 13.5.6; AK (Dawes) 297, 336; AK (Sewter) 366, 409–10. 33 AK 2.8.5; AK (Dawes) 62; AK (Sewter) 94. 34 AK 3.12.2, 4.1.2, 4.4.4, 4.5.1; AK (Dawes) 95, 98, 105; AK (Sewter) 131, 135, 141, 143. 35 AK 6.1.1, 6.1.2; AK (Dawes) 138; AK (Sewter) 181. 36 AK 6.13.1; AK (Dawes) 163; AK (Sewter) 210. 37 AK 7.3.2; AK (Dawes) 173; AK (Sewter) 222. 38 AK 7.8.3, 7.8.10; AK (Dawes) 183, 187; AK (Sewter) 234, 237. 39 AK 9.3.3; AK (Dawes) 219; AK (Sewter) 275.

82

INVENTION OF THE COUNTERWEIGHT TREBUCHET

ers, “her story of imperial siege machinery at Nicea appears primarily intended to extol her father’s military knowledge and concern for the crusaders.” 40 France also believes that Anna sought “to praise her father at the expense of the hated western barbarians.” Again France provides a fallacious argument to prove his point. Anna, he tells us, “would have us believe that the Latins were quite unable to face a city of the strength of Nicaea and only captured it because he sent them siege engines of his own devising.” 41 Anna’s account of the siege of Nicaea is altogether different from France’s characterization of it. The city capitulated; it was not taken by assault. It is true that Anna does exhibit biases and displays a tendency to eulogize her father, but do such inclinations play a part in her account of the siege of Nicaea? And if so, do they render her version of the siege untrustworthy, particularly with respect to Alexios’s helepoleis? Anna’s allegedly biased narrative is belied by her praise of crusader engineering expertise at Nicaea (see below). John Birkenmeier has observed that Anna’s tendency to extol her father appears to be “in direct proportion to Anna’s lack of detailed information” of the events she describes.42 For the siege of Nicaea, Anna could rely on ample data to construct her narrative. She doubtless made use of eyewitness accounts and may even have used contemporary written sources.43 We are certain that the episode pertaining to Alexios’s helepoleis comes from a contemporary account, because it appears in the preface of the Panoplia written ca. 1100 by the learned monk Euthymios Zygabenos, who was closely associated with Alexios.44 Zygabenos adds that the siege-engines invented by Alexios (mhcanhma´ twn ejpi´noiai, me¯khane¯mato¯n epinoiai) were not inferior to the machines designed by Archimedes and Palamedes. The testimony of Zygabenos is significant and supports Anna’s account of her father’s artillery innovations at Nicaea. The question now turns on whether the assertion of Zygabenos is credible or not. A number of factors support the reliability of his claim. The crusaders were woefully lacking in artillery and siege equipment when they arrived on the shores of the Bosporos. At Constantinople, Anna relates that the crusaders failed in an assault on the city walls because they possessed no helepoleis.45 Apparently the crusaders proceeded on their expedition with little or no ordnance and relied on opportunistic methods for assembling batteries of machines when the situation required it. When the need for artillery arose, the crusaders depended on the initiative of their Rogers, Latin Siege Warfare, 22. J. France, “Technology and the Success of the First Crusade,” in War and Society in the Eastern Mediterranean, 7th–15th Centuries, ed. Y. Lev (Leiden, 1997), 171. 42 J. W. Birkenmeier, “The Development of the Comnenian Army” (Ph.D. diss., Catholic University of America, 1998), 5. On the credibility of Anna as a historian, see Buckler, Anna Comnena, 225–478; J. Chrysostomides, “A Byzantine Historian: Anna Comnena,” in Medieval Historical Writing in the Christian and Islamic Worlds, ed. D. O. Morgan (London, 1982), 30–46; G. A. Loud, “Anna Komnena and Her Sources for the Normans of Southern Italy,” in Church and Chronicle in the Middle Ages. Essays Presented to John Taylor, ed. I. Wood and G. A. Loud (London, 1991), 41–57. 43 AK 14.7.3–7; AK (Dawes) 380–82; AK (Sewter) 459–61. In a recent analysis of Anna’s sources, James Howard-Johnston has proposed that much of the Alexiad was the work of Anna’s husband, Nikephoros Bryennios: J. Howard-Johnston, “Anna Komnene and the Alexiad,” in Alexios I Komnenos: Papers of the Second Belfast Byzantine International Colloquium, 14–16 April 1998, ed. M. Mullett and D. Smythe, Belfast Byzantine Texts and Translations 4.1 (Belfast, 1996), 260–302. 44 Euthymios Zygabenos, Panoplia, PG 130:20; Buckler, Anna Comnena, 414 n. 3. On Zygabenos, see AK 15.9.1; AK (Dawes) 415; AK (Sewter) 500. 45 AK 11.1.7; AK (Dawes) 271; AK (Sewter) 335. 40 41

PAUL E. CHEVEDDEN

83

commanders and sought the assistance of the Byzantines for an adequate supply of ordnance for their siege operations. The crusader commanders could rely upon their own expert technicians or could hire native engineers to construct artillery. Anna, for example, credits Raymond of St. Gilles with setting up helepoleis for an assault on the walls of Nicaea. She also praises the cat-castle that he built as being constructed “very scientifically.” 46 There is no question that the crusaders could draw upon specialist technicians to construct and operate siege-engines. This does not preclude Byzantine assistance in the provision of artillery at Nicaea. The critical problem for the crusaders was assembling a sufficient number of artillery pieces and siege-engines in order to overcome a number of well-fortified cities that lay in their path. The city of Nicaea was second only to Constantinople in terms of the strength of its defenses.47 Anna relates that Alexios “had thoroughly investigated Nicaea and . . . judged that it could not possibly be captured by the Latins, however overwhelming their numbers. . . . The great strength of its walls, he was sure, made Nicaea impregnable; the Latins would never take it.” 48 So the emperor had helepoleis constructed and sent to the crusaders. The crusaders did not need additional manpower; they needed machines. Even so, Alexios did provide the crusaders with additional manpower—2,000 Byzantine shock troops ( peltast", peltasts). If the emperor responded to needs that were nonessential—additional manpower—would he not have addressed vital needs as well? Both France and Rogers argue otherwise and dismiss substantive evidence that conflicts with their position.49 Fairly objective Latin authors, such as Fulcher of Chartres, acknowledge that Alexios’s aid and counsel was essential to the crusader expedition.50 Thus, there is no reason to doubt that his assistance extended to the supply of artillery, even though Latin sources are silent on the matter. By the time the Latin sources for the First Crusade were written, relations between Byzantium and the West had soured, and this fact may explain why Latin accounts are not forthcoming regarding the full nature of Byzantine support at AK 11.1.3, 11.1.6; AK (Dawes) 270–71; AK (Sewter) 334–335. The cat-castle built by Raymond of St. Gilles was a combination of a “cat” or mantlet and a “castle” or mobile siege-tower. Inside the “cat,” expert sappers loosened “the tower at its foundations with iron instruments,” while above on the “castle” other soldiers battered the wall and fought with defenders. Because Greek, like Arabic, had no word for this composite device, Anna calls it both a “tower” and a “tortoise.” This instrument of war was not unknown in antiquity. In 211 B.C., Philip V of Macedon besieged Echinus with a cat-castle (Polybius, Histories 9.41). The evidence for mobile siege-towers at the siege of Nicaea found in Anna’s text and in the Gesta Francorum is discounted by Rogers who contends that “the short period of construction involved and the description of the attack make it very doubtful that wall-dominating siege-towers were employed” (Rogers, Latin Siege Warfare, 22). France also insists that the assaults on Nicaea did not involve the use of siege-towers (France, “Technology,” 172). 47 On the fortifications of Nicaea, see A. M. Scheider and W. Karnapp, Die Stadtmauer von Iznik-Nicaea (Berlin, 1938); Foss and Winfield, Byzantine Fortifications, 79–117, 261–81; S. Eyice, Iznik-Nicaea: The History of the Monuments (Istanbul, 1991). 48 AK 11.2.1, 11.2.3; AK (Dawes) 271–72; AK (Sewter) 335–36. 49 AK 11.2.4; AK (Dawes) 272; AK (Sewter) 336; France, Victory in the East, 122, 144; Rogers, Latin Siege Warfare, 19. 50 Fulcher of Chartres, Historia Hierosolymitana (1095–1127), ed. H. Hagenmeyer (Heidelberg, 1913), 178–79; idem, A History of the Expedition to Jerusalem, 1095–1127, trans. F. R. Ryan (New York, 1973), 80. France also admits that the crusaders may have relied on Byzantine poliorcetic expertise: “The crusaders did not lack knowledge of siege-equipment, but in the early stages they lacked technical grasp and may have owed something to Byzantine engineers” (France, “Technology,” 173). 46

84

INVENTION OF THE COUNTERWEIGHT TREBUCHET

Nicaea. It is also important to take into consideration the schematic nature of all of the accounts of the siege of Nicaea. None of our sources, Greek or Latin, provides full details of the siege. Information lacking confirmation from other sources, particularly nonLatin ones, should not be discounted out of hand.51 Zygabenos’s account of Alexios’s siege machines, even allowing for adulation and exaggeration, is important and credible. He does not adhere to the usual topoi in his praise of Alexios, such as his bravery, justice, or philanthropy, but instead emphasizes the emperor’s mechanical genius. The comparison of Alexios’s machines with the inventions of Archimedes and Palamedes indicates that the devices introduced by the emperor at Nicaea were a major innovation, not a minor step in an evolutionary chain of artillery development. While the term me¯khane¯mato¯n that Zygabenos employs to refer to Alexios’s siege-engines may designate artillery as well as other siege machines, Anna clearly links her father’s innovative efforts at Nicaea to heavy artillery (helepoleis). The Byzantine military tradition emphasized innovation and the development of new technologies applicable to warfare. Byzantine military treatises cite a number of new implements of war introduced by Byzantium: Greek fire; a portable, hand-held device for shooting this deadly substance known as a cheirosiphon (ceirosi´fwn); a new type of testudo called a laisa (lai´sa), and protection devices for expeditionary camps.52 In the field of artillery, the Byzantines reached a highly advanced stage of development. The great age of Byzantine conquest under Leo VI (886–912), Constantine VII (945–959), Nikephoros II Phokas (963–969), John I Tzimiskes (969–976), and Basil II (976–1025) placed a premium on the development of artillery, and historical sources record remarkable achievements attained by Byzantine artillery specialists. During the reconquest of Crete in 960– 961, a Byzantine trebuchet is said to have hurled a live ass over the walls of Chandax (Heraklion) to the starving Muslim inhabitants inside. Kalervo Huuri has estimated that the ass must have weighed between 120 and 200 kilograms (an adult Asian ass can weigh up to 290 kg).53 The two most powerful trebuchets of the eleventh century were Byzantine machines. The first, built by Emperor Basil II, launched stone-shot weighing between 111 and 200 kilograms; and the second, used by Emperor Romanos IV Diogenes in 1071, discharged missiles weighing 96 kilograms.54 Clearly, the Byzantine Empire was 51 For recent assessments of Anna Komnene’s account of the “First” Crusade, see R.-J. Lilie, “Der Erste Kreuzzug in der Darstellung Anna Komnenes,” in VARIA II, Poikila Byzantina 6 (Bonn, 1987), 49–148; idem, “Anna Komnene und die Lateiner,” BSl 54 (1993): 176–79; idem, Byzantium and the Crusader States, 1096–1204, trans. J. C. Morris and J. E. Ridings (Oxford, 1993), 4, 26, 34–37, 46, 51–53; J. Shepard, “When Greek Meets Greek: Alexius Comnenus and Bohemond in 1097–98,” BMGS 12 (1988): 185–277; idem, “CrossPurposes: Alexius Comnenus and the First Crusade,” in The First Crusade: Origins and Impact, ed. J. Phillips (Manchester, 1997), 107–29; idem, “‘Father’ or ‘Scorpion’? Style and Substance in Alexios’s Diplomacy,” in Mullett and Smythe, Alexios I Komnenos (as above, note 43), 1:68–132; R. D. Thomas, “Anna Comnena’s Account of the First Crusade: History and Politics in the Reigns of the Emperors Alexius I and Manuel I Comnenus,” BMGS 15 (1991): 269–312. 52 On Byzantine military innovations, see E. McGeer, Sowing the Dragon’s Teeth: Byzantine Warfare in the Tenth Century (Washington, D.C., 1995); Sullivan, “Byzantine Siege Warfare,” 179–200; G. T. Dennis, S.J., “Were the Byzantines Creative or Merely Imitative?” ByzF 24 (1997): 1–9. 53 Theodosios Diaconos, Theodosii Diaconi de Creta capta, ed. H. Criscuolo (Leipzig, 1979), 28, v. 718 ff; ¨tzwesen,” 91. Huuri, “Geschu 54 Basil II’s hybrid trebuchet (baban) was installed in the citadel of Bitlis. In 1054 the Seljuq sultan T . oghrı¨l Beg Muh. ammad transported it to Manzikert where he used it to bombard the city with stone-shot weighing between 111 and 200 kilograms (Matthew of Edessa, Patmut
iwn [Jerusalem, 1869], 142–45; A. E. Dostourian,

PAUL E. CHEVEDDEN

85

second to none in the development of artillery. Alexios built upon a legacy of advanced military technology devoted to artillery when he constructed at Nicaea helepoleis of several types, “and most of them were not fashioned according to conventional designs for such machines.” The conclusion reached by Rogers that “there is no clear evidence indicating that the Byzantines provided the crusaders with siege technology or specific machines” 55 rests upon a fallacious method of proof. To disprove that the Byzantines provided the crusaders with siege machines, one must do more than demonstrate that Byzantine nonassistance was possible. One must show that Byzantine non-assistance was more probable than their assistance. Both Rogers and France fail to make a credible case for Byzantine non-assistance with respect to artillery and other siege-engines. Their arguments rest on assertion, not evidence. Considering the affirmative evidence for Byzantine assistance, and taking into account that there is no evidence for their non-assistance, it appears exceedingly likely that the Byzantines furnished the crusaders besieging Nicaea with siege-engines, particularly heavy artillery. What kind of heavy artillery was provided? The hybrid trebuchet was the capital artillery weapon of this period. However, at the time Nicaea was besieged, it was underArmenia and the Crusades, Tenth to Twelfth Centuries: The Chronicle of Matthew of Edessa [Lanham, 1993], 87–88; Aristake¯s Lastiverts
i, Patmut
iwn Aristakisi Lastivertc
woy, ed. K. N. Juzbashan [Erevan, 1963], 92–93; idem, Aristakes Lastivertc
i’s History, trans. R. Bedrosian [New York, 1985], 103–5; Chevedden, “Hybrid Trebuchet,” 187–88; Chevedden et al., “Traction Trebuchet,” table 1). Arabic sources describe the large trebuchet of Romanos IV Diogenes in the siege train of the emperor just prior to the battle of Manzikert in 1071 (Fath. ibn
Alı¯ al-Bunda¯rı¯, Zubdat al-nus.rah wa-nukhbat al-
us.rah, ed. M. T. Houtsma, Recueil de textes relatifs `a l’histoire des Seljoucides 2 [Leiden, 1889], 42; Sibt. Ibn al-Jawzı¯, Mir¯at al-zama¯n fı¯ tarı¯kh al-a
ya¯n: al-h. awa¯dith alkha¯.s.sah bi-tarı¯kh al-Sala¯jiqah bayna al-sanawa¯t, 1056–1086 [Ankara, 1968], 148; Chevedden, “Hybrid Trebuchet,” 188; Chevedden et al., “Traction Trebuchet,” table 1). 55 Rogers, Latin Siege Warfare, 25. A tendency to discount or ignore Greek sources has impeded our understanding of the siege of Nicaea. Rogers examines “the almost supernatural appearance” of a Lombard engineer who designed a mantlet in the final stages of the siege that was successful “because the sides of its roof were better sloped.” According to Albert of Aachen and William of Tyre, it was this device that brought about the surrender of the city (Rogers, Latin Siege Warfare, 19–20, 23–24). At the time of the siege of Nicaea, the Byzantines were renowned for their construction of mantlets. The mysterious Lombard appears to have been thoroughly conversant with Byzantine methods of siegecraft and construction of siege machines. During the 12th century, a particular type of mantlet used by the crusaders was identified by a Latin compound derived from Middle Greek laisa, indicating that the crusaders were indebted to Byzantine expertise in the construction of certain siege machines. The Latin term was cercleia. The first part of the word (cerc-) means “hooped” (from circus), and the second part (-leia) is the Latin cognate of laisa, a Greek term for a mantlet “made by plaiting vine twigs of fresh cut willow branches in the form of arches” (Heron of Byzantium, Poliorketika, ed. C. Wescher, Poliorce´tique des Grecs [Paris, 1867], 207, quoted in Sullivan, “Byzantine Siege Warfare,” 196). The Greek term was itself borrowed from Slavic during the 9th century and appears in a number of Byzantine sources from the 10th through the 11th century. Neither Rogers nor France identifies this mantlet in the Latin sources. On Byzantine sapping operations with mantlets, see G. T. Dennis, S.J., Three Byzantine Military Treatises (Washington, D.C., 1985), 34–43, 316–19; E. McGeer, “Tradition and Reality in the Taktika of Nikephoros Ouranos,” DOP 45 (1991): 129–40; idem, “Byzantine Siege Warfare in Theory and Practice,” in Corfis and Wolfe, The Medieval City under Siege (as above, note 6), 123–29; idem, Sowing the Dragon’s Teeth, 159, 161, 163, 167; Sullivan, “Byzantine Siege Warfare,” 196–97, 200. On the Latin term cercleia, see Ricardus, Itinerarium peregrinorum et gesta Regis Ricardi, ed. W. Stubbs, 2 vols., Rolls Series, Rerum Britannicarum Medii Aevi Scriptores 38 (London, 1864), 1:220, 224; H. J. Nicholson, Chronicle of the Third Crusade: A Translation of the Itinerarium peregrinorum et gesta Regis Ricardi (Aldershot, 1997), 210, 213; Ambroise, L’Estoire de la Guerre Sainte, ed. G. Paris, Collection de Documents Ine´dits sur l’Histoire de la France (Paris, 1897), vv. 3203, 4688, 4817, 4827, 4833, 4931, 4936.

86

INVENTION OF THE COUNTERWEIGHT TREBUCHET

going a change. Anna relates that most of Alexios’s pieces of heavy artillery (helepoleis) were not constructed according to the usual designs. This suggests that Alexios was following a new design. Considering that the next creative approach in artillery was to result in the invention of the counterweight trebuchet, it is highly probable that the new pieces of artillery at Nicaea, “which amazed everyone,” were gravity-powered machines. The trebuchet is a complex device, an amalgamation of interdependent component parts that, when working in proper relation, produce a successful machine. The invention of the counterweight trebuchet was most probably the product of a sustained interaction of engineering techniques, creative abilities, and experimentation. The refinements incorporated in the design of the hybrid trebuchet that finally resulted in a workable counterweighted machine lie obscured from the historian’s gaze, but the process of invention doubtless involved a conceptual leap in the fundamental operation of the trebuchet, new design choices, and a trial-and-error method. One of the critical innovations that singled the transition from a traction-powered machine to a gravity-powered one was the lengthening of the sling. With the elimination of the pulling crew, space was created directly under the beam for the placement of a trough, or runway, for the sling. The trough made it possible to increase the length of the sling, and a longer sling made it possible to greatly increase the range of the machine. To enhance the power of his artillery at Nicaea, Alexios may have replaced human power with gravity power and lengthened the sling. Whatever steps he did take in the development of the trebuchet at Nicaea, they were significant enough to be ranked with the works of Archimedes, the most famous inventor of ancient Greece. The association of Alexios with Archimedes indicates that an important breakthrough in the design and construction of the trebuchet occurred at the siege of Nicaea. Given the imminent appearance of gravity-powered artillery, this breakthrough is most likely to have been the development of the first counterweight trebuchet. EVIDENCE FOR THE EXISTENCE OF THE COUNTERWEIGHT TREBUCHET Niketas Choniates Following the siege of Nicaea, evidence for the appearance of the counterweight trebuchet begins to mount. The earliest historical narrative to provide descriptive details of the counterweight trebuchet is a twelfth-century Byzantine account of Niketas Choniates. In describing the siege of Zevgminon in 1165, Niketas states that, “Andronikos [Komnenos] took charge of a ‘rock-throwing engine’ [ petrobolous me¯khanas] and by using the sling [sphendone¯], the windlass [stro´ ␸alo", strophalos], and the beam [lu´ go", lugos], shook the section of the wall between the two towers violently.” 56 This machine is unquestionably a counterweight trebuchet because it is equipped with a winch or windlass (strophalos), which neither the traction nor hybrid trebuchets require to draw down the beam in order to prepare it for launch. The word lugos, here 56 Choniates, ed. van Dieten, 134. Harry J. Magoulias has translated this passage as “Andronikos took charge of a stone-throwing engine and by using the sling, winch, and screwpress, shook the section of the wall between the two towers violently” (O City of Byzantium, 76).

PAUL E. CHEVEDDEN

87

translated as “beam,” can be interpreted as a withe, or branch,57 or several branches bound together, and so may refer to the rotating beam of the trebuchet, which can either be composed of a single shaft of wood or of several bound together. The large hybrid trebuchet of Romanos IV Diogenes had a large composite beam composed of eight spars.58 The verb from which lugos is derived has a secondary meaning of “to throw,” or “to turn,” or “play, as a joint in the socket.” 59 These actions correspond perfectly with the operation of a rotating beam of a trebuchet, and hence the identification of lugos with this component of the trebuchet appears to be certain. During the siege of Prosakos in 1197, Choniates writes that the man responsible for discharging the stone missile from the “rock-throwing engine” (petro´ bola mhcanh´ mata, petrobola me¯khane¯mata) turned the lugos around and aimed the sling, which also suggests that the lugos is the rotating beam of the trebuchet.60 Earlier, at the siege of Nicaea in 1184, Choniates states that Andronikos “positioned the helepoleis, carefully examining the sling [sphendone¯], [and] secured both the beam [lugos] and the windlass [strophalos].” 61 The De cerimoniis, which provides details of the provisioning of the failed Byzantine expedition against Crete under Constantine Gongyles in 949, indicates that block-andtackle equipment (troci´lia, trochilia) was required for the various trebuchets (manganika) used for the expedition.62 It is uncertain whether the block-and-tackle equipment was used for the assembly of the machines or for the hauling down of the throwing arm to prepare the artillery for launch. Even if the block-and-tackle equipment was used as a pullback mechanism for the throwing arm, such as we find on T . arsu ¯ sı¯’s counterweight trebuchet (see below, Fig. 1 and App. 2), there is no evidence for a windlass being used in the tenth century for the hauling down of the beam of a trebuchet. The use of such a powerful hauling apparatus for the launching of projectiles from a trebuchet in the twelfth century is a strong indicator of the existence of the counterweight trebuchet. The fact that the beam and the windlass of the helepolis used at Nicaea in 1184 must be secured also indicates that this trebuchet is a counterweight, not a traction or hybrid, machine. T. arsu ¯sı¯’s Double-Purpose Machine The earliest full-length description and illustration of a counterweight trebuchet is found in T . arsu ¯b, written around 583/1187 (Fig. 1 ¯ sı¯’s military manual, Tabs.irah fı¯ al-h. uru and App. 2).63 The portion of the treatise that deals with the counterweight trebuchet has LSJ, 1063. Chevedden, “Hybrid Trebuchet,” 188. 59 LSJ, 1063. 60 Choniates, Historia, ed. van Dieten, 506. 61 Ibid., 282; O City of Byzantium, 156. 62 De cerimoniis aulae byzantinae, ed. J. J. Reiske, CSHB, 2 vols. (Bonn, 1829–30), 1:670. Sullivan believes that the word trochilia is suggestive of a windlass and asserts that the presence of a windlass supports his contention that torsion-powered artillery was used by the Byzantines (Sullivan, “Byzantine Siege Warfare,” 199). 63 The full title of T . arsu ¯b min al-aswa¯ ¯ sı¯’s military manual is Tabs.irat arba¯b al-alba¯b fı¯ kayfı¯yat al naja¯h fı¯ al-h. uru





 wa-nashr a la¯m al-i la¯m fı¯al- udad wa-al-a¯la¯t al-mu ¯nah ı ala´ liqa¯ al-a da¯ (Instruction of the masters on the means of deliverance in wars from disasters, and the unfurling of the banners of information: Equipment and engines which aid in encounters with enemies) (hereafter Tabs.irah fı¯ al-h. uru ¯b). The autograph manuscript of this text in the Bodleian Library at Oxford University (MS Hunt. 264), hereafter cited as B, has been partially 57 58

88

INVENTION OF THE COUNTERWEIGHT TREBUCHET

unfortunately been incorrectly edited and translated by Claude Cahen.64 All subsequent scholars have drawn upon Cahen’s flawed text for an understanding of T . arsu ¯ sı¯’s counterweight machine. His defective text was widely disseminated by Bernard Lewis, who translated it into English and published it in an anthology of original source materials edited by Claude Cahen (“Un traite´ d’armurerie compose´ pour Saladin,” Bulletin d’e´tudes orientales 12 [1947– 48]: 103–63) and A. Boudot-Lamotte (Contribution `a l’e´tude de l’archerie musulmane [Damascus, 1968]). A later manuscript of this text, copied by Muh. ammad ibn Salma¯n in 790/1388, is in the Su ¨ leymaniye Library in Istanbul (Ayasofya Collection MS 2848 mu ¨ ), hereafter cited as A. An Arabic edition of the entire text, utilizing the above manuscripts, has recently been published by Karen S. ¯adir (Murd.¯ı b.
Alı¯ b. Murd.¯ı al-T . arsu ¯ sı¯, Mawsu ¯
at al-aslih. ah al-qadı¯mah: al-mawsu ¯m Tabs.irat arba¯b al-alba¯b fı¯ kayfı¯yat al-naja¯h fı¯ al-h. uru ¯b min al-anwa¯ [sic] wa-nashr a
la¯m al-i
la¯m fı¯ al-
udad wa-al-a¯la¯t al-mu
ayyanah [sic]
ala´ liqa¯ al-a
da¯, ed. Karen S. ¯adir [Beirut, 1998]). The earliest extant illustration of a counterweight trebuchet appears in Tabs.irah fı¯ al-h. uru ¯b (Oxford, Bodl. MS Hunt. 264, fols. 134v–135r) and is reproduced in Cahen, “Traite´,”pl. 3, fig. 14, and in Chevedden, “Artillery of King James I,” pl. 7. 64 The section pertaining to the counterweight trebuchet in the Bodleian manuscript of this text, which Cahen edited, contains two facing pages (fols. 134v and 135r); the upper portion is filled with lines of text and the lower portion is embellished with an illustration of a counterweight trebuchet (Fig. 1). Because the illustration stretches across both of the pages, the text does not follow the usual reading order, as Cahen supposed, with the successive lines of folio 134v being followed by the successive lines of folio 135r. Instead, the lines are arranged to be read straight across both pages, so that the continuation of the first line of folio 134v is the first line of folio 135r, and so forth. Cahen’s error was compounded by the fact that he inserted three nonexistent lacunas in the text on fol. 134v. This impeded possible correction by subsequent scholars who credited any lack of sense in Cahen’s edition and translation to lacunas in the original manuscript (Cahen, “Traite´,” 120, 142–43; Chevedden, “Citadel of Damascus,” 298–301). I thank Emilie Savage-Smith for alerting me to Cahen’s editorial error. Cahen translated T . arsu ¯ sı¯’s description of the counterweight trebuchet as follows: ˆ l-H Description d’un mangonneau persan telle que me l’a faite le shaı¨kh Abu . asan b. al-Abraqıˆ al-Iskandaraˆnıˆ. La puissance de jet en est de cinquante rat.l environ; `a la racine de son montant il y a une arbale`te jarkh; le tout est actionne´ par un seul homme, qui fait le lancement: lorsque l’homme tire la fle`che, les cordes de chanvre qui tendent la corde de l’arc atteignent son verrou; alors l’homme agence la cuiller dans un anneau fixe´ `a un socle qui retient la fle`che; puis il prend l’arc et en tire, puis de´gage la fle`che, ce qui produit le jet de la pierre. ˆte´ du Prends donc un mangonneau persan et dresse-le pour faire un lancement; creuse `a co montant une cuvette dont la profondeur soit approximativement ´egale `a la longueur des cordes de chanvre qui sont attenantes `a la fle`che; puis prends un filet de chanvre aux mailles ´etroites, et place `a ses extre´mite´s trois caˆbles de chanvre solide dont la longueur soit telle qu’ils aillent du sommet ` est la truie, jusqu’au fond de la cuvette; qu’il y ait au bout de la fle`che un anneau de la fle`che, ou de fer auquel soient accroche´s les caˆbles attache´s au filet; et que dans le filet on place des pierres ˆte´ en quantite´ correspondant `a la force des hommes qui tirent la fle`che. Au bout de la fle`che, `a co du caˆble de la cuiller, il doit y avoir un mıˆkhaˆn place´ dans une poulie pendue `a la fle`che; lorsque l’homme tire cette fle`che, apre`s avoir place´ la pierre dans la cuiller, et accroche´ le caˆble de celle-ci au crochet place´ au bout de la fle`che, il [agence] la cuiller, `a l’aide d’un crochet de fer place´ en son fond, dans un anneau fixe´ `a un socle, qui retient ainsi l’action du filet; [lacune?. . . . . . . . . . . ] sa corde avec les caˆbles qui soule`vent le filet dans un crochet fixe´ `a ces caˆbles; et lorsque les caˆbles montent en soulevant le filet, [lacune?. . . . ] la fle`che dans son conduit; il en tire, puis sur-le-champ revient `a la cuiller et la de´place suivant son seul jugement. Pour ce qui est de la traction, il y a diverses fac¸ons. Voici le dessin. On peut tirer le filet en tirant la teˆte de la fle`che, puisque celle-ci revient comme une balance (romaine) et qu’une fois tire´e elle est fixe´e. Quant `a l’arc jarkh, il doit ˆetre place´ `a la base du pied du mangonneau dans deux crochets de fer qui le tiennent; on tire la corde de l’arc, et l’ame`ne dans le verrou de son conduit. Lorsque l’homme a agence´ la cuiller dans le socle, il saisit l’arc et le de´tend (?); le filet tire la fle`che en la rappelant `a sa position primitive, et la traction est plus forte que ne serait celle des hommes, car le filet tire selon sa proportion.

PAUL E. CHEVEDDEN

89

on Islamic civilization.65 Donald Hill was so puzzled by the flawed text that he considered T . arsu ¯ sı¯’s description of the counterweight trebuchet as “too confused for us to identify it for certain as a counterweight machine.” 66 C. E. Bosworth may have been baffled as well, but that did not stop him from speculating about the design and function of the machine. He identified it as an arrow-projecting catapult. “The arrow or projectile,” he argues, “is placed vertically on the left, a heavy counter-weight on the right.” No arrow is shown in the illustration, and the crossbow of the machine is positioned facing downward. If Bosworth’s arrow or projectile were positioned on the crossbow of the machine, it would be set facing the ground, not facing upward so that the missile might be hurled into the air. Bosworth mistook the actual projectile of the machine—the stone-shot in the sling—for ´ also misunderstood the functhe counterweight of the trebuchet.67 Jose´ Frederico Fino tion of the machine by suggesting that the crossbow attached to it served as a kind of acceleration device to impart initial momentum to the beam during the first phase of the launch cycle.68 Cahen’s flawed text is not the only problem confronting those attempting to understand T . arsu ¯ sı¯’s description of this machine. Even though the depiction of his machine is B. Lewis, Islam: From the Prophet Muhammad to the Capture of Constantinople, 2 vols. (New York, 1974), 1:221–22. Lewis’s English translation of Cahen’s French rendition of T . arsu ¯ sı¯’s description of the counterweight trebuchet reads: 65

Description of a Persian Mangonel, Made for Me by Shaykh Abul-H . asan ibn al-Abraqı¯ al-Iskandara¯nı¯, with a Throwing Power of Fifty Pounds More or Less Its base is a crossbow [ jarkh], and it is operated by a single man who makes the launch. When the man pulls the shaft, the hemp cords, which stretch the bowstring, reach its bolt; then the man catches the cup in a ring fixed to a strut which holds the shaft. Then he takes the bow and shoots and releases the shaft so that the stone is thrown. Take a Persian mangonel and set it up to make a launch. Dig a hole by the side of the pole, to a depth equal to the length of the hemp cords on the shaft. Then take a close-meshed hemp net and place at its ends three strong hemp ropes, long enough to reach from the top of the shaft, where the axle is, to the bottom of the hole; at the end of the shaft there should be an iron ring, to which the ropes attached to the net are tied; and in the net stones should be placed in a quantity corresponding to the strength of the men who pull the shaft. At the end of the shaft, by the cup rope, there should be two nails placed on a windlass hanging from the shaft. When the man pulls this shaft, after having placed the stone in the cup and tied the cup rope to the hook placed at the top of the shaft, he . . . the cup with an iron hook placed at its end in a ring fixed to a strut which supports the action of the net . . . its cord with the ropes which raise the net in a hook fixed to the ropes, and when the ropes rise with the net . . . the arrow in its course. He shoots and then immediately returns to the cup and releases it according to his judgement. There are various ways of pulling it. Here is a picture of it. One may pull the net by pulling the top of the shaft, since it swings back like a steelyard and can be pulled and caught. The crossbow should be placed at the bottom of the strut of the mangonel, on two iron hooks which hold it. One draws the bow string and pulls it toward the bolt on its course. When the man catches the cup in the strut, he takes the bow and holds it so that the net pulls the shaft. He brings it back to its position. This traction is stronger than that of men, since the net draws according to its proportion. 66 Hill, “Trebuchets,” 104. Hill later modified his original assessment of T . asu ¯ sı¯’s description of the counterweight trebuchet: “although the passage is obscure, it is possible that this was a counterweight machine” (D. R. Hill,“Mandjanı¯k. ,” EI2, 6:405–6). Rogers declares the text a “confusing description of a counterweight lever device” (Rogers, Latin Siege Warfare, 263). 67 C. E. Bosworth, “Armies of the Prophet: Strategy, Tactics and Weapons in Islamic Warfare,” in The World of Islam: Faith, People, Culture, ed. B. Lewis (London, 1976), 214–15 and fig. 4. 68 ´, “Machines de jet me´die´vales”; idem, Forteresses, 149–63. J.-F. Fino

90

INVENTION OF THE COUNTERWEIGHT TREBUCHET

very clear and is accompanied by explanatory caption titles, which have been properly edited by Cahen, the nature and function of his ingenious contrivance has proved elusive. The difficulties of understanding this machine are increased by the hybrid nature of the device. T . arsu ¯ sı¯ depicts not one machine but two: a counterweight trebuchet and a large siege crossbow. In addition, the counterweight trebuchet itself performs two functions: it is both a stone-projector and a spanning device for a large siege crossbow. The wizardry of the machine is made even more compelling by the fact that it is operated by a single man. The author’s interest in gravitational energy is not motivated by a realization that the fall of a large mass can exert more force than human traction-power. Gravity power is commended only because it exerts “a constant force, whereas men differ in their pulling force.” The text clearly states that this machine has the same power as that of a traction trebuchet with a pulling-crew of about fifty men. Hence, T . arsu ¯ sı¯’s counterweight trebuchet is comparable to a medium-size traction machine. T . arsu ¯ sı¯ appears not to be a forward-looking visionary intent upon extolling the benefits of a new energy source, but a technical wizard who seeks to overawe his audience with a display of machine-magic: a multipurpose, laborsaving powered mechanism that requires only one person for its operation. The hybrid nature of T . arsu ¯ sı¯’s machine is further demonstrated by its retention of design features of the traction trebuchet: it is equipped with a short sling, similar to that of a traction machine, and it has no trough. Its retention of these features may even have been necessitated by its restriction to a single operator. Although the treatise in which this machine appears was written for Saladin around 1187, its description of the counterweight trebuchet is based on an account of a machine made for the author by an Armenian arms manufacturer, Shaykh Abu . asan ibn al-Abraqı¯, who worked in Egypt ear¯ al-H lier in the twelfth century (see App. 2). Hence, T . arsu ¯ sı¯’s counterweight trebuchet may not reflect artillery technology contemporary with Saladin’s reign (1169–93), but of an earlier period. It may even be argued that T . arsu ¯ sı¯’s idiosyncratic selection of material from his informant does not necessarily reflect the real advances made in the field of military engineering during the twelfth century. Both T . arsu . arsu ¯ sı¯ and Ibn al-Abraqı¯ (perhaps with T ¯ sı¯’s prompting) show special interest in multipurpose weapons with little or no practical utility: several variations of a lance fitted with a crossbow are described and illustrated in the treatise, as well as a crossbow with a circular shield. Hence, T . arsu ¯ sı¯’s counterweight trebuchet may reflect a Rube Goldberg tendency to produce an ingenious device that may never have found any practical use. It may not tell us much about the true progress being made in artillery design at the time the treatise was written. Nevertheless, T . arsu ¯ sı¯ presents the only pictorial evidence for the existence of the counterweight trebuchet in the twelfth century. New Terminology In the twelfth century, the emergence of the counterweight machine is confirmed by the use of new terminology in the historical sources to refer to this piece of artillery. Arabic sources of the twelfth century identify the counterweight trebuchet as a “big” trebuchet (manjanı¯q kabı¯r), a “great” trebuchet (manjanı¯q
az.¯m), ı or as a “huge” or “frightful” trebuchet (manjanı¯q ha¯il). Syriac sources use the term “great” trebuchets (manganı¯qe¯

PAUL E. CHEVEDDEN

91

rawrbe¯) to denote counterweight artillery, as does Eustathios of Thessalonike (mega´ lh mhcanh´ , “great siege-engine”). In the Latin West, the new artillery is designated by the term trebuchet, a diminutive form derived from the medieval Latin word trabuc[h]us. The word first appears in 1189 as trabuchellus, following a standard diminutive construction, and later as trabuchus (see below and App. 1).69 Scholars agree that the term trebuchet designates the new gravity-powered artillery. However, can we be certain that the other words that refer to “big,” “great,” and “huge” or “frightful” machines also denote the counterweight trebuchet? These terms may simply refer to oversized traction or hybrid trebuchets, not to counterweight trebuchets. The evolution of artillery nomenclature suggests, however, that this is not the case. In the ninth century, the term “big” trebuchet (manjanı¯q kabı¯r) had been used in Arabic sources to denote the hybrid trebuchet. During this century, this new technical term appears only in the accounts of two sieges: the siege of Amorion in 223/838 and the siege of Baghdad in 251/865. T . abarı¯ and the anonymous author of Kita¯b al-
uyu ¯n wa-al-h. ada¯iq initially used a new term, manjanı¯q kabı¯r, to designate the hybrid trebuchet, but they then reverted back to the old term, manjanı¯q, to refer to the gravity-assisted machine. The term manjanı¯q had been used in Arabic sources to denote the trestle-framed traction trebuchet since the appearance of this machine in Arabia during the lifetime of Muh. ammad. When the hybrid machine was introduced, manjanı¯q took on a broader meaning and referred to any type of trestle-framed trebuchet. The broad use of the term manjanı¯q does not reflect any reluctance on the part of Islamic societies to employ a new term for a new machine but rather the reluctance on the part of the educated elite writing the literary texts to recognize or employ new terminology. Arabic oral culture applied the term al-Ghad. ba¯n (The Furious One) to the hybrid trebuchet, but Arabic chroniclers ignored it.70 The failure of the Arabic sources to identify the hybrid trebuchet with a change of appellation was in part due to the nature of the nomenclature for artillery. When the hybrid machine appeared, the terminology for the trebuchet was already established, and this terminology was related to the configuration of the framework of the machine, a component that remained unchanged regardless of whether the machine was a traction or a hybrid model. The hybrid machine made use of gravitational energy for the first time, amplifying the muscular force of the pulling crew. Although this led to a change in the operation of the trebuchet, it did not alter the configuration of the machine’s framework. Once the novelty of the hybrid trebuchet had worn off, the term manjanı¯q was applied to both models of the trestle-framed trebuchet, and the technical term “big” trebuchet (manjanı¯q kabı¯r) was no longer applied to it. After a three-century hiatus, this technical term emerged again. Its re-emergence is not likely to indicate the existence of familiar technology but the introduction of something new: the counterweight trebuchet. In addition, other terms were applied to artillery in the twelfth century that emphasized either great size or great power (see App. 1). This terminological change is reflected not 69 For references to counterweight trebuchets employed during the 12th century, see App. 1. For a discussion of the system of nomenclature for the trebuchet employed during the 13th century, see below and Chevedden, “Artillery of King James I,” 61–63, 68–76. 70 T . abarı¯, Tarı¯kh, 3:1247–48, also 1238, 1245; and Kita¯b al-
uyu ¯n wa-al-h. ada¯iq, ed. M. J. de Goeje and P. de Jong (Leiden, 1871), 491; Chevedden, “Hybrid Trebuchet,” 188–90, 201.

92

INVENTION OF THE COUNTERWEIGHT TREBUCHET

only in Arabic, but also in Syriac and Greek, and clearly signals the emergence of new artillery, not the continuation of already existing forms. Gravity-Powered Artillery in Siege Accounts of the Twelfth Century The second siege of Tyre by the crusaders in 518/1124 provides the first instance of the counterweight trebuchet being identified by these new terms. The Chronicle of 1234 speaks of the crusaders using “great trebuchets” (manganı¯qe¯ rawrbe¯) and “many hybrid trebuchets” (qalqu ¯me¯ sagı¯
¯e) during this siege. The artillery breached the walls and destroyed many towers.71 William of Tyre’s account of the siege mentions that the crusaders sent to Antioch for an Armenian artillery expert named Havedic, who came to Tyre and “displayed so much skill in directing the machines and hurling the great stone missiles that whatever was assigned to him as a target was at once destroyed without difficulty.” With the aid of this Armenian specialist and his battery of artillery, the siege “assumed the aspect of a new war in the eyes of the Tyrians.” The crusaders pressed the siege with renewed strength, and the city was forced to surrender on 7 July 1124.72 The fact that the crusaders drew upon native expertise during this siege for the operation of counterweight artillery suggests that they may have been in the process of assimilating this new technology at this time. The Armenian engineer who was employed to operate the artillery most probably acquired knowledge of gravity-powered artillery from the Byzantines, either directly or indirectly. The Byzantine Empire not only had the requisite power and wealth to develop and exploit new ideas in artillery design, it had a tradition of innovation that encouraged the development of new devices of mechanized siegecraft. ¨rid In 519/1125 Aq Sunqur al-Bursuqı¯, the Seljuq governor of Mosul, and the Bo Atabeg T . ughtigin of Damascus battered the walls of
Aza¯z day and night with counterweight trebuchets, referred to as manganı¯qe¯ rawrbe¯ (great trebuchets).73 In 532/1138, the Byzantine emperor John II Komnenos bombarded Shayzar with counterweight trebuchets, identified as maja¯nı¯q ha¯ilah (huge trebuchets) by Usa¯mah ibn Munqidh and as Anon., Anonymi auctoris Chronicon ad annum Christi 1234 pertinens, ed. J.-B. Chabot, CSCO, ser. 3, vol. 37, Scriptores Syri 15 (Paris, 1916), 94. Partially translated into English by A. S. Tritton, with notes by H. A. R. Gibb, “The First and Second Crusade from an Anonymous Syriac Chronicle,” JRAS (1933): 95; and partially translated into French by A. Abouna, with notes by J.-M. Fiey, CSCO, 354, Scriptores Syri 154 (Paris, 1974), 71. On Syriac terms for artillery, see Chevedden, “Hybrid Trebuchet,” 204–7, 219–22. 72 William of Tyre, Historia rerum in partibus transmarinis gestarum, 13.10, ed. R. B. C. Huygens (Turnhout, 1986); idem, A History of Deeds Done Beyond the Sea, trans. E. A. Babcock and A. C. Krey, 2 vols. (New York, 1943), 2:15. Rogers believes that Havedic’s expertise “was required, not for knocking down fortifications, but for directing counter-battery fire [sic],” and that “his specialty seems to have been accuracy rather than the use of heavy artillery” (Rogers, Latin Siege Warfare, 83, 242, 247 n. 22). France also believes that Havedic “was brought in to direct counter-fire [sic] at the defenders’ machines” (France, Western Warfare, 120). The crusaders may have requested the services of Havedic because defensive artillery was hampering their attack on the city, but William of Tyre emphatically states that Havedic hurled “great stone missiles” and “that whatever was assigned to him as a target was at once destroyed without difficulty.” Nowhere does he state that Havedic restricted his targets to enemy artillery. The artillery employed by Havedic doubtless knocked down fortifications. On Armenians and crusaders, see J. H. Forse, “Armenians and the First Crusade,” JMedHist 17 (1991): 13–22. 73 Chronicle of 1234, ed. Chabot, 98; Anon, Anonymi auctoris Chronicon ad annum Christi 1234 pertinens, trans. Abouna, 73; Tritton “Anonymous Syriac Chronicle,” 97. 71

PAUL E. CHEVEDDEN

93

maja¯nı¯q
iz.¯am (great trebuchets) by Ibn al-
Adı¯m. These machines hurled stone-shot weighing between 43 and 54.5 kilograms a minimum distance of 150 meters. Commenting on the power of the Byzantine artillery, Usa¯mah mentions that a single stone-shot discharged by one of the trebuchets destroyed a house and that the bombardment breached the wall of the fortress.74 In 551–552/1157 the Seljuq sultan Muh. ammad II, reacting to a resurgence of
Abba¯sid power in Iraq, bombarded Baghdad with two counterweight trebuchets, identified as manjanı¯qayn
az.¯mayn ı (great trebuchets).75 The Normans used three counterweight trebuchets, termed maja¯nı¯q kiba¯r (big trebuchets), in their abortive attack on Alexandria in late July of 1174. These machines bombarded the city with black stones transported from Sicily especially for this purpose. A Muslim sally in force destroyed three large cat-castles advanced against the city, and a successful night attack and the imminent approach of Saladin’s army forced the Normans to withdraw after a three-day siege.76 In 572/1176 Saladin utilized counterweight trebuchets, identified as al-manjanı¯qa¯t alkiba¯r (big trebuchets), at the siege of Mas.ya¯f. Because of nearby crusader activity, Saladin was unable to prosecute the siege and settled for a negotiated peace settlement.77
Ima¯d ¯ mid in 579/1183 identifies all three types of trebual-Dı¯n’s account of Saladin’s siege of A chets in use in the twelfth century: the pole-framed trebuchet (
arra¯dah), the trestleframed traction trebuchet (manjanı¯q), and the trestle-framed counterweight trebuchet, identified as a “huge trebuchet” (manjanı¯q ha¯il). The counterweight machine used in this Usa¯mah ibn Munqidh, Kita¯b al-i
tiba¯r, ed. Qa¯sim al-Sa¯marra¯¯ı (Riyadh, 1987), 134; Kama¯l al-Dı¯n
Umar ibn Ah. mad Ibn al-
Adı¯m, Zubdat al-h. alab min tarı¯kh H . alab, ed. Sa¯mı¯ al-Dahha¯n, 3 vols. (Damascus, 1954), 2:268. Usa¯mah ibn Munqidh states that the Byzantine forces under emperor John II Komnenos set up trebuchets which hurled stone-shot weighing between 20 and 25 rat.ls a distance farther than a bow-shot. Since Usa¯mah was a native of Shayzar, the measure of weight that he utilized was most probably the Shayzar rat.l, which was equivalent to 2.137 kilograms (W. Hinz, Islamische Masse und Gewichte [Leiden, 1970], 31). The weight of the stone-shots launched by the Byzantine army was, therefore, between 43 and 54.5 kilograms (94–120 pounds). The information provided by Usa¯mah also allows us to gauge the distance of the launch of the stone-shot. If the projectiles were hurled farther than a bow-shot, this distance would probably be beyond the effective range of a military, rather than a sporting, bow. According to Marsden, the maximum effective range of a composite bow for military purposes was in the region of 150 to 200 yards (137–183 m) (Marsden, Greek and Roman Artillery: Historical Development, 12). Denys Pringle, utilizing information contained in Maurice’s Strategicon, estimates the maximum effective range of a bow-shot to be 140 meters (The Defence of Byzantine Africa from Justinian to the Arab Conquest, 2 vols. [Oxford, 1981], 1:150). Considering that the Byzantine trebuchets hurled stone-shot beyond the range of a bow-shot, the distance reached by these stones would probably have been at least 150 meters and most probably closer to the 200-meter mark or beyond. 75 Bunda¯rı¯, Zubdat al-nus.rah, 249. 76
Ima¯d al-Dı¯n Muh. ammad ibn Muh. ammad al-Ka¯tib al-Isfaha¯nı¯, Kita¯b al-fayh al-qussı¯ fı¯ al-fath. al-Qudsı¯, ed. C. de Landberg (Leiden, 1888), 227–28; idem, Sana¯ al-Barq al-Sha¯mı¯, 562/1166–583/1187, abridged by alFath. ibn
Alı¯ al-Bunda¯rı¯, ed. Fath. iyah al-Nabara¯wı¯ (Cairo, 1979), 76–79;
Abd al-Rah. ma¯n ibn Isma¯
¯l ı Abu ¯ Sha¯mah, Kita¯b al-rawd. atayn fı¯ akhba¯r al-dawlatayn, ed. M. H. M. Ah. mad and M. M. Ziya¯dah (Cairo, 1962), 1.2:598–600;
Alı¯ ibn Muh. ammad Ibn al-Athı¯r, al-Ka¯mil fı¯al-tarı¯kh, ed. C. J. Tornberg, 13 vols. (Beirut, 1965), 11:412–14; Baha¯ al-Dı¯n Yu ¯sufı¯yah, aw, ¯ suf ibn Ra¯fi
Ibn Shadda¯d, al-Nawa¯dir al-sult.¯anı¯yah wa-al-mah. ¯asin al-Yu Sı¯rat S. ala¯h. al-Dı¯n, ed. Jama¯l al-Dı¯n al-Shayya¯l (Cairo, 1964), 48–49; Jama¯l al-Dı¯n Muh. ammad ibn Sa¯lim Ibn ¯ shu Wa¯s.il, Mufarrij al-kuru ¯b fı¯ akhba¯r Banı¯ Ayyu ¯b, ed. Jama¯l al-Dı¯n al-Shayya¯l, S. A. F.
A ¯ r, and H. M. Rabı¯
, 5 vols. (Cairo, 1953– ), 2:11–16; A. S. Ehrenkreutz, Saladin (Albany, 1972), 124–25; M. C. Lyons and D. E. P. Jackson, Saladin: The Politics of the Holy War (Cambridge, 1982), 76–77; Rogers, Latin Siege Warfare, 120–21; Y. Lev, Saladin in Egypt (Leiden, 1999), 164. 77 Ka¯tib al-Isfah ¸ ¯anı¯, Sana¯ al-Barq al-Sha¯mı¯, 105–6; Abu ¯ Sha¯mah, Rawd. atayn, 669–70; Ibn al-Athı¯r, Ka¯mil, 11:436; Ibn Wa¯s.il, Mufarrij, 2:47–48; Ehrenkreutz, Saladin, 152; Lyons and Jackson, Saladin, 108–9. 74

94

INVENTION OF THE COUNTERWEIGHT TREBUCHET

siege was given the name al-Mufattish (The Examiner), indicating that this piece of artillery was treated with special regard. Heavy bombardment and mining over a three-day period discouraged resistance and led to a negotiated settlement, by which Saladin gained possession of the city and Ibn Nı¯sa¯n withdrew with his possessions.78 Saladin employed a total of nine al-manjanı¯qa¯t al-kiba¯r at the siege of Kerak in Juma¯da´ I 580/August 1184. These machines delivered an around-the-clock bombardment, while three digging-mantlets were constructed across the deep ditch on the eastern front of the fortress to dismantle the castle wall. Filling-mantlets were also employed to protect men engaged in filling up the ditch in preparation for an escalade. Saladin was on the verge of taking the castle when a Frankish relief force moved on Kerak and forced him to withdraw.79 To capture Thessalonike in 1185 the Normans bombarded the city with a number of counterweight trebuchets. Eustathios, the archbishop of the city, admiringly describes the artillery employed by the Normans and explains their siege tactics. On the western side of the city, the Normans “moved up newly invented heavy artillery (helepoleis).” These machines, Eustathios tells us, “did not perform brilliantly because they were difficult to manage due to their large size.” On this side of the city, the Normans appear to have attempted an attack using massed heavy artillery to breach the fortifications. The contrast drawn by Eustathios between the siege techniques employed on the western side of the city with those used on the eastern side supports this conjecture. On the eastern side, Eustathios tells us that the Normans “concentrated on more conventional methods” of siegecraft. Here, artillery was used to neutralize the defense while mining operations effected a breach in the city wall. A battery of trestle-framed trebuchets (mhcana` " petrobo´ lou", me¯khanas petrobolous), “most of them of small size,” bombarded defensive artillery mounted on the city walls and “caused damage to the city with their accurate aim,” while a pair of larger trebuchets tore away the battlements and caused severe damage to the interior of the city. One of the two larger trebuchets on the eastern side of the city was called “The Daughter of the Earthquake” (seismou' quga´ thr, seismou thygate¯r) because its missiles impacted with such violence. On two occasions, Eustathios refers to this machine, or perhaps its mate, as “the great siege-engine” (megale¯ me¯khane¯) and once as “their largest siegeengine” (megi´sth mhcanh´ , megiste¯ me¯khane¯). The projectiles that it discharged were so heavy—“fully as much as a man could lift”—that Eustathios likened the barrage of missiles flung by the smaller petroboloi to “the efforts of children compared with those hurled by the great siege-engine.” Eustathios was so fond of his comparison of the smaller pieces of artillery to children that he called the “great” trebuchet “Their Mother” (th` n mhte´ ra, te¯n me¯tera). The commander of the city had another name for this powerful piece of 78
Ima¯d al-Dı¯n Muh. ammad ibn Muh. ammad al-Ka¯tib al-Isfaha¯nı¯, al-Barq al-Sha¯mı¯, ed. Fa¯lih. H . usayn (Amman, 1987), 5:85; idem, Sana¯ al-Barq al-Sha¯mı¯, 218–21; Abu ¯ Sha¯mah, Rawd. atayn, 2:38–42; Ibn Shadda¯d, Nawa¯dir, 58; Ibn al-Athı¯r, Ka¯mil, 11:493–94; Ibn Wa¯s.il, Mufarrij, 2:134–36; Ehrenkreutz, Saladin, 180–81; Lyons and Jackson, Saladin, 190–91. 79 Ka¯tib al-Isfaha¯nı¯, Sana¯ al-Barq al-Sha¯mı¯, 241–43; Ibn al-Athı¯r, Ka¯mil, 11:506–7; Abu ¯ Sha¯mah, Rawd. atayn, 2:54–56; Ibn Shadda¯d, Nawa¯dir, 66; Ibn al-
Adı¯m, Zubdah, 3:78–79; Ibn Wa¯s.il, Mufarrij, 2:157–60; Ehrenkreutz, Saladin, 184; Lyons and Jackson, Saladin, 217–18; M. Barber, “Frontier Warfare in the Latin Kingdom of Jerusalem: The Campaign of Jacob’s Ford, 1178–79,” in The Crusades and Their Sources. Essays Presented to Bernard Hamilton, ed. J. France and W. G. Zajac (Aldershot, 1998), 9–22.

PAUL E. CHEVEDDEN

95

artillery: “The Old Woman” (grai'a, graia). As stones rained down upon Thessalonike, the dispirited commander cried out impassively, “Listen to the Old Woman,” and now and then he was heard to mutter, “The Old Woman is getting tired.” The use of heavy artillery, on its own, along the western side of the city to batter the defenses strongly suggests that the Normans attempted an artillery assault to breach the wall. The helepoleis used in this operation, identified by Eustathios as newly invented, were doubtless counterweight trebuchets. On the eastern side of the city, the Normans tested defenses that were “not free of defects” and “not well maintained” with conventional siege techniques. Artillery provided a continuous barrage while the besiegers filled in the ditch, demolished the outer wall, and undermined the main wall. The battery of artillery on the eastern side of the city included only two large trebuchets. These machines were most likely counterweight trebuchets, and one of them, judging from the attention Eustathios devotes to it, was an especially large and powerful stone-projector.80 In September 1187, Saladin besieged Jerusalem with counterweight trebuchets, identified by Bar Hebraeus as manganı¯qe¯ rawrbe¯ (“great trebuchets”).81 According to the Old French continuation of William of Tyre, one of Saladin’s trebuchets ( perriere) discharged three shots against the city in a single day. Such a restrained performance is indicative of a counterweight trebuchet, which has a much slower sequence of discharge than the traction or the hybrid machine. The account also mentions that among the trestle-framed and pole-framed trebuchets ( perrieres et mangoniaus) that Saladin set up were eleven large ones. The great size of these eleven machines suggests that they were counterweight trebuchets.82 In Saladin’s most ambitious siege operation, against Tyre from 12 November 1187 to 1 January 1188 (9 Ramad. ¯an–29 Shawwa¯l 583), pole-framed and trestle-framed trebuchets (
arra¯da¯t and manjanı¯qa¯t) were employed, according to Bunda¯rı¯’s abridgement of alBarq al-Sha¯mı¯.
Ima¯d al-Dı¯n’s more detailed account of the siege in his Fath. al-Qudsı¯ cites the use of trestle-framed traction and counterweight trebuchets (al-maja¯nı¯q al-s.igha¯r waal-kiba¯r). The Frankish fleet in Tyre made a nocturnal attack on the Muslim vessels that blockaded the harbor, capturing some ships and putting the rest to flight. This setback, coupled with the difficulties of a winter siege campaign and lack of enthusiasm in the Muslim camp to press the siege, forced Saladin to withdraw.83 80 Eustathios, Archbishop of Thessalonike, The Capture of Thessaloniki, trans. J. R. Melville Jones (Canberra, 1987), 72–99. Huuri speculated that the Norman siege of Thessalonike in 1185 may have witnessed the first use of the counterweight trebuchet, but Rogers judged the evidence for this conjecture to be inconclusive ¨tzwesen,” 91–92; Rogers, Latin Siege Warfare, 121–23). (Huuri, “Geschu 81 Gregorius Bar Hebraeus, Makhtebha¯nu ¯th Zabhne¯, ed. P. Bedjan (Paris, 1890), 375; idem, The Chronography ˆ’l Faraj, the Son of Aaron, the Hebrew Physician, Commonly Known as Bar Hebraeus, trans. E. A. W. of Gregory Abu Budge (London, 1932), 1:325. 82 M. R. Morgan, ed., La Continuation de Guillaume de Tyr (1184–1197), Documents relatifs `a l’histoire des croisades publie´s par l’Acade´mie des Inscriptions et Belles-Lettres 14 (Paris, 1982), 64–65: “Le jor que Saladin se remua de la Porte David et vint a la porte Saint Estiene, celui jor fist adrecier une perriere qui geta .iij. feis as murs de la cite´. Et la nuit fist adrecier tant de perrieres et de mangoniaus que l’endemain en trova l’en .xj. tous grans et getant as murs de la cite´.” For an English translation of this text, see P. W. Edbury, The Conquest of Jerusalem and the Third Crusade: Sources in Translation (Aldershot, 1996), 56. 83 Ka¯tib al-Isfaha¯nı¯, Sana¯ al-Barq al-Sha¯mı¯, 317–23; idem, Fath. al-Qudsı¯, 73–86, 110; Ibn al-Athı¯r, Ka¯mil, 11:553–55; Abu ¯ Sha¯mah, Rawd. atayn, 2:119–120; Ibn Shadda¯d, Nawa¯dir, 83–84; Ibn Wa¯s.il, Mufarrij, 2:242–45; Ibn al-
Adı¯m, Zubdah, 3:100; Ehrenkreutz, Saladin, 206–8; Lyons and Jackson, Saladin, 279–83.

96

INVENTION OF THE COUNTERWEIGHT TREBUCHET

The Crusader Siege of Acre (585–587/1189–91) An abundance of information on artillery is provided in the accounts of the siege of Acre by the crusaders (12 Rajab 585–17 Juma¯da´ II 587/26 August 1189–12 July 1191). Scholars examining this siege for evidence of the use of gravity-powered artillery have been unable to determine with certainty that such artillery was employed.84 I would argue, however, that the vocabulary used for heavy artillery in the sources and descriptive details that are provided of the machines confirm the use of counterweight trebuchets.
Ima¯d al-Dı¯n mentions that the garrison of the city launched a sally in force on 1 Sha
ba¯n 586/3 September 1190 that destroyed two counterweight trebuchets (manjanı¯qayn kabı¯rayn) of Henry of Champagne, one costing 1,500 dinars.85 Ibn Shadda¯d records the same event but states that one of the counterweight trebuchets was much bigger than the other. He identifies the large machine costing 1,500 dinars as “a big trebuchet of great magnitude” (manjanı¯q kabı¯r
az.¯m ı al-shakl) and the other as a trestle-framed trebuchet 86 (manjanı¯q). Later during the month of Sha
ba¯n/September, Ibn Shadda¯d records that two fire-arrows shot from a large crossbow set fire to two pieces of artillery, both identified as “huge trebuchets” (manjanı¯qa¯t ha¯ilah), indicating that these machines were both counterweight trebuchets.87 The Itinerarium Peregrinorum et Gesta Regis Ricardi describes a powerful and very effective trebuchet used by Acre’s defenders against the trebuchets of Henry of Champagne: There were plenty of trebuchets [ petrariae] in the city, but one of them was unequalled for its massive construction and its effectiveness and efficiency in hurling enormous stones. Nothing could stand against the power of this machine. It hurled huge stone-shot: “violent action, far-hurled stones; the blow smashed everything, whatever it struck.” If the stones met no obstruction when they fell, they sank a foot deep into the ground. This machine struck some of our trebuchets and smashed them to pieces or at least rendered them unusable. Its shots also destroyed many other siege machines, or broke off what it hit. It shot with such force, and its blows were so effective, that no material or substance could withstand the unbearable impact without damage, no matter how solid or wellbuilt it was.88

Rogers, Latin Siege Warfare, 227, 234. Ka¯tib al-Isfaha¯nı¯, Fath. al-Qudsı¯, 282. 86 Ibn Shadda¯d, Nawa¯dir, 134–35. 87 Ibid., 136; Abu ¯ Sha¯mah, Rawd. atayn, 2:162; Ibn Wa¯s.il, Mufarrij, 2:335. 88 Ricardus, Itinerarium, ed. Stubbs, 1:98: “Petrariarum hostilium, quarum fuit in civitate copia, una fuit incomparabilis, et magnitudine compactae machinae, et pro voto torquentium, inaestimabilis molis lapides jaculando efficax. Hujus nihil potuit resistere vehementiae. Incredibilis quippe molis lapides jaciebat; 84 85

Emissos etiam lapides procul impetus egit, Omnia comminuit jactus quaecunque feriret. Lapides nihilominus, quoties nullo retardarentur obstaculo, unius pedis longitudine agebantur in terram cadentes. Nonnullas petrariarum nostrarum percutiens in particulas dispersit, vel certe inutiles effecit; machinas quoque alias plures vel ictu dissolvit, vel particulam quam attigerat abscidit. Tanta nimirum erat vehem entia jaculandi, et impetus tam pertinax, quod nihil tam solidum, vel ita fuit compactum, cujuscunque materiae vel substantiae, quod posset incolume tam intolerabilis percussurae sustinere injuriam.” The translation of this passage is taken from Nicholson, Chronicle of the Third Crusade, 103, with the following changes: the term petrariae is translated as “trebuchets” (not “stonethrowers”), the phrase molis lapides in the second sentence is rendered “huge stone-shot” (not “really incredible lumps of rock”); and the verb jaculor is translated as “to shoot” (not “to fire”). Cf. Ambroise, Estoire, vv. 3535–60.

PAUL E. CHEVEDDEN

97

The Muslim trebuchet was doubtless a counterweight machine. Features that characterize this ordnance—pinpoint accuracy and great power—are highlighted in the Latin account of it. The Anglo-Norman poet known as Ambroise describes the same machine but adds an important detail that confirms that it is gravity-powered. He states that “all of two men’s strength it took to draw its sling.” 89 The drawing of the sling refers to the pulling or stretching out of the sling to its full length in preparation for launch. Fixed to the bedplate of the counterweight trebuchet is a trough, placed on the same vertical plane as the throwing arm. This device serves as a runway for the projectile during the initial phase of the launch cycle. Once the beam has been wound back to its launch position, the sling is stretched out along the trough, and the projectile is loaded into the pouch of the sling. At the completion of this operation, the machine is ready for discharge (Figs. 2 and 3). The destructive effects of the artillery bombardment against Acre, particularly during the final stages of the siege, indicate that most of the machines described in the historical accounts were counterweight trebuchets. The Itinerarium Peregrinorum, which provides a detailed description of eleven trebuchets deployed by the crusaders against the Maledicta Tower and its adjacent walls in June and July of 1191, is worth quoting at length: The king of France . . . concentrated on constructing siege machines and placing trebuchets [ petrariae] in suitable places. He arranged for these to shoot continually day and night. He had one excellent one which he called “Bad Neighbor” [Malvoisine] [1]. The Turks in the city had another which they called “Bad Relation” [Mal Couisine] and often used to smash “Bad Neighbor” with its violent shots. The king kept rebuilding it until its continual bombardment partly destroyed the main city wall and shattered the Cursed Tower. On one side the duke of Burgundy’s trebuchet [2] had no little effect. On the other the Templars’ trebuchet [3] wreaked impressive devastation, while the Hospitallers’ trebuchet [4] also never ceased hurling, to the terror of the Turks. Besides these, there was a trebuchet that had been constructed at general expense, which they called “God’s Stone-Thrower” [5]. A priest, a man of great probity, always stood next to it preaching and collecting money for its continual repair and for hiring people to gather the stones for its ammunition. This machine at last demolished the wall next to the Cursed Tower for around two perches’ length [11 yards or 10 meters]. The count of Flanders had had a choice trebuchet [6], which King Richard had after his death, as well as another trebuchet [7] which was not so good. These two constantly bombarded the tower next to a gate which the Turks frequently used, until the tower was half-demolished. Besides these, King Richard had two new ones made [8 and 9] with remarkable workmanship and material which would hit the intended target no matter how far off it was. . . . He also had two mangonels [traction trebuchets] [10 and 11] prepared. One of these was so swift and violent that its shots reached the inner streets of the city meat market. King Richard’s trebuchets hurled constantly by day and night. It can be firmly stated that one of them killed twelve men with a single stone. That stone was sent for Saladin to see, with messengers who said that the diabolical king of England had brought from Messina, a city he had captured, sea flint and the smoothest stones to punish the Saracens. Nothing could withstand their blows; everything was crushed or reduced to dust.90 Ambroise, Estoire, vv. 3545–46: “Que dous genz coveneit a metre // En la funde” (The Crusade of Richard Lion-Heart, trans. M. J. Hubert with notes and documentation by J. L. La Monte, Records of Civilization, Sources and Studies 34 [New York, 1941], 160). 90 Ricardus, Itinerarium, ed. Stubbs, 1:218–19: “Rex Franciae . . . machinis intendebat conficiendis, et petrariis locis aptis applicandis, quas et nocte dieque incessabiliter instituit jaculari. Quarum unam habuerat 89

98

INVENTION OF THE COUNTERWEIGHT TREBUCHET

Arabic sources provide details of Philip II’s artillery assault against Acre launched on 4 Juma¯da´ I 587/30 May 1191. He employed seven trebuchets (maja¯nı¯q) that caused considerable damage to the walls before Muslim defenders were able to destroy the machines. Burj
Ayn al-Baqar (Maledicta Tower) sustained heavy damage in this attack.91 Philip II and Richard I joined together for a massive artillery assault against Acre in June–July 1191 (Juma¯da´ I–Juma¯da´ II 587). According to
Ima¯d al-Dı¯n, the crusaders “brought every type of trebuchet (manjanı¯q) as large as a mountain” against the city. “They set up thirteen trebuchets (manjanı¯q) in one place,” he recounts, “and stones brought down the wall, with each stone more of it.” The outer wall was destroyed and the main wall was breached.
Ima¯d al-Dı¯n recounts that the walls were reduced to an elevation “no higher than a man’s height.” His final damage assessment details the devastation: “They [the crusaders] engaged in the discharge of trestle-framed trebuchets [manjanı¯qa¯t], the erecting of siege machines and pole-framed trebuchets [
arra¯da¯t], and the loading of stones until the wall was on the verge of destruction. The walls were shaken loose, breaches were visible, and the curtains were demolished.” 92 Ibn Shadda¯d’s account echoes
Ima¯d al-Dı¯n’s description: “Trebuchets [manjanı¯qa¯t] destroyed the walls to no higher than a man’s height. Trebuchets [manjanı¯qa¯t] pounded the walls continually until the stones became detached.” 93 The historical sources offer compelling evidence for the use of counterweight trebuchets at the siege of Acre. The technical vocabulary used in the Arabic sources for the artillery at Acre offers very strong evidence for the machine’s use, as do the operational details provided by Ambroise that are specific to counterweight artillery and the degree of damage inflicted by artillery bombardment in this siege. Trebuchet and Its Cognates In the Latin West, the term trabuchellus first appears in a fealty document issued in Vicenza on 6 April 1189. In this document, villagers in the region of Vicenza profess peroptimam quam vocavit Malam Vicinam. Turci vero infra civitatem alteram habebant quam vocabant Malam cognatam, quae vehementibus jactibus frequentius dissipare solebat Malam Vicinam: quam rex reaedificabat, quousque jugiter jaciendo, principalem civitatis murum in parte diruit, et Turrim Maledictam conquassavit. Hinc et petraria ducis Burgundiae non in vanum jaciebat: illinc Templariorum petraria Turcos vastabat egregie, et Hospitalariorum quoque Turcis metuenda nequaquam cessabat a jactibus. Praeter has quoque fuit quaedam petraria, communicatis impensis compacta, quam vocitabant petrariam Dei. . . . Per ipsam demum ad aestimationem duarum perticarum, juxta turrim Maledictam, conquassatus est murus. Comes Flandriae petrariam habuerat electam, quam post ejus mortem habuit rex Ricardus, et praeterea minorem electam. Hae duae sine intermissione jaciebant versus turrim juxta portam quandam, quam Turci frequentabant, donec turris medietatem diruerunt. Et praeter has, facerat rex Ricardus alias novas duas, operis electi et materiae, ineffabiliter destinatam percutientes quorumcunque metam locorum. . . . Duos etiam praeparaverat mangunellos quorum unus tantae fuerat agilitatis et vehementiae, quod jactus ejus pervenirent in interiores macelli civitatis plateas. Petrariae itaque regis Ricardi die noctuque jugiter jaciebant, de quarum una certissime constat, quod unius lapidis jactu prostraverit in mortem duodecim homines.” The translation of this passage is taken from Nicholson, Chronicle of the Third Crusade, 208–9, with the following changes: the term petraria is translated as “trebuchet” (not “stonethrower”), and the verb jaculor is translated as “to shoot” (not “to fire”). Cf. Ambroise, Estoire, vv. 4741–4808. 91 Ka¯tib al-Isfaha¯nı¯, Fath. al-Qudsı¯, 330–35; Ibn Shadda¯d, Nawa¯dir, 158–60; Ibn al-Athı¯r, Ka¯mil, 12:64; Abu ¯ Sha¯mah, Rawd. atayn, 2:184; Ibn Wa¯s.il, Mufarrij, 2:350. 92 Ka¯tib al-Isfaha¯nı¯, Fath. al-Qudsı¯, 345, 346, 349, 350, 366. 93 Ibn Shadda¯d, Nawa¯dir, 165–66.

PAUL E. CHEVEDDEN

99

fealty to the city and pledge among other things “not to transport a mangonel, a trebuchet, a petraria, a crossbow or a bow, or permit anyone to do so . . . up to the boundary of its [Vicenza’s] consulate.” 94 A decade later, the standard nominal form trabuchus is used. In October 1199, the Cremonans besieged Castelnuovo Bocca d’Adda with petrariai and trebuchets (trabuchi).95 The origin and precise meaning of trebuchet is a subject of uncertainty among scholars. The conventional etymology of trabuc[h]us (OF. trabuc, trebuc, and triboc; Cat. trabuc; Cast. trabuco; It. trabucco; Ger. driboc[k], tripoch, or triboc; and OE. trabuch) suggests that the term is derived from the Latin prefix trans-, expressing displacement, and the Old French word buc, referring to a trunk (of a body).96 This etymology infers that the word is related to the machine’s operation rather than to its shape or structural configuration—the trunk (of a body) being displaced or discharged by the machine. But since the machine generally throws stones, this etymology is unlikely, even if buc can be interpreted allegorically to refer to any large projectile. An etymology related to the form of the machine appears to be more plausible than one related to its function. Such an etymology is readily apparent. The first part of the word is the prefix tri- (three), and the second part is the Latin noun bracchium (arm) or its cognates in medieval Latin (bracco), romance languages, or German. A tribracchium, or tribracho, literally means a device having three arms or branches. The trestle-framed trebuchet probably received this designation because its three main components were conceived of as being three large arms: a beam (one arm) pivoting on an axle that was supported by two triangulated trusses (the second and third arm) forming the trestle frame. The description of trestle-framed hybrid trebuchets by Abbo of Saint-Germaindes-Pre´s at the siege of Paris in 885–886 confirms that Europeans of the Middle Ages did reduce the machine to its three main components: a throwing arm and the trusses of the trestle frame.97 The Chronic¸on Sampertrinum supports the derivation of trebuchet from tribracchium by equating a tribracho with a tribock in its description of the campaign of Emperor Otto IV in Thuringia in 1212: “Otto, having come into Thuringia, besieged and reduced the castle of the Landgrave at Salza with a trebuchet (tribracho) called a tribock.” 98 94 Giambasta Verci, Storia degli Ecelini, 3 vols. (Brassano, 1779), 3:97: “nec unum mangano nec trabuchello aut cum prederia vel balista vel archu traham nec aliquem trahere permittam nec faciam si vetare potero usque ad finem sue consularie.” I thank Donald J. Kagay for translating this passage. 95 Giovanni Codagnello, Iohannis Codagnelli Annales Placentini, ed. O. Holder-Egger, MGH, ScriptRerGerm (Hanover-Leipzig, 1901), 25: “Cremonenser . . . quoddam fossatum munierunt cariolis, quas posuerunt super ripam illius, et gladiis et beltreschis et predariis et trabuchis.” 96 Oxford English Dictionary, s.v. “trabuch”; A. J. Greimas, Dictionnaire de l’ancien franc¸ais jusqu’au milieu du XIVe (Paris, 1968), 639. 97 Abbo, Le Sie`ge de Paris par les Normands, 1.360–67, ed. and trans. H. Waquet (Paris, 1942), 42–43; Chevedden, “Hybrid Trebuchet,” 196–98. 98 Chronicon Sampertrinum: “Otto veniens in Thuringiam cum tribracho illo, cognomento tribock, castrum lantgravii in Salza obsedit et expugnavit,” quoted in R. Schneider, Die Artilleri des Mittelalters (Berlin, 1910), 28. In 1212 the Annales Marbacenses records Otto’s subsequent use of the tribok following the siege of Salza. Annales Marbacenses, ed. H. Bloch, MGH, ScriptRerGerm 9 (Hanover, 1907), 81–82: “Et inde [from Salza] progrediens obsedit oppidum Wizense quod similiter expugnavit usque ad arcem. Ibi tunc primo cepit haberi usus instrumenti bellici, quod vulgo tribok appellari solet” (From Salza he advanced to besiege Weissensee which he similarly reduced. There was used for the first time that war engine which is commonly called the tribok). Trans. P. Contamine, War in the Middle Ages; trans. M. Jones (Oxford, 1984), 104. The medieval Latin verb trabucare (to bombard) and its romance cognates are derived from trabuchus (J. F. Niermeyer, Mediae latinitatis lexicon minus [Leiden, 1976], 1034).

100

INVENTION OF THE COUNTERWEIGHT TREBUCHET

The term trebuchet (OF. trebuchet, also trebuket, trebusket, trabuchet; Prov. trabuquet; Cat. trabuquet; Cast. trabuquete; It. trabocchetto; med. L. tra-, trebuchetum) is the diminutive form of trabuchus. Because the diminutive form is used, some have argued that the trebuchet must have been a smaller machine than the trabuchus. France even contends that a number of historical sources indicate that the term trebuchet denotes a “light weapon altogether.” 99 The counterweight trebuchet is, however, the largest artillery weapon of its class. Villard de Honnecourt’s massive trestle-framed trebuchet is estimated to have had a counterweight weighing 30 tons.100 Clearly, there must be some other reason, unrelated to size, that explains the use of a diminutive form to designate such an enormous machine. If it does not arise from irony, it may have been used to domesticate the terror of the new weapon. J. R. Hale has suggested this explanation as the reason behind the naming of cannon in the era of early gunpowder weaponry, and it seems to be equally applicable to the era of pre-gunpowder artillery.101 This is a very human response to the destructive power of new weaponry and can be seen in our own day in the naming of nuclear devices, such as “Little Boy” that destroyed Hiroshima.102 The Catalan Grand Chronicles shed light on this question. The chronicles of King James I of Aragon-Catalonia (1208–76) and Bernat Desclot always employ the term trabuquet (or trebuquet), never trabuc,103 whereas the chronicles of Ramon Muntaner and King Peter IV of Aragon (III in Catalonia) (1336–87) always use the term trabuc, never trabuquet.104 Three of the four chroniclers—James, Desclot, and Muntaner—describe the siege of the city of Majorca (modern Palma) in 1229 by King James. While the accounts of this siege by James and Desclot are very detailed and make abundant references to the use of trabuquets by both besieger and besieged,105 Muntaner’s account is abbreviated and has only one reference to the Muslims’ siege machines, identified as trabucs.106 This appears to indicate that there was no difference between these machines; the Catalan chroniclers simply used the terms interchangeably. If this supposition is correct, the machine designated as the trabuquet (⫽ trebuchet) corresponds to the trabuc; both terms refer to a trestle-framed gravity-driven machine. Some scholars argue that the trebuchet can be further defined structurally. Joseph Goday y Casals and Jordi Bruguera believe that the trebuchet had a hinged counterweight.107 99 W. Giese, “Waffen nach den katalanischen Chroniken des XII. Jahrhunderts,” Volkstum und Kultur der Romanen 1 (1928): 147; J. France, Western Warfare in the Age of the Crusades, 1000–1300 (Ithaca, 1999), 122. 100 Bechmann, Villard de Honnecourt, 265. 101 J. R. Hale, “Gunpowder and the Renaissance: An Essay in the History of Ideas,” in From the Renaissance to the Counter-Reformation. Essays in Honor of Garrett Mattingly, ed. C. H. Carter (New York, 1965), 131. 102 H. Jaffe, “What’s in a Name? Just a Funny Little Nuke Test,” Los Angeles Times, 8 August 1983. 103 James I, King of Aragon, Llibre dels fets del rei En Jaume, ed. J. Bruguera, 2 vols. (Barcelona, 1991), 2:69.3, 69.5, 69.9, 69.11, 69.13, 69.15, 69.18, 69.25, 69.44, 69.46, 125.10, 126.5, 126.13, 157.13, 157.20, 158.4, 194.25, 262.4; Bernat Desclot, Cro`nica de Bernat Desclot, in Les quatre grans cro`niques, Jaume I, Bernat Desclot, Ramon Muntaner, Pere III, ed. F. Soldevila (Barcelona, 1971), chaps. 39, 40, 42, 49, 110, 118. 104 Ramon Muntaner, Cro`nica de Ramon Muntaner, in Soldevila, Les quatre grans cro`niques (as above, note 103), chaps. 7, 44, 93, 107, 165, 169, 223, 229, 246, 258, 259, 272; Peter IV of Aragon (III in Catalonia), Cro`nica de Pere el Cerimonio´s, in ibid., chaps. 1.11, 1.16, 5.32, 6.23, 6.24. 105 James I, Llibre dels fets, chaps. 69, 157, 158; Bernat Desclot, Cro`nica, chaps. 39, 40, 42. 106 Ramon Muntaner, Cro`nica, chap. 7. 107 `nica del Rey D. Jaume,” in Congre´s d’historia Joseph Goday y Casals, “Medis d’atach y de defensa en la Cro de la corona d’Arago´, dedicat al rey En Jaume I y a la seua ´epoca, 2 vols. (Barcelona, 1909–13), 2:803; J. Bruguera, `nica de Jaume I,” Estudis de llengua i literatura catalanes 1 (1980): 54–55. “Vocabulari militar de la Cro

PAUL E. CHEVEDDEN

101

Ada Bruhn Hoffmeyer, however, contends that the counterweight was fixed.108 The hinged-counterweight proponents can find justification for their view in Villard de Honnecourt’s thirteenth-century description of a trestle-framed trebuchet (trebucet) that had such a counterweight.109 In his reconstruction of Villard’s machine, EugeneEmmanuel Viollet-le-Duc was the first to propose that the machine designated by the term trebuchet had a hinged, rather than a fixed, counterweight. He believed that the term mangonel was applied to the machine with a fixed counterweight.110 Fixed-counterweight proponents, such as Bruhn Hoffmeyer, look to Giles of Rome (Aegidius Colonna) to support their argument. In his De regimine principium libri tres, produced in Italy around 1275, Giles of Rome identifies three different types of counterweight trebuchets: the machine with a fixed counterweight is called a trabucium, the one with a hinged counterweight a biffa, and the machine with both a fixed and a hinged counterweight is termed a tripantium.111 His classification, however, appears to be of his own making and is not corroborated by any other source. His technical vocabulary may be relevant to Italy in the late thirteenth century, but it does not appear applicable to other regions of Europe, where no distinction is made between the trestle-framed trebuchet with a fixed counterweight and the one with a hinged counterweight. The difficulties of interpreting written sources enhance the value of illustrated sources. Trebuchets are depicted in a large number of medieval and Renaissance manuscripts. A crude illustration of a trebuchet with a counterweight in the form of a box, indicating that it was attached by a hinge to the short arm of the beam, appears in an early thirteenth-century manuscript of Wolfram von Eschenbach’s Willehalm and is referred to as a tripochen.112 A later manuscript of the same text has two corrupt illustrations of a trebuchet, both of which have a hinged counterweight, and the second machine is identified as a driboch.113 The description accompanying the illustration of Villard de Honnecourt’s trebuchet indicates that this machine also had a hinged counterweight. Francesco di Giorgio Martini, however, identifies both the trebuchet with a hinged counterweight and the one with a fixed counterweight as a traboccho.114 The evidence from these illustrated sources offers two possible interpretations. The term trebuchet may have been applied to both the fixed-weight and the hinged-weight, trestle-framed gravitypowered trebuchet, or the term may have been used initially to refer to one type only— most likely the hinged-weight version—and over time its usage changed. The differences between the two types of trestle-framed counterweight trebuchets may have become inA. Bruhn Hoffmeyer, Arms and Armour in Spain: A Short Survey (Madrid, 1982), 2:102, 106, 112. Paris, Bibliothe`que nationale de France, MS fr. 19093, fol. 30r. 110 Dictionnaire raisonne´ de l’architecture du XIe au XVIe sie`cle, 10 vols. (Paris, 1854–68), 5:233. 111 De regimine principum libri tres, quoted in R. Schneider, Die Artillerie des Mittelalters (Berlin, 1910), 163–64; trans. J. Hewitt, Ancient Armour and Weapons in Europe from the Iron Period of the Northern Nations to the End of the Thirteenth Century, 3 vols. (Oxford, 1855–60), 1:349–50. 112 Munich, Bayerische Staatsbibliothek, MS cod. germ. 193 III, fol. 4v; reproduced in K. von Amira, Die ¨cke der grossen Bilderhandschrift von Wolframs Willehalm (Munich, 1921), 20 and pl. 10. Bruchstu 113 ¨ sterreichische Nationalbibliothek MS cod. 2670, fols. 33v and 81v; reproduced in W. von Vienna, O ¨rlin und der Fortsetzung des Ulrich von Tu ¨rheim Eschenbach, Willehalm mit der Vorgeschichte des Ulrich von dem Tu (Graz, 1974), fols. 33v and 81v. 114 Trattati di architettura ingegneria e arte militare, ed. C. Maltese, 2 vols. (Milan, 1967), 1:272–73; fols. 60r, 61v, 62r; pls. 111, 114, 115. 108 109

102

INVENTION OF THE COUNTERWEIGHT TREBUCHET

creasingly blurred during the Renaissance, and this may account for a broader usage of the term trebuchet at this time. White has suggested that the term trebuchet was derived from the term for a duckingstool.115 It is far more likely, however, that the ducking-stool took its name from the counterweight trebuchet, rather than the other way around. Both machines are similar in design, but the siege machine was introduced earlier and probably served as the model for the ducking-stool. INVENTION AND DIFFUSION OF THE COUNTERWEIGHT TREBUCHET The driving force behind the development of the counterweight trebuchet was most probably the increased military activity in the eastern Mediterranean during the later part of the eleventh century and the early part of the twelfth century. This period witnessed Seljuq and crusader incursions and a Byzantine military resurgence under the Komnenoi. It also saw Armenian settlement in Cilicia and the employment of Armenians in the military services of Islamic, Byzantine, and crusader states in the region. The talent and economic resources committed to siegeworks naturally increased with rising militarization, and this condition provided a great impetus to the development of artillery. Rogers has suggested that “it is likely that more effective artillery was developed by expert siege crews in the regular employment of a commander who could afford them.” 116 If this was the case, artillery specialists employed by the Byzantine Empire may well have developed the counterweight trebuchet. Strangely, a Byzantine origin for gravity-powered artillery has never been proposed. ¨hler and Kalevero Huuri drew attention to an explicit reference Ever since Gustav Ko to a trabuchus that was used in northern Italy in 1199,117 scholars have regarded this as evidence for the first appearance of the counterweight trebuchet in Western Europe, and many have consequently favored a European origin for the machine.118 Even those familiar with Byzantine source materials, such as Bruhn Hoffmeyer, Arnold W. Lawrence, and Robert W. Edwards, have overlooked the possibility that Byzantium may have developed the weapon, and have instead credited the machine to the Latin West.119 The favoring of the Latin West may also stem from Cahen’s conflation of two distinct Arabic terms that were applied to two different types of gravity-powered artillery: the manjanı¯q maghribı¯ (the western Islamic trebuchet) and the manjanı¯q firanjı¯ (the Frankish or European trebuchet). Cahen argued that at the end of the thirteenth century the manjanı¯q maghribı¯ was designated by the term manjanı¯q firanjı¯. The manjanı¯q maghribı¯, however, was a trestle-framed counterweight trebuchet, while the manjanı¯q firanjı¯ was the European bricola, a gravityWhite, Medieval Technology and Social Change, 102; Oxford English Dictionary, s.v. “trebuchet.” Rogers, Latin Siege Warfare, 247. 117 ¨hler, Entwickelung, 194; Huuri, “Geschu ¨tzwesen,” 64 n. 2, 171. Ko 118 White, Medieval Technology and Social Change, 102–3; Hill, “Trebuchets,” 104; Gillmor, “Introduction,” 2 n. 6; R. W. Edwards, review of C. Foss, Survey of Medieval Castles of Anatolia, Speculum 62.3 (1987): 678–79; France, Western Warfare, 122. 119 A. Bruhn Hoffmeyer, “Military Equipment in the Byzantine Manuscript of Scylitzes in Biblioteca Nacional in Madrid,” Gladius 5 (1966): 136; A. W. Lawrence, “A Skeletal History of Byzantine Fortifications,” BSA 78 (1983): 222; Edwards, review of Survey of Medieval Castles of Anatolia, 678–79. 115 116

PAUL E. CHEVEDDEN

103

powered pole-framed trebuchet invented in the Latin West (Fig. 5). The manjanı¯q firanjı¯ is clearly of European origin; the manjanı¯q maghribı¯ is not.120 Not all scholars who have dealt with the issue of the origin of the counterweight trebuchet have credited it to Western Europe. White, who earlier had favored a European origin for the machine, raised the possibility that it may have originated in Iran.121 Hill, who initially stated that the counterweight trebuchet was first developed “somewhere in Mediterranean Christendom or western Islam, towards the end of the twelfth century,” also considered a Muslim origin for the machine.122 Joseph Needham and Robin Yates present a variety of views on the invention of the counterweight trebuchet in Science and Civilisation in China. Although they contend that engineers from Arab countries introduced the machine to China, they nevertheless maintain that “the invention [of the counterweight trebuchet] was quite probably made in several places about the same time.” They propose that “one inventor may have been Chhinang Sheˆn, the Chin commander who defended Lo-Yang against the Mongols in ⫹ 1232.” Needham and Yates also assert that the appearance of the word trebuchet in three German chronicles simultaneously in 1212 “marks the entry of the counterweight trebuchet.” They further argue that the counterweight trebuchet was “an Arab modification,” “an Arab invention,” was derived “from Arab practice,” or was “developed in West Asia.” 123 No scholar has suggested that the Byzantines, who were well known for their military inventiveness, developed the counterweight trebuchet. Despite the achievements made by Byzantium in the development of artillery, misunderstandings persist about Byzantine knowledge of and use of these weapons. Carroll Gillmor states that the traction trebuchet did not appear definitely in the Byzantine Empire until 1071.124 Edwards has explicitly denied that the counterweight trebuchet was known in twelfth-century Byzantium.125 Clive Foss and David Winfield have suggested that the counterweight trebuchet “was introduced into Europe in the twelfth century, [was] taken up by the crusaders, and rapidly spread through the Mediterranean and Near East,” but they sidestep the question of the machine’s origin: “The history of the transmission of the trebuchet is not at all clear, and it could as well have spread from Byzantium to the West as vice versa.” They contend that gravity-powered ordnance came into common use during the reign of Manuel I Komnenos (1143–80), but they provide no textual evidence for this conclusion.126 The emergence of the counterweight trebuchet is sparsely documented. A number of languages—Syriac, Arabic, Greek, and Latin—use new terminology for artillery in the twelfth century. These changes in terminology reflect changes in technology: the introduction of the counterweight trebuchet. Although the documentation for this machine during the first century of its development is random and somewhat meager, it provides dramatic evidence of cultural diffusion. Cahen, “Traite´,” 158. On the nomenclature of trebuchets, see Chevedden, “Artillery of King James I.” White, Medieval Religion and Technology, 284. 122 Hill, “Trebuchets,” 104; Hassan and Hill, Islamic Technology, 101. 123 Needham and Yates, Science and Civilisation in China, 218, 233, 237, 238, 240, 573. 124 Gillmor, “Traction Trebuchet,” 1–2. 125 Edwards, review of Survey of Medieval Castles of Anatolia, 678. 126 Foss and Winfield, Byzantine Fortifications, 48 and 181 n. 27. 120 121

104

INVENTION OF THE COUNTERWEIGHT TREBUCHET

Once gravity-powered artillery was invented, the dynamics of conflict and contact spurred reactive technological adaptation of this new technology over a broad area. The documentation for the counterweight trebuchet in the twelfth century comes from a wide variety of sources across a vast geographical landscape, from Mesopotamia to the central Mediterranean. The transmission of technologies and techniques in the region of western Asia and the Mediterranean world was mutual and continuous, and this was especially the case with artillery. Earlier forms of artillery had likewise experienced rapid diffusion in western Asia and the Mediterranean. Following the death of Alexander the Great, the catapult spread throughout the Mediterranean world and became common even in relatively unimportant cities.127 The traction trebuchet first appeared in the eastern Mediterranean at the end of the sixth century and soon spread westward to Spain. In the early part of the seventh century, it caught the attention of Isidore, bishop of Seville (ca. 570–636), who referred to it as a fundibulus in his Etymologiae.128 Shortly thereafter, the trebuchet was transmitted over a wide expanse of territory from the Atlantic to the Indus by the Islamic conquest movements. These conquests spurred innovations in military technology that led to the development of the hybrid trebuchet. By the ninth century, this machine was being used in western Asia, the Mediterranean world, and in northern Europe.129 The rough outlines of the divergent patterns of transmission of the counterweight trebuchet can be culled from the sources. By the 1120s, gravity-powered artillery had been transmitted to Outremer and to the sultanate of the Great Seljuqs. Seljuq dependencies and local Atabeg dynasties may also have acquired the machine at this time. At mid-century, the Seljuq sultan used counterweight trebuchets to bombard the
Abba¯sid capital. In the 1130s, John II Komnenos used counterweight trebuchets with great effect in his siege campaigns in Cilicia and Syria.130 These campaigns mark the full incorporation of the counterweight trebuchet into the Byzantine armed forces, and there is no doubt that subsequent Byzantine campaigns under John Komnenos and his successor Manuel I Komnenos (1143–80) employed the new artillery.131 If gravity-powered artillery Marsden, Greek and Roman Artillery: Historical Development, 73–85. Isidore equated the traction trebuchet ( fundibulus) with the two-armed stone-projecting torsion catapult (ballista), the most powerful artillery piece of classical antiquity (Isidore, Etymologiae, 18.10). By the 4th century, this stone-projector had been replaced by the onager, and in the 6th century the onager was superseded by the traction trebuchet. By likening the fundibulus to the defunct ballista in terms of its “hurling and throwing” capabilities, Isidore was presenting new information on artillery to suit the realities of his own day. Medieval texts of Vegetius likewise present new information on artillery to conform to circumstances of a later period. The fundibulus, or traction trebuchet, was the principal artillery piece of Isidore’s day. This term does not refer to the fustibalus (staff-sling), or to the onager, the one-armed torsion machine of late antiquity, as suggested by Bruhn Hoffmeyer, but to the traction trebuchet (Bruhn Hoffmeyer, Arms and Armour in Spain: A Short Survey, 1:95, 141). For more details on the diffusion of the traction trebuchet in the Mediterranean world, see Chevedden, “Hybrid Trebuchet,” 192–95, and Chevedden et al., “Traction Trebuchet.” 129 For more details on the diffusion of the hybrid trebuchet in the Middle East, the Mediterranean world, and Europe, see Chevedden, “Hybrid Trebuchet,” 179–222, and Chevedden et al., “Traction Trebuchet.” 130 See App. 1 and John Kinnamos, Epitome rerum ab Ioanne et Alexio [sic] Comnenis gestarum, ed. A. Meineke, CSHB (Bonn, 1836), 16–20; idem, Deeds of John and Manuel Comnenus, trans. C. M. Brand (New York, 1976), 21–24; Choniates, ed. van Dieten, 21–31; O City of Byzantium, 13–18. 131 John Kinnamos, Epitome, 21–22; idem, Deeds, 25–26; Choniates, ed. van Dieten, 32–39; O City of Byzantium, 19–22. Counterweight trebuchets (helepoleis) were most likely used in the joint Byzantine/crusader at127 128

PAUL E. CHEVEDDEN

105

was used in the Latin East in the 1120s, knowledge of it was certainly transmitted to the Latin West in rapid order. The Normans of Sicily were well acquainted with the new artillery when they attacked Alexandria in 1174 and Thessalonike in 1185, suggesting previous familiarity with the counterweight trebuchet that probably dates back to the campaigns of Roger II in southern Italy in the 1130s.132 The appearance of the counterweight trebuchet in northern Italy toward the end of the twelfth century also suggests an earlier acquaintance with the machine, perhaps dating back to Frederick Barbarossa’s siege campaigns in Lombardy from 1155 to 1175.133 Before Saladin’s arrival in Egypt in 559/1164, the counterweight trebuchet must have been known in the Nile Valley, since all of T . arsu ¯ sı¯’s information on the machine was gleaned from an Armenian arms manufacturer who worked under the Fa¯.timids in Alexandria.134 From the 1170s to the early 1190s, Saladin utilized gravity-powered artillery against his Muslim and crusader foes. The great display of counterweight artillery at the siege of Acre (1189–91) was the culmination of a long process of development that had begun nearly a century earlier. Rogers regards the siege of Acre as the beginning of a transition to a new stage in the development of siege warfare in the Middle Ages. For the first time, according to Rogers, artillery was used to breach enemy fortifications, not merely to support “escalades or siege-tower assaults by harassing defenders and neutralizing their artillery,” as had been the role of ordnance in earlier siege operations. “The effectiveness of artillery at Acre,” Rogers claims, “anticipates developments in thirteenth-century poliorcetics, in which bombardment played an important role. . . . In this regard, the siege of Acre can be seen as ushering in the great age of pre-gunpowder artillery in the medieval west.” 135 The turning point in the tactical use of artillery, however, had already taken place prior to the siege of Acre. Eustathios’s account of the siege of Thessalonike in 1185 implicitly recognizes that artillery played two principal roles in siege warfare: (1) it was used to neutralize the defense so that besiegers could fill in ditches and batter and undermine fortifications in order to facilitate a successful infantry assault; (2) it was used to breach the walls of strongpoints in order to pave the way for assaulting infantry. By identifying the first tactic as the “more conventional” scheme of siege operations, Eustathios confirms what historical accounts of twelfth-century sieges indicate: that artillery was also used to breach fortifications. This new role for artillery did not begin with the introduction of the counterweight trebuchet. Powerful traction trebuchets, especially hybrid ones, also had this capability. But proficiency at breaching fortifications was greatly expanded by the employment of gravity-powered artillery. By the end of the twelfth century, the countertack on Damietta in 565/1169. On this expedition, see Choniates, ed. van Dieten, 163–68; O City of Byzantium, 92–95; William of Tyre, Historia, 20.15 (pp. 929–30); idem, Deeds Done Beyond the Sea, 20.15 (2:364–65); Ibn Shadda¯d, Nawa¯dir, 41–43; Ehrenkreutz, Saladin, 79–81; Lyons and Jackson, Saladin, 37; Rogers, Latin Siege Warfare, 84–86; R.-J. Lilie, Byzantium and the Crusader States, 1096–1204, trans. J. C. Morris and J. E. Ridings (Oxford, 1993), 200–202. 132 Rogers, Latin Siege Warfare, 105–20. 133 Rogers, Latin Siege Warfare, 132–53. 134 On Armenians in Fa¯.timid Egypt, see S. B. Dadoyan, The Fatimid Armenians: Cultural and Political Interaction in the Near East (Leiden, 1997). 135 Rogers, Latin Siege Warfare, 234–35.

106

INVENTION OF THE COUNTERWEIGHT TREBUCHET

weight trebuchet had reached an advanced stage of development, and defensive planning was soon to undergo revolutionary changes in order to counter the greater destructive power of the new artillery and to exploit this artillery for use in the defense of strongpoints. The transmission of the counterweight trebuchet during the twelfth century is far easier to trace than engineering data on the machine itself. Only two twelfth-century sources provide structural details of the counterweight trebuchet: Niketas Choniates’ Historia and T . arsu ¯b. Both of these authors used old terminology to ¯ sı¯’s Tabs.irah fı¯ al-h. uru refer to the new machine. This indicates a general reluctance to coin new technical terms for new technology. The history of technology is rife with examples of technological changes that are not reflected by terminological ones.136 In the twelfth century, language seldom kept pace with technological advances in the field of artillery. The references to the counterweight trebuchet that we do have from this period probably represent a mere fraction of the instances in which the machine was used. Moreover, the counterweight trebuchet was probably not recognized by new nomenclature until it had evolved to a certain stage of development and was employed on a regular basis. One would scarcely expect a distinct nomenclature for the machine to develop when it was still new and unfamiliar. Thus, the first occurrence of a new name for the counterweight trebuchet probably does not signal the date of its introduction. The counterweight trebuchet doubtless emerged prior to the use of new terms to denote it. If so, the siege of Nicaea in 1097 may mark the introduction of the counterweight trebuchet. Yet, the new terms for this machine, which refer to “big,” “great,” and “huge” or “frightful” pieces of artillery, all disappear during the thirteenth century and are replaced by a different set of names for gravity-powered ordnance. What does this terminological shift signify? Are we right to assume that the terms that arise in the twelfth century denote a new technological reality (the counterweight trebuchet), given the dramatic changes in nomenclature that occur in the thirteenth century? Perhaps the shift to new terms in the thirteenth century signifies the actual emergence of gravity-powered artillery, while the transition in terms a century earlier may reflect only an evolution to more powerful versions of existing artillery devices. An understanding of the real nature of the terminological shift can be achieved by examining the terms. WHAT’S IN A NAME? During the thirteenth century, the counterweight trebuchet was designated by three new terms in Arabic: the manjanı¯q maghribı¯ or manjanı¯q gharbı¯ (the western Islamic trebuchet), the manjanı¯q qara¯bughra¯ (the “Black Camel” trebuchet), and the manjanı¯q firanjı¯ or manjanı¯q ifranjı¯ (the Frankish or European trebuchet). The manjanı¯q maghribı¯ and manjanı¯q qara¯bughra¯ were both trestle-framed machines. The manjanı¯q maghribı¯ generally launched stone-shot,137 while the manjanı¯q qara¯bughra¯ was specifically designed to launch 136 On problems in terminology, with particular reference to artillery, see J. Needham, Science and Civilisation in China, vol. 5, Chemistry and Chemical Technology, pt. 7, Military Technology: The Gunpowder Epic (Cambridge, 1986), 11–12, 130, 131, 248, 310, 311, 373. 137 The historical sources mention the manjanı¯q maghribı¯ or manjanı¯q gharbı¯ (the Western Islamic trebuchet) as being employed in the sieges of the following cities and castles: Damietta in 615/1218 (Abu ¯ Bakr ibn
Abd

PAUL E. CHEVEDDEN

107

bolts.138 The manjanı¯q firanjı¯ was a large pole-framed machine, obviously derived from Western Europe as its name indicates, and known in Latin sources as the bricola (Fig. 5).139 In the Latin West, the term trebuchet was joined by a legion of new terms after 1200: blida or blide (also bleide, bidda, pleiden, pleyde, plaid, plijde and pleite), biffa (also biffe and Alla¯h ibn Aybak al-Dawa¯da¯rı¯, known as Ibn al-Dawa¯da¯rı¯, Kanz al-durar wa-ja¯mi
al-ghurar, vol. 7, al-Durr almat.lu ı
Ashu ¯b fı¯ akhba¯r mulu ¯k banı¯ Ayyu ¯b, ed. Sa
¯d ¯ r [Cairo, 1972], 195; Ibn al-Muqaffa
, History of the Patriarchs, 3.2:129 [Arabic text], 218 [trans.]; Oliver of Paderborn, The Capture of Damietta, trans. J. J. Gavigan [Philadelphia, 1948], 25, 26, 34, 44, 54, 56), Kakhta¯ in 623/1226 (Na¯s.ir al-Dı¯n al-H . usayn ibn Muh. ammad Ibn Bı¯bı¯, Histoire des Seldjoucides d’Asie Mineure d’apres l’abre´ge du Seldjouknameh d’Ibn-Bibi, ed. T. Houtsma [Leiden, 1902], 118; and idem, Die Seltschukengeschichte des Ibn Bibi, trans. H. W. Duda [Copenhagen, 1959], 122), Baalbek in 627/1229 (Muh. ammad ibn Ibra¯hı¯m ibn Muh. ammad al-Ans.¯arı¯, Tarı¯kh dawlat al-akra¯d wa-al-atra¯k, Istanbul, ˘lu Ali Pas¸a Collection, MS 695, fol. 124v), Shayzar in 630/1232–33 (Ibn Wa¯s.il, Su ¨ leymaniye Library, Hekimog Mufarrij, 5:65; Ibn al-
Adı¯m, Zubdah, 3:214), H . arra¯n and Edessa in 633/1236 (Ibn al-Muqaffa
, History of the Patriarchs, 4.1, 70 [Arabic text], 147 [trans.]; Ans.¯arı¯, Tarı¯kh, fols. 138v–139r), Homs in the winter of 646/ 1248–49 (Ibn Wa¯s.il, Mufarrij al-kuru ¯b fı¯ akhba¯r Banı¯ Ayyu ¯b, Paris, Bibliothe`que nationale, fonds arabe, MS 1703, fol. 60v–61r; Ans.¯arı¯, Tarı¯kh, fol. 165r), Caesarea in 663/1265 (Muh. yı¯ al-Dı¯n Ibn
Abd al-Z. ¯ahir, al-Rawd. al-za¯hir fı¯ sı¯rat al-Malik al-Z. ¯ahir, ed.
Abd al-
Azı¯z al-Khuwayt.ir [Riyadh, 1976], 230; Na¯s.ir al-Dı¯n Muh. ammad ibn
Abd al-Rah.¯m ı Ibn al-Fura¯t, Ayyubids, Mamlukes and Crusaders: Selections from the “Tarı¯kh al-Duwal wa’lMulu ¯k,” ed. and trans. U. Lyons and M. C. Lyons, 2 vols. [Cambridge, 1971], 1:85; 2:69), and S. afad in 664/ 1266 (Ibn
Abd al-Z. ¯ahir, Rawd. , 257–58; Ibn al-Fura¯t, Ayyubids, Mamlukes and Crusaders, 1:116–17; 2:91–92). For illustrations of this siege machine, see Chevedden, “Artillery of King James I,” figs. 8, 9, 13. 138 This siege machine is identified by Ibn Urunbugha¯ al-Zaradka¯sh as the “Black Camel” (qara¯bughra¯) trebuchet (Kita¯b al-anı¯q fı¯ al-mana¯janı¯q, ed. Nabı¯l Muh. ammad
Abd al-
Azı¯z Ah. mad [Cairo, 1981], 26; ibid., ed. Ih. sa¯n Hindı¯ [Aleppo, 1985], 45), but all Arabic historical sources, save for Nasawı¯’s Sı¯rat Jala¯l al-Dı¯n, which was produced in a Persian cultural milieu, designate it as the “Black Bull” (qara¯bugha¯) trebuchet, a corruption of the original term (see D. Ayalon, “H . is.¯ar,” EI2, 3:473). Arabic historical chronicles cite this weapon as being used in the sieges of the following cities and castles: Akhla¯.t in 626/1229 (Muh. ammad ibn Ah. mad al-Nasawı¯, Sı¯rat Jala¯l al-Dı¯n, ed. H . ¯afiz. H . amdı¯ [Cairo, 1953], 303), Homs in the winter of 646/1248–49 (Ibn Wa¯s.il, Mufarrij [Bibliothe`que nationale, MS 1703], fol. 61r), Marqab in 684/1285 (Muh. yı¯ al-Dı¯n Ibn
Abd al-Z. ¯ahir, Tashrı¯f al-ayya¯m wa-al-
usu ¯r, ed. Mura¯d Ka¯mil and Muh. ammad
Alı¯ al¯r fı¯ sı¯rat al-Malik al-Mans.u Najja¯r [Cairo, 1961], 78), Tripoli in 688/1289 (Abu ¯ Bakr ibn
Abd Alla¯h ibn Aybak al-Dawa¯da¯rı¯, known as Ibn al-Dawa¯da¯rı¯, Kanz al-durar wa-ja¯mi
al-ghurar, vol. 8, al-Durrah al-zakı¯yah fı¯ akhba¯r al-Dawlah al-Turkı¯yah, ed. U. Haarmann [Cairo, 1971], 283; Qut.b al-Dı¯n Mu ¯ sa´ ibn Muh. ammad al-Yu ¯ nı¯nı¯, Dhayl mir¯at al-zama¯n fı¯ tarı¯kh
al-a ya¯n, Istanbul, Topkapı Sarayı Mu ¨ zesi Ku ¨ tu ¨ phanesi, Ahmet III Collection, MS 2907/I.3, fol. 254r; Na¯s.ir al-Dı¯n Muh. ammad ibn
Abd al-Rah.¯m ı Ibn al-Fura¯t, Tarı¯kh Ibn al-Fura¯t, ed. C. K. Zurayk and N. Izzedin, [Beirut, 1936–42], 8:80), Acre in 690/1291 (Shams al-Dı¯n al-Jazarı¯, H . awa¯dith al-zama¯n wa-anba¯uhu wa-wafaya¯t
 al-aka¯bir wa-al-a ya¯n min abna¯ ihi [ Jawa¯hir al-sulu ¯k], Paris, Bibliothe`que nationale, fonds arabe, MS 6739, fol. 24v; idem, La Chronique de Damas d’al-Jazarı¯ [Anne´es 689–698 H.], summary translation of Jawa¯hir al-sulu ¯k by J. Sauvaget [Paris, 1949], 5; Badr al-Dı¯n al-
Aynı¯,
Iqd al-juma¯n fı¯ tarı¯kh ahl al-zama¯n, Istanbul, Topkapı Sarayı Mu ¨ zesi Ku ¨ tu ¨ phanesi, Ahmet III Collection, MS 2912/4, fol. 144v; Qut.b al-Dı¯n Mu ¯ sa´ ibn Muh. ammad alYu ¯nı¯nı¯s [Freiburg im ¯ nı¯nı¯, Dhayl mir¯at al-zama¯n, ed. A. Melkonian, Die Jahre 1287–1291 in der Chronik al-Yu Breisgau, 1975], 86 [Arabic text]; Ibn al-Fura¯t, Tarı¯kh, 8:112), and Qal
at al-Ru ¯ m in 691/1292 (Ibn alDawa¯da¯rı¯, Kanz, 8:333; Mufad. d. al Ibn Abu ¯ al-Fad. ¯ail, al-Nahj al-sadı¯d wa-al-durr al-farı¯d fı¯ma¯ ba
da tarı¯kh Ibn al-
Amı¯d, ed. and trans. E. Blochet, Histoire des sultans mamlouks, PO [Paris, 1919–28], 14:553; K. V. Zetterste´en, ed., Beitra¨ge zur Geschichte der Mamlu ¯kensultane in den Jahren 690–741 der Higra, nach arabischen Handschriften [Leiden, 1919], 16; Shiha¯b al-Dı¯n al-Nuwayrı¯, Niha¯yat al-arab fı¯ funu ¯n al-adab, quoted in Ah. mad ibn
Alı¯ al-Maqrı¯zı¯, Kita¯b al-sulu ¯k li-ma
rifat duwal al-mulu ¯k, 4 vols. [Cairo, 1934–73], 1.3:778 n. 2; Ibn al-Fura¯t, Tarı¯kh, 8:136). For an illustration of this siege machine, see Chevedden, “Artillery of King James I,” fig. 10. 139 The manjanı¯q firanjı¯ or manjanı¯q ifranjı¯ (the Frankish or European trebuchet) is first cited in Arabic historical sources at the beginning of the Mamlu ¯ k period and is mentioned as being employed in the sieges of the following cities and castles: Caesarea in 663/1265 (Ibn
Abd al-Z. ¯ahir, Rawd. , 230; Ibn al-Fura¯t, Ayyubids, Mamlukes and Crusaders, 1:85; 2:69), al-Bı¯rah in 674/1275 (
Izz al-Dı¯n Muh. ammad Ibn Shadda¯d, Tarı¯kh alMalik al-Z. ¯ahir, ed. Ah. mad Hut.ayt. [Wiesbaden, 1983], 125; Qut.b al-Dı¯n Mu ¯ sa´ ibn Muh. ammad al-Yu ¯ nı¯nı¯, Dhayl mir¯at al-zama¯n, 4 vols. [Hyderabad, 1954–61], 3:114; Na¯s.ir al-Dı¯n Sha¯fi
ibn
Alı¯ al-
Asqala¯nı¯, H . usn al-

108

INVENTION OF THE COUNTERWEIGHT TREBUCHET

buffa), bricola (also bricole, brı`ccola, brigola, and brigolete), carabaga (also caraboha, carabouha, carabaccani, and carabachani), couillart (also coillard, coyllar, and collarde), tribulus, tribuculus, and tripantium, to mention the most prominent.140 Scholars have been baffled by this profusion of terms, calling the terminology “confused,” “contradictory,” “uncertain,” “vague and imprecise,” “inconsistent,” exhibiting “no uniformity,” “very complicated,” and “all but impenetrable.” 141 Summing up the views of many historians, Rogers asserts that the set of terms used for medieval artillery in the Latin West is characterized by an absence of clarity and coherence.142 Although the technical vocabulary of medieval artillery is complicated and somewhat elusive, clarification can be achieved. Today, a resolution of the terminological confusion is hampered mainly by a proclivity for rehashing old arguments and a reluctance to bring a wider range of historical evidence to bear on the question. The difficulties of understanding the nature of the terminological shift, as well as the terms themselves, are reduced by an awareness of three important factors: (1) the terms reflect a diversification of the counterweight trebuchet into different forms; (2) the terms embody regional variations in terminology; and (3) the terms represent changes in nomenclature over a period of time. Had gravity-powered artillery not diversified into different forms, there would have been no need for new terms. The old terms—such as manjanı¯q in Arabic, helepolis mana¯qib al-sirrı¯yah al-muntaza
ah min al-sı¯rah al-Z. ¯ahirı¯yah, ed.
Abd al-
Azı¯z al-Khuwayt.ir [Riyadh, 1976], 158), Marqab in 684/1285 (Ibn
Abd al-Z. ¯ahir, Tashrı¯f, 78), S. ahyu ¯ n in 685/1286 (Ibn
Abd al-Z. ¯ahir, Tashrı¯f, 149–50), Tripoli in 688/1289 (Ibn al-Dawa¯da¯rı¯, Kanz, 8:283; Yu ¨ zesi Ku ¨ tu ¨¯ nı¯nı¯, Dhayl [Istanbul, Topkapı Sarayı Mu phanesi, Ahmet III Collection, MS 2907/I.3], fol. 254r; Ibn al-Fura¯t, Tarı¯kh, 8:80), Acre in 690/1291 (Jazarı¯, H . awa¯dith al-zama¯n, fol. 24v; idem, Chronique de Damas, 5; Badr al-Dı¯n al-
Aynı¯,
Iqd al-juma¯n fı¯ tarı¯kh ahl alzama¯n, Istanbul, Topkapı Sarayı Mu ¨ zesi Ku ¨ tu ¨ phanesi, Ahmet III Collection, MS 2912/4, fol. 144v; Yu ¯ nı¯nı¯, Dhayl [ed. Melkonian], 86 [Arabic text]; Ibn al-Fura¯t, Tarı¯kh, 8:112), and Qal
at al-Ru ¯ m in 691/1292 (Ibn alDawa¯da¯rı¯, Kanz, 8:333; Zetterste´en, Beitra¨ge, 16; Jazarı¯, Chronique de Damas, 16; Ibn Abu ¯ al-Fad. ¯ail, Nahj,  14:553; Nuwayrı¯, Niha¯yah, quoted in Ibn Abu ¯k 1.3:778 n. 2; ¯ al-Fad. ¯a il, Nahj, 14:553 n. 1, and Maqrı¯zı¯, Sulu Ibn al-Fura¯t, Tarı¯kh, 8:136). For illustrations of this siege machine, see Chevedden, “Artillery of King James I,” fig. 11; Ibn Urunbugha¯, Anı¯q, ed. Ah. mad, 47, 51; idem, Anı¯q, ed. H . indı¯, 97–98. In his account of the siege of Hsiang-Yang by the Mongols in the late 13th century, the Chinese historian Cheˆng Ssu-Hsiao refers to “Muslim trebuchets” (hui-hui phao) used against the city: “The design of the Muslim trebuchets came originally from the Muslim countries, and they were more powerful than ordinary trebuchets. In the case of the largest ones, the wooden framework stood above a hole in the ground. The projectiles were several feet in diameter, and when they fell to the earth they made a hole three or four feet deep. When [the artillerists] wanted to hurl them to a great range, they added weight [to the counterpoise] and set it further back [on the arm]; when they needed only a shorter distance, they set it forward, nearer [the fulcrum]” (Needham and Yates, Science and Civilisation in China, 5.6:221). The largest trebuchets at the siege of Hsiang-Yang appear to have been bricolas since the framework of the bricola was commonly mounted in a single hole in the ground. 140 On the terminology of the counterweight trebuchet, see Huuri, “Geschu ¨ tzwesens,” 63–67, 132, 171–74, 187–89, 213–15; Chevedden, “Artillery of King James I,” 61–63, 71–76, 84, 86, 87, 91, figs. 7–15; idem, “Hybrid Trebuchet,” 180, 181, 183, 184–86, 189, 190, 192, 197, 200, 201, 206–9, 211, 212, 214–16, figs. 2, 4, 5. 141 ´, Forteresses, 150; idem, These remarks are from the following studies cited in the order listed: Fino “Machines de jet me´die´vales,” 25; H.-P. Eydoux, Les chaˆteaux de soleil: forteresses et guerres des croise´s (Paris, 1982), 279; R. A. Brown, English Castles (London, 1976), 175; France, Western Warfare, 118; France, Victory in the East, 48; Bruhn Hoffmeyer, Arms and Armour in Spain: A Short Survey, 2:104; Hacker, “Greek Catapults and Catapult Technology,” 41. For other comments on the question of nomenclature, see L. Monreal y Tejada, Ingenierı´a militar en las cro´nicas catalanas, Discurso de Ingreso en la Real Academia de Buenas Letras (Barcelona, 1971), 18. 142 Rogers, Latin Siege Warfare, 258.

PAUL E. CHEVEDDEN

109

or petrobolos in Greek, and petraria in Latin—may have taken on a broader meaning as they had done in the past, and the new terms that referred to “big,” “great,” and “huge” or “frightful” machines may have faded away. Since the term trebuchet arises long after the introduction of the counterweight machine, its usage probably does not coincide with the first appearance of this weapon in Europe. Rather, it probably reflects the beginning of the process of diversification of gravity-powered artillery into multiple forms in the Latin West. The process of diversification of gravity-powered artillery in the Latin West was most likely underway by the 1180s. During this decade, the term trebuchet appears for the first time, and “newly invented heavy artillery [helepoleis]” are cited at the siege of Thessalonike. These “newly invented” machines were not likely to have been trestle-framed counterweight trebuchets, but gravity-powered artillery considerably different from the trestle-framed engine. Eustathios identifies two types of capital ordnance used by the Normans at Thessalonike: “newly invented” heavy artillery employed on the western side of the city and two other large trebuchets on the eastern side of the city, which were not described as newly made or created or put to use for the first time. He regarded the pair of machines on the eastern side of the city as conventional in every way, save for the great size and power of one of them, a trebuchet so immense and powerful that it was called “The Daughter of the Earthquake.” The two conventional pieces of artillery were doubtless trestle-framed counterweight trebuchets, the earliest form of gravity powered artillery. By the 1180s, the trestle-framed counterweight trebuchet was no longer either a novelty or a rarity, and thus Eustathios does not identify the two large “rock-throwers” on the eastern side of the city as new machines. His “newly invented” helepoleis on the western side of the city must have represented an innovation in the design of the counterweight trebuchet to warrant this characterization. If so, these machines were most probably European bricolas. The design of the bricola was noticeably different from that of the trestle-framed trebuchet (Fig. 5). The massive pole that was used for its frame and the two hinged-counterweights that hung from its bifurcated throwing arm gave it a unique appearance that justified its name, the “two-testicle machine” (from Latin bi-coleus). The innovative and distinctive design of the bricola explains why such artillery would have attracted the attention of Eustathios and would have prompted him to single it out from the other artillery used by the Normans. The new artillery brought against Thessalonike, like so many new technologies, apparently had design or performance problems during its first employment. The newly invented trebuchets performed poorly, Eustathios tells us, because their large size made them difficult to manage. Historical sources indicate that the bricola first emerged in the lands of the western Mediterranean basin. The earliest mention of the bricola in an historical source records its use in 1238 by Emperor Frederick II of Hohenstaufen at the siege of Brescia. During the 1240s, bricolas appear in a number of sieges in Italy. In 1242 Frederick II sent bricolas and other siege machines to the Levant, and soon thereafter the Muslims incorporated this versatile piece of artillery into their siege arsenal, calling it the Frankish or European trebuchet (manjanı¯q ifranjı¯ or manjanı¯q firanjı¯). The bricola went on to become the most widely used piece of naval ordnance in the Mediterranean. It was mounted on the poops

110

INVENTION OF THE COUNTERWEIGHT TREBUCHET

of ships and was used to bombard coastal cities and fortresses. The early history of the bricola makes a Norman origin for this machine quite likely. At Thessalonike, Eustathios recognized its distinctive design as an innovation in the design of heavy artillery.143 BYZANTIUM AND THE COUNTERWEIGHT TREBUCHET If the process of diversification of the counterweight trebuchet began during the 1180s, development of this machine was probably underway much earlier. Although it is impossible to determine with certitude when gravity-powered artillery was invented, due to the tenuous nature of the evidence, a late eleventh-century dating appears to be a reasonable estimate. Such a dating for the emergence of the counterweight trebuchet favors a Byzantine origin. The Byzantines were leaders and innovators in the field of artillery. The first Mediterranean civilization to utilize the trebuchet was Byzantium. The Byzantines not only improved upon the artillery technology that they had acquired from Asia, but they also found new tactical uses for it. They adopted the “hand-trebuchet,” which they dubbed the cheiromangana, and employed this weapon in field battles. Byzantine engineers brought heavy artillery to the peak of its development during the eleventh century. They produced the two most powerful trebuchets that are recorded in the historical sources for that century, and at Nicaea they startled everyone by constructing trebuchets fashioned according to entirely new design principles.144 Economic and political factors also suggest a Byzantine origin for the machine. When Alexios I Komnenos came to the throne in 1081, the Byzantine state was in a desperate 143 For a discussion of the bricola, see Chevedden, “Artillery of King James I,” 72–76. On Frederick II’s siege of Brescia, see Annales placentini gibellini, MGH, SS, ed. G. H. Pertz (Hanover, 1863), 18:479; T. C. van Cleve, The Emperor Frederick II of Hohenstaufen, Immutator Mundi (Oxford, 1972), 415–16; and D. Abulafia, Frederick II, a Medieval Emperor (London, 1988), 308–9. For other sieges in Italy during the early 1240s that featured the bricola, see Annali genovesi di Caffaro e de’ suoi continuatori, ed. L. T. Belgrano and C. Imperiale di Sant’Angelo, 4 vols. (Rome, 1890–1929), 3:100, 120, 121; Chevedden, “Artillery of King James I,” 74–75. On Frederick II’s shipment of bricolas to the Levant in 1242, see Annali genovesi, 3:128: “Et cum inimici mari et terra cum machinis, prederiis [⫽ petrariis], bricolis, scalis et aliis hedifficiis eorum infortunio ad locum Levanti pervenissent.” 144 Needham and Yates claim that the “hand-trebuchet” was invented by a Chinese military engineer named Liu Yung-Hsi in 1002 (Needham and Yates, Science and Civilisation in China, 5.6:214). Byzantine sources, however, mention the “hand-trebuchet” (cheiromangana) in the 10th century. The Praecepta militaria of the emperor Nikephoros II Phokas (963–969), dating from ca. 965, recommends the use of “handtrebuchets” in field battles to break up enemy formations. The Taktika of Nikephoros Ouranos, dating from ca. 1000, also recommends the use of “hand-trebuchets” in field operations. The Anonymus De obsidione toleranda includes the “hand-trebuchet” in a list of artillery (ed. H. van den Berg [Leiden, 1947], 48, lines 3–4: tetrare´ a", magganika` , kai` ta` " legome´ na" hjlaka´ ta", kai` ceiroma´ ggana . . . ). On the Byzantine cheiromangana, see McGeer, Sowing the Dragon’s Teeth, 21, 65, 97. McGeer incorrectly identifies the cheiromangana as a portable arrow launcher, similar to a crossbow, that was mounted on a stand. The Byzantines probably came to realize the usefulness of the trebuchet in field operations after their defeat in the battle of Anzen in 223/838. In this battle, the Byzantine army under Emperor Theophilus faced
Abba¯sid forces under the caliph’s general Afshı¯n. On the afternoon of 22 July, Turkish archers isolated and surrounded the emperor and a band of 2,000 Khurramı¯ refugees from al-Jiba¯l (modern Lurista¯n) and were closing in for the kill when a rainstorm rendered their bows useless. The Muslims quickly brought up traction trebuchets and hurled stones on the Byzantine forces, which then dispersed in panic (Chronique de Michel le Syrien, ed. and trans. J.-B. Chabot, 4 vols. [Paris, 1905], 3:95, 4:535; Bar Hebraeus, Chronography, ed. Bedjan, 149; W. Treadgold, The Byzantine Revival, 780–842 [Stanford, 1988], 300). An illustration in the Cantigas de Santa Maria, dating from ca. 1280, depicts a “hand-trebuchet” being operated alongside a counterweight trebuchet in the siege of a city (Chevedden, “Artillery of James I,” fig. 12; Chevedden, “Hybrid Trebuchet,” 213, fig. 4).

PAUL E. CHEVEDDEN

111

situation. The Seljuqs were in occupation of most of Anatolia, the Pechenegs were menacing the Danubian provinces, and Constantinople was under threat of a Norman attack. By 1095 Alexios had overcome the Normans, defeated the Pechenegs, and was about to begin his reconquest of Asia Minor. As the economic and political fortunes of the Byzantine Empire revived under Alexios, greater resources became available to the army. These resources were used to attract a wide range of talent to advance Byzantine military ventures, and fresh talent was doubtless used to develop and exploit new ideas relating to mechanized siegecraft. Alexios’s project to reconquer Anatolia would have provided a tremendous incentive for the development of artillery, since the peninsula could only be won back by utilizing a powerful siege-train well equipped with ordnance. Alexios had both the foresight and the finances to hire the technicians and establish the workshops to build batteries of heavy artillery for his campaign of reconquest. The Byzantines under the Komnenoi built upon a military tradition that encouraged innovation, and their creative efforts must have concentrated on the development of heavy artillery, a major pursuit of the Byzantine army throughout the eleventh century. The machine that marked the high point in the development of mechanical warfare—the counterweight trebuchet—probably emerged under Byzantine auspices, and the invention “which amazed everyone” at Nicaea in 1097 may well have been the first gravity-powered piece of artillery.145 Central Washington University On Alexios I Komnenos’s attempt to reconquer Anatolia, see H. Ahrweiler, Byzance et la mer: la marine de guerre, la politique et les institutions maritimes de Byzance aux VIIe–XVe sie`cles (Paris, 1966), 175–225; C. Cahen, Pre-Ottoman Turkey: A General Survey of the Material and Spiritual Culture and History, c. 1071–1330, trans. J. Jones-Williams (New York, 1968), 72–96; M. Whittow, “How the East Was Lost: The Background to the Komnenian Reconquista,” in Mullett and Smythe, Alexios I Komnenos (as above, note 43), 1:55–67; M. Mullett, “1098 and All That: Theophylact Bishop of Semnea and the Alexian Reconquest of Anatolia,” Peritia 10 (1996): 237–52. For a discussion of the development of the Komnenian army during this period, see Birkenmeier, “Development of the Comnenian Army.” For political and economic developments of this period, see F. Chalandon, Essai sur le re`gne d’Alexius Ier Comne`ne (1081–1118) (Paris, 1900); P. Magdalino, The Empire of Manuel I Komnenos, 1143–1180 (Cambridge, 1993); C. M. Brand, Byzantium Confronts the West, 1180–1204 (Cambridge, Mass., 1968); A. Harvey, Economic Expansion in the Byzantine Empire, 900–1200 (Cambridge, 1990). 145

Appendix 1 References to Counterweight Trebuchets Employed during the Twelfth Century

Abbreviations A.2 A.2 [E] A.2 [F] AS B, Zubdah BH BH [E]

C C [E] Cod. Ecel. E GC IA ID, Barq ID, Fath. ID, Sana¯

Anon. Anonymi auctoris Chronicon ad annum Christi 1234 pertinens. Ed. J.-B. Chabot. CSCO, ser. 3, vol. 37, Scriptores Syri 15. Paris, 1916. ———. “The First and Second Crusade from an Anonymous Syriac Chronicle.” Trans. A. S. Tritton, with notes by H. A. R. Gibb. JRAS (1933): 69–101, 273–305. ———. Anonymi auctoris Chronicon ad annum Christi 1234 pertinens. Vol. 2. Trans. A. Abouna, CSCO 354, Scriptores Syri 154. Paris, 1974. ı Kita¯b al-rawd. atayn fı¯ akhba¯r al-dawlatayn. Abu ¯ Sha¯mah,
Abd al-Rah. ma¯n ibn Isma¯
¯l. Vol. 1, pt. 2. Ed. M. H. M. Ah. mad and M. M. Ziya¯dah. Cairo, 1962. al-Bunda¯rı¯, Fath. ibn
Alı¯. Zubdat al-nus.rah wa-nukhbat al-
us.rah. Ed. M. T. Houtsma. Recueil de textes relatifs `a l’histoire des Seljoucides 2. Leiden, 1889. Bar Hebraeus, Gregorius. Makhtebha¯nu ¯th Zabhne¯. Ed. P. Bedjan. Paris, 1890. ˆl Faraj, the Son of Aaron, Bar Hebraeus, Gregorius. The Chronography of Gregory Abu the Hebrew Physician, Commonly Known as Bar Hebraeus. Vol. 1. Trans. E. A. W. Budge. London, 1932. Choniates, Niketas. Historia. Ed. J. van Dieten. CFHB 11. Berlin, 1975. ———. O City of Byzantium, Annals of Niketas Choniate¯s. Trans. H. J. Magoulias. Detroit, 1984. Codex diplomaticus Ecelinianus. In Giambatista Verci, Storia degli Ecelini. 3 vols. Bassano, 1779. Eustathios, Archbishop of Thessalonike. The Capture of Thessaloniki. Trans. J. R. Melville Jones. Canberra, 1987. Codagnello, Giovanni. Iohannis Codagnelli Annales Placentini. Ed. O. Holder-Egger. MGH, ScriptRerGerm. Hanover-Leipzig, 1901. Ibn al-
Adı¯m, Kama¯l al-Dı¯n
Umar ibn Ah. mad. Zubdat al-h. alab min tarı¯kh H . alab. 3 vols. Ed. Sa¯mı¯ al-Dahha¯n. Damascus, 1951–68. al-Ka¯tib al-Isfaha¯nı¯,
Ima¯d al-Dı¯n Muh. ammad ibn Muh. ammad. Al-Barq al-Sha¯mı¯. Vol. 5. Ed. Fa¯lih. H . usayn. Amman, 1987. ———. Kita¯b al-fayh al-qussı¯ fı¯ al-fath. al-Qudsı¯. Ed. Carlo de Landberg. Leiden, 1888. ———. Sana¯ al-Barq al Sha¯mı¯, 562/1166–583/1187. Abridged by al-Fath. ibn
Alı¯ alBunda¯rı¯. Ed. Fath. iyah al-Nabara¯wı¯. Cairo, 1979.

PAUL E. CHEVEDDEN IS IW

U

113

Ibn Shadda¯d, Baha¯ al-Dı¯n Yu ¯ suf ibn Ra¯fi
. Al-Nawa¯dir al-sult.¯anı¯yah wa-al-mah. ¯asin al-Yu ¯sufı¯yah, aw, Sı¯rat S. ala¯h. al-Dı¯n. Ed. Jama¯l al-Dı¯n al-Shayya¯l. Cairo, 1964. Ibn Wa¯s.il, Jama¯l al-Dı¯n Muh. ammad ibn Sa¯lim. Mufarrij al-kuru ¯b fı¯ akhba¯r Banı¯ Ayyu ¯b. Vols. 1–3. Ed. Jama¯l al-Dı¯n al-Shayya¯l. Cairo, 1953–60; Vols. 4–5. Ed. S. A. F.
A ¯ shu ¯ r and H. M. Rabı¯
. Cairo, 1972–77. Usa¯mah ibn Munqidh. Kita¯b al-i
tiba¯r. Ed. Qa¯sim al-Sa¯marra¯¯. ı Riyadh, 1987.

Date

Site

Besieged by

Term

Source

518/1124

Tyre

Crusaders

manganı¯qe¯ rawrbe¯ (great trebuchets)

A.2, 94 A.2 [F], 71 A.2 [E], 95

519/1125


Aza¯z

Aq Sunqur al-Bursuqı¯, Seljuq governor of Mosul, and T . ughtigin, ¨rid Atabeg of Bo Damascus

manganı¯qe¯ rawrbe¯ (great trebuchets)

A.2, 98 A.2 [F], 73 A.2 [E], 97

532/1138

Shayzar

John II Komnenos

maja¯nı¯q ha¯ilah (huge trebuchets)

U, 134

maja¯nı¯q
iz.¯am (great trebuchets)

IA, 2: 268 B, Zubdah, 249

551–552/1157 Baghdad

Sultan Muh. ammad II

manjanı¯qayn
az.¯mayn ı (two great trebuchets)

1165

Andronikos I Komnenos

petrobo´ lou" mhcana´ " C, 134 ( petrobolous me¯khanas) C [E], 76 (rock-throwing engines)

Normans

maja¯nı¯q kiba¯r (big trebuchets)

ID, Fath. , 227–28 ID, Sana¯, 78 AS, 598–600 IW, 2: 11–16

al-manjanı¯qa¯t al-kiba¯r al-maja¯nı¯q al-kiba¯r (big trebuchets)

ID, Sana¯, 105 AS, 669–70 IW, 2: 47–48

Zevgminon

569–570/1174 Alexandria

572/1176

Mas.ya¯f

Saladin

579/1183

¯ mid A

Saladin

580/1184

Kerak

Saladin

al-manjanı¯qa¯t al-kiba¯r (big trebuchets)

ID, Sana¯, 241

1184

Nicaea

Andronikos I Komnenos

eJle´ poli" (helepolis) (city-taker)

C, 282 C [E], 156

manjanı¯q ha¯il (huge trebuchet), called al-Mufattish (The Examiner)

ID, Barq, 85

114

INVENTION OF THE COUNTERWEIGHT TREBUCHET

Date

Site

Besieged by

Term

Source

1185

Thessalonike

Normans

eJlepo´ lei" (helepoleis) (city-takers)

E, 72–73

seismou' quga´ thr E, 74–75; 96–99 (seismou thugate¯r), (The Daughter of the Earthquake); also known as mega´ lh mhcanh´ (megale¯ me¯khane¯) (great siege-engine), th´ n mhte´ ra (te¯n me¯tera) (Their Mother), and grai'a ( graia) (The Old Woman) 583/1187

Jerusalem

Saladin

manganı¯qe¯ rawrbe¯ (great trebuchets)

BH, 375 BH [E], 325

583/1187

Tyre

Saladin

al-maja¯nı¯q al-kiba¯r (big trebuchets)

ID, Fath. , 73–86, 110

1189

Vicenza

trabuchellus

Cod. Ecel., 97

586/1190

Acre

manjanı¯qayn kabı¯rayn (two big trebuchets) of Henry of Champagne, one costing 1,500 dinars, destroyed by sally in force; manjanı¯qa¯t ha¯ilah (two huge trebuchets) destroyed by fire-arrows shot from a large crossbow

ID, Fath. , 282

1199

Crusaders

Castelnuovo Cremonans Bocca d’Adda

trabuchus

IS, 136 AS, 2: 162 IW, 2: 335

GC, 25

Appendix 2 T . arsu ¯ sı¯’s Description of the Counterweight Trebuchet

T . arsu ¯ sı¯’s account of a counterweight trebuchet comprises not one but two descriptions. The first is merely a summary of the longer description that follows it. Both may have been composed by T . arsu ¯ sı¯ from information provided by Ibn al-Abraqı¯, or the first may be a short paraphrase of the original description received from Ibn al-Abraqı¯, which then follows it. By calling his counterweight machine a “Persian” trebuchet (manjanı¯q fa¯risı¯), T . arsu ¯ sı¯ does not mean to imply that it is of Iranian origin, as Cahen and White supposed (Cahen, “Traite´,” 158; White, Medieval Religion and Technology, 284). He is simply instructing the reader to take a traction trebuchet of the same design as one he has previously identified as a “Persian” or “Turkish” trebuchet (manjanı¯q al-fa¯risı¯ wahuwa al-turkı¯) and to make certain modifications to it in order to produce a counterweight machine. Here is his description: [B133v/A100r] A Description of a Persian Trebuchet, Made for Me by Shaykh Abu . asan ibn al¯ al-H Abraqı¯ al-Iskandara¯nı¯, with a Throwing Power of Fifty Men More or Less At the bedplate of its framework is a bow of a crossbow. The entire device [trebouchet and crossbow] is operated by a single man who discharges it [the double-purpose machine]. When the man hauls down the beam, the hemp cords, which are pulled by it [the beam], cause the bowstring to reach its catch. The man secures the pouch [of the sling] to a ring fixed to a base which holds the beam. He takes the [cross]bow and shoots it. Then he releases the beam, and the stone is thereby discharged. [B134r] Take a Persian trebuchet [manjanı¯q fa¯risı¯] and set it up to shoot. Dig a hole next to its framework to a depth equal to the length of the hemp cords on the beam. Then take a net of close-meshed hemp and place at its ends three strong hemp cords [A100v], long enough to reach from the end of the beam, where the axle is, to the bottom of the hole. At the extremity of the beam there should be an iron ring to which the cords of the net are attached. The quantity of stones placed in the net should be equal to the power of the men who [would be required to] pull the beam [of a traction trebuchet]. At the extremity of the beam, next to the cord of the pouch [of the sling], there should be a system of pulleys placed in a pulley block that hangs from the beam. After the man hauls down [the beam] [B134v], he places the stone in the pouch [of the sling] and hangs its cord on the hook [or the style] fixed to the end of the beam. The man [B135r] is able to pull the net [i.e., the counterweight] by pulling the end of the beam, since it swings back like a steelyard. After he hauls down [the beam], he secures [B134v] the pouch [of the sling], with an iron hook placed at its [the pouch’s] lower end, to a ring fixed to a base which holds the power [qu ¯wah, i.e., the weight] of the net [i.e., the counterweight] [B135r]. The bow of the crossbow

116

INVENTION OF THE COUNTERWEIGHT TREBUCHET

is placed at the bottom of the base of the trebuchet in two iron hooks that hold it [B134v]. Its [bow]string, with the cords which raise the net [i.e., the counterweight], is placed in a drawingclaw fixed to the cords, so when the cords rise with the net [i.e., the counterweight] [B135r], they pull the bowstring and convey it to the catch on its stock. After the man secures the pouch [of the sling] to the base, he takes the [cross]bow and places [B134v] the bolt on its course and shoots it. Then he returns immediately to the pouch [of the sling] and discharges it all by himself [B135r]. When the net [i.e., the counterweight] pulls the beam, it brings it back to its upright position; and this is better than the traction[-power] of men, because it [i.e., the net or counterweight] pulls with a constant force [B134v], [whereas] men differ in their pulling force. Here is an illustration of it [pictured on fols. B134v-135r/A101v-102r; see Fig. 1].

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Tarsu . ¯ sı¯, Tabs. irah fı¯ al-h. uru¯ b, Oxford, Bodleian, MS Hunt. 264, fols. 134v–135r. Tarsu¯ sı¯’s counterweight trebuchet, the earliest extant illustration of a gravity-powered trebuchet, dating from ca. 1187. This machine, operated by a single man, served two functions: it was a medium-size trebuchet, having the throwing power of a “fifty-man” traction machine, and it functioned as a spanning device for a large siege crossbow, shown on the left. Captions read: left (top to bottom), “the bow”; “the net which has in it bases [sic]” [stones]; “the hole in which the net [the counterweight] descends”; center, “the position of the man for hauling down the beam and shooting the stone”; right (top to bottom), “the pouch”; “the stone”; “the U-bolt”; “the base.”

2

Rashı¯d al-Dı¯n, Ja¯ m i‘ a l-tawa¯rı¯kh (1306–1307), Edinburgh, Edinburgh University Library, MS Arab 20, fol. 130v. The army of Mah.mu¯d ibn Sebuktegin captures the fortress of Muh.ammad ibn Abu¯ Nas. r Muh.ammad in 1012. The operator of the counterweight trebuchet—depicted in Arab or Persian attire—is about to discharge the machine by striking a blow to the pin holding down the beam with his mallet. Bolts and metal plates create strong and effective joints on the framework of the machine and on the box holding the counterweight.

3

Conrad Kyeser of Eichstätt, Bellifortis (ca. 1405), Göttingen, Niedersächsische Staats- und Universitätsbibliothek, Cod. MS philos. 63, fol. 30r. A large counterweight trebuchet. The main beam, counterweight box, sling, projectile, trough, windlass, and framework are all clearly visible with dimensions given for some of the component parts. The main beam measures 54 “workfeet,” or 15.55 m, with a throwing arm of 46 “workfeet,” or 13.248 m. The distance from the axle of the beam to the axle of the hinged counterweight box is 8 “workfeet,” or 2.304 m, dividing the beam in the ratio 5.75:1. A ratio of 6:1 is designated for the beam of a trebuchet in the Innsbruck manuscript of Bellifortis (Fig. 4), indicating that the dimensions given here may have been miscalculated. The trestle frame is composed of two linked supporting trusses, each forming an equilateral triangle with base and sides measuring 46 “workfeet,” or 13.248 m. The main axle is placed at the apex of the trusses 11.47 m above the ground. The main beam is banded by metal plates to withstand splitting, and the trusses of the framework are reinforced by horizontal braces that are bolted to the structure. The prong at the end of the long arm, which is essential for the release of the sling, is not depicted. Instead, both cords of the sling are incorrectly shown as being attached to a ring at the extremity of the long arm. This massive machine used a simple peg-and-hole, catch-and-trigger device to retain and release the beam. A hole drilled in the base of one of the trusses of the machine contains the peg. A restraining rope, attached to the base of the other truss is drawn over the long arm of the beam at a point just above the windlass and is looped over the bottom end of the peg. When the peg is lifted out of its socket, the looped end of the rope is released, and the beam flies free. This type of machine was identified in Arabic historical sources as the Western Islamic trebuchet (manjanı¯q maghribı¯ ) and in Byzantine sources as the helepolis (city taker).

4

Conrad Kyeser of Eichstätt, Bellifortis, Innsbruck, Tiroler Landesmuseum Ferdinandeum, MS 16.0.7, fol. 21r. A large counterweight trebuchet noticeably different from its counterpart in the Göttingen manuscript of Bellifortis (Fig. 3). The main beam, counterweight box, and trestle framework are clearly shown with some components dimensioned. The main beam measures 56 “workfeet,” or 16.12 m, with a throwing arm of 48 “workfeet,” or 13.82 m. The distance from the axle of the beam to the axle of the hinged counterweight box is 8 “workfeet,” or 2.30 m, dividing the beam in the ratio 6:1. The trestle frame is composed of two linked supporting trusses, each forming an equilateral triangle with base and sides measuring 48 “workfeet,” or 13.82 m. The main axle is placed at the apex of the trusses 11.97 m above the ground. The rounded beam is composed of three spars of wood that are banded by rope and reinforced by three giant wooden pegs. A large strut is used to prop the counterweight box, and the truss of the framework is reinforced by horizontal braces that are bolted to the structure. The dimensions provided for this trebuchet revise some of the measurements given for the machine in the Göttingen manuscript of Bellifortis (Fig. 3). The measurement for the throwing arm in this illustration was originally written as “46,” corresponding to the figure given in the Göttingen manuscript, and was then changed to “48.” The draftsman realized that the dimensions for the machine in the Göttingen manuscript would not divide the beam in the ratio 6:1 and made the necessary corrections.

5

Mariano di Jacopo Taccola, Liber Tertius de ingeneis ac edifitiis non usitatis (1433), Florence, Biblioteca Nazionale Centrale, Cod. palat. 766, fol. 41r. The “brichola” (bricola), a pole-framed trebuchet with two hinged counterweights. This machine was identified in Arabic historical sources as the Frankish or European trebuchet (manjanı¯q firanjı¯/ifranjı¯ ). In Byzantium, the bricola was designated by the Latin load word praikoula or prekoula.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Paris, Bibliothe`que Nationale de France, Codex Grec 54: Modus Operandi of Scribes and Artists in a Palaiologan Gospel Book KATHLEEN MAXWELL

uch can be learned about the working methods of scribes and artists when they leave their product unfinished.1 This is true of the late thirteenth-century illustrated Gospel book, Paris, Bibliothe`que Nationale de France, cod. gr. 54. It was intended to be a deluxe, bilingual (Greek and Latin) manuscript featuring full-page evangelist portraits and an extensive narrative cycle of some fifty-two miniatures. In the end, only about half of its Latin text and twenty-two narrative miniatures were actually completed. Five other miniatures remain unfinished, and space was reserved in the text for twentyfive additional miniatures that were never even begun.2 No ornamental headpieces were painted (although space was reserved for them by the scribe), and the Latin text lacks many initials.3 These unfinished elements offer insight into the working methods of Paris 54’s scribes and artists, and they allow us to recognize some degree of interaction between members of these two groups.

M

This article is derived in part from my dissertation, “Paris, Bibliothe`que Nationale, Codex Grec 54: An Analysis of the Text and the Miniatures” (Ph.D. diss., University of Chicago, 1986). Professor Robert S. Nelson suggested that I research Paris 54 for his Byzantine manuscript illumination seminar during the winter quarter of 1978 and served as my dissertation advisor. I am happy to take this opportunity to thank Professor Nelson for the interest and support he has demonstrated over the years in my work on this topic. I also wish to thank Kathryn Wayne, and her predecessor, Ann Gilbert, art history librarians of the University of California at Berkeley, for the generous access accorded me to the art history seminar since 1982 when I moved to San Francisco. Thank you also to Dr. Alfred Bu ¨ chler, Professor Christine Havice, and Professor Richard Rouse, who read various drafts of this manuscript. I have also benefitted from the comments of the anonymous readers for Dumbarton Oaks Papers to whom I am much indebted. Finally, very special thanks to my husband, Paul Missiroli Sullam, whom I met shortly after I began my study of Paris 54 and who has shown unconditional support for my work ever since. 1 See J. Lowden, The Octateuchs: A Study in Byzantine Manuscript Illumination (University Park, Pa., 1992), 23–26, 64 (re: the Seraglio Octateuch); J. C. Anderson, “The Seraglio Octateuch and the Kokkinobaphos Master,” DOP 36 (1982): 101. 2 See the Appendix for a description of the contents of Paris 54. 3 Neither the Greek nor Latin texts of any of the four Gospels ever received an opening decorated initial. Moreover, Luke’s Gospel contains no titles for either the Greek or Latin texts, and John’s Gospel has no title for its Latin text (see Figs. 1 and 17).

118

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

Our understanding of the circumstances in which illustrated manuscripts of the early Palaiologan era were produced remains limited. There is no evidence of an imperial scriptorium,4 and the execution of text and miniatures is thought to have occurred in separate locales.5 Typically, the scribe appears to have been responsible for the basic design of the manuscript, including text format and decisions concerning the size and location of decoration and/or illustration. That is, he wrote the text, reserving space as appropriate for headpieces, illuminated initials, and illustrations. Once his task was accomplished, he presumably turned the completed text over to an artist (or a group of artists) for ornament and miniatures. However, even when significant numbers of manuscripts can be grouped together based on script, ornament, and/or illustration, they often reveal surprisingly complex relationships between scribes and artists that do not conform to any convenient definitions evoked by the term “atelier.” 6 The results of my study of Paris 54 underscore the highly variable nature of the artist/scribe relationship. We shall see that the relative autonomy of the primary scribe responsible for the Greek text is upheld, but correspondences between anomalies in the Latin text and the miniatures of the narrative cycle indicate that some scribes and artists worked more intimately in the production of Paris 54 than appears to have been typical of Byzantine manuscripts of this or other periods.7 I will begin with a review of the literature related to Paris 54 and proceed gospel by gospel, analyzing texts and miniatures in an effort to determine how Paris 54 was produced. See H. Belting, Das illuminierte Buch in der spa¨tbyzantinischen Gesellschaft (Heidelberg, 1970), 51–60. See H. Buchthal and H. Belting, Patronage in Thirteenth-Century Constantinople: An Atelier of Late Byzantine Book Illumination and Calligraphy, DOS 16 (Washington, D.C., 1978), 91–93. Robert S. Nelson writes that “a division of labor between the writing and decorating of a Greek manuscript appears to have long existed in Byzantium, and examples of the disassociation of script and decoration have been noted from the tenth to the fourteenth centuries.” See further, R. Nelson, Theodore Hagiopetrites: A Late Byzantine Scribe and Illuminator (Vienna, 1991), 118 ff and n. 23, for a helpful summary of these issues and related bibliography. 6 Nelson writes: “Thus deluxe Greek manuscripts were produced by no single method, and varying combinations of individuals might collaborate on the three basic components, script, ornament, and figural miniatures.” See further, Nelson, Theodore Hagiopetrites, 120. The difficulties encountered in trying to define how manuscripts were produced in the late 13th century are echoed in R. Nelson and J. Lowden, “The Palaeologina Group: Additional Manuscripts and New Questions,” DOP 45 (1991): 62–63 and 67–68. See also ¨ B 21 (1972): 48 ff; L. Brubaker, H. Buchthal, “Illuminations from an Early Palaeologan Scriptorium,” JO “Life Imitates Art: Writings on Byzantine Art History, 1991–1992,” BMGS 17 (1993): 173–223; and esp. S. Dufrenne, “Proble`mes des ateliers des miniaturistes byzantins,” Akten, I/2: XVI Internationaler Byzantinistenkon¨ B 31.2 (1981): 445–70. gress ⫽ JO 7 Exceptions include an unusual example of close collaboration between a scribe and an illuminator in a Greek New Testament from the Escorial (cod. X.IV.17) noted by Buchthal: “They were members of a single team. There are comparatively few instances in Byzantine illumination where this point—a point which we take for granted in the Latin West—is so obvious.” See H. Buchthal, “A Greek New Testament Manuscript in the Escorial Library: Its Miniatures and Its Binding,” in Byzanz und der Westen: Studien zur Kunst des europa¨ischen Mittelalters, ed. I. Hutter, SBWien, vol. 412 (Vienna, 1984), 92–93. Irmgard Hutter describes the differing relationship between art and text in Byzantium and the Latin West in “Decorative Systems in Byzantine Manuscripts, and the Scribe as Artist: Evidence from Manuscripts in Oxford,” Word and Image 12.1 (1996): 4 and 18 n. 12 for further bibliography. There are cases in Byzantine manuscripts where the scribe is also the illuminator and/or the artist, see Nelson, Theodore Hagiopetrites, esp. 116 ff; and K. Corrigan, “Constantine’s Problems: The Making of the Heavenly Ladder of John Climacus, Vat. gr. 394,” Word and Image 12.1 (1996): 61–93, where she demonstrates that the scribe, Constantine, was also responsible for the miniatures. 4 5

KATHLEEN MAXWELL

119

HISTORIOGRAPHICAL REVIEW Paris 54 attracted the attention of New Testament scholars almost a century before its citation in the publications of paleographers or art historians.8 Paris 54’s bilingual text led many to propose that it was produced on Italian soil.9 This hypothesis attracted Gabriel Millet’s opposition as early as 1916 and was opposed again in 1931 by Viktor Lazarev.10 It was not until 1929, when Henri Omont fully published Paris 54’s illustrations, that the manuscript began to be consistently noted by art historians.11 In 1944 Kurt Weitzmann related the evangelist portraits of Paris 54 to those of four other Byzantine codices: Athos, Iviron 5; Athens, National Library, cod. 118; Princeton, University Library, MS Garrett 2 (olim Andreaskiti 753), and Philotheou 5.12 Iconographic parallels between some of these Byzantine portraits and several evangelists in the Wolfenbu ¨ ttel Sketchbook, a Western product datable to the second quarter of the thirteenth century, as well as the presence of Latin scripts in some of the portraits of the Byzantine group, led Weitzmann to suggest that the latter must date from the first half of the thirteenth century, a period coinciding with the Latin Occupation of Constantinople (1204–61).13 He proposed further that these manuscripts were produced in Constantinople following tenth-century models and had been commissioned, or perhaps adjusted, for a Latin customer.14 Weitzmann was apparently unaware that a link had already been made between 8 See, for example, Johan Jakob Wetstein, Novum Testamentum Graecum editionis receptae cum lectionibus varientibus Codicum MSS., Editionum aliarum, Versionum et Patrum, nec non commentario pleniore ex scriptoribus veteribus Hebraeis, Graecis et Latinis historiam et vim verborum illustrante, vol. 1, Continens Quattuor Evangelia (Amsterdam, 1751), 47, where he dates Paris 54 to the 14th century based on its Latin script. 9 The following thought that Paris 54 was probably produced in Italy: J. P. P. Martin, Description technique des manuscrits grecs relatifs au Nouveau Testament conserve´s dans les bibliothe`ques de Paris (Paris, 1884), 29; N. V. Pokrovskij, Evangelie v pamjatnikach ikonografii preimusˇˇcestvenno vizantijskich i russkich (St. Petersburg, 1892), xx; A. Vaccari, “La Grecia nell’Italia meridionale,” Orientalia Christiana 3.13 (1925): 273 ff (I thank Susan Pinto Madigan for this reference.); H. Gerstinger, Die griechische Buchmalerei (Vienna, 1926), 38; A. Grabar, La peinture religieuse en Bulgarie (Paris, 1928), 170; and H. Omont, Miniatures des plus anciens manuscrits grecs de ¨ber die Farbengebung la Bibliothe`que nationale du VIe `a XVIe sie´cle (Paris, 1929), 47 n. 1. J. J. Tikkanen, Studien u in der mittelalterlichen Buchmalerei, Commentationes Humanarum Litterarum, Societas Scientiarum Fennica (Helsinki, 1933), 132 and 163, suggested that Paris 54 may have come from Italy. 10 ´ vangile aux XIVe, XVe et XVIe sie`cles (Paris, 1916), 646; V. Lazarev, G. Millet, Recherches sur l’iconographie de l’E “Duccio and Thirteenth Century Greek Icons,” The Burlington Magazine 59 (1931): 159. 11 Omont, Miniatures des plus anciens manuscrits grecs, pls. XC–XCVI. 12 It has since been suggested that the texts of two of these manuscripts date from prior to the 13th century and that their current portraits were added during the Palaiologan period. Belting supports an early 12thcentury date for the text of Princeton, Garrett 2, while Prato has argued that the text of Athens 118 can be dated to the late 10th century on paleographical grounds. Prato also notes that both Paris 54’s and Iviron 5’s Greek texts betray their 13th-century date through the intrusion of numerous majuscule letters, especially epsilon and eta. See Belting, Das illuminierte Buch, 61; G. Prato, “Scritture librarie arcaizzanti della prima eta` dei Paleologi e loro modelli,” Scritture e civilta` 3 (1979): 169 and pl. 5; and idem, “La produzione libraria in area greco-orientale nel periodo del regno latino de Costantinopoli (1204–1261),” Scritture e civilta` 5 (1981): 105–7. 13 Few dated 13th-century manuscripts were known at the time Weitzmann was writing his article, see K. Weitzmann, “Constantinopolitan Book Illumination in the Period of the Latin Conquest,” GBA 25 (1944): 193–214, reprinted in Studies in Classical and Byzantine Book Illumination, ed. H. L. Kessler (Chicago, 1971), 314 ff. 14 Ibid., 200 f. Weitzmann was following Millet’s lead here with regard to both Paris 54’s date and locale. ´ vangile, 646. See Millet, Recherches sur l’iconographie de l’E

120

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

Paris 54 and a manuscript dated to 1298. In 1933 J. J. Tikkanen noted that two of the evangelist portraits of Paris 54 recur in the Smyrna Lectionary.15 In 1949 Lazarev reviewed Weitzmann’s article and insisted that Weitzmann’s whole group be assigned to the second half of the century on the basis of stylistic comparisons with such dated thirteenth-century manuscripts as London, British Library, Burney 20 (dated 1285).16 Less than a decade later, Otto Demus wrote his landmark article on the origins of the Palaiologan style and supported Lazarev’s later date for most members of Weitzmann’s manuscript group. He also cautioned that stylistic analysis of these portraits was extremely difficult, since many follow the style of their tenth-century models while others were executed in a wholly Palaiologan mode. Demus opted for a date in the 1260s for both Paris 54 and Iviron 5.17 These dates were revised just a few years later when Demus proposed a date within the rule of Michael VIII (1261–82) for Iviron 5. Demus contrasted its “painterly” style with the “cubist” and “extreme solution” of Paris 54, which he dated ca. 1285–89.18 A date for Paris 54 late in the thirteenth century was implied by Rudolph Naumann and Hans Belting in their study of the Euphemia Church in Constantinople,19 when they arrived at a date in the fourth quarter of the thirteenth century for the church’s fresco cycle after comparing its ornament to that of many of the monuments discussed by Lazarev.20 In his 1967 survey of Byzantine art, Lazarev devoted considerable attention to both Iviron 5 and Paris 54 and their respective dates. He dated Iviron 5 to the last third of the thirteenth century and Paris 54 to the last quarter of the thirteenth century. Lazarev thought that the former was a first-rate Constantinopolitan product and that Paris 54 was based on the same model but provincial in execution. He believed Paris 54 to be the product of a Macedonian master working for a high-ranking Crusader or Latin prelate. Lazarev also argued that Paris 54 was later in date because its style was more representative of mature Palaiologan trends than Iviron 5.21 A few years later, Belting dated Iviron 5 to the mid-thirteenth century and suggested that Paris 54 was about two decades later (ca. 1280).22 In 1975/76 an important contribution concerning the chronology of Weitzmann’s ¨ber die Farbengebung, 201 n. 1. See Tikkanen, Studien u See V. N. Lazarev, review of Weitzmann, “Constantinopolitan Book Illumination in the Period of the Latin Conquest,” in VizVrem 2 (1949): 368 f (in Russian). Later, in an article supporting a date ca. 1285 for Leningrad 101, Lazarev again argued in favor of his later date for Weitzmann’s group. Stylistic comparisons were also made with Jerusalem, Taphou 5, and with the Mokwi Gospels (dated 1300); see V. N. Lazarev, “Novyj pamiatnik konstantinopol’skoj miniiatiury XIII v.,” VizVrem 5 (1952): 178–90. 17 O. Demus, “Die Entstehung des Palaeologenstils in der Malerei,” Berichte zum XI. Internationalen Byzantinisten-Kongress (Munich, 1958), 4.2:20. 18 O. Demus, “The Style of the Kariye Djami and Its Place in the Development of Palaeologan Art,” in The Kariye Djami, vol. 4, Studies in the Art of the Kariye Djami and Its Intellectual Background, ed. P. Underwood (Princeton, 1975), 145–46. (N.B.: Demus’s manuscript was submitted in 1960.) 19 R. Naumann and H. Belting, Die Euphemia-Kirche am Hippodrom zu Istanbul und ihre Fresken, IstForsch 25 (Berlin, 1966), 153–68. 20 See above, note 16. 21 V. Lazarev, Storia della pittura bizantina (Turin, 1967), 278–81. 22 Belting, Das illuminierte Buch, 35 and 40 n. 134. 15 16

KATHLEEN MAXWELL

121

group was made by Stella Papadaki-Oekland, when she republished photos of the Smyrna Lectionary (dated 1298). The Mark and John portraits of this manuscript are of the same type as their counterparts in Iviron 5 and Paris 54. If the Smyrna manuscript’s portraits can also be assumed to date to 1298,23 then Papadaki-Oekland has provided all the evidence necessary to push back, by perhaps as much as a decade, the date of some of the later members of Weitzmann’s group.24 Papadaki-Oekland suggests that both Athens 118 and Iviron 5 should be dated before the Smyrna Lectionary, while Princeton, Garrett 2 and Philotheou 5 immediately post-date it.25 Unfortunately, Papadaki-Oekland was not aware that connections between the Smyrna Lectionary and members of Weitzmann’s Constantinopolitan group had already been made by Tikkanen long before even Weitzmann’s 1944 publication.26 Hugo Buchthal reassessed the Weitzmann group in his comprehensive study of the Wolfenbu ¨ ttel Sketchbook.27 He argued for dates as early as the 1250s for the Athens Gospels and after 1300 for later manuscripts. On the basis of stylistic comparisons with the frescoes of Sopoc´ani, Iviron 5 was dated ca. 1265 and Paris 54 was assigned to the last quarter of the thirteenth century. As a result of this fifty-year span, Buchthal argued that these manuscripts should no longer be considered as a group: They do not, in fact, form a real group; the miniatures are not homogeneous enough to be assigned to a single scriptorium or school of illumination. The manuscripts present a rather diversified picture, in which it is only the constant repetition of certain clearly recognizable types which reappear in various combinations, that gives the impression of a fairly uniform style. In other words, the miniatures are transmitting the style of their iconographical models, even more than in earlier periods of Byzantine art. And when the model is somewhat antiquated and the development extends over several decades, the survival of isolated figure types should not be taken for real continuity.28

Buchthal only acknowledges the possibility of a one-to-one copy relationship between Iviron 5 and Paris 54 and isolates the two manuscripts as a distinct subgroup.29 With this much effort to place Paris 54 in the third or even fourth quarter of the thirteenth century, it is surprising to note that Weitzmann’s mid-thirteenth-century date The manuscript was apparently destroyed in the early 1920s. S. Papadaki-Oekland, “OiJ mikrografi´e" eJno` " came´ nou ceirogra´ fou tou' 1298,” Delt.Crist. Arc. j JEt., ser. 4, 8 (1975–76): 29–54 (in Greek with English summary). More recently, Nelson and Lowden, “The Palaeologina Group,” 62, have linked the scribe of the Smyrna Lectionary to manuscripts associated with the “Atelier of the Palaeologina.” For the latter, see Buchthal and Belting, Patronage in Thirteenth-Century Constantinople. 25 Papadaki-Oekland, “OiJ mikrografi´e",” 51–54. The exact relationship of the Paris 54 portraits to the Smyrna portraits is not discussed in the English summary. 26 Ibid. In her English summary, Papadaki-Oekland writes that the Smyrna manuscript has not been mentioned in the literature since 1909. No one concerned with the issues of “Weitzmann’s Constantinopolitan manuscript group” seems to have been aware of Tikkanen’s contribution—least of all Weitzmann. See Tikka¨ber die Farbengebung, 201 n. 1. nen, Studien u 27 ¨ttel, ByzVindo 12 (Vienna, 1979), 45–46. See now also, I. SpathH. Buchthal, The “Musterbuch” of Wolfenbu arakis, The Left-handed Evangelist: A Contribution to Palaeologan Iconography (London, 1988). The Wolfenbu ¨ ttel Sketchbook plays a central role in Spatharakis’ discussion. 28 ¨ttel, 45–46. Buchthal, The “Musterbuch” of Wolfenbu 29 Ibid., 47. A few years later, Nelson pushed back the date of Iviron 5 to ca. 1285 on the basis of comparisons with London, British Library, Burney 20 (dated 1285). See R. S. Nelson, “Paris, Gr. 117 and the Beginnings of Palaeologan Illumination,” WJKg 37 (1984): 18. 23 24

122

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

is upheld in the Louvre catalogue that accompanied the 1992 exhibition Byzance: l’art byzantin dans les collections publiques franc¸aises.30 As the foregoing demonstrates, a definitive chronology for thirteenth-century Byzantine illustrated manuscripts remains elusive. However, great strides have been made in the last twenty-five years, and much of that progress is built upon the foundations provided by Lazarev in his Russian publications of more than a half century ago. Paris 54 remains something of an anomaly. Extensively illustrated Gospel books are rare in the Palaiologan period. As has been pointed out by others, it appears that monumental painting became the preferred venue for extensive narrative cycles in this period.31 However, as we shall see in the next section, there are many aspects of Paris 54 that contribute to its singular status, above and beyond its narrative cycle. DESCRIPTION Paris 54 is a generously proportioned manuscript (335 by 250 mm), with a handsome double-column layout (Figs. 1 and 7),32 a format dictated by its bilingual text.33 The Greek text takes precedence over the Latin through its location in the left column. Paris 54 is bilingual in another aspect as well. It contains two different sets of quire signatures.34 Greek quire signatures are found through almost two-thirds of the manuscript (through quire 22; see Appendix). Armenian quire signatures appear throughout the entire manuscript (i.e., through quire 35).35 The Armenian quire signatures in Paris 54 are likely to Muse´e du Louvre, 3 November 1992–1 February 1993 (Paris, 1992), 450 (with color illustration of the Evangelist Mark). The bibliography in Paris 54’s catalogue entry does not cite, however, any more recent publication by an art historian than the 1971 reprint of Weitzmann’s “Constantinopolitan Book Illumination” (above, note 13). 31 See Nelson and Lowden, “The Palaeologina Group,” 64. 32 The two texts were not allotted equal space; the Greek text column averages 9.3 cm in width and always continues through the first part of the space between the two text columns, which itself measures 1.4 cm in width (the distance from b to c in Fig. 7). The Latin text, on the other hand, averages only 7 cm in width. 33 See B. M. Metzger, Manuscripts of the Greek Bible: An Introduction to Greek Paleography (New York, 1981), 56, for a list of bilingual New Testament manuscripts. However, a more fruitful comparison, from a paleographical standpoint, is provided by the bilingual (Greek and Latin) Hamilton Psalter. See C. Havice, “The Marginal Miniatures in the Hamilton Psalter (Kupferstichkabinett 78.A.9),” JbBM 26 (1984): 79–142. 34 For more information on quires, see E. Gamillscheg and R. Browning, “Quire,” ODB 3:1767–68. 35 C. R. Gregory, Textkritik des Neuen Testaments (Leipzig, 1909), 1:132, writes that the gatherings were first numbered in Greek and then in Armenian. This is unlikely, however, since, unlike the Greek quire signatures, the Armenian quire signatures extend throughout the entire manuscript. Gregory also wrote: “I have no doubt that it was at least in part the work of an Armenian; there are Armenian as well as Greek numbers for quires.” Idem, Canon and Text of the New Testament (New York, 1907), 372. Other manuscripts that feature Armenian quire signatures include Brescia, A. VI.26, a Greek Gospel book whose evangelist portraits were repainted by an Armenian artist from the Crimea in the 14th century. See R. S. Nelson, The Iconography of Preface and Miniature in the Byzantine Gospel Book (New York, 1980), 42; K. Maxwell, “Armenian Additions to a Greek Gospelbook: Brescia, Biblioteca Civica Queriniana A. VI.26,” REArm 25 (1994–95): 337–52; and H. Buschhausen and H. Buschhausen, “Die Halbinsel Krim, ein wenig beachtetes Zentrum der byzantinischen Buchmalerei im 14. Jahrhundert,” Byzantine East, Latin West. Art-Historical Studies in Honor of Kurt Weitzmann, ed. C. Moss and K. Kiefer (Princeton, 1995), 339–43. In addition, Millet noted long ago that Florence, Laurentian Library, Plut. VI. 23 was paginated in Armenian, presumably when an Armenian copy of it (Erevan, Matenadaran 7651) was produced in the 13th ´ vangile, 569. See also S. Der Nersessian, Miniature Painting century. See Millet, Recherches sur l’iconographie de l’E 30

KATHLEEN MAXWELL

123

date from the manuscript’s production, since they share the somewhat unusual red ink color of the Greek quire signatures and the running text. Moreover, brief notations in Armenian are also found in the lower margins of folios 53r–79v.36 Paris 54’s quires are also unusual in that each contains ten folios rather than eight, as was the rule in Greek manuscripts.37 There is yet another aspect of Paris 54 that distinguishes it from virtually every other Byzantine manuscript: its ink is color coded according to the speaker in the text.38 Bright red ink was used for the simple narrative text, while a darker red or crimson ink was reserved for Jesus’ words, the genealogy of Christ, and the words of the angels. Old Testament passages, the words of the disciples, Zachariah, Mary, Elizabeth, Simeon, and John the Baptist are in blue, while dark brown ink was used for the speech of the Pharisees, people from the crowd, Judas Ischariot, the Centurion, the devil, shepherds, and the scribes.39 The Greek text, written in a fine archaizing hand based on the eleventhcentury pearl script, is particularly impressive in appearance.40 in the Armenian Kingdom of Cilicia from the Twelfth to the Fourteenth Century, DOS 31 (Washington, D.C., 1993), 1:173 f, where an instruction to the scribe in Armenian on fol. 27v plot. VI. 23 is also noted. Finally, Francis Wormald noted the presence of Armenian quire signatures in a 12th-century Latin Crusader manuscript known as the Missal of the Church of the Holy Sepulchre (Paris, Bibliothe`que nationale, cod. lat. 12056). He concluded that its scribe “was an Armenian who could write Latin.” See H. Buchthal, Miniature Painting in the Latin Kingdom of Jerusalem, with paleographical and liturgical chapters by F. Wormald (Oxford, 1957), 135. 36 I have not succeeded in translating these notations. They are comprised primarily of a few indecipherable letters. 37 Quinions are not typical of Armenian manuscripts either, as they also usually feature eight folios per quire. However, quinions (i.e., gatherings containing ten folios) were apparently typical in late antiquity and both the Rossano Codex and the Vienna Dioscurides feature quires made up of quinions. See G. Cavallo, G. Gribomont, and W. C. Loerke, Codex Purpureus Rossanensis, vol. 1, Commentarium, Codices Selecti, facs. 81 (Rome-Graz, 1987), 24. C. De Hamel, Scribes and Illuminators (Toronto-Buffalo, 1992), 18, writes that gatherings of ten leaves were also common in early Irish manuscripts and 15th-century Italian books. See ibid., 20, for the difficulties in constructing a quire of ten leaves relative to that of eight leaves. 38 Gregory, Canon and Text, 372, dubbed Paris 54 the “rainbow manuscript.” The only other Greek manuscript that I have encountered that offers a comparable effect is Oxford, Bodleian Library, Lincoln MS gr. 35 where on fols. 14r and 15v (only) the text is varied in blue, gold, and red. See A. Cutler and P. Magdalino, “Some Precisions on the Lincoln College Typikon,” CahArch 27 (1978): 179–98; and I. Spatharakis, The Portrait in Byzantine Illuminated Manuscripts (Leiden, 1976), 190–207. F. Mu ¨ therich and J. E. Gaehde, Carolingian Painting (New York, 1976), 27, no. XVI, note that the first three pages of the sacramentary fragment from Metz (ca. 870) “are written in alternating lines of gold, green, and red capital letters on blank parchment.” See also the facsimile Sakramentar von Metz, Ms. Lat. 1141, Bibliothe`que Nationale, Paris, intro. F. Mu ¨ therich, 2 vols. (Graz, 1972). Nelson indicates that varied ink colors are also a characteristic of Gothic manuscripts: see Iconography of Preface and Miniature in the Byzantine Gospel Book, 42. An example of the discriminating use of varied ink colors to enrich the effect of the script is also seen in the Armenian realm. The text on the initial page of each gospel of the Glajor Gospels at the University of California at Los Angeles is written in splendid zoomorphic initials. These are followed in Matthew by several folios of gold script and in Mark, Luke, and John by two pages of red, black, and blue ink, respectively. See T. F. Mathews and A. K. Sanjian, Armenian Gospel Iconography: The Tradition of the Glajor Gospels, DOS 29 (Washington, D.C., 1991). 39 That is, dark brown ink was used by mere humans, both of a positive and negative disposition towards Christ. Gregory, Canon and Text, 373, was the first to describe correctly the hierarchy of ink color in Paris 54. 40 Pearl script is extremely regular and legible. See further Prato, “Scritture librarie arcaizzanti,” 151–91. See also H. Hunger, “Archaisierende Minuskel und Gebrauchsschrift zur Blu ¨ tezeit der Fettaugenmode: Der Schreiber des Cod. Vindob. Theol. gr. 303,” in La Pale´ographie grecque et byzantine, Colloques Internationaux du Centre National de la Recherche Scientifique 559 (Paris, 1977), 283–90, esp. 285 f.

124

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE MATTHEW (FOLS. 11R–109V)

Text An analysis of the Greek text of Matthew suggests that the use of blue ink, at least, may have been an afterthought. Matthew’s text begins with Christ’s genealogy and was executed in the rarer crimson ink. Following the genealogy (Matt. 1:17), the text reverts to the bright red of the running text.41 The first instance of blue ink, on folio 13r (Matt. 1:20), is superimposed directly over the bright red ink of the running text. For the next six folios, every example of blue ink was written over an erasure of red ink (e.g., Fig. 3). Subsequent appearances of blue ink in the Greek text show no trace of erasure beneath them.42 Therefore, the use of blue ink appears to have been conceived, and perfected, after the Greek text was initiated. Since the erasures occur only in the Greek text, I conclude that the Greek text was written in large part, if not in total, before the Latin text was begun.43 Several scribes executed the Latin text of the Gospel of Matthew, and they were careful to maintain absolute color symmetry with the Greek text.44 Undoubtedly, this symmetry was motivated by aesthetic aims. It was also necessary that both texts break for the narrative miniature at the same point, since the miniature extends across both text columns (Fig. 2).45 As a result, an almost line-for-line color parallelism was attained throughout the text of Matthew. The scribal machinations through which this symmetry was achieved are of some interest. If, for example, the scribe was not able to stretch the Latin text to fill the requisite number of lines (this having been determined by the Greek version on the left), he might then repeat regular minims (mmmmmmmm) to avoid leaving a gap. This horror vacui mentality is best illustrated, perhaps, on folio 16v (Fig. 3), where these minims continue for five and one-half lines in red ink in order to keep the following passage in blue ink parallel with the Greek text. The last four lines of folio 86r (Fig. 4) exhibit the same repetition of minims, but here they are randomly grouped to resemble words more closely (mmmmmmmm mmm mmmmm mm mmmm). Folio 16r offers an example of a more decorative means of maintaining color parallelism. Here, a series of flourishes

on lines 15–17 (Fig. 5) accomplishes the same result. Other text-stretching techniques include the repetition of the final words in a verse (e.g., ieiunantes, fol. 25v), the writing out of a word followed by its abbreviation (e.g., loquamini, fol. 38v; hominibus, fol. 39r), or even the repetition of an entire verse. Matthew 9:11, for example, was repeated at the 41 Most Byzantine manuscripts are written in brown or black ink. However, Metzger writes that “whole manuscripts were sometimes written with red ink.” See further, idem, Manuscripts of the Greek Bible, 17. 42 Gregory, Textkritik, 1:132, also noted these erasures. 43 These observations do not suggest a copy relationship with another manuscript (at least regarding this aspect of the text); in any event, there are no surviving manuscripts to support this assumption. 44 I wish to thank Albert Derolez, Distinguished Visiting Professor, University of California, Berkeley, fall 1985, for reviewing a microfilm of Paris 54 with me and for very generously sharing his observations on its Latin script. 45 See Lowden, Octateuchs, 70, where he demonstrates the importance of the correct location of the illustration vis-a`-vis the text in his analysis of Vat. gr. 747.

KATHLEEN MAXWELL

125

top of folio 34v to maintain color harmony with the Greek text, even though this verse had already been written at the bottom of folio 34r.46 Far more frequently encountered than such text-stretching devices, however, is the resort to abbreviation, crowding, and awkward spacing to fit the Latin text in a space predetermined by the Greek. Obvious examples of this are seen on folios 22v (Fig. 6) and 35r, but many more could be cited. This problem was certainly exacerbated by the unequal amount of space allotted to the Latin text columns (see Fig. 7). Narrative Miniatures Matthew’s Gospel contains twelve narrative miniatures in addition to a full-page portrait of the evangelist inserted on a separate folio. All of these miniatures were completed by one of two artists, whom I have dubbed Artist A and Artist B.47 The evidence derived from Matthew’s Gospel allows only limited conclusions concerning the production of this manuscript. Based on the evidence of the initial blue ink passages in the Greek text, however, we can deduce that Matthew’s Greek text was completed before its Latin counterpart. We do not know, however, when the miniatures were executed. An analysis of the next three gospels indicates that, there, the miniatures were completed at the same time as, or even before the Latin text.48 MARK (FOLS. 111R–171V) Text The Greek text of Paris 54 was written in the varied ink colors throughout all four Gospels, but this practice was maintained in the Latin text only through the first thirty folios of Mark (up to fol. 142r).49 In fact, already by the end of the first quire of Mark, the color correspondence begins to break down.50 On folio 121r in the Latin text (Fig. 8), the top half of the page and two lines at the bottom have been written by a different hand from those seen thus far in the Latin text of Paris 54. Moreover, these passages, which should have been written in the rarer crimson ink, were instead written in black ink. The central portion of the folio was executed in bright red and dark brown inks, and in both cases the colors correspond appropriately, and the script is in the style of the Latin scribe. The new scribe’s handiwork can also be identified on folio 122v, and here again passages that should have been written in crimson were incorrectly executed in the black ink. The repetition of biblical passages in order to synchronize with extensive catenae is also noted by Lowden with regard to Vat. gr. 747. See ibid., 11–12. 47 Paris 54’s narrative cycle will be discussed in more detail in a future publication. 48 While Matthew’s Gospel is the most complete in Paris 54, it still lacks the headpiece and the initials at the beginning of the Greek and Latin texts, as well as numerous initials within the Latin text. 49 An exception to this is on fol. 105r, where the complete page in Greek is written in bright red ink, but in the Latin text column one verse (Matt. 27:37: “This is Jesus the King of the Jews”) is written in blue ink. Alfred Bu ¨ chler alerted me to its significance as this is the inscription written over the cross in the Latin West. In contemporary Byzantine crosses, the inscription reads “King of Glory.” See further, A. Bu ¨ chler, “King of Glory and King of the Jews: The Titulus of the Cross in the Christian East,” BSCAbstr 16 (1990): 68–69. 50 That is, on the last folio (fol. 121r) of quire 11 (the first quire of Mark), which corresponds to Mark 3:27. See Appendix. 46

126

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

The problems with the crimson passages in the Latin text continue intermittently through the third quire of Mark (fols. 132r–141v). On those folios that did not require the use of crimson ink, there is no evidence of the new scribe,51 except on folio 142r (i.e., the beginning of quire 14), where the Latin text is executed wholly in black ink by the new Latin scribe (Fig. 9). The Latin text continues in this fashion until it terminates on folio 150. The problems encountered in the Latin text on folios 121–133 are so consistent as to clarify the manner in which the text was executed. The Greek text was completed first in the polychromatic ink colors. Then the Latin text was filled in but not from beginning to end. Instead, passages requiring the same color were completed at the same time over a number of folios. In the folios where there are problems with the crimson ink passages, the blue and dark brown passages appear to have been completed before the rarer crimson passages. The original scribe responsible for the Latin text apparently never had the opportunity to finish the crimson passages. It is difficult to gauge how much time may have elapsed before the new scribe began to finish those passages using the inappropriate black ink. Miniatures The scribe responsible for the Greek text reserved space for nine text miniatures throughout the Gospel of Mark. The first four miniatures were completed; the last five were not even begun. The four completed miniatures are by Artist B and occur in the first two quires.52 The first empty frame reserved for an uninitiated miniature is found at the beginning of the fourth quire of Mark, on folio 142r, which coincides with the beginning of major anomalies in the Latin text of Mark (Fig. 9).53 It is at folio 142r that the Latin text is, for the first time, executed by the new hand exclusively in black ink. On folio 150v, the Latin text stops, not to be resumed again until the Gospel of John.54 The evidence concerning the modus operandi of the scribes and artists working on the Gospel of Mark re-confirms that the Greek text was completed before the Latin. Here, however, we learn also that the miniatures were executed at the same time as the Latin text. Anomalies in Mark’s unfinished Latin text coincide with the interruption of the 51 On fols. 121v, 122r and 122v, 123r and 123v, and 124r the passages in the Latin text that should have been written in crimson are also in black and by the new hand. On fols. 124v and 125r, most of the Latin text is written in the original hand, but these are examples, too, of places where very little crimson ink was required. The problems with the crimson passages in the Latin text continue through fol. 129v. On these folios almost every example of black ink was written by the new scribe and should have been in crimson. Fols. 130r, 131r, 132v, 133v, and 134r are wholly by the first scribe. This is logical, in that none of the passages on these pages called for the use of crimson, whereas both fols. 132r and 133r have examples of black ink by the new hand where crimson-colored ink should have been used. The crimson passages on fols. 134v–140v appear to have been executed in the appropriate color and by the first hand. The only exception is on fol. 138r and verso; there in the Latin text, the script that should have been crimson, was written in a new blue-grey color and seems to be the work of the new hand. 52 They are located in both quires on the versos of the third and fourth folios (fols. 114v, 115v, 124v, and 125v). 53 Spaces reserved for uninitiated miniatures in Mark are found on fols. 142r, 162v, 166r, 167v, and 168v. Significant blank spaces extending across both text columns or empty red frames indicate that a miniature was intended but never executed. 54 With the exception of one brief passage on fol. 194r.

KATHLEEN MAXWELL

127

miniature cycle. This raises the possibility, at least, that one of the Latin text scribes and Artist B were the same individual.55 While this issue will not be pursued in the present paper, a close collaboration is suggested between Artist B and the Latin scribes at work before the final Latin hand appears. LUKE (FOLS. 173R–276V) Text The Greek text of Luke is complete, but now the Latin text column is blank. However, examination of the miniatures of Luke’s Gospel offers more detail about the ways in which scribes and artists created Paris 54. Miniatures Luke’s Gospel was scheduled to receive twenty miniatures (see Appendix).56 Only five were completed, five remain unfinished, and ten were never begun. These miniatures were executed by a new artist, Artist C, whose hand is not detected in the miniatures of Matthew and Mark.57 Thus, although no progress was made on the Latin text in Luke, the execution of the narrative miniatures was well underway. Codicological examination of Luke’s narrative cycle reveals certain trends. Three miniatures were scheduled for the first quire of Luke (Fig. 10). The first miniature, the Annunciation (fol. 176r), was completed, but the other two—the Visitation (fol. 177v; Fig. 11) and the Presentation (fol. 182r; Fig. 12)—were left unfinished. Three miniatures were also intended for the second quire (Fig. 13). Here also the first miniature the Baptism (fol. 186v), was completed, while the second miniature, the Miraculous Draught of Fishes (fol. 191v), which is located on the same bifolio, was not even begun. This is less surprising than might first appear. Although the Baptism and the Miraculous Draught of Fishes are both located on the versos of their respective folios, they are, in fact, located on opposite sides of the bifolio that they share (see Appendix).58 The third miniature, the Cure of the Paralytic (fol. 193v), located on the last folio of the quire, was completed. The third quire in Luke contains two miniatures (Fig. 14). The first, the Parable of the Widow of Naim (fol. 201r), was completed, while the second, the Anointing of Christ 55 See Hutter, “Decorative Systems in Byzantine Manuscripts,” 15; and Corrigan, “Constantine’s Problems.” 56 The uninitiated miniature at the bottom of fol. 247r in Paris 54 may represent a scribal error. The red frame measures only 5.8 ⫻ 16.9 cm. The scribe has reserved another space at the top of the very next folio and its dimensions are more in keeping with the other narrative miniatures in Paris 54. Thus, there is the possibility that the Greek scribe intended to go back and erase the red frame on the bottom of fol. 247r but never got around to doing it. 57 See above, note 47. 58 A similar situation exists with regard to several miniatures in Matthew’s text. The Denial of Peter and the Holy Women at the Sepulchre share the same bifolio and are located on the rectos of their respective folios. The same is true of the Remorse of Peter and the Descent from the Cross. The miniatures in both cases have distinctly different color schemes. In fact, the Denial and Remorse miniatures are very similar in coloration and style, as are the miniatures of the Descent and the Holy Women at the Sepulchre. This makes perfect sense given their respective themes and environments. It also makes sense from a codicological view, since, once again, even though both sets of miniatures are located on the rectos of their respective folios, they are not located on the same side of the bifolio they share.

128

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

(fol. 203v), is the most advanced of the unfinished miniatures (Fig. 15). Two miniatures are also found in the fourth quire (Fig. 16). The first, Christ Asleep in the Boat (fol. 207r), remains unfinished, while the second, the Transfiguration (fol. 213r), is complete. No miniatures were scheduled for Luke’s fifth quire. The Cure of the Man with Dropsy (fol. 233v) is the only miniature in the sixth quire of Luke; it is located on the last folio of the quire and was never finished. The remaining nine miniatures planned for Luke are distributed among the next five quires. None of these was ever begun.59 The pattern of completed and uninitiated miniatures in Mark,60 taken in conjunction with the location of unfinished miniatures in Luke, suggests that an artist did not complete one miniature at a time, or even the miniatures of one quire at a time. Instead, he worked on a number of miniatures at the beginning of each Gospel, and these might be dispersed over as many as a half dozen quires.61 One might assume, because the Greek text of Luke is complete and the Latin text blank, that no further evidence can be gleaned associating anomalies in the text with those in the narrative miniatures. In truth, however, surprising parallels can be demonstrated. Beginning with the seventh quire of Luke (quire 23; fols. 234r–243v) a number of anomalies of all types appear. Up until this point (i.e., through the sixth quire of Luke), both finished and unfinished miniatures occur (with only one example of an uninitiated miniature on fol. 191v; see Appendix). However, beginning with the seventh quire and for the remainder of Luke, only uninitiated miniatures are found. It is at the seventh quire that the Greek quire signatures suddenly disappear and are never resumed. (The Armenian quire signatures continue without interruption.) A different ruling system is also initiated at this point.62 And, it is with this quire that the parchment suddenly becomes thinner and more transparent. Moreover, the Greek script now looks different, perhaps suggesting the first appearance of a second Greek scribe.63 Finally, the canon The tenth quire of Luke is the penultimate quire of his Gospel. See Appendix. The four completed miniatures in Mark are found in the first two quires. See notes 52 and 53 above. 61 Der Nersessian, Miniature Painting in the Armenian Kingdom of Cilicia, 169, arrived at similar conclusions in her study of Erevan, Matenadaran 7651. She notes that in the original 13th-century campaign of work on this manuscript “the work was carried on more or less simultaneously in all four gospels, but in none of them was it completed.” Sargis Pidsak completed the unfinished miniatures of this manuscript in the next century, and with one exception the miniatures attributed to him are located in later quires of each gospel. 62 Inconsistencies in the lineation can be seen beginning on fol. 234 and continuing through approximately fol. 290. At fol. 234, the left margin is not easily verified (compare a in Fig. 7); the far right column e has been frequently omitted, and the width of the space between the two text columns (the space between b and c) is increased by 0.4 cm. The upper and lower margins also differ, measuring 4.1 and 7.1 cm, respectively. Fol. 243, which is the same bifolium as fol. 234, is even more puzzling, in that no lines can be distinguished at all. On fol. 235, e is missing and no horizontals (including f and g) are visible. The character of the space between the two text columns also differs considerably, in that it now contains four lines—two narrowly spaced, double lines delineate a wider intercolumniation. Fol. 235’s counterpart, fol. 242, is identical. Fol. 236 is similar to fol. 234, while its counterpart, fol. 241, has very unevenly defined horizontals and no discernible vertical lines. Fol. 237 contains the newly defined space between the two text columns of fol. 235 and lacks e. The horizontals of fol. 237 have been drawn beyond d in an uncharacteristically sloppy manner. At the beginning of the next quire, on fol. 244, the newly defined space between the two text columns is continued and the right margin, e, has been reinstated. The next forty-five or so folios continue in this vein and are distinguished by their careless ruling or by the fact that it is almost impossible to distinguish horizontals or verticals or both types of lines at any given time. 63 The Greek text does continue in the polychromatic colors, however, suggesting that if this is a new scribe, little time elapsed between his work and that of the first Greek scribe. 59 60

KATHLEEN MAXWELL

129

numbers beginning on folio 235r are inexplicably shifted to the right side of the column for the folio rectos only. JOHN (FOLS. 278V–361V) Text The Greek text of John is fully executed in the four different ink colors (Fig. 17). The Latin text is also partially complete,64 now in yet another (untrained?) hand that cannot compete with the beautiful script of the Greek text. The Latin scribe writes in dark brown ink only and does not justify the right margin. Miniatures John’s miniature cycle seems, at first glance, to run counter to the pattern encountered in Mark and Luke. John’s Gospel was scheduled to receive eleven narrative scenes, of which only one was executed. This single painted miniature, however, is not the first narrative miniature, as one would expect; rather, it is the third in the sequence. Closer scrutiny reveals, however, that the first two miniatures planned for John (and, of course, the evangelist portrait) are located in quire 27 (Fig. 18; see Appendix), the quire in which Luke’s Gospel ends and John’s Gospel begins.65 The evangelist portrait, an insert, is complete, but the two miniatures were not executed. Thus, the only executed miniature in John, Christ and the Samaritan Woman at the Well (fol. 289r), is the first miniature of the succeeding quire (fols. 285–294), that is, the first quire devoted exclusively to John’s text. THE UNFINISHED MINIATURES: TECHNICAL CONSIDERATIONS The five unfinished miniatures of Luke allow us to reconstruct the means by which an artist painted them.66 They are listed below, not in the order they are encountered in the text, but in descending degree of completion: Christ Asleep in the Boat, folio 207r Presentation, folio 182r (Fig. 12) The Latin text stops permanently at fol. 329r. This is unusual, according to Nelson. Most scribes would begin a new quire with the start of each Gospel. See R. Nelson, “Theoktistos and Associates in Twelfth-Century Constantinople: An Illustrated New Testament of A.D. 1133,” J. Paul Getty Museum Journal 15 (1987): 58 and n. 30. In Carr’s discussion of the manuscripts of the decorative style, of which sixty are Tetraevangelia, she notes that the “quires are modified to permit the opening of major text divisions on new gatherings.” See A. Weyl Carr, Byzantine Illumination, 1150–1250: The Study of a Provincial Tradition, ed. O. Grabar and H. L. Kessler, Studies in Medieval Manuscript Illumination 1 (Chicago, 1987), 2. 66 These miniatures have fascinated scholars ever since they were first mentioned by Silvestre in his 1841 publication, Pale´ographie Universelle (no pagination). Some scholars actually prefer the drawings to the completed miniatures in Paris 54. H. R. Willoughby, The Four Gospels of Karahissar, vol. 2, The Cycle of Text Illustrations (Chicago, 1936), 91, wrote, “Practiced students nearly always prefer the five free drawings in wash in Paris 54 to the . . . finished and polished paintings in the same manuscript.” I do not share this preference, although I have seen other Byzantine manuscripts where I do (e.g., the Vatican Job, Vat. gr. 751). J. Labarte, Histoire des arts industriels, 2nd ed. (Paris, 1875), 189, was critical of the general style of the miniatures but felt, unlike Willoughby, that the “coloring lacks less than the drawing.” 64 65

130

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE Cure of the Man with Dropsy, folio 233v Visitation, folio 177v (Fig. 11) Anointing, folio 203v (Fig. 15)

The first three miniatures have thin, red-ink rectangular frames, inside which the general features are sketched in a pale yellow wash (Fig. 12).67 The composition is sketched in fully enough to indicate where gold needed to be applied. The gold backgrounds and nimbi are also complete, with the exception of the nimbus in Christ Asleep in the Boat on folio 207r.68 This exception makes readily visible the white gesso base upon which the artist applied his gold.69 The fourth miniature on the list, the Visitation (Fig. 11), shows the next stage of execution, in which brown ink was used to delineate more specific details of both figures and architecture. The facial features and the hair of the figures, as well as some of the folds of their garments are now indicated, as are the stairs, windows, and the archway of the building on the right side of the composition. The fifth and most advanced of the unfinished miniatures, the Anointing of Christ’s Feet (Fig. 15), is perhaps the most revealing. Here, the artist has reinforced most of the red-ink frame with a thicker line of red paint. Significant portions of the color have been applied. Apparently, the artist worked with one color at a time, painting all areas in the composition intended to be that color before moving on to the next hue. Several colors were completed in this fashion, including red, green, dark brown, and mustard. A dark ochre ground was applied to all of the faces, except that of the kneeling figure. Was this ochre the last color applied by the artist, and was he interrupted before applying it to the face of the kneeling figure? Not necessarily, since the kneeling figure is the only female depicted and women’s faces may not have received the dark ochre ground.70 The evidence provided by these five miniatures indicates the following steps: Conversely, the line drawings of the manuscripts of the decorative style were executed in magenta. See Carr, Byzantine Illumination, 1150–1250, chap. 1, where she maintains that many of the members of this group were produced in Cyprus or Palestine. In an unusual deluxe member of the decorative style group, Malibu, J. Paul Getty Museum, Ludwig II 5 (olim Phillipps 3887), however, Buchthal notes that some of the miniatures show no evidence of underdrawing. See H. Buchthal, “An Unknown Byzantine Manuscript of the Thirteenth Century,” Connoisseur 155 (1964): 221, and Carr, Byzantine Illumination, 1150–1250, 252–53. 68 See J. J. G. Alexander, Medieval Illuminators and Their Methods of Work (New Haven–London, 1992), 40. M. P. Brown, Understanding Illuminated Manuscripts: A Guide to Technical Terms (Malibu, Calif., 1994), 59, writes “Gilding formed the first stage in the painting processes of illumination, since it was a messy activity, the gilded area often requiring trimming with a knife.” 69 The most thorough discussion of Paris 54’s technique is found in H. Bordier, Descriptions des peintures et autres ornements contenus dans les manuscrits grecs de la Bibliothe`que nationale (Paris, 1883), 228. He proposed that this gesso base was probably made up of a combination of albumen and starch. See “Goldgrund” RBK ´ vangile, 7, notes that the gold of the marginal miniatures 2:882–83. Millet, Recherches sur l’iconographie de l’E ¨ber in Paris, Bibliothe`que nationale, suppl. gr. 914 is painted over a plaster base. See also Tikkanen, Studien u die Farbengebung, 182. O. Demus, “Die Farbe in der byzantinischen Buchmalerei,” Palette 26 (1967): 9–11 notes that most of the gold in Byzantine manuscripts was applied in powder form. See also Brown, Understanding Illuminated Manuscripts, 58–59, for the use of gesso in gilding. 70 Der Nersessian, Miniature Painting in the Armenian Kingdom of Cilicia, 1:92, notes that in some of the miniatures of New Julfa, Monastery of the Holy Savior, 57/161, all but the facial features have been depicted. She assumed that they were to be completed by another artist. 67

KATHLEEN MAXWELL

131

1. The scribe reserved space for the miniature and may or may not have drawn a rectangular frame in red ink. The frame was executed in one of the two red inks used in the multicolored text (Fig. 9). 2. The artist sketched the most general features of the composition in a pale yellow wash. The composition had to be sufficiently advanced so that all portions requiring gold could be determined (Fig. 12). 3. A thick white gesso base was applied to those areas designated to receive gold (background, nimbi, etc.). 4. Gold was applied. 5. The composition was further detailed with brown ink (Fig. 11). 6. The composition was filled in by applying one color at a time (Fig. 15). A study of flaked areas in completed miniatures in Paris 54 indicates that not all narrative scenes were executed in this fashion, however. For example, it is difficult to confirm whether the gold of the miniatures of Matthew, Mark, and John is applied over the raised white gesso ground that is so apparent in the miniatures from Luke. In addition, flaked areas of completed miniatures of Matthew indicate that a variety of colors was used for the underdrawings. In the Nativity, a pale green color is visible beneath the flaked area on Joseph’s chest, and, in the scene of the Holy Women at the Tomb, light mauve underpainting can be seen where the paint of the wall has flaked. The evidence provided by the Wedding Feast of the King’s Son is probably the most striking—here, in the flaked areas of the architecture on the left side of the composition, a hot pink-orange hue was used to sketch the buildings. These variations suggest that several artists were responsible for the miniatures of Paris 54 but that the five unfinished miniatures of Luke are the work of one master. The unfinished miniatures of the narrative cycle provide surprising parallels with the working methods utilized by the Latin scribe in folios 120v–141v. The filling-in of the Latin text one color at a time is analogous to the way in which the artist responsible for the unfinished miniatures in Luke colored his miniatures (Fig. 8). Both methods probably reflect time-saving devices utilized by professionals. SUMMARY AND CONCLUSIONS Text Figure 1 summarizes the status of Paris 54’s Greek and Latin texts and the degree of completion of the narrative miniatures. Matthew’s Gospel is the only one complete in all three categories. All evidence indicates that the Greek text was written before the corresponding Latin text. The scribe responsible for the Greek text apparently refined the initial concept of the varied ink colors as he proceeded. Our observations concerning the initial passages in blue ink on folios 13r–19r suggest this. This same scribe was responsible for reserving space throughout the manuscript for the narrative miniatures. This was accomplished by leaving a number of blank lines in the text, typically just before the relevant passage was to be written. Frequently these blank spaces were further defined by a frame drawn in one of the two red ink colors used for the text. The evidence from Mark’s Gospel indicates that the Latin text and narrative miniatures were produced simultaneously. The point at which the Latin text is taken over

132

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

exclusively by the new hand using the inappropriate monochromatic black ink coincides exactly with the appearance of the first uninitiated miniature (Cure of the Man with the Dumb Spirit). Both occur on folio 142r, the first folio of the fourth quire of Mark (Fig. 9). In Luke’s Gospel, by contrast, the Latin text column is blank. Surprisingly, however, work on its narrative cycle was well advanced, with five completed miniatures, five unfinished miniatures, and ten uninitiated ones. Moreover, even though Luke’s Latin text was not executed, there are notable parallels between textual and codicological anomalies and the location of Luke’s unfinished miniatures. This is seen at the seventh quire of Luke beginning on folio 234. From here until the end of Luke, none of the miniatures was even begun. The Greek quire signatures disappear and a new system of lineation is introduced. Moreover, starting on folio 235, the canon numbers are suddenly placed in the right margin as opposed to the left for the folio rectos. Finally, in John’s Gospel, some fifty folios of the Latin text were completed by yet another, perhaps untrained hand in dark brown ink. Only one miniature, Christ and the Samaritan Woman at the Well, was executed. This is the third, not the first miniature in John, but it is the first miniature of the first quire devoted entirely to John’s text. The foregoing summary suggests that work on Paris 54 progressed in several phases. The first phase saw the completion of Matthew in all three major aspects: Greek text, Latin text, and narrative miniatures. This phase also saw the completion of all of Mark’s Greek text. The first four miniatures of Mark (i.e., through fol. 125v) were also executed at this time and are attributed to Artist B, the same artistic personality responsible for the majority of the miniatures of Matthew. However, only that portion of Mark’s Latin text written in polychromatic ink can be assigned to this phase. I think that the Greek texts of Luke and John were also executed during the first phase. While another scribe may have written quire 23 (quire 7 of Luke), the fact that he worked in a style very similar to the primary Greek scribe and used the polychromatic ink colors suggests that, despite codicological inconsistencies from quire 23 through the beginning of John’s Gospel, the Greek texts of Luke and John are likely to be products of the first phase. The second phase may have followed quite quickly on the heels of the first, with only a brief lull between them. Phase Two had no impact on the texts but introduced a new artist, Artist C, who executed the five completed and five unfinished miniatures of Luke, as well as the one completed miniature in John’s Gospel. The fact that Artist C did not attempt to execute the uninitiated miniatures of Mark suggests that Artist B was expected to return to finish his task. It is more difficult to explain why work on the Latin text was allowed to lapse, even as Artist C worked on the miniatures of Luke and completed one in John. Was no Latin scribe available? This seems unlikely, since a number of scribes apparently participated in the polychromatic Latin texts of Matthew and Mark.71 A seemingly greater time gap separated Phase Two and Phase Three. Phase Three was limited to text additions in the inappropriate black ink found sporadically in folios 121v–141v of Mark’s text and exclusively in folios 142r–150r. The fifty or so folios of 71 Many complex paleographical issues are raised by the Latin scribes of Paris 54 and it would take a specialist to address them adequately.

KATHLEEN MAXWELL

133

John’s Latin text, executed in dark brown ink only from its beginning on folio 279r through folio 329r, may also date from this phase. The fact that these scribes never attempted to execute their texts in the polychromatic ink colors suggests that circumstances had changed so radically that the role originally envisioned for the manuscript was no longer relevant. Perhaps, the patron had died and resources become so limited that the only course available was to finish the manuscript as expeditiously as possible so that it could, at least, serve some purpose, if not precisely the one for which it was intended. Yet, even that was not accomplished, and Paris 54, for reasons we may never really comprehend, remained unfinished in its Latin text, its narrative miniatures, and its ornament. Ironically, we have gained the least insight into the modus operandi of scribes and artists from the first Gospel of Paris 54. Matthew’s Gospel, essentially complete in its text and miniatures, offers few clues concerning the process by which this manuscript was written and illustrated. It was, instead, the unfinished text and illustrations of the Gospels of Mark and Luke that revealed the most information about the way in which this work was produced. Codicological and textual anomalies in these two Gospels can be linked rather closely with similar disruptions in miniature production. Taken together, the evidence makes apparent that artists and scribes worked more intimately in the production of this manuscript than is generally recognized as common in the Byzantine sphere. Santa Clara University

Appendix Description of Contents of Paris, Bibliothe`que Nationale de France, Cod. Gr. 54

i, ii, iii, I

Three unnumbered parchment folios (which I refer to as i, ii, iii) and a fourth labelled “I” precede folio 1.

Folios 1–10

These folios do not include any quire signatures. Folios 1–9r contain a Latin set of unornamented canon tables on the left side of each page. Space was reserved on the right for canon tables in Greek.

Quire 172

Folios 11r–20v, Matt. 1:1–5:12 Folio 10v, portrait of the Evangelist Matthew Folio 11r, beginning of Matthew’s text: missing headpiece and both initials at beginning of texts Folio 13v, Nativity (7.6 ⫻ 17.2 cm)

Quire 2

Folios 21r–30v, Matt. 5:12–8:10

Quire 3

Folios 31r–40v, Matt. 8:10–11:1 Folio 32v, Exorcism of Demons (7.1 ⫻ 17.4 cm) Folio 35v, Cure of the Woman with the Issue of Blood (7.3 ⫻ 17.3 cm)

Quire 4

Folios 41r–50v, Matt. 11:1–13:31

Quire 5

Folios 51r–60v, Matt. 13:31–16.9 Folio 55r, Miracle of the Loaves (10.4 ⫻ 17.1 cm)

Quire 6

Folios 61r–70v, Matt. 16:9–19:9

Quire 7

Folios 70r73–79v, Matt. 19:10–22:1

Greek quire signatures are found through quire 22. Armenian quire signatures are found throughout the manuscript. 73 An error was made during the pagination of the manuscript. There are two folios numbered 70 and no folio 71. 72

KATHLEEN MAXWELL

135

Quire 8

Folios 80r–89v, Matt. 22:2–24:38 Folio 80r, Marriage Feast of the King’s Son (7.5 ⫻ 17.3 cm)

Quire 9

Folios 90r–99v, Matt. 24:39–26:55 Folio 91r, Wise and Foolish Virgins (7.3 ⫻ 17.4 cm) Folio 96v, Last Supper (9.5 ⫻ 17.3 cm) Folio 99r, Betrayal (8.9 ⫻ 17.1 cm)

Quire 10

Folios 100r–109v, Matt. 26:55–28:20 Folio 101r, Denial of Peter (7.8 ⫻ 17.2 cm) Folio 102r, Remorse of Peter (8.9 ⫻ 17.2 cm) Folio 107r, Deposition (7.7 ⫻ 16.8 cm) Folio 108r, Marys at the Sepulchre (8.2 ⫻ 17.4 cm) Folio 109v, end of Matthew’s text Folios 110–111, bifolium insert Folios 110r and 110v, blank Folio 111r, portrait of the Evangelist Mark Folio 111v, blank

Quire 11

Folios 112r–121v, Mark 1:1–4:5 Folio 112r, beginning of Mark’s text: missing headpiece Folio 114v, Peter’s Mother-in-Law (8.3 ⫻ 17.5 cm) Folio 115v, Cure of the Leper (7.7 ⫻ 17 cm) Folio 121r, end of perfect color parallelism of Latin text and first appearance of new Latin hand using inappropriate black ink for passages that should have been crimson

Quire 12

Folios 122r–131v, Mark 4:5–6:34 Folio 124v, Christ Stilling the Water (8.9 ⫻ 17.4 cm) Folio 125v, Exorcism of the Demoniac (8.7 ⫻ 16.8 cm) Crimson Latin text by a new scribe

Quire 13

Folios 132r–141v, Mark 6:34–9:18 Folios 132r and 133r, evidence of new Latin hand using wrong color ink

Quire 14

Folios 142r–151v, Mark 9:18–11:29 Folio 142r, Cure of the Dumb Spirit (uninitiated; red-ink frame only) Folios 142r–150r, Latin text exclusively in black ink by new Latin hand first seen on folio 121r Folios 150v ff, no Latin text

Quire 15

Folios 152r–161v, Mark 11:29–14:25 No Latin text

Quire 16

Folios 162r–171v, Mark 14:25–16:14 Folio 162v, Gethsemane (uninitiated; space reserved) Folio 166r, Denial of Peter (uninitiated; red-ink frame)

136

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE Folio 167v, Mocking of Christ (uninitiated; red-ink frame) Folio 168v, Crucifixion (uninitiated; space reserved) No Latin text Folio 172r, blank Folio 172v, blank Folio 173r, portrait of the Evangelist Luke (an insert) Folio 173v, blank

Quire 17

Folios 174r–183v, Luke 1:1–2:44 Folio 174r, beginning of Luke’s text: missing headpiece, rubrics, and initials Folio 176r, Annunciation (7.9 ⫻ 17.3 cm) Folio 177v, Visitation (unfinished) (8.3 ⫻ 17.1 cm) Folio 182r, Presentation (unfinished) (8.3 ⫻ 17 cm) No Latin text

Quire 18

Folios 184r–193v, Luke 2:45–5:21 Folio 186v, Baptism (8.7 ⫻ 17.4 cm) Folio 191v, Miraculous Draught of Fishes (uninitiated; red-ink frame) Folio 193v, Cure of the Paralytic (7.8 ⫻ 17.3 cm) No Latin text

Quire 19

Folios 194r–203v, Luke 5:21–7:39 Folio 201r, Son of Widow of Naim (8.7 ⫻ 17.5 cm) Folio 203v, Anointing of Christ’s Feet (unfinished) (7.7 ⫻ 17.1 cm) Folio 194r, brief resumption of Latin text

Quire 20

Folios 204r–213v, Luke 7:39–9:33 Folio 207r, Christ Asleep in the Boat (unfinished) (8.2 ⫻ 17cm) Folio 213r, Transfiguration (10.2 ⫻ 17.4 cm) No Latin text

Quire 21

Folios 214r–223v, Luke 9:33–11:41 No Latin text

Quire 22

Folios 224r–233v, Luke 11:41–14:1 Folio 233v, Cure of the Man with Dropsy (unfinished; red-ink frame) No Latin text

Quire 23

Folios 234r–243v, Luke 14:2–17:9 Folio 235v, Meal of the Rich Man (uninitiated; red-ink frame) Folio 241v, Rich Man and Lazarus (uninitiated; red-ink frame) No Latin text; Greek quire signatures cease New lineation schema Canon numbers for folio rectos now placed in far right margin (continues through the end of Luke)

KATHLEEN MAXWELL

137

Quire 24

Folios 244r–253v, Luke 17:9–19:44 Folio 247r, unknown miniature subject/scribal error? (uninitiated; red-ink frame at bottom of the page measuring only 5.8 ⫻ 16.9 cm) Folio 247v, Parable of the Pharisee and the Publican (uninitiated; red-ink frame measuring 7.8 ⫻ 17.2 cm) Folio 248v, Parable of the Rich Man and Eternal Life (uninitiated; red-ink frame) No Latin text; no Greek quire signature

Quire 25

Folios 254r–263v, Luke 19:44–22:28 Folio 255r, Parable of the Master of the Vineyard (uninitiated; red-ink frame) Folio 258v, Parable of the Widow’s Mite (uninitiated; red-ink frame) No Latin text; no Greek quire signature

Quire 26

Folios 264r–273v, Luke 22:28–24:20 Folio 269v, Simon Carries the Cross (uninitiated; red-ink frame) No Latin text; no Greek quire signature

Quire 27

Folios 274r–284v, Luke 24:20–John 2:17 Folio 276v, Ascension (uninitiated; red-ink frame) Folio 276v, end of Luke’s Gospel Folio 277r, kephalaia listing for Gospel of John Folio 277v, blank Folio 278r, blank Folio 278v, portrait of the Evangelist John (an insert) Folio 279r, beginning of John’s Gospel: no headpiece or initials. Rubric for Greek text only; resumption of the Latin text by a different hand in dark brown ink Folio 280r, Anastasis (uninitiated; red-ink frame) Folio 283v, Wedding at Cana (uninitiated; red-ink frame) No Greek quire signature

Quire 28

Folios 285r–294v, John 2:17–5:11 Folio 289r, Samaritan Woman at the Well (7.5 ⫻ 17 cm) Folio 294r, Cure at the Pool (uninitiated; red-ink frame) No Greek quire signature

Quire 29

Folios 295r–304v, John 5:11–7:3 No Greek quire signature

Quire 30

Folios 305r–314v, John 7:4–9:2 No canon numbers; no kephalaia No Greek quire signature

Quire 31

Folios 315r–324v, John 9:3–11:32 Folio 315r, Cure of the Man Born Blind (uninitiated; red-ink frame) Folio 322, Raising of Lazarus (uninitiated; red-ink frame) No Greek quire signature

138

` QUE NATIONALE DE FRANCE, CODEX GREC 54 PARIS, BIBLIOTHE

Quire 32

Folios 325r–334v, John 11:32–13:33 Folio 328r, Entry into Jerusalem (uninitiated; red-ink frame) Folio 332v, Washing of the Feet (uninitiated; red-ink frame) Folio 329r, Latin text stops No Greek quire signature

Quire 33

Folios 335r–344v, John 13:33–17:12 Folios 335r–339v, no Latin text; no Greek quire signature

Quire 34

Folios 345r–354v, John 17:12–19:41 No Latin text; no Greek quire signature

Quire 35

Folios 355r–364v, John 19:41–21:25 (end of John) Folio 356r, Noli Me Tangere (uninitiated; red-ink frame) Folio 357v, Doubting of Thomas (uninitiated; red-ink frame) Folio 359v, Appearance to Disciples (uninitiated; red-ink frame) No Latin text; no Greek quire signature

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Text Fig. Status of the Greek and Latin texts and of miniatures of the narrative cycles in Paris, Bibliothèque nationale de France, cod. gr. 54

1

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 112r, beginning of Mark (photo: Bibliothèque nationale de France, Paris)

2

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 55r, Miracle of the Loaves and Fishes (photo: Bibliothèque nationale de France, Paris)

3

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 16v (photo: Bibliothèque nationale de France, Paris)

4

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 86r (photo: Bibliothèque nationale de France, Paris)

5

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 16r ( photo: Bibliothèque nationale de France, Paris)

6

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 22v (photo: Bibliothèque nationale de France, Paris)

7

Paris, Bibliothèque nationale de France, cod. gr. 54, lineation schema, fols. 11–233 (not to scale)

8

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 121r ( photo: Bibliothèque nationale de France, Paris)

9

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 142r ( photo: Bibliothèque nationale de France, Paris)

10

11

Paris, Bibliothèque nationale de France, cod. gr. 54, Luke, quire 1

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 177v, Visitation (photo: Bibliothèque nationale de France, Paris)

12

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 182r, Presentation (photo: Bibliothèque nationale de France, Paris)

13

Paris, Bibliothèque nationale de France, cod. gr. 54, Luke, quire 2

14

15

Paris, Bibliothèque nationale de France, cod. gr. 54, Luke, quire 3

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 203v, Anointing of Christ’s Feet (photo: Bibliothèque nationale de France, Paris)

16

Paris, Bibliothèque nationale de France, cod. gr. 54, Luke, quire 4

17

Paris, Bibliothèque nationale de France, cod. gr. 54, fol. 279r, beginning of John’s Gospel ( photo: Bibliothèque nationale de France, Paris)

18

Paris, Bibliothèque nationale de France, cod. gr. 54, Luke, last quire

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Imagined Images: Visions of Salvation and Intercession in a Double-Sided Icon from Poganovo BISSERA V. PENTCHEVA

To the memory of my father f the many icons created during the Palaiologan period, perhaps the most arresting and enigmatic is the double-sided panel from the monastery of Poganovo in southeastern Serbia. The fourteenth-century icon shows a vision of Christ in Majesty; he sits on a rainbow set within an aureole of blue light (Fig. 1). The four mythical symbols of the Evangelists emerge from the mandorla. A mountainous terrain and a lake appear below the vision of Christ. To the left stands the prophet Ezekiel with hands raised in a gesture of amazement, while to the right the prophet Habakkuk sits on the rocks and holds an open book on his lap. The Theophany is paired with another enigmatic scene on the other side of the icon. Here the Virgin, mournful and silent, casts her gaze toward the gesturing hand of the aged John the Theologian (Fig. 2). The disciple, while engaging the Mother of God in conversation, addresses the viewer with his eyes and body. A dedicatory inscription is placed on the gold background between the two figures. The scenes are painted on a large wood panel, which is more than three feet high (93 ⫻ 61.5 cm). The background and beveled frame are entirely covered with gold leaf. A hole on the lower edge of the panel suggests that the icon was carried in processions affixed to a pole. The panel was transferred from Poganovo to Sofia, Bulgaria, before the outbreak of World War I. It is currently displayed in the crypt of the cathedral of Alexander Nevsky. Many studies have been dedicated to the iconography of the scenes and the identity of the donor.1 The Theophany has been treated by most scholars as a copy of the fifthcentury mosaic in the apse of the church of Hosios David at the Latomos monastery in Thessalonike. The scene on the other side has been interpreted as a Crucifixion. Andre´

O

I am grateful to my advisor Ioli Kalavrezou for her insightful comments, to Nicholas Constas for providing me with references for the liturgical texts, and to Christina Maranci for encouraging me to write about this icon and for greatly improving the style of the paper. I have also benefited from the comments of Elka Bakalova at the Frick Symposium, 9 April 1999, where I presented parts of this research, and to Christo Matanov, Liliana Mavrodinova, and Georgy Gerov, with whom I discussed many of the problematic issues involved with the donor of the icon. 1 Several studies concentrated on the identity of the donor in the dedicatory inscription. Todor Gerasimov argued that the panel was commissioned in 1395 by Helena Dragasˇ, wife of Emperor Manuel II Palaiologos (1391–1425) to commemorate the death of her father, Constantine Dejan in 1394. Gerasimov supported his theory with the evidence of two stone inscriptions in the church of Poganovo mentioning the names of a

140

DOUBLE-SIDED ICON FROM POGANOVO

Grabar argued that the two sides were unified by the idea of Christ as the life-giving spring.2 According to his interpretation, the image of the Virgin and John alluded to the moment of the Crucifixion when the Evangelist proclaimed the water and blood spurting from Christ’s wound as the life-giving spring. Gordana Babic´ suggested that the Theophany represented the Second Coming, or the Parousia, of Christ.3 Another approach was offered by Andre´ Xyngopoulos, who explored the connection between the Poganovo icon and the art of Thessalonike.4 He argued that the Theophany did not directly copy the prototype at Hosios David but was modeled after some hypothetical miniature of the twelfth-century Diegesis text, which told the story of the miraculous appearance of the mosaic at Latomos. Xyngopoulos also contended that the name Virgin Kataphyge, meaning “refuge,” referred to the homonymous hypogeum in Thessalonike where St. Demetrios found his violent death. The iconography of the Poganovo panel has also been discussed certain Constantine and Helena. Based on the previous research of Kru ˘ stju Mijatev, Gerasimov argued that ˆne the icon was commissioned for the family monastery of the Dragasˇi at Poganovo. T. Gerasimov, “L’ico bilate´rale de Poganovo au Muse´e arche´ologique de Sofia,” CahArch 10 (1959): 279–88; K. Mijatev, Poganovskijat Manastir (Sofia, 1936), 20. Mijatev’s and Gerasimov’s theory is now considered problematic, since the monastery and the land around Poganovo probably did not belong to the Dragasˇ family. While Christo Matanov has argued in his earlier studies that the Poganovo monastery was included in the territory of Constantine Dragasˇ, he has recently abandoned this position. He discovered that the lands around Poganovo were not recorded as part of the ˘kraten timarski opis na Kjustendilski sandzˇak ot 1517 domain of Dragasˇi in the 16th-century Ottoman defter, Su godina: Tapu defter 74, microfilm in the Oriental Department of the Bulgarian National Library Sts. Kiril and Methodii. I am grateful to professor Matanov for sharing this information with me. For his earlier theory, ˘lgarski zemi prez XIV vek (Sofia, 1986), 122; and idem, Knjazˇestvoto na Dragasˇi. see C. Matanov, Jugozapadnite bu ˘m istorijata na severoiztocˇna Makedonia v predosmanskata epoha (Sofia, 1997), 177. Ku Andre´ Xyngopoulos proposed another donor for the Poganovo icon: Helena, the wife of John V Palaiologos. In Xyngopoulos’s view, she commissioned only the obverse of the icon with the vision of Christ during her sojourn in Thessalonike in 1354, while the other side of the icon was painted at a later date. Xyngopoulos’s theory was immediately rejected by Andre´ Grabar, first because it did not account for the name of Constantine, who was among the donors of the Poganovo monastery, and, second, because technical analysis ˆne bilate´rale proved that the two sides of the panel were painted at the same time. A. Xyngopoulos, “Sur l’ico de Poganovo,” CahArch 12 (1961): 341–50, esp. 348; A. Grabar “Sur les sources des peintres byzantins des XIIIe et XIVe sie`cle,” CahArch 12 (1961): 351–89, esp. 363, 366, 372. An important hypothesis about the identity of the donor was later offered by Gordana Babic´. She argued that the title basilissa used in the inscription could not refer to a Constantinopolitan empress but instead was applied to a wife of a despot. Therefore, Babic´ proposed that Helena Ugljesˇa of Serres commissioned the icon in 1371 to commemorate both her dead son and husband. Babic´ then tried to interpret the iconography ˆne de Poganovo et la vasilissa of the scenes in connection with the life of Helena Ugljesˇa. G. Babic´, “Sur l’ico He´le`ne,” L’art de Thessalonique et des pays balkaniques et les courants spirituels au XIVe sie`cle: Recueil des rapports du ´ dition speciale 31 (Belgrade, 1987), IVe colloque serbo-grec, Belgrade, 1985, Institut des ´etudes balkaniques, E 57–66. The same idea has been further developed by Goiko Subotic´, “Ikona vasilise Jelene i osnivacˇi manastira Poganova,” Saopsˇtenja—Rebublicˇki zavod za zasˇtitu spomenika kulture Srbije 25 (1993): 25–40. Recently, Ivan Djuric´ offered another hypothesis for the identity of the donor; he proposed reading the name as Eudokia and argued that the donor was the second wife of Constantine Dragasˇ, Eudokia Komnene. I. Djuric´, “Evdokija Komnina i nein muzˇ Konstantin Dragasˇ,” ZRVI 12 (1983): 259–72. The problem surrounding the identity of the donor cannot be resolved because paint losses at exactly this spot on the inscription render the name of the donor illegible. Even the recent infrared photographs cannot help clarify the name definitively. However, the faint traces of two letters “NH” support a reading of ELENH. 2 ˆne byzantine du XIVe sie`cle au Muse´e de Sofia,” CahArch 10 (1959): 289– A. Grabar, “A propos d’une ico 304; idem, “Sur les sources des peintres byzantins,” 351–89. 3 ˆne de Poganovo,” 63–66. Babic´, “Sur l’ico 4 ˆne bilate´rale de Poganovo,” 341–50. Xyngopoulos, “Sur l’ico

BISSERA V. PENTCHEVA

141

in connection with the liturgy. Demetrios Pallas argued that the images relate to the symbolism of the Easter celebrations.5 Similarly, Edmond Voordeckers linked the scenes on the icon to another liturgical feast, the Metastasis, or Assumption, of John the Evangelist.6 He argued that the figure of John was prominent, that this side of the icon should be considered as the front, and that the panel was set on the iconostasis as a title icon, since the monastery itself was dedicated to John the Evangelist. This paper reconsiders previous interpretations. The latter assumed that the iconography of the Poganovo icon could be understood in terms of standard compositions such as the Crucifixion and the Second Coming of Christ. Yet, the images are not so explicit; they only allude to these normative scenes. I argue that the ambiguous and opaque representations of the Poganovo icon give only visual clues that are meant to stimulate the viewer to recall a series of standard scenes. The understanding of the icon thus depends on visions imagined and remembered. I explore how meaning is created out of this interaction between represented and imagined images. The broad spectrum of associations triggered by the two scenes shows that the icon had a complex liturgical use. In addition to the celebrations of Easter and the Metastasis of John, I argue that the panel was also involved in the commemorative services for the dead donors. Finally, on the basis of the evidence from monastic typika, I insert the Poganovo icon in the broader context of other panels with a similar liturgical function. EASTER SYMBOLISM The visions depicted on the two sides of the icon evoke the Crucifixion and the Resurrection of Christ without actually reproducing the standard iconography characteristic of these scenes. By alluding to the Passion, the Poganovo images respond to the liturgical role of the icon during Easter celebrations. The Virgin and John the Evangelist are identified by red inscriptions placed on the gold background (Fig. 2). Mary appears as Mother of God with the name Kataphyge,7 while John is identified as the Theologian.8 In addition, a dedicatory inscription on the gold background between the two figures reads “[Helena ?] the pious basilissa in Christ the God.” 9 The pairing of the Virgin and John the Evangelist recalls the iconography of the Crucifixion. An early example in which the two figures stand next to each other under the same arm of the cross is offered by the tenth-century fresco at the apse of the New Church at Tokali Kilise (Fig. 3).10 Christ is nailed to the cross; his right side is pierced and spurts blood. The Mother of God raises her covered hands, while John the Evangelist places his right hand in front of his chest in a gesture of prayer and reverence. His figure resembles closely that of John on the Poganovo icon. In the more common depictions of the Crucifixion, the Virgin and John flank Christ, D. Pallas, Die Passion und Bestattung Christi, MiscByzMonac 2 (Munich, 1965), 147–60. ˆnes byzantines,” Byzantion 53.1 (1983): 52–68, E. Voordeckers, “L’interpre´tation liturgique de quelques ico esp. 59–62. 7 MR QU H KATAFUGH (Mh´ thr Qeou' hJ Katafugh´ ). 8 O AG IW O QEOLOGOS (oJ a”gio" jIwa´ nnh" oJ Qeo´ logo"). 9 ˆne bilate´rale de Poganovo,” 284. [ JEle´ nh] ejn Cristv' tv' Qev' pisth` basi´lissa, from Gerasimov, “L’ico 10 A. W. Epstein, Tokali Kilise: Tenth-Century Metropolitan Art in Byzantine Cappadocia, DOS 22 (Washington, D.C., 1986), 73–74, figs. 83–87, color pl. 5. 5 6

142

DOUBLE-SIDED ICON FROM POGANOVO

as in a fourteenth-century icon at the Byzantine Museum in Athens (Fig. 4).11 Christ’s dead body hangs from the cross. The Virgin stands on the left side. Her hands rest underneath her chin; one is covered in the cloth, the other clasped on top. Her head inclines toward Christ, but her face is downcast. On the right side of the cross, John is depicted in a mournful gesture, his right hand resting on his cheek. His other arm droops down heavily. His whole body sags under the weight of mourning. The figure of the Virgin on the Athens icon is similar to the Theotokos of the Poganovo panel; both are attenuated, completely wrapped in their garments, and selfcontrolled in their deep sorrow. They place their hands in the same modest and vulnerable gesture. In both cases, the delicate body of the Virgin is contrasted to the corpulent figure of John. However, the two icons differ in the depiction of the Evangelist; in the Athens panel he is a mournful youth, but in the Poganovo icon John has an aged face and makes a gesture of speech. I will return to the meaning of this departure in the section on John’s Metastasis below. Many years ago, Grabar observed that the figures of the Virgin and the evangelist were “cut out” of a Crucifixion scene.12 The cross is no longer there, yet, the two figures remain standing; the Virgin has preserved her posture and gestures of mourning. The pairing of the grieving Theotokos and John immediately triggers the memory of the original setting from which the two images have been taken. We perceive the Poganovo scene as a Crucifixion; yet, there is no cross. Its very absence relates symbolically to the loss that occurs with death. While the side with John and the Virgin alludes to a Crucifixion scene, the other face of the icon presents an image that relates symbolically to the Resurrection. Christ sits on the rays of a gold rainbow in an aureole of blue light (Fig. 1). Youthful,13 he wears a golden chiton and raises his right arm while holding an open scroll in his left hand. His scroll reads, “Behold our God in whom we hope and rejoice in our salvation, he will give rest to this house” (Isa. 25:9).14 His hands and feet bear the marks of the Crucifixion. Next to his head is the inscription, “Jesus Christ of the miracle of Latomos.” 15 The halo of blue light surrounding his figure is composed of seven rings. The four apocalyptic symbols of the Evangelists emerge from the fifth ring of the mandorla. Clockwise from the top left, these are Matthew, John, Luke, and Mark; their initials are inscribed above their heads.16 Below this vision is a mountainous landscape that gradually descends into a water basin, where seven fish swim. On the two sides of the rocks stand the figures of the prophets Ezekiel and Habakkuk, identified by inscriptions above their heads.17 Ezekiel is The icon measures 103 ⫻ 84 cm. Conversation with God: Icons from the Byzantine Museum in Athens (9th–10th centuries), exh. cat., The Hellenic Center, London, 22 May–20 June 1998, ed. C. Baltoyanni (Athens, 1998), 76–79, with the most recent bibliography. 12 ˆne,” 300. Pallas, Die Passion und Bestattung Christi, 91–97. Grabar, “A propos d’une ico 13 On the reappearance of the youthful Christ in the scenes of Majesty in the 11th century and its meaning, see A. W. Carr, “Gospel Frontispieces from the Comnenian Period,” Gesta 21.1 (1982): 3–20, esp. 7–10. 14 jIdou` oJ qeo` " hJmw'n ejf∆ v» ejlpi´zomen kai` hjgalliw´ meqa ejpi` th' swthri´a hJmw'n aujto` " dw´ sei ajna´ pausin tv' oi“kv tou´ tv. 15 [t]o` ejn Lato´ mou qau'ma. 16 IW[a´ nnh"], M[a'rko"], M[atqai'o"], and L[ouka'"]. 17 PRO[fh´ t]H["] IEZEKIHL and [profh´ th"] ABBAKOUM. 11

BISSERA V. PENTCHEVA

143

depicted as an elderly man dressed in a chiton, standing with arms raised and palms open to the viewer in a gesture of awe and amazement. Habakkuk is represented as a youth seated on the rocks; he has propped his chin in his right hand and on his lap holds an open book with the written message, “Son of Man, eat this scroll” (Ezek. 3:1).18 The whole scene is imbued with a blue light that emanates from the halo above and highlights the draperies, the rocky mountains, and the waters. The setting, lyrical with gold and blue light, conveys the essence of the scene: a theophany. As noted above, the vision of Christ copies the mosaic in the church of Hosios David of the monastery of Latomos in Thessalonike (Fig. 5).19 The connection between the two works is revealed not only through their similar iconography but also in the inscription of the Poganovo icon, which identifies the vision as that “of the miracle of Latomos.” 20 In the prototype, Christ appears in an aureole of light, seated on a rainbow.21 He also holds a scroll with a similar inscription, reading, “Behold our God, in whom we hope and rejoice in our salvation, that he may grant rest to this house.” 22 There also we find the four mythical beasts and the two prophets. The mosaic relies heavily on the visions of Ezekiel, Habakkuk, Isaiah, Joel, and Zacharias.23 Yet, the theophany cannot be securely associated with any particular prophetic account. The ambiguity stems from the mosaic’s lack of identifying inscriptions above the two figures flanking Christ. Grabar interpreted these prophets as Ezekiel and Zacharias;24 later on he proposed that the two figures were Peter and Paul.25 James Snyder argued that the standing prophet is Ezekiel, while the seated is John the Evangelist.26 Unlike the mosaic, the Poganovo icon clearly identifies the vision as that of Ezekiel and Habakkuk (Fig. 1). This clarity of authorship together with the Crucifixion marks on Christ’s hands and feet and the written command “Eat this scroll” on the open book create a new meaning for the scene. Babic´ has proposed that the wounds signify that Christ is depicted after the Resurrection in his Parousia, or Second Coming.27 UiJe` ajnqrw´ pou kata´ fage th` n kefali´da tau´ thn. ˆne,” 289–304. R. Cormack, Writing in Gold: Byzantine Society and Its Icons Grabar, “A propos d’une ico (London, 1985), 132–33; and E. Tsigaridas, Latomou Monastery: The Church of Hosios David (Thessalonike, 1988), 85–86. For a discussion of the aureole of light and the ambiguous gender of the face of Christ, see T. Mathews, The Clash of Gods: A Reinterpretation of Early Christian Art (Princeton, 1993), 116–23. 20 The miracle refers to the supernatural appearances of the mosaic image of Christ in the apse, as related in Ignatios’ 12th-century text, Diegesis. According to this text, Theodora, the daughter of the emperor Maximianus, secretly converted a bath on the imperial properties into a church and displayed a mosaic of the Virgin in the apse. This mosaic miraculously transformed itself into an image of Christ in Majesty. A. Papadopoulos-Kerameus, ed., Varia graeca sacra (St. Petersburg, 1909; repr. Leipzig, 1975), 102–13, esp. 107. For a discussion of the date of the text, see V. Grumel, “La mosaı¨que du Dieu Sauveur au Monaste`re de Latome `a Salonique,” EO 33 (1930): 157–75; and Cormack, Writing in Gold, 132–33. 21 On the significance of the rainbow and divinely generated light, see L. James, Light and Color in Byzantine Art (Oxford, 1996), 91–109. 22 ijdou` oJ qeo` " hJmw'n ejf∆ v» ejlpi´zomen kai` hjgalliw´ meqa ejpi` th' swthri´a hJmw'n o”ti ajna´ pausin dw´ sei ejpi` to` n oi«kon tou'ton, from Tsigaridas, Latomou Monastery, 37. 23 Isa. 12:3, 25:9, 44:3; Ezek. 3:1, 2:10, 47:9; Joel 3:18; Zech. 14:8. 24 A. Grabar, Martyrium: recherches sur le culte des reliques et l’art chre´tien antique, 3 vols. (Paris, 1943–46), 2:201. 25 ˆne,” 289. Grabar, “A propos d’une ico 26 J. Snyder, “The Meaning of ‘Maiestas Domini’ in Hosios David,” Byzantion 37 (1967): 143–52, esp. 150– 53. Snyder’s theory is not supported by the evidence, especially since the text of the Apocalypse did not enter the Byzantine liturgy; it existed only in separate commented versions. 27 ˆne de Poganovo,” 63. Babic´, “Sur l’ico 18 19

144

DOUBLE-SIDED ICON FROM POGANOVO

Alternately, the vision can be interpreted as a Resurrection, an association also triggered by the presence of the two prophets. Habakkuk’s vision is linked to Christ’s Crucifixion and Resurrection, as demonstrated by Gregory of Nazianzus’ Second Oration for Easter. A miniature from the ninth-century manuscript Paris, Bibliothe`que nationale de France, cod. gr. 510, folio 285r, shows Habakkuk standing to the right and pointing to a vision of a youthful angelic creature surrounded by the rays of a mandorla and flanked by the Heavenly Host (Fig. 6).28 Gregory of Nazianzus stands next to Habakkuk and gestures in amazement. On the left are two female figures: St. Paraskeve holding the instruments of the Passion and the empress Helena carrying a miniature model of the tomb of Christ.29 The lances, the sponge, and gall, and the tomb all evoke the Crucifixion and implicitly anticipate the Resurrection. Thus, they influence the perception of the divine creature in the mandorla as a prefiguration of the Resurrected Christ. The miniature gives a visual interpretation of Gregory’s homily, which starts with a quotation from Habakkuk: I will stand upon my watch, saith the venerable Habakkuk; and I will take my post beside him today. . . . Well, I have taken my stand and looked forth; and behold a man riding on the clouds and he is very high, and his countenance is as the countenance of an Angel, and his vesture as the brightness of piercing lightning; and he lifts his hand towards the east and cries with a loud voice. His voice is the voice of a trumpet; and round about him is, as it were, a multitude of the Heavenly Host; and he saith, “Today is salvation come into the world, to that which is visible and to that which is invisible. Christ is risen from the dead, rise ye with Him. Christ has returned again to himself, return ye.” 30

Habakkuk foretells the Resurrection of Christ, and his quoted words at the beginning of Gregory’s homily became firmly associated with Easter symbolism. For this reason Habakkuk’s portrait became a standard element of the iconography of preface miniatures to Gregory’s Second Oration for Easter.31 This homily made popular Habakkuk’s vision of the Resurrection of Christ among the Byzantine audience. Therefore, the prophet’s presence on the Poganovo icon prompted the Byzantine viewer to perceive the Theophany as a vision of Christ’s Resurrection.32 The figure of Ezekiel offers a similar association with the Passion of Christ. Since his prophecy is read on Holy Monday and Holy Wednesday,33 his Theophany is assimilated within the Easter symbolism. At a point during his prophecy, Ezekiel is ordered to eat L. Brubaker, Vision and Meaning in Ninth-Century Byzantium (Cambridge, 1999), 205–7, 284–86; eadem, “Politics, Patronage, and Art in Ninth-Century Byzantium: The Homilies of Gregory of Nazianzus in Paris (B.N. Gr. 510),” DOP 39 (1985): 1–13, esp. 10; S. Der Nersessian, “The Illustrations of the Homilies of Gregory of Nazianzus, Paris gr. 510: A Study of the Connection between Text and Image,” DOP 16 (1962): 197–228, esp. 201; H. Omont, Miniatures des plus anciens manuscrits grecs de la Bibliothe`que nationale du VIe au XIVe sie`cle (Paris, 1929), pl. XLIII; and Pallas, Die Passion und Bestattung Christi, 154. 29 Brubaker, “Politics, Patronage, and Art in Ninth-Century Byzantium,” 10. 30 PG 36:624A-B, translation from Der Nersessian, “The Illustrations,” 201. 31 Brubaker, “Politics, Patronage, and Art in Ninth-Century Byzantium,” 10; G. Galavaris, The Illustrations of the Liturgical Homilies of Gregory Nazianzus, Studies in Manuscript Illumination 6 (Princeton, 1969), 120–25. 32 Grabar interpreted the scene as a theophany of Ezekiel, and this concept has persisted to the extent that the icon is even identified under the name of “The Vision of Ezekiel.” As Pallas has shown the image is equally tied to the prophecy of Habakkuk; Pallas, Die Passion und Bestattung Christi, 152–55. 33 Ibid., 155–56; and The Lenten Triodion, trans. Mother Mary and Archimandrite Kallistos Ware (LondonBoston, 1977; repr. 1984), 517, 541. 28

BISSERA V. PENTCHEVA

145

the scroll with the revelations. By consuming it, he gains the power to prophesy. The same command, “Eat this scroll,” appears on the icon and alludes to Ezekiel’s act of conceiving and conveying the vision of the Resurrected Christ. The two sides of the Poganovo icon, then, represent the Crucifixion and the Resurrection. The allusion to the Crucifixion is not unusual; in fact, many of the extant bilateral icons represent this moment of Christ’s life.34 Pallas counted fifty-two existing double-sided panels, of which half include a Crucifixion or such related scenes as a Deposition, Basileus tes doxas, or a Hetoimasia on the reverse.35 Pallas linked this particular choice of subject-matter to the role of bilateral icons in the liturgical celebrations of Pascha. Similarly, by alluding to Christ’s Crucifixion and Resurrection, the images of the Poganovo panel responded to the Easter symbolism. THE METASTASIS OF JOHN In addition to their associations with Easter, the images on the two sides of the Poganovo icon also relate to the Metastasis of John.36 Voordeckers has convincingly argued that John is the principal figure of the panel and that his aged face and designation as “Theologian” signify that he is represented in his role as the author of both the Gospel and the Apocalypse.37 It is now necessary to explore further how John’s preeminence is conveyed though the visual clues. Without diminishing its associations with Easter, the Theophany of Christ represented on the icon’s other side can also be read as a scene of the Assumption of John the Evangelist. The two faces of the panel can thus be viewed as a diptych, with an author-portrait on one side and his vision on the other. In this way, the icon also partakes in the liturgical celebration of the Metastasis of John. John’s importance is expressed by simple visual clues: the Evangelist addresses the viewer with direct eye contact, and his figure, unlike the unlit figure of the Virgin, is bathed in light (Fig. 2). A similar contrast is conveyed through the two figures’ postures. John’s head gently tilts to the Virgin, while his face and heavy body are directly frontal. His contrapposto stance expresses energy and movement. By contrast, the Virgin is static, her entire columnlike body is enveloped in cloth. An opposition is also established through their faces and expressions. John has a wrinkled countenance, bald head, and a long, curly grey beard; his appearance conveys serenity and wisdom. By contrast, the face of the Virgin is youthful and emanates a suppressed and pensive sorrow. While the Grabar, “Sur les sources des peintres byzantins,” 366–80. Pallas, Die Passion und Bestattung Christi, 308–23. The Hodegetria icon in Constantinople with the Virgin and Child on the front and the Crucifixion on the back probably served as a model for these bilateral icons. Several 14th-century icons with the Virgin and Child on the front and the Crucifixion on the back survive in Ohrid and Athens. V. Djuric´, Icoˆnes de Yougoslavie (Belgrade, 1961), 85–86, pl. 4; Tre´sors me´die´vaux de la re´publique de Mace´doine, exh. cat., Paris, Muse´e national du Moyen Age, Thermes de Cluny, 9 February–3 May 1999 (Paris, 1999), 62–63, with the most recent bibliography; W. Volbach and J. Lafontaine-Dosogne, Byzanz und der christliche Osten, Propula¨en Kunstgeschichte 3 (Berlin, 1968), 238, pl. 26; and Conversation with God: Icons from the Byzantine Museum in Athens, 76–82, 96–102. 36 Voordeckers, “L’interpre´tation liturgique,” 58–62. 37 Subotic´ has argued that John is represented in his role of author. Yet, Subotic´ never analyzed how this meaning is conveyed through such visual clues as the aged face of the Evangelist; Subotic´, “Ikona vasilise Jelene,” 30–32. 34 35

146

DOUBLE-SIDED ICON FROM POGANOVO

Virgin is passive and withdrawn, John is active, open, and direct. Therefore, he becomes the figure singled out to address the viewer and to lead him/her into the icon. The Evangelist is given a position of prominence because the panel was meant to be used in the celebrations of his Metastasis. John’s feast, celebrated on September 26, commemorates his death and Assumption.38 Ioannis Spatharakis and Emanuel Klinkenberg have argued that because there was no established iconographic model for representing the Evangelist’s ascension many different representations were employed.39 In one, for example, on a sixteenth-century Cretan icon (Fig. 7),40 John is represented in a mandorla of light, being lifted up to Heaven by angelic creatures. The visual formula resembles the standard way of depicting Christ’s ascension within a mandorla of light lifted by angels. Such iconographic similarities made the Assumption of John analogous to that of Christ. For this reason, the Poganovo Theophany, by representing Christ in a mandorla of light hovering over the mountainous landscape, alludes to the representations of John’s Assumption. In addition to these purely iconographic parallels between the ascensions of Christ and John, the two figures were also very closely linked in Byzantine theology. John is assimilated into Christ at the moment of the Crucifixion. Christ addresses his mother and asks her to receive John as her only son: her Christ. Similarly, he commands John to accept Mary as his mother. The moment of adoption is manifested in the words that often accompany representations of the Crucifixion, as, for example, in the Tokali fresco (Fig. 3). The standard address reads, “Mother, behold your son! Son, behold your Mother!” The transformation of John at the Crucifixion was discussed by Origen already in the third century: We might dare say, then, that the Gospels are the first fruits of all Scriptures, but that the first fruits of the Gospels is that according to John, whose meaning no one can understand who has not leaned on Jesus’ breast nor received Mary from Jesus to be his mother also. But he who would be another John must also become such as John, to be shown to be Jesus, so to speak. For if Mary has no son except Jesus, in accordance with those who hold a sound opinion of her, and Jesus says to his mother, “Behold your son,” and not “Behold, this man also is your son,” he has said equally, “Behold, this is Jesus whom you bore.” For indeed everyone who has been perfected “no longer lives, but Christ lives in him,” and since “Christ lives in him,” it is said to Mary, “Behold your son, the Christ.” 41

Origen argues that of all the disciples, John was the closest to Christ and thus became assimilated to him at the moment of the Crucifixion. Since the two figures of the Virgin and John on the Poganovo icon are excerpted from a Crucifixion scene, they evoke in the mind of the viewer the particular moment when John is perfected and becomes Christ. The aged face of John the Evangelist is another significant element in his portrayal. One would expect to see the youthful Evangelist at the scene of the Crucifixion. The old 38 M. Jugie, La mort et l’assomption de la Sainte Vierge (Vatican City, 1944), 710–26. Synaxarion of September 26 in Menaion tou Ianouariou [-Dekemvriou], 12 vols. (Venice, 1852), 2:445–58, esp. 453. Voordeckers, “L’interpre´tation liturgique,” 58–62. 39 I. Spatharakis and E. Klinkenberg, “The Pictorial Cycle of the Life of St. John the Evangelist in Crete,” BZ 89.2 (1996): 420–40. 40 To kallos tes morphes metabyzantines eikones, 15–18 aionon, apo tis sylloges ton poleon Moscha, Sergieph Posant (Zagorsk), Tver kai Riazan, ed. E. Ankroskina et al. (Athens, 1995), 192–93, icon 6. 41 Origen, Commentary on the Gospel according to John, trans. R. E. Heine, 2 vols. (Washington, D.C., 1993), 1:38.

BISSERA V. PENTCHEVA

147

John on the Poganovo icon represents a departure from the standard iconography. This change signifies that he is portrayed as the author of the Gospel and Revelation.42 His aged face alludes to the scenes in which the apostle dictates his Gospel to Prochoros at Patmos, as represented, for example, in the late tenth-century miniature from a manuscript in the Dionysiou monastery at Mount Athos (Dionysiou cod. gr. 588, fol. 225v) (Fig. 8).43 The viewer’s perception of this image of John as author of the Gospel was also supported by the liturgy. Lections from his writings were read during Easter and placed at the beginning of the Synaxarion.44 The liturgical celebration of Christ’s Passion thus drew on the Gospel of John. Since the icon alludes to both the Crucifixion and the Resurrection, and was used in the Easter celebrations, the portrait of John could be easily perceived as the witness to Christ’s Passion.45 His aged face and upright standing pose further portray John as the prophetic writer of the Apocalypse. A more explicit pairing of the text of Revelation with a portrait of the old John is offered by a miniature from a twelfth-century manuscript with the Gospels, psalter, and odes in Moscow (State Historical Museum gr. 407, fol. 288v).46 Fourteenthcentury frontispieces were added to this codex; one of these inserted miniatures, portraying the old John, was placed at the beginning of the Apocalypse text (Fig. 9). In one sense, this image resembles John’s frontispiece portrait at the beginning of his Gospel in the same manuscript, but in the Apocalypse frontispiece his standing position and the agitation and fervor in his posture and clothing have greater affinity with the portraits of the Old Testament prophets in the same codex. John is thus portrayed as a prophet of Revelation. If John is depicted as the author of the Apocalypse, then the Theophany on the icon’s other side can be understood as a representation of his Revelation. Moreover, the seven layers of the mandorla, the seven fish, and the four mythic creatures suggest an Apocalyptic scene.47 The Theophany thus presents an image of the end of time uncovered and 42 ˇevcˇenko, “The Cave of the Apocalypse,” in Praktika tou Pallas, Die Passion und Bestattung Christi, 149. N. S diethnous symposiou me thema: Hi. Mone Hag. Ioannou tou Theologou—900 Chronia historikes martyrias (1088–1988) ˇev(⫽ Hetaireia byzantinon kai metabyzantinon meleton, Diptychon Paraphylla 2) (Athens, 1989), 169–80. S ˇcenko has argued that while the text tradition presented Patmos as the place where both the Apocalypse and the Gospel were written, the visual tradition continued to link Patmos and the cave exclusively with the Gospel text. 43 S. M. Pelekanides, The Treasures of Mount Athos: Illuminated Manuscripts (Athens, 1974–), 446–48, fig. 287; and A. A. Karakatsanis, ed., The Treasures of Mount Athos, 2nd ed. (Thessalonike, 1997), 228. 44 For a discussion of Byzantine lectionaries see, M.-L. Dolezal, “The Middle Byzantine Lectionary: Textual and Pictorial Expression of Liturgical Ritual” (Ph.D. diss., University of Chicago, 1991). 45 Moreover, the image of John on the icon’s back can be related to the Menologion section of a Gospel lectionary, which concentrates on the saints’ feast days. In this way, the two sides of the Poganovo panel function as a visual Gospel lectionary, with its Synaxarion side bearing the Theophany and its Menologion side, the portrait of John. 46 John is portrayed once as the author of the Gospel and a second time as the prophetic writer of the Apocalypse. The second image places John in the ranks of the Old Testament prophets. V. Lazarev, Storia della pittura bizantina, Biblioteca di storia dell’arte 7 (Turin, 1967), 371, figs. 509, 511; and A. Cutler, The Aristocratic Psalters, Bibliothe`que des cahiers arche´ologiques 13 (Paris, 1984), no. 33. I thank the anonymous reviewer for pointing out this reference to me. 47 ˆne de Poganovo,” 63–64. According to the apocryphal accounts, John was believed to Babic´, “Sur l’ico have remained alive until Christ’s Second Coming. The same idea is expressed in the Gospel of John: “When Peter saw him [John] he said to Jesus, ‘Lord, what about him?’ Jesus said to him, ‘if it is my will that he remain until I come, what is that to you? Follow me!’ So the rumor spread in the community that this disciple

148

DOUBLE-SIDED ICON FROM POGANOVO

recorded by John the Theologian. Similarly, the presence of Ezekiel and Habakkuk reveals the Old Testament sources on which John’s prophetic vision draws. The two faces of the icon function as a diptych depicting the author on one side and his vision on the other. The two sides are connected through the gesturing hand of John, a standard sign of both speech and direction. He points to the scene of the end of time and pronounces his prophecy. In addition, front and back are also united by the inscription written on the open book, commanding, “Son of Man, eat this scroll!” In a manner similar to Ezekiel, John is presented a scroll by an angel and asked to eat it (Rev. 10:9). By consuming it, he is given the knowledge to prophesy what is to come at the end of time. The command “Eat!” written on Habakkuk’s open book can thus be perceived as an address to John as well. He is the one who both consumes the divine revelation and prophesies its contents. In the prayers written in the fifteenth-century by Symeon of Thessalonike for the feast of the Metastasis, the Evangelist is eulogized for his prophecy about the end of time, the resurrection of the dead, and the salvation of the righteous.48 The same concepts are expressed through the images of the Poganovo icon, which reveal the Apostle foretelling the end of time and offering the view of salvation in Christ. Symeon of Thessalonike’s writings, like the Poganovo icon, record a contemporary perception of the prophetic role of John the Theologian. His cult intensified significantly during the Palaiologan period; many new sanctuaries were dedicated to the saint and more scenes of his life were represented in art.49 A CALL FOR SALVATION Along with its use during the celebrations of Easter and John’s Metastasis, the Poganovo icon also played an important role as an intercessor for the salvation of its donors, and its images were designed to address this function. Both the Virgin and John the Evangelist were selected as ideal intercessors, forming a Deesis on the donors’ behalf. The Virgin is addressed as Kataphyge. This name has been interpreted in a variety of ways. Babic´ argued that it referred to the moment when the Mother of God sought refuge in John after Christ’s death.50 Other scholars have perceived the name as designating a particular iconographic type of the Theotokos. I do not believe that Kataphyge exemplifies John’s act of offering refuge to the Virgin, but rather that the name is addressed to the Theotokos and qualifies her function as intercessor on behalf of the would not die” (John 21:20–23). See also Jugie, La mort et l’assomption, 711–26; and Menaion tou Ianouariou [-Dekemvriou], 455–56. 48 Nekrw'n hJ ajna´ stasi" eij" zwh` n hJma'" e“geiron, UiJe` tou' Qeou', o”te fwnh' sou nekrou` " ejgei´rei" kai` kri´nei" a”panta", eujcai'" tou' sou' zw'nto" maqhtou' tou' se´ kataggei´lanto", th` n zwh` n th` n aijw´ nionÚ I. Phountoules, Symeon archiepiskopou Thessalonikes: ta leitourgika syngrammata, Etaireia makedonikon spoudon 10 (Thessalonike, 1968), 162; Qeath` " ejdei´cqh tw'n ajrrh´ twn, th'" do´ xh" Patro` " kai` tou' ajrni´ou Cristou', ajgge´ loi" wJmi´lhse, ta` me´ llonta e“gnwke, th` n ca´ rin th` n eJpta´ fwton ei«de tou' Pneu´ mato", th' neu´ sei tou' uiJou' sou, parqe´ ne, oJ ajpokalu´ yei" ijdw` n mu´ sth" kai` fi´lo"Ú ibid., 164. 49 Many of the churches around Poganovo are also dedicated to the same saint. See L. Mavrodinova, ˘rkva (Sofia, 1980). Spatharakis and Klinkenberg, “The Pictorial Cycle of the Life of St. John Zemenskata Zu the Evangelist,” 420–40. The intensification of the cult of John the Evangelist in the Palaiologan period is a phenomenon that needs further study. 50 ˆne de Poganovo,” 61; and Subotic´, “Ikona vasilise Jelene,” 29. Babic´, “Sur l’ico

BISSERA V. PENTCHEVA

149

suffering and weak.51 The name does not identify any iconographic type of the Virgin and instead evokes an ideal quality of the Mother of God to offer refuge.52 The name appears very often in the Marian canons.53 The earliest use of this word—as, the verb katapheugo—is attested in an early fourth-century prayer to the Virgin.54 In addition, the Virgin is called Kataphyge in the liturgy for Holy Week.55 It is also possible that Kataphyge refers to the famous sanctuaries of the Virgin. As mentioned earlier, the hypogeum in Thessalonike, where St. Demetrios was killed, was dedicated to the Virgin Kataphyge.56 An even more illustrious religious site was the kataphyge of the Virgin Zoodochos Pege in Constantinople.57 The sanctuary was restored and brought to prominence once again in the end of the thirteenth century. Most of the miracles performed at the shrine were healings through water and the intercession of the Virgin.58 At the Constantinopolitan kataphyge, the power of the Virgin was clearly associated with healing waters. The Poganovo panel suggests a similar connection between the Mother of God and water through the image of the life-giving lake depicted on the icon’s other side. The Virgin is also a chosen intercessor because she has received proleptic immortality. The Mother of God bodily ascended to Heaven at her Koimesis, or Dormition. By seeking refuge in her, believers hope to find similar salvation. This desire is sometimes graphically represented in Palaiologan manuscripts. One of the most striking examples is offered by a late fourteenth-century psalter at Oxford (Christ Church gr. 61, fol. 102v) (Fig. 10). In this miniature, the Virgin actually pulls the monk and donor of the manuscript, Kaloeidas, from the grave, leading him to Christ represented on the next page. However, unlike the active and vigorous figure of the Virgin in the manuscript, Mary of the Poganovo icon is passive: a refuge of grief and a receptacle of internalized sorrow. Like the Virgin, John is also an intercessor who has been given proleptic immortality, since he bodily ascended at his Metastasis. Thus both figures could be seen as witnesses to the Theophany depicted on the icon’s other side. If we place the two faces of the icon on one plane in our mind’s eye, the constructed image conjures up a Deesis scene with the Virgin and John standing beside the vision of Christ. An example of the standard ˆne byzantine,” 302–4, and later on by E. BakA similar idea is suggested by Grabar, “A propos d’une ico alova, “Aspekti na su ˘ otnoshenieto na slovesen text-izobrajenie v bu ˘ lgarskoto srednovekovie,” Problemi v izkustvoto 1 (1991): 3–20, esp. 14–17. 52 ˆnes Grabar, “Les images de la Vierge de Tendresse: type iconographique et the`me (a` propos des deux ico `a Decˇani),” Zograf 6 (1975): 25–30; idem, “Remarques sur l’iconographie byzantine de la Vierge,” CahArch 26 (1977): 169–78, esp. 171, 176; both repr. in A. Grabar, L’art pale´ochre´tien et l’art byzantin (London, 1979). 53 S. Eustratiades, He Theotokos en te hymnographia (Chennevie`res-sur-Marne, 1930), 33. The canons are published in idem, Theotokarion (Chennevie`res-sur-Marne, 1931). The Virgin is called Kataphyge in the following canons: 2:208, 10:149, 15:112, 18:20, 20:169, 38:202, 55:16, 67:9. 54 F. Mercenier, “L’antienne Mariale grecque: la plus ancienne,” Le Muse´on 52 (1939): 229–33; and O. Steg¨ berlieferung,” Zeitschrift fu ¨r Kirchliche Theologie 74 mu ¨ ller, “Sub Tuum Praesidium: Bemerkung zur ¨alteste U (1952): 76–82; and G. W. H. Lampe, A Patristic Greek Lexikon (Oxford, 1961–68), 752. 55 The Lenten Triodion, 572–73. 56 ˆne bilate´rale de Poganovo,” 347; and idem, Symbolai eis ten topographian tes byzanXyngopoulos, “Sur l’ico tines Thessalonikes, Etaireias Makedonikon Spoudon 2 (Thessalonike, 1949), 7–20. 57 ˇ udotvornye obrazy v Konstantinopol’skom khrame ‘Z ˇ udotvorˇ ivotonosnogo istocˇnika,’” C A.-M. Talbot, “C naja Ikona v Vizantii i Drevnei Rusi, ed. A. M. Lidov (Moscow, 1996), 117–22; and eadem, “Epigrams of Manuel Philes on the Theotokos tes Peges and Its Art,” DOP 48 (1994): 135–65. 58 AASS, Nov. 3: 885. 51

150

DOUBLE-SIDED ICON FROM POGANOVO

composition is offered by the fourteenth-century fresco in the parekklesion of Kariye Camii (Fig. 11).59 Here Christ appears in an aureole of light, flanked by the Virgin and John the Baptist. In the Poganovo icon, the Evangelist has become the second mediator, replacing the Baptist. This substitution reveals once again the importance placed on John the Evangelist in the Poganovo panel.60 Moreover, both the Virgin and John have been assimilated into the figure of Christ. For the Evangelist, this transformation occurs at the moment of the Crucifixion. For the Virgin, the change is associated with her image in the apse mosaic of the Latomos monastery. According to the Diegesis, the original representation of the Virgin in the church of Hosios David miraculously turned into an image of Christ.61 Christ’s portrait on the Poganovo icon recalls this moment of transformation because it is identified as “the one of the miracle of Latomos.” Thus the inscription calls forth the memory of the legendary metamorphosis of the image of the Virgin into one of Christ. Consequently, both the Mother of God and John, depicted on one side of the panel, could be perceived as assimilated into the One—into Christ in the mandorla of light shown on the other side. By portraying the ideal intercessors, the Virgin and John the Evangelist, and by offering a vision of Christ as the savior and judge at the end of time, the Poganovo icon ideally fulfilled its function to intercede on behalf of the souls of the deceased donors. Furthermore, the vision of Christ at Latomos was specifically associated with funerary spaces already in the middle Byzantine period. The relationship is exemplified by the use of the Vision of Latomos scene in a twelfth-century fresco in the narthex of the ossuary of the monastery at Bacˇkovo.62 The functional meaning63 of the Vision of Latomos was thus clearly associated with death and salvation. The presence of the same scene on the Poganovo icon would have immediately triggered the funerary connotations of the subjectmatter. The icon was the focus of litanies and prayers for the deceased, and its images played an active role. The panel gave a tangible form to ideas of death and salvation. The Poganovo icon not only shows Christ at the Second Coming, but presents John as the guide who will lead the viewer to the Theophany. Salvation is conveyed through visible form: the figure of John, his gesture, and his vision. In a recent study, Marie-Jose´ Mondzain has argued that the difference between theology and economy, or God’s salvific plan, is the difference between believing without seeing versus believing while seeing.64 The PoP. Underwood, The Kariye Djami, 4 vols. (New York–Princeton, 1966–75), 1:202 and 3:374. While John the Baptist judges the human race sternly and harshly, John the Evangelist appears to be a more sensitive and empathetic figure. In the visual tradition, he is the apostle who mourns Christ at the Crucifixion, the Deposition, and the Lamentation; similarly, he is the figure most emotionally engaged at the Koimesis of the Virgin, leaning over her bier to kiss the dead body. 61 kai` ejn pro` " ajnatola` " aJyi'di grafh'nai th` n aJgnh` n Qeomh´ tora, th` n tou' ge´ nou" hJmw'n dhlono´ ti mesi'tin kai` sw´ teiran. grafome´ nh" ou«n th'" eijko´ no" kai` h“dh pro` " pe´ ra" tou' e“rgou ble´ ponto", wJ" th' uJsterai´a oJ gra´ fwn parage´ neto, h kai` pe´ ra" ejpiqh´ sein th' grafh' ejbouleu´ eto [ . . . ] oJra' grafh` n eJte´ ran kai` ouj th` n aujth´ n, polu` to` dia´ foron e“cousan kai` parhllagme´ non ajpo` th'" oJmoio´ thto", ejn ei“dei ajndrikv' to` n ku´ rion hJmw'n jIhsou'n Cristo` n ejpi` nefe´ lh" fwteinh'" kai` ejpi` pteru´ gwn ajne´ mwn ojcou´ menon kai` peripatou'ntaÚ Varia Graeca Sacra, 107 (see above, note 20). 62 E. Bakalova, Bacˇkovskata kostnizˇa (Sofia 1977), 67–72, figs. 34–37. 63 H. Belting, “An Image and Its Function in the Liturgy: The Man of Sorrows in Byzantium,” DOP 34/35 (1980–81): 1–16. 64 M.-J. Mondzain, Image, icoˆne, ´economie: les sources byzantines de l’image contemporain (Paris, 1996), 40. 59 60

BISSERA V. PENTCHEVA

151

ganovo icon participates in the economy, because it offers a visible manifestation of the promise of salvation. Since the salvific plan is revealed at the Annunciation, it is the female body that allows for the transformation of the invisible divine into the visible human form. Therefore, the economy relates to the female body of the Virgin.65 The Poganovo icon extends this idea to include John as well. Both the Virgin and John have consumed the invisible divine Word. The Virgin is the receptacle where Christ acquires a human form, while John, by consuming the Word, becomes the container of the prophecy of salvation at the end of time. Each has given a visible form to the divine Logos and, thus, has made manifest the salvific plan to the human race. It is this assurance in the coming salvation that is expressed in John’s sad but knowing face. The Virgin’s inconsolable grief is juxtaposed with John’s sorrow transfigured by knowledge. While for John the “Word is God” (John 1:1), and this Word—Christ—becomes manifest again at the time of the Last Judgment, for the Virgin the “Word became flesh” (John 1:14), and she relates to Christ in human terms as her child. Her maternal relationship to Christ is contrasted to the theological relationship defined by John. The two figures express the dichotomy of human suffering at the loss of a loved one and the religious dogma of salvation: feeling is superseded by understanding. Just as the eyes of the Virgin follow the gesturing hand of John, so the believers/viewers are asked to follow and accept the vision of salvation represented on the icon’s other side. THE ICON IN THE CELEBRATIONS OF LITURGICAL FEASTS The complex iconography of the Poganovo icon and the interaction of the images on the two sides unveil a variety of meanings: the Crucifixion and the Resurrection, the Metastasis of John, and a Deesis with the Virgin Kataphyge and John the Theologian. These scenes speak directly to the icon’s use during the celebrations of Easter, the Metastasis of the monastery’s patron saint, John the Evangelist, and the commemorative services conducted on the death anniversaries of the donors. Viewers engaged the icon in both the public sphere, as the visual focus of the main feast days of the monastery, and in the private sphere, as a site for personal contemplation of death and salvation.66 The double function of the Poganovo icon is characteristic of a number of other panels, as attested by the foundation documents of several earlier monasteries. The liturgical use of icons is discussed at great length in the typika of the Pantokrator monastery in Constantinople founded by Emperor John II Komnenos (1118–43)67 and the Kosmosoteira monastery at Pherrai built by the emperor’s brother, Isaakios Komnenos.68 The Mondzain, Image, icoˆne, ´economie, 199 ff. Carr, “The Vita Icon of St. Basil: Note on a Byzantine Object,” in Four Icons in the Menil Collection, ed. B. Davezac, Menil Collection Monographs 1 (Houston, 1992), 94–105, esp. 102. 67 Byzantine Monastic Foundation Documents: A Complete Translation of Surviving Founder’s Typika and Testaments, ed. J. Thomas and A. C. Hero (Washington, D.C., 2000), 725–81. None of the later typika of the 14th and 15th centuries provides information on the subject. They focus on the administration and the economic structure of the monastery rather than on the way the church and its icons were prepared and used during the important feast days. 68 Byzantine Monastic Foundation Documents, 782–858. 65 66

152

DOUBLE-SIDED ICON FROM POGANOVO

texts specify how the icons were prepared for the special celebrations and in what services they were employed. On important feast days, the principal icons of the church were to be provided with additional oil lamps and candles.69 The light reflected from their gold backgrounds further increased the icons’ radiance. Such illumination would have played an important role in the perception of the Poganovo panel, because it would have enhanced the experience of the divine vision of Christ emerging out of the radiance of a gold and blue aureole. Light is also symbolic of salvation and thus associated with the idea of Christ offering his Grace. The icons were carried in procession during the celebration of Easter and the feast days of the patron saints, as attested in the typikon of the Kosmosoteira: However, since I am never satiated in my longing for an honorable and comprehensive performance with regard to the feast of my benefactress, I wish, over and above all that has been said, for the following: after matins on that day on which we celebrate the Dormition, I wish for a number of priests and deacons to don the priestly vestments. The superior with them should carry in his hands the Gospel, the other in line the icon of the Mother of God, and another the great cross which is set at the sanctuary. The appropriate ektenes having been celebrated therein, the priests should come out of the church bearing these objects, with appropriate illumination preceding them, and go out the gate of the enclosure in a procession. They should go down to the cemetery of the monks, and circle thus towards the other entrance. . . . They should enter the church and return the sacred objects they have been carrying to their required holy places.70 On Easter Day, the day of the Resurrection of Christ, my God, there must be a procession like the one proposed for the Dormition of my ever-pure Lady, Mother of God.71

The Kosmosoteira monastery was dedicated to the Virgin. On her main feast day, the Koimesis, the icon of the Virgin was taken in procession together with the cross and Gospel book. An identical ceremony was also performed at Easter. Since the Poganovo monastery was dedicated to Saint John the Evangelist, its double-sided panel was most likely involved in similar processions on the patron’s feast day, the Metastasis of John, and at Easter. Finally, the typika also prescribe the role icons played in commemorative rites. In both the Pantokrator and the Kosmosoteira monasteries, the icons became the focus of prayers and active intercession on behalf of the souls of the deceased. In the Pantokrator monastery, processional icons, or signa,72 were brought to the tomb at the heron.73 The commemorative services at this monastery were very elaborate, befitting the imperial status of its patrons. By contrast, at the Kosmosoteira monastery a more personal commemorative rite was performed that probably more closely reflects the services at the Poganovo monastery. The typikon of the Kosmosoteira records the following: Ibid., 753 (Pantokrator), 802 (Kosmosoteira). Ibid., 827. 71 Ibid., 843. 72 ˇevcˇenko, “Icons in the Liturgy,” DOP 45 For a discussion of the meaning of the word signon, see, N. P. S (1991): 45–57, esp. 46; and M. N. Butyrskij, “Vizantijskoe bogosluzˇenie u ikony soglasno tipiku monastyrja ˇ udotvornaja Ikona v Vizantii i Drevnei Rusi (as above, note 57), 145–58. Pantokratora 1136 goda,” in Lidov, C 73 Byzantine Monastic Foundation Documents, 739, 755. For a discussion of the development of the imperial commemorative cult at the Pantokrator monastery out of the traditional Friday night presbeia in the Blacherˇevcˇenko, “Icons in the Liturgy,” 52 ff. nai and the Friday monastic memorial services, see, N. S 69 70

BISSERA V. PENTCHEVA

153

Instead of any kind of adornment of fantastic glory to my tomb, [I want] the icon from Rhaidestos of the Mother of God as the Kosmosoteira, [which was] sent down to me from heaven, and which I framed with ornament of gold and silver. [I want it] to be placed at one end of my tomb in its projected form. It should remain resting at that spot throughout all time, preserved without change, to mediate for my wretched soul. Furthermore, I wish [the icon of ] Christ, which is the same size, to rest alongside it, the placement of these icons being appropriate for them, and pleasing as well, and illumination suitable. . . . At any rate I wish the tomb to be divided from the entire narthex by the bronze railing I mentioned earlier, but the access to the tomb [should be] through [this railing]. Every evening, after the dismissal of vespers, I want the superior and the rest of the monks to enter, and in front of the holy icons standing there to pronounce the trisagion and say a certain number of Kyrie eleisons for mercy upon my soul. They [must] not fail to make the dismissal in this way, but [must] propitiate God and the Mother of God for me, with these [prayers].74

In the Kosmosoteira monastery the icons of the Virgin Kosmosoteira from Rhaidestos and the icon of Christ were moved from the iconostasis and set permanently on the donor’s tomb.75 The supplications were performed in front of these two panels. There is no surviving evidence about the location of the Poganovo icon in the church. Grabar suggested that the panel was placed behind the altar,76 while Voordeckers argued that it was set in the iconostasis next to the despotic icons.77 By contrast, I believe that the Poganovo panel functioned as a signon; it was placed on a proskynetarion, a stand in front of the iconostasis, and on special occasions moved to the donor’s tomb.78 The icon was thus in public view, while remaining accessible for personal contemplation.79 CONCLUSION The scenes of the Poganovo icon trigger the recollection of numerous images in the mind of the viewer. This process is further stimulated by the texts of the liturgical rites in which the icon had an active role. The allusions to the Crucifixion and Resurrection are prompted by the liturgy for Easter, while John’s apocalyptic vision and his proleptic immortality are called forth during the celebration of his Metastasis. Similarly, associations with the Parousia of Christ, the Deesis, and the life-giving waters are alluded to in the commemorative services held on the death anniversaries of the donors. The Poganovo panel offers a visual focus for these liturgical celebrations. The enigma of its elusive images dissipates once the icon becomes part of the liturgical rite, the processions, and prayers. On these occasions, the images represented and the images imagined merge to form a vision of salvation. Harvard University Byzantine Monastic Foundation Documents, 839. Ibid., 802. 76 ˆne,” 299. Grabar, “A propos d’une ico 77 Voordeckers, “L’interpre´tation liturgique,” 62. 78 This hypothesis is supported by the evidence of similar icons at the Pantokrator and the Kosmosoteira monasteries, Byzantine Monastic Foundation Documents, 739, 753–56, 839. 79 Carr, “The Vita Icon of St. Basil,” 102. 74 75

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Sofia, crypt of the cathedral of Alexander Nevsky, the Poganovo icon. Theophany of Christ ( photo: National Painting Gallery, Sof ia)

2

Sofia, crypt of the cathedral of Alexander Nevsky, the Poganovo icon. The Virgin Kataphyge and John the Theologian (photo: National Painting Gallery, Sofia)

3

Göreme Valley, Tokali Kilise, Crucifixion, fresco in the apse

4

Athens, the Byzantine Museum, icon of the Crucifixion (photo: Byzantine Museum, Athens)

5

Thessalonike, church of Hosios David, monastery of Latomos, Vision of Christ in Majesty, mosaic in the apse (after N. Chatzidakis, Byzantine Mosaics, Greek Art [Athens, 1994], 40, fig. 11)

6

Paris, Bibliothèque nationale de France, cod. gr. 510, fol. 285r, Vision of Habakkuk (after L. Brubaker, Vision and Meaning in NinthCentury Byzantium [Cambridge, 1999], fig. 29)

7

Moscow, Tretjakov Gallery, triptych, right wing, the Metastasis of John the Evangelist (after To kallos tes morphes metabyzantines eikones, 15–18 aionon, apo tis sylloges ton poleon Moscha, Sergieph Posant (Zagorsk), Tver kai Riazan, ed. E. Ankroskina et al. [Athens, 1995], no. 6)

8

Mt. Athos, Dionysiou cod. gr. 588, fol. 225v, John receiving divine inspiration (after The Treasures of Mount Athos, ed. A. A. Karakatsanis, 2nd ed. [Thessalonike, 1997], 228, no. 5.2)

9

Moscow, State Historical Museum, gr. 407, fol. 288v, frontispiece to John’s Apocalypse (after V. Lazarev, Storia della pittura bizantina, Biblioteca di storia dell’arte 7 [Turin, 1967], fig. 509)

10

Oxford, Christ Church gr. 61, fol. 102v, the Virgin pulling the monk Kaloeidas from the tomb ( photo: courtesy of the Governing Body of Christ Church)

11

Istanbul, Kariye Camii, fresco of the Last Judgment

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Constantinople: The Fabric of the City Dumbarton Oaks Symposium 1998

INTRODUCTION Constantinople: The Fabric of the City O how great is that noble and beautiful city! How many monasteries, how many palaces there are, fashioned in a wonderful way! How many wonders there are to be seen in the squares and in the different parts of the city!

ike Fulcher of Chartres, writing nine hundred years ago, we still find great fascination in the city of Constantinople. Although it is the center around which our Byzantine studies revolve, our knowledge of the Byzantine city remains rather limited. It can hardly compare with what our counterparts in classics know about ancient Athens or Rome. Constantinople has not benefited from the decades of careful, systematic excavations these cities have enjoyed. Moreover, Constantinople lacks the equivalent of the American School in Athens, where even the most bookish of philologists are regularly subjected to in situ lectures on archaeology, so that they come away with a vivid image of the settings for the important events of antiquity. Confronting the modern megalopolis of Istanbul unaided can be a daunting experience for the Byzantinist. The population has swelled in recent decades to more than twelve million, and remnants of the past diminish in visibility as the city grows. Our impression of Byzantine Constantinople is consequently limited to isolated “monuments of unageing intellect,” which present, at best, the distilled essence of the city’s historic greatness. That is to say, we know Constantinople more as a concept than as a reality. Or, to put it another way, we know a great deal in detail but very little in general. We are familiar with the intricate movements of the Byzantine liturgy, but we are completely lost when we set foot outside the church door. Similarly, we know the stations and acclamations and even the costume changes involved in an imperial triumph, but we are unsure as to the size and shape of the public spaces it traversed. And, if we step off the Mese, we are in terra incognita. Moreover, it is perhaps ironic that the two most influential institutions in Byzantine society, and the institutions that have left us the best documentation— monasticism and the imperial court—are perhaps the most difficult for us to visualize today. With these issues in mind, we organized the 1998 Dumbarton Oaks Symposium, “Constantinople: The Fabric of the City.” Our intention was to address the shape and development of Constantinople, its planning, its structures, its image and character— that is, the city as a living, three-dimensional entity. The symposium was interdisciplinary,

L

158

CONSTANTIOPLE: FABRIC OF THE CITY

with scholars of different backgrounds and interests addressing the city from their own perspectives, reconstructing in words and images the settings for the events that shaped the Byzantine world. We are pleased that many of the papers presented at the symposium could be included in this volume. Several themes emerged in the course of the symposium. The first is the fruitful conjunction of archaeology, that is, the study of the physical remains of the city, with the study of texts. By themselves, a brief reference in a text or a fragmentary remnant of a structure may tell us relatively little, but in combination they can be very informative indeed. The texts give life to the masonry, and the masonry gives credence to the texts. A second theme is the usefulness of comparative studies. The better preserved monuments in other cities can throw light on the exiguous remains of Constantinople. This has been demonstrated for all aspects of the city, whether it be the reconstruction of the Hippodrome and its functions, the placing of commercial activities in their settings, or the reconstruction of churches, palaces, and other buildings. A third theme concerns the physical remnants of the Byzantine city and their incorporation into the expanded city of Istanbul. In spite of modern urban renewal, a significant amount of the Byzantine city still remains visible: at the same time, however, these survivals are insufficiently surveyed and recorded. There are, for example, many vestiges of residential buildings in the city, in addition to its better known churches. And the churches themselves, through careful observation of the Byzantine remains in their vicinity, can be reintegrated into their Byzantine urban settings. A fourth theme has been the vision of Constantinople as a mosaic of daily life, a city made up of many neighborhoods, each with its local churches, monasteries, shops, and baths. Even the cloistered monastics interacted with their local communities. These different neighborhoods were linked through commercial and other activities, and also through the streets and water systems, the evolution of which throws light on the residential development of the city. The examination of the fabric of the city provides a vivid picture of its daily life—the best places to live, where to buy a book or a pig, where to take a bath, and even how to build a vault. Finally, in spite of significant changes, elements of the Byzantine city survived into Ottoman times. The gardens of the Mangana were preserved, in part, under the ramparts of Topkapı Palace; churches were converted to mosques; the Turkish bath continued the tradition of the Byzantine bath, even if not in the same structures; shopping took place in similar locations, but in covered bazaars instead of porticoed streets. Even today, the Hippodrome can serve as a focal point for public ceremonies and entertainments. If it is no longer possible to have chariot races, it is still possible to see folk dancers and musicians performing in front of the Obelisk of Theodosius, before the marble gazes of the Byzantine emperor and his court. Istanbul has always been a city with ghosts, ghosts of its past that have lingered and have refused to be eradicated by future ages. For the Byzantines themselves, Constantinople was haunted by the ghosts of pagan antiquity in the form of statues and other monuments. For the present-day inhabitants of Istanbul, it is the ghosts of Byzantium that, in turn, haunt the city. When we first planned this symposium, we were afraid that the ghosts might be fading away, as the vestiges of monuments were swept under the tide of progress and modernization. Hence, our initial aim was to reinvigorate the study of

CONSTANTINOPLE: FABRIC OF THE CITY

159

the physical remains of the Byzantine city and, to summon, as it were, the ghosts back to life. However, as the symposium and subsequent events have demonstrated, the ghosts have proved to be more robust than we initially thought. Since the 1998 Dumbarton Oaks symposium took place, Byzantine Constantinople has witnessed a flurry of activity, ranging from new archaeological discoveries dramatically announced by the international press, to new programs—less dramatic, yet no less important—for survey and documentation, and new initiatives for the preservation and restoration of the historic monuments. In addition, a major collection devoted to Byzantine studies has been bequeathed to the library of the American Research Institute in Turkey, broadening the scope of research now possible in Istanbul. Several international scholarly organizations are now planning to hold their conferences in Istanbul, and scholarly studies and discussions continue. Notably, our symposium was followed by a workshop on “Byzantine Constantinople: Monuments, Topography, and Daily Life,” ˘azic¸i University and the Institut which took place in situ in Istanbul, organized by Bog ´ Franc¸ais des Etudes Anatoliennes in April 1999. The papers from the workshop are now being prepared for publication. In conclusion, we wish to thank those responsible for facilitating the symposium and this resulting publication. We are grateful to the Senior Fellows of the program in Byzantine studies, who adopted the topic and provided valuable advice. In addition, thanks are owed to the Director of Byzantine Studies, Alice-Mary Talbot, for her continual support, and to Allison Sobke and Sarah Gordon, who shouldered most of the practical arrangements for the meeting. Most of all, however, our gratitude is due to the former Director, Angeliki Laiou, who created a fertile ground for the project to grow. Henry Maguire Johns Hopkins University Robert Ousterhout University of Illinois, Urbana-Champaign

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

The old town of Byzantium and the new quarter next to the Golden Horn (c = cistern)

2

3

Streets around the church of the Holy Apostles, oriented on the northwestern overland road. The long, solid black line is the aqueduct. The short black line is the foundation wall.

Streets around the church of the Holy Apostles, oriented on the aqueduct (the long, solid black line). The short black line is the foundation wall.

4

Streets in the Constantinian town (c = cistern)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Triumphal Way of Constantinople and the Golden Gate CYRIL MANGO

he title of this paper requires a few words of justification. To my knowledge the designation Via Triumphalis (or whatever its Greek equivalent may have been) does not occur with reference to Constantinople. Nor do we find there a gate or arch named Porta Triumphalis. Yet, as Michael McCormick has shown in detail,1 triumphs continued to be celebrated at Constantinople until the second half of the twelfth century, and it would be pedantic to argue that they should not be called triumphs because they differed in many significant respects from their Roman prototype and from one another. The question we wish to address is whether Byzantine triumphs followed a predetermined and specially designed route that may properly be called a “triumphal way.” If it is granted that they did, when and how was such a route established? In Rome the term triumphus referred to an archaic and highly regulated rite that was decreed by the Senate upon the fulfilment of certain strict preconditions. Scholars have disagreed whether the triumphal procession, which could be held only in Rome, always followed the same itinerary, but the chances are that it did.2 It was a circuitous route from the Campus Martius and the Circus Flaminius, through the Porta Triumphalis (close to which stood a temple of Fortuna Redux), across the Circus Maximus, round the Palatine hill, along the Via Sacra, and finally up to the Capitol, where sacrifice was offered. Whether the Byzantine ceremonial mirrored any of these stations is a question that naturally arises. For the present it is enough to note that the Roman triumphus was a welldefined and inscriptionally recorded event and that there also existed a more modest rite called ovatio, which could be accorded to a general if the victory he had won did not meet all the required criteria for a triumph. Such clear definitions are lacking in the Byzantine period. When we turn to our chief source, the Book of Ceremonies, we find that it describes a variety of different occasions, which either celebrated a victory or similar event. What may be called full triumphs, involving an emperor’s entry through the Golden Gate, exhibition of captives and booty, and procession down the main street, are those of Theophilus in 831 (or 837)3 and Basil

T

Eternal Victory (Cambridge, 1986). So F. Coarelli, “La Porta Trionfale e la Via dei Trionfi,” Dialoghi di archeologia 2 (1968): 58. 3 De cerimoniis aulae byzantinae, ed. J. J. Reiske, CSHB (Bonn, 1829), 503 ff (⫽ Constantine Porphyrogenitus: Three Treatises on Imperial Military Expeditions, ed. J. F. Haldon [Vienna, 1990], 146 ff). We are told (De cerim., ed. Reiske, 507.22 ff) that the same ceremonial was observed during Theophilus’s second triumph. The 1 2

174 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE I in 878.4 In addition, there is a sixth-century chapter devoted to the emperor’s return from a campaign or long journey, more of an adventus than a triumph, that envisages alternative itineraries. If coming from Thrace, the emperor could make his entry by way of the Hebdomon and, presumably, the Golden Gate;5 if arriving by water, he could choose to disembark at the forum of the Strategion or go by ship straight to the imperial palace, the last case involving no public exposure. The same chapter contains the wellknown account of Justinian’s return from Selymbria in 559, which, exceptionally, took place through the gate of Charisios (the Adrianople Gate) because the emperor wanted to pay his respects to Theodora’s tomb at the church of the Holy Apostles.6 Justinian had gone to Selymbria to supervise the repair of the breached Long Walls. He had engaged in no military activity and won no victory. The carefully staged ceremony did not include the exhibition of either captives or spoils (since there were none), yet when Justinian entered the palace, a triumphal address or acclamation (qriambeuta´ lion) was recited in a loud voice, and we are told that this had been expressly arranged by the Master of Offices (the patrician Peter) because the entry had not been made through the Golden Gate. Other pertinent accounts in the Book of Ceremonies are those of epinikia celebrated in both the Forum and the Hippodrome, with exhibition of captives and booty after a victory won by a general.7 Finally, an emperor’s entry into the city after his elevation to the throne could assume a triumphal character, as in the case of Nicephorus II. What emerges from this brief survey is that the distinction between a triumph and an adventus was rather blurred. An emperor’s entry into his city was normally made through the Golden Gate, but this practice could be varied in the light of particular circumstances.8 The mention of the Strategion as an alternative point of arrival in the sixth century is, as we shall see, of some interest. For victories won by generals, no procession through the city is recorded. Even Belisarius in 534, we are told, walked on foot from his house to the Hippodrome.9 We should now turn our attention to the route. Before Constantine the main highway leading to Byzantium, which we conventionally call the Via Egnatia, did not follow the Marmara coast all the way, but turned inland after Heraclea in order to avoid the deep accepted dates of the two triumphs are 831 and 837. W. T. Treadgold, “The Problem of the Marriage of the Emperor Theophilus,” GRBS 16 (1975): 332, argues that the Porphyrogenitus conflated the description of the two triumphal entries. 4 Bonn, De cerim., ed. Reiske, 498 ff (⫽ Haldon, 140 ff). 5 The text (De cerim., ed. Reiske, 496.9 [⫽ Haldon, 138, C687]) speaks only of a gate (po´ rth). 6 De cerim., ed. Reiske, 497–98 (⫽ Haldon, 138 ff). 7 De cerim., ed. Reiske, 607 ff, 612 ff. Both chapters mention ejpinika´ rioi. For the historical circumstances, see McCormick, Eternal Victory, 161, 166. Cf. also the acta in the Hippodrome after a victory over the Arabs, probably in 863: De cerim., ed. Reiske, 332–33. 8 Theophanes, Chronographia, ed. C. de Boor, 2 vols. (Leipzig, 1883–85; repr. Hildesheim, 1963), 469, records that on 24 June 793 Constantine VI brought in through the Blachernai Gate 1000 soldiers of the ¨ B 35 (1985): 8, Armeniac theme who had mutinied. M. McCormick, “Analyzing Imperial Ceremonies,” JO and Eternal Victory, 142–43, makes the ingenious suggestion that the choice of that particular gate was inspired by the circumstance that 25 June was the anniversary of the delivery of Constantinople from the Arab blockade in 678, the service of thanksgiving being held in the Blachernai church. Undoubtedly, the parade of 793 celebrated an imperial success, but can it be called a triumph? 9 Procopius, Wars 4.9.3.

CYRIL MANGO

175

lagoons of Bu ¨ yu ¨ k and Ku ¨¸cu ¨k C ¸ ekmece. After passing through a place called Cenofrurio and another called Melantias, it reached Byzantium,10 no doubt at the main land gate, which we know to have been just short of Constantine’s forum.11 Melantias, eighteen miles distant from Byzantium, has not to my knowledge been exactly localized, but it is usually placed at the north end of the two lagoons. By 333, as we learn from the Bordeaux Itinerary, a coastal road had been built. From Heraclea it proceeded to Selymbria, Athyras (Bu ¨ yu ¨k C ¨¸cu ¨k ¸ ekmece), and Rhegion (Ku C ¸ ekmece), and so reached Constantinople.12 Thus there were now two roads, the old and the new,13 and we have to ask at what points they entered Constantine’s city. The old road probably did so just northwest of the church of the Holy Apostles because it was there that we find the Melantias Gate,14 and we have seen that Melantias was the station closest to Byzantium on the old road. The fact that Constantine sited his mausoleum there, at the highest point of his city, suggests that he still considered that artery to be very important. As for the new road, it probably terminated at “Constantine’s Golden Gate.” Inside the city the two roads would theoretically have met at an acute angle in the general area of modern Beyazit or Laˆleli. I have put “Constantine’s Golden Gate” in quotation marks because it cannot be taken for granted that Constantine built it. The Notitia of ca. 425 offers its earliest attestation and is also the only text that calls it “Porta aurea.” 15 Thereafter it is seldom mentioned and appears under several names, including the gate of At(t)alus.16 It was decorated with statues, including, allegedly, one of Constantine, which fell down in 740.17 Our only description of the gate is by Manuel Chrysoloras (early 15th century), who specifies that it was built of big marble blocks, had a very wide and lofty opening, and was crowned by some kind of stoa.18 Strangely enough, the Patria does not refer to it, unless it is the enigmatic tetrapylon with a chamber on top of columns (whatever that may mean) indicated in that part of the city.19 (In Rome, too, the Porta Triumphalis is represented as a quadrifrons.)20 Itineraria romana, ed. O. Cuntz (Leipzig, 1929), 1:20, 33, 50. Zosimus, 2:30.2, 4. 12 Itin. romana, ed. Cuntz, 90–91. 13 Cf. the mention of the strata vetus by Eutropius, Breviarium, 9.15.2, with reference to the murder of Aurelian at Cenofrurio. 14 The decisive indication about the Melantias Gate is that it was at the Deuteron, whose situation was established by A. M. Schneider, “Deuteron und Melantiastor,” BNJ 15 (1939): 181–86. 15 Notitia dignitatum, ed. O. Seeck (Berlin, 1876), 239.8 and again at 243.56. 16 The identity of Attalus is unclear. The only prominent person of that name in the period that concerns us, Priscus Attalus (PLRE, II, Attalus 2), was a usurper supported by the Visigoths (409–410, 414–415), who would certainly not have been honored at Constantinople. Indeed, his fall was celebrated by a theater spectacle and chariot races: Chronicon Paschale, ed. L. Dindorf, CSHB (Bonn, 1832), 573. 17 The great earthquake of that year threw down “the statue of Constantine the Great that stood above the gate of Atalos as well as that of Atalos himself,” Theophanes, Chronographia, ed. de Boor, 412. 18 PG 156:45C-D. 19 Scriptores originum Constantinopolitanarum, ed. T. Preger (Leipzig, 1989), 181 (we cite this edition by page). A. Berger, Untersuchungen zu den Patria Konstantinupoleos, Poiki´la Buzantina´ 8 (Bonn, 1988), 359, thinks that the entry is misplaced because the silly story it tells about the corpse of a dead emperor being placed there prior to burial implies a situation between the Great Palace and the church of the Holy Apostles. The author of the Patria does not, however, shine by his logic. The existence of a tetrapylon near the Sigma is confirmed by the Interpretation of the Oracles of Leo the Wise, PG 107:1145B, which Berger quotes. 20 Coarelli, “La Porta Trionfale,” 68 and figs. 1, 2, 5, 6, 8. 10 11

176 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE The gate stood until 1509, when it was toppled by earthquake.21 It was then known as the gate of Jesus (I˙ sakapı) after a painting of the Crucifixion that was placed upon it, perhaps in the late Byzantine period.22 Its situation, as generally acknowledged, is marked by the small Byzantine church converted into a mosque called I˙ sakapı mescidi, and corresponds pretty well to what we know about the line of the Constantinian land walls. The latter, however, were left unfinished by Constantine and were completed by Constantius II.23 It may even be, as we shall see below, that Julian had a hand in their construction. The anecdotal evidence outlined above suggests that a great ceremonial gate erected either by Constantine or one of his immediate successors would naturally have marked the southern avenue as the triumphal way. The choice may have been influenced by the fact that the Byzantine equivalent of the Campus Martius (where triumphs traditionally started in Rome) was located at coastal Hebdomon at the seventh milestone. The Hebdomon, however, is first attested in 364.24 Our triumphal way can still be traced for much of its course on the map of Istanbul. From the vicinity of I˙ sakapı mescidi, a straight street called Cerrahpas¸a Caddesi (Fig. 1) proceeds past the pedestal of the Column of Arcadius as far as the Murad Pas¸a mosque (built in 1471–78), in front of which were found in 1957–58 the piers of a monumental arch and part of a colossal statue representing a river god.25 After a gap corresponding to the valley of Aksaray, we pick up the street again at the Theodosian arch, and we know that from there it proceeded in a perfectly straight line, past Constantine’s forum, to the Milion, a distance of 1,400 meters. It need not be assumed that the entire extent of this avenue was laid out by Constantine’s engineers, except on paper as a long-range project. The ceremonial stations along what became the triumphal way are known from several medieval itineraries in the Book of Ceremonies.26 Although we might argue about the exact location of some of them, the sequence is established, and there are enough fixed or nearly fixed points to localize the remainder in broad terms (Fig. 2). We know much The relevant Ottoman texts unfortunately give no particulars about the appearance of the gate. See ¨ tu S. Y. O ¨ ken, “Isa kapı Mescidi ve Medresesi in Istanbul” (diss., Bonn, 1974), 134 ff. 22 See T. Papazotos, “To Isa kapı¯ mescidi sthn Kwnstantinou´ polh,” Delt.Crist. Arc. j JEt., 4th ser., 18 (1995): 44 ff. 23 Julian, Or. 1.33, ed. J. Bidez (Paris, 1932), 1:59. 24 When Valens was proclaimed there: Fasti Hydatiani (Consularia CP), ed. T. Mommsen, MGH, AA IX, 240; Chron. Pasch., ed. Dindorf, 556. It appears from Themistius, Or. 6.83a, that at the time the Tribunal of the Hebdomon had not yet been given an appropriate architectural setting. Thenceforth the Hebdomon became the normal locus for imperial proclamations, which meant that the emperor’s entry into the city would naturally have taken place through the Golden Gate as recorded in De cerim., ed. Reiske, 414 ff. 25 ¨zesi Yıllıg˘ı 1 (1959): 3, 19, says that the piers, which were immediately deF. Dirimtekin, Ayasofya Mu stroyed, were of limestone, measured 5 ⫻ 6 m, stood 6 m apart and were preserved to a height of 2 m. R. Janin, REB 21 (1963): 256, on the other hand, states that the piers measured 3 ⫻ 4 m and were separated by a space of 3 m. No photograph of them appears to have been published. For the river god, see N. Fıratlı, La sculpture byzantine figure´e au Muse´e arche´ologique d’Istanbul (Paris, 1990), no. 507, giving the date of discovery as 1959, and my remarks in Le de´veloppement urbain de Constantinople (IVe–VIIe sie`cles), TM, Monographies 2, 2nd ed. (Paris, 1990), 70. 26 The more important passages for the stretch between the Golden Gate and the Great Palace are De cerim., ed. Reiske, 1.8 (pp. 55 ff), 1.17 (pp. 100, 105 f), and the triumph of Basil 1 (pp. 501 f). For the Holy Apostles–Great Palace stretch, see 1.5 (pp. 49 ff), 1.10 (pp. 74 ff). 21

CYRIL MANGO

177

less about the northerly street. The wide, nearly straight avenue that today links Beyazit, Fatih, and the Adrianople Gate is a modern creation that did not exist even in the nineteenth century.27 Of the stations named in the Book of Ceremonies, only the gate of Charisios (Adrianople Gate), and the churches of the Holy Apostles and St. Polyeuktos are fixed. We do not even know for sure where the north and south streets met, although we are told it was at the Capitolium, which modern scholarly opinion places at Laˆleli. I believe that view to be, broadly speaking, correct. Marcian’s column, which is not mentioned in Byzantine texts, may have stood off rather than on the northern street.28 For our purpose the Capitolium is potentially of great interest. It is said to have been erected by Constantine, and its situation at a major urban junction identifies it as an important landmark. It seems that no capitolia were built in the empire after the Severan period,29 but even in the fourth century the term must have been understood as a temple of the Capitoline Triad or, at any rate, a temple dedicated to the tutelary deity of Rome. It is hard to imagine to whom or what Constantine’s Capitolium would have been dedicated, unless it was to the victorious sign of the cross. We know that it had a prominent cross or chrismon that fell down in a storm in 40730 and was therefore in an exposed position. But was it an original feature or a later addition? Setting aside Professor Speck’s imaginative speculations,31 we know little about the building apart from there having been exedras on three sides and several porphyry monuments in front of it, including the Tetrarchs that are now in Venice. It is not certain whether Constantine ever celebrated a triumph at Constantinople. He may have done so after his victory over the Goths in 332,32 and it may be that the socalled Column of the Goths, which still stands in the gardens of the Seraglio, was erected or adapted to celebrate that occasion. The inscription on its pedestal, Fortunae reduci ob devictos Gothos, replaces, as Professor Peschlow has shown,33 an earlier one, also in Latin but of unknown content. We know that the column supported a statue of Tyche,34 and it is possible that there was next to it a temple of Fortuna Redux, as there was near the Porta Triumphalis in Rome. If Constantine did plan a triumphal way, it may even be argued on the Roman analogy that the Capitolium would have been its terminal point, meaning that the procession would have moved not from west to east, as it did in later times, but from east to west, starting perhaps at the Strategion, the agora of ancient Byzantium, where Constantine dedicated an equestrian statue of himself.35 The Strategion, later upgraded to a forum Theodosiacum, also boasted an Egyptian obelisk and an 27 As shown by the earliest detailed street map of the city (1882), E. H. Ayverdi, ed., 19. Asırda I˙stanbul haritası (Istanbul, 1958). 28 In my De´veloppement urbain, 30, 46, I may have erred in assuming that Marcian’s column lay on the northern street. Its pedestal, if I am not mistaken, is aligned very nearly north-south. 29 See M. Cagiano de Azevedo, “I ‘capitolia’ dell’impero romano,” MemPontAcc 5 (1941): 66. 30 Chron. Pasch., ed. Dindorf, 570. 31 P. Speck, “Urbs, quam Deo donavimus . . . ,” Boreas 18 (1995): 143–73. He even offers (p. 152) two alternative reconstructions. 32 See McCormick, Eternal Victory, 39. 33 “Betrachtungen zur Gotensa¨ule in Istanbul,” in Tesserae. Festschrift J. Engemann (Mu ¨ nster, 1991), 215–28. 34 As clearly stated by John Lydus, De mensibus, 4.132. 35 Socrates, 1.16.1.

178 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE arch surmounted by a bronze statue of Tyche holding a cornucopia.36 In the same general area, we find something called “the Victories,” presumably a monument.37 Finally, in Region IV, which corresponds to the valley between the first and second hills, stood a Liburna marmorea, navalis victoriae monumentum,38 which may have commemorated Constantine’s victory over Licinius in the Hellespont in 324 or possibly that won over Gainas in 400. In short—setting aside the Column of the Goths, which stands somewhat apart— there is evidence for a concentration of victory monuments in the area of the Strategion, which was certainly linked by a major street to the Augustaion/Great Palace center. That probably explains the choice of the Strategion for the staging of adventus ceremonies in the sixth century. The situation of the Strategion is not, however, known exactly (see Appendix), and I would not like to speculate here about the probable course of the connecting street, which need not have been laid out according to a rectilinear pattern. That street must have terminated at a maritime gate, probably the one called the gate of Eu¨¸sk kapısı.39 genius in the Byzantine period and usually identified with the Turkish Yalıko ¨¸sku Unfortunately, the gate in question was destroyed along with the Yalıko ¨ when the 40 railroad was built in 1871. It seems to have been of marble and was adorned with a statue of Julian, the subject of an epigram in the Palatine Anthology (9. 689): “Julian, whom you see (ou»to" Ij ouliano´ "), in setting up walls for the protection of the people, has erected a trophy, symbol of his vigilance. He is eager to slaughter his enemies at a distance rather than waging war in front of the city.” Assuming this was the emperor Julian, who was, in fact, planning a Persian campaign during his stay at Constantinople (361–362), one may well ask what were the peoplesaving walls (laosso´ a tei´cea) he built and why the inscription was placed at the gate of 36 For the obelisk, see Notitia, ed. Seeck, 233, and the Patria in Scriptores, ed. Preger, 183. For the arch, see Marcellinus Comes, a. 510; in the Middle Ages it was known as the Arch of Urbicius (Patria, 141.8). 37 Chron. Pasch., ed. Dindorf, 494, with reference to the city founded by Byzas, e“nqa nu'n kalou'ntai aiJ Ni'kai, plhsi´on tw'n Kili´kwn (khli´kwn cod.). “The Cilicians,” if that is the right reading, appears to refer to a locality. We may note that one of the plotters in the conspiracy against Justinian in 562, the banker Marcellus, is described as ajrguropra´ th" oJ tw'n Kili´kwn oJ e“cwn to` ejrgasth´ rion plhsi´on th'" aJgi´a" Eijrh´ nh" th'" ajrcai´a" kai` ne´ a" (Malalas, frag. 49, Excerpta de insidiis, ed. C. de Boor [Berlin, 1905], 173). One may wonder whether “the Victories” had anything to do with the group of three 16-cubit statues that were set up ejn tv' Bosporei´v in acknowledgment of Athenian help against Philip of Macedon (340–339 B.C.). These are said to have represented the demos of Athens being crowned by those of Byzantium and Perinthus. The source is a “decree of the Byzantines” appended to Demosthenes, De corona, 90–91. The text of the decree is now held to be spurious, but it does not follow that the gigantic statues did not exist. 38 Notitia, ed. Seeck, 232. Its only other mention is in the context of the Nika Riot, when the insurgents departing from the Octagon, e“balon pu'r ejpi` to` Li´burnon ejpi` th` n Magnau´ ran, but the fire was put out: Chron. Pasch., ed. Dindorf, 623. If the Magnaura was the palace hall of that name, we may conclude that the Liburna was on or near the Augustaion, which did form part of the Region IV. 39 ¨¸sk kapısı (at the junction of the Golden Horn walls with those of the Seraglio), For the situation of Yalıko see “Plan particulier du Se´rail,” in A. I. Melling, Voyage pittoresque de Constantinople et des rives du Bosphore (Paris, 1819), album, pl. 49; and A. G. Paspates, The Great Palace of Constantinople, trans. W. Metcalfe (London, 1893), map at the end. For the identification with the gate of Eugenius, A. van Millingen, Byzantine Constantinople: The Walls of the City (London, 1899), 227. 40 MM 2:467, 564. C. G. Curtis and M. A. Walker, Broken Bits of Byzantium (London, 1869–91), 1: no. 2, record in the neighborhood of the gate a relief of a Victory standing on a globe and in the same area (nos. 4–7) fragments of a “frieze” representing a Roman soldier holding a shield, a drowning man and striding horses.

CYRIL MANGO

179

Eugenius. The walls referred to would hardly have been the maritime walls, since in the fourth century the city faced no threat from the sea.41 If the reference is to the completion of the land walls, it would follow that the gate of Eugenius was at the time the main point of disembarkation for visitors arriving at Constantinople—a suitable place for exhibiting such an inscription. Tentative as they are, the observations above may provide some background to what has been considered an innovation of the Komnenian period, namely the rerouting of triumphs along the relatively short stretch from the “eastern gate of the Acropolis” to Hagia Sophia and the Great Palace.42 Unfortunately, the new itinerary is not described in sufficient detail to make confident conclusions.43 As we have seen, there is considerable uncertainty both about the existence and the direction of a Via Triumphalis during the first half century of the capital’s history. It is only in the Theodosian period that its path becomes established by the creation of the ceremonial fora of Theodosius and Arcadius, with their famous historiated columns, triumphal arches, and statues. The fiction that Theodosius was descended from the emperor Trajan no doubt inspired reproduction of Trajan’s column and even some features of his forum in Rome.44 Of these much-discussed monuments I have nothing new to say. The erection of the Theodosian walls and the new Golden Gate45 added about 1,800 meters to the length of the Via Triumphalis. I do not know why the Golden Gate was placed so close to the seashore instead of being more or less in line with the old gate, but it is worth pointing out that at no time was there a straight avenue between the two. I believe that some attempt at a monumental linkage was made but not followed up. In 435 yet another Theodosian forum was constructed in loco qui Heliane dicitur,46 clearly The ancient city was, of course, walled all round, but the Constantinian extension had no maritime defences and it was only in 439 that sea walls were built. For the identification of Julian as the emperor, see A. M. Schneider, “Mauern und Tore am Goldenen Horn zu Konstantinopel,” NachrGo¨tt (1950): 96. 42 See P. Magdalino, The Empire of Manuel I Komnenos, 1143–1180 (Cambridge, 1993), 240 ff. 43 Nicetae Choniatae Historia, ed. J. L. van Dieten, CFHB (Berlin–New York, 1975), 18–19 (triumph of 1133), ¨randner (Vienna, 1974), no. 157–58 (triumph of 1167). Theodoros Prodromos, Historische Gedichte, ed. W. Ho VI, p. 221, v. 24 ff, refers rather obscurely to a colonnade on the shore. The editor translates (p. 226 f) “auf der grossen Landzunge an der Enge der Propontis, wo zahlreiche Sa¨ulen, kreisformig angeordnet, in Kreuzform zusammengefu ¨ gt sind,” which I find difficult to visualize. The “eastern gate” is usually identified as that of St. Barbara (Topkapı), which was flanked by two marble towers, destroyed in ca. 1816. Situated just round the point of the promontory, facing east, where the current is strongest, and lacking a harbor (Ioannis Cantacuzeni eximperatoris historiarum libri quattuor, ed. L. Schopen, 3 vols. [Bonn, 1828–32], 3:232), the gate of St. Barbara would not have been a convenient place of disembarkation. 44 On the Forum of Theodosius, see now A. Berger, “Tauros e Sigma: Due piazze di Costantinopoli,” in Bisanzio e l’Occidente: Arte, archeologia, storia. Studi in onore di Fernanda de’Maffei, ed. M. Bonfioli, R. Farioli Companati, and A. Garzya (Rome, 1996), 19–24, who argues that, contrary to previous opinion, it was a mere 55 m across. 45 Opinion has long been divided on the question whether the existing marble gate was put up by Theodosius I or II. The majority view favors the latter, but the claim of Theodosius I was still championed by M. Wheeler, “The Golden Gate of Constantinople,” in Archaeology in the Levant. Essays for K. Kenyon, ed. R. Moorey and P. Parr (Warminster, 1978), 238–41. What is reasonably certain is that the triple arch and flanking pylons were built as a unit and that the abutting curtain wall on the south represents a second phase, as observed by O. Davies, JRS 34 (1944): 74–75. That does not necessarily imply that the first and second phases were separated by an appreciable lapse of time. It seems obvious to me that the gate, with its massive pylons, was planned in the context of the new land walls, which, for all we know, may have been on the drawing board already in the reign of Theodosius I. 46 Marcellinus Comes, s.a. 41

180 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE referring to Helenianae. A palace of Helenianae is known to have stood in the area of Samatya/Sulumanastır and adjoined or included a semicircular portico called Sigma.47 Precisely at the Sigma, the Patria records a statue of Theodosius II on a column that was erected by the eunuch Chrysaphios Tzoumas.48 While normally the historical veracity of the Patria inspires little confidence, in this case there is no need to doubt the information. Chrysaphios Tzoumas was indeed the most influential person at court in the latter years of Theodosius’ reign, and the author of the Patria, who mentions him only here, could not have pulled his name out of a hat. In the ninth/tenth century, the triumphal way led from the Theodosian Golden Gate to the Sigma, but there one had to turn left along a street leading to the church of St. Mokios in order to reach the Exokionion or (H)exakionion, the plaza outside the “Constantinian” Golden Gate.49 Extending about 5.5 kilometers from the Theodosian Golden Gate to the Milion and the Hippodrome, the triumphal way took one hundred years to create and remained basically unchanged after 435. Some but probably not all of it was colonnaded. The Porticus Troadenses, so named after their columns of Hellespont granite, bordered it from the “Constantinian” Golden Gate to perhaps the Forum of Arcadius,50 and there were certainly continuous porticoes between the Capitol and the Milion.51 I find no clear reference to porticoes in the valley of the Lycus or outside the “Constantinian” Golden Gate. Even so, compared to the circuitous route of the Via Triumphalis in Rome, Constantinople’s must have had an undeniable grandeur. The last really flamboyant addition to the triumphal monuments of Constantinople was made by Justinian—I am referring to his column cum equestrian statue at the Augustaion. Its position was significant, standing as it did both at the terminal point of the triumphal way and in the closest proximity to Hagia Sophia. It was also the first and only time, if I am not mistaken, that an equestrian rather than a standing statue was placed on top of a monumental column. But did it, strictly speaking, commemorate a victory? I Berger (as in note 44) argues that the Sigma was not a forum or part of a forum, but the entrance porch of the Helenianae, which he places on the site of St. Mary Peribleptos (Sulumanastır). I am not entirely convinced. 48 Scriptores, ed. Preger, 182. 49 Combine De cerim., ed. Reiske, 501.19 and 506.7. 50 It is reasonably certain that the Troadesian porticoes, which were in Region XII (Notitia, ed. Seeck, 239), terminated at the Constantinian Golden Gate. Hesychius (Scriptores, ed. Preger, 16.5 ⫽ Patria, 138.1) says that Constantine moved the land walls “to the so-called Troadesian porticoes,” and Chron. Pasch., ed. Dindorf, 590, refers to “the Troadesian walls.” The Vita Isaacii in cod. Monac. gr. 366, as reported by V. Tiftixoglu in ¨hgeschichte Konstantinopels, ed. H.-G. Beck (Munich, 1973), 57, 96 n. 66, speaks of “the so-called Studien zur Fru Troadesian gate” (probably the Constantinian Golden Gate). Marcellinus Comes, a. 448, states that a sudden fire destroyed utramque porticum Troadensem turresque portarum utrasque, the damage being immediately repaired by the Praetorian Prefect Antiochus. The towers in question may have been the flanking towers of the Golden Gate. The west starting point of the porticoes is less certain. The statement that the earthquake of 447 caused damage between the Troadesian porticoes and the Bronze Tetrapylon (on the Mese) is not particularly helpful: Chron. Pasch., ed. Dindorf, 589 (misdating the earthquake to 450); Malalas, ed. L. Dindorf (Bonn, 1831), 363 f; Synaxarium CP, 425.9. Berger, Untersuchungen, 353, may not be right in suggesting that the tetrapylon in question was the one near the Sigma. Fig. 3 shows a set of granite columns, which may well have belonged to the Troadesian porticoes, dug up in 1982 a short distance west of the Column of Arcadius. They have since disappeared. 51 Cod. Just. 8.10.12.6 (undated law of Zeno). 47

CYRIL MANGO

181

have long been intrigued by the reference to “three pagan kings, also bronze and on columns, kneeling before Emperor Justinian and offering their cities into his hands.” This description occurs only in two Russian pilgrim texts, of 1390 and 1420 respectively,52 whose authors may simply have picked up a legend they were told. The three kings can easily be explained as those of the Persians, Ostrogoths, and Vandals. But if Justinian did create such a three-dimensional tableau instead of representing the barbarians in relief on the pedestal of his column, as would have been normal, how is it that this striking arrangement is not mentioned in any description of the monument from Procopius onwards? I hope to discuss on another occasion an independent, if also late, attestation of the three kings transformed into the three Magi. What appears fairly certain is that tribute-bearing barbarians, whether in relief or as statues, did form part of Justinian’s monument. The dedication of triumphal columns, statues, and arches ceased at Constantinople in the early years of the seventh century, but triumphs, as we have said, continued to be celebrated until the twelfth. We know that the tradition of depicting imperial victories was continued in mosaic or painting in palace halls and private mansions, such as those of Basil I and Manuel I, for example. But was any attempt made at a more public expression? For an answer we must turn to the Golden Gate. It can be described as being double. The outer, or propylaeic, gate opens through the forward wall, which presumably had existed in one form or another since the Theodosian period. The outer gate led into a paved courtyard flanked by two marble pylons, between which stood the Theodosian gate with its three arched openings, originally supporting a quadriga of elephants (as did the Porta Triumphalis in Rome). Between the outer gate and the moat ran a walkway protected by a battlemented parapet. When the Castle of the Seven Towers was built by Mehmed the Conqueror, the outer gate became redundant and was blocked up. It seems that, at about the same time, the walkway in front of it was barred by a transverse wall. Visitors could then gain only a distant view across the moat. The outer gate is still partly preserved: two ancient columns of Carystus marble are surmounted by Theodosian capitals supporting an arch. Columns, pedestals, and capitals are all reused and have been added to a pre-existing gate structure behind them. On either side of the gate, better preserved on the right than on the left, are remnants of two tiers of marble frames (only one tier remaining on the left) inserted into a straight wall of fairly regular masonry (Figs. 3, 4). The frames, placed at somewhat uneven intervals, once enclosed reliefs and are themselves reused (Fig. 5). They include a cornice fragment not earlier than the sixth century, to judge by its decoration of joined upright palmettes (Fig. 6). Soundings carried out by Theodore Macridy and Stanley Casson in 1927 shed some light on this ensemble and led to the discovery of small fragments of reliefs, both antique and Byzantine, which are now in the Istanbul Archaeological Museum.53 Further study of the standing structure was carried out by B. Meyer-Plath and 52 G. P. Majeska, Russian Travelers to Constantinople in the Fourteenth and Fifteenth Centuries, DOS 19 (Washington, D.C., 1984), 136, 184 and commentary on p. 240. Cf. my remarks in Studies on Constantinople (Aldershot, 1993), study X, 3. 53 “Excavations at the Golden Gate, Constantinople,” Archaeologia 81 (1931): 63–84.

182 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE A. M. Schneider, who came to the view that the outer gate and straight curtain wall on either side of it dated from the second half of the fourteenth century,54 a judgment that has been endorsed by Wolfgang Mu ¨ ller-Wiener55 and further developed by Sarah Gu56 berti Bassett. However, the reasoning of Meyer-Plath and Schneider was based not on archaeological evidence, which merely shows a succession of undated building phases, but rather on a desire to link these phases with textual information. It so happens that Byzantine writers have little to say about any alterations of the Golden Gate complex for the greater part of the Middle Ages57 but they do record considerable activity in the fourteenth century, namely under John VI Cantacuzenus (1347–54) and at the end of the reign of John V (1389–90). The relevant texts, however, have no demonstrable connection with the propylaeic gate and forward curtain wall,58 while the careful building technique of the observable phases59 has none of the characteristics of late Palaiologan work and, to my eye, looks middle Byzantine. It is, however, the missing reliefs that interest us, and here, without reproducing the entire dossier of travelers’ accounts, we must enter into some detail. The reliefs are first specifically mentioned in 1411 by Manuel Chrysoloras, who describes them as the Labors Die Landmauer von Konstantinopel (Berlin, 1943), 2:54 ff. Bildlexikon zur Topographie Istanbuls (Tu ¨ bingen, 1977), 297 and fig. 388, stating that the outer gate was thickened by John V and provided with a new facade decorated with reliefs. 56 “John V Palaiologos and the Golden Gate in Constantinople,” in To` JEllhniko´ n. Studies in Honor of Speros Vryonis, Jr. (New Rochelle, 1993), 1:117–33. 57 Choniates, ed. van Dieten, 570.27, reports that the Byzantine notables and armed men, in their haste to flee the city in 1204, broke down the “newly-built wall” at the Golden Gate (to` neo´ dmhton ejkei'se kataspw'si tei'co"). The identity of that wall remains unclear. In the Bonn edition (ed. I. Bekker [1835], 754.15), instead of tei'co", we read tw'n pulw'n ejpitei´cisma, translated as nuper structo propugnaculo diruto, which has led to the unwarranted belief that the outer wall was meant. Thus Meyer-Plath and Schneider, Landmauer, 41, 57: “das neugebaute Vorwerk des Goldenes Tores.” 58 As for Cantacuzenus, he states himself (Bonn ed., 3:292–93, 304) that on gaining power (1347) he repaired the top part of the two marble towers, which had long been neglected. In 1354 this fort (␸rou´ rion) was surrendered to John V, who undid the repairs and left it unguarded. This, clearly, has nothing to do with the problem that concerns us. The later activity of John V is described by Ducas (ed. V. Grecu [Bucharest, 1958], 75), writing a good 70 years after the event, and represented as a reaction to mounting pressure from Sultan Beyazid (who acceded to the Ottoman throne on 15 June 1389), hence hardly before the latter part of that year. John’s intention was not merely to strengthen the Golden Gate but to construct there a “little town” (poli´cnion). To that end he built “two towers on either side [or on one of the two sides?] of the gate that is made of white marble” (kataskeua´ sa" pu´ rgou" du´ o ejn tv' qate´ rv [th qate´ ra cod.] tw'n merw'n th'" pu´ lh" leukv' marma´ rv sunhrmosme´ nv [read sunhrmosme´ nh"?]). He did so not with new materials, which evidently could not be procured in a hurry, but by demolishing three old churches (All Saints, Forty Martyrs, and what remained of the Basilica of St. Mokios). He also fenced off part of the city from the Golden Gate as far as the seashore and made a harbor there as a place of refuge. The exact situation of the two towers is not made clear, but it is likely that they were on the inside of the Theodosian fortifications to provide a point of attachment to the wall extending to the seashore (a distance of some 350 m), which was what constituted the “little town.” Soon thereafter Beyazid demanded the demolition of the new fort, and John V had no choice but to comply, a humiliation that hastened his death. These events have often been discussed, e.g., by G. Kolias, JEllhnika´ 12 (1952): 55–57; by myself, BZ 45 (1952): 383–84; by J. W. Barker, Manuel II Palaeologus (New Brunswick, 1969), 467–68, 546. I find it quite incredible that in the desperate circumstances in which he found himself John V would have created a vast decorative ensemble, nor can I credit his ability to procure unusually large mythological reliefs. 59 See sketch-plan in Meyer-Plath and Schneider, Landmauer, 54, fig. 17, phases C and D. Phase D, which goes behind the marble jambs of the gate, exhibits very meticulous work with weathered joints plastered over and ruled horizontally and vertically. 54 55

CYRIL MANGO

183

of Hercules and the punishment of Prometheus,60 although they may be the same works referred to in the Patria to the effect that the small idols (xoana) above and below a statue of Victory (of which more later) conveyed much knowledge of the future to those skilled in interpreting such matters.61 Our next witness is Gyllius (1544–50), whose account may be rendered as follows: . . . the Seven Towers are enclosed by the city wall, having a gate, once open, but now obstructed, whose jambs ( parastades) are two Corinthian columns of mottled [sic] marble marked with green veins, supporting eight colonnettes forming three arches (sustinentes octo columellas efficientes tres arcus). On the left side of the gate are preserved six marble plaques, each framed by colonnettes (some rounded, others square), containing nude figures fighting with clubs, artistically carved in relief. The upper ones have above them cupids, as if rushing to fly. On the right side are likewise six plaques framed in the same fashion by colonnettes. In the first plaque of the lower [row] is a youth holding a musical instrument, lying on his back, his legs crossed. Above him is suspended a little figure, like that of a cupid, and above the cupid is a woman. In the upper plaque, is a nude figure holding an upright club, a lion skin wrapped round his arm, leading dogs with his left hand. Above this figure is a she-lion with swollen paps. Another plaque contains two peasants carrying baskets full of grapes. In another is a winged horse, a woman holding its bridle. At the back are two other women. In the upper part of the plaque, another woman is lying down, and near her lies a youth. I have set down these particulars because of the ancient and high artistry of these plaques.62

The statement that the two Carystian columns carried eight colonnettes forming three arches troubled Josef Strzygowski, who, after toying with the unlikely idea that columellas should be emended to columbellas (the little doves at the corners of the Theodosian capitals), had to admit that Gyllius made a mistake: the eight colonnettes were on either side of the gate.63 It is, of course, entirely possible that Gyllius erred in this respect or that his text was garbled in the process of editing, but unless we are sure that what he says is impossible, we are left with the statement that there was a superstructure above the gate consisting of either paired colonnettes or, more probably, two rows of four each, subtending three arches. In the 1620s a notorious attempt was made on behalf of two English noblemen, the Earl of Arundel and the Duke of Buckingham, to acquire the reliefs for their private collections of antiquities. The task fell to the British ambassador, Sir Thomas Roe, who was less than enthusiastic about the project but did his best to please the noble lords. Besides, Arundel had sent his chaplain, a certain William Petty, who proved extremely persistent. Writing to Buckingham in 1625, Roe describes the reliefs of the Golden Gate in the following terms: . . . upon the sides and over yt [my emphasis] twelve tables of fine marble, cutt into historyes, some of very great relevo, sett into the wall, with small pillars, as supporters. Most of the figures are equall, some above the life, some less. They are, in my eye, extremly decayed; but Mr. Petty doth so prayse them, as that he hath not seene much better in the great . . . collections of Italye. . . . There are of them but sixe that are woorth the PG 156:45. Scriptores, ed. Preger, 183. 62 De topographia Constantinopoleos (Lyon, 1561), iv.9, pp. 217–18. 63 “Das Goldene Thor in Konstantinopel,” JDAI 8 (1893): 19–21. 60 61

184 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE taking downe, the other being flatt Gothish bodyes, lame, and of later tymes, sett up only to fill places of the other sixe. Two, in my opinion (though Mr. Petty like them) want much of excellence, great, but brute; and, as I coniecture, are some storye of Hercules, not mentioned in his labors. The fower [four] to which I have most affection, are fuller of woorke: the one is . . . an Endimion carelessly sleeping by his sheepe; Luna descending from the sckye with a torch in her hand, representing night; and a Cupid hovering in the ayre. . . . The next is an historye I understand not, eyther of some race or game; in the middest is a horse, a young man naked running by yt, and reaching to pull another off. Some other figures there are, which I remember not; but it hath beene a peice of great bewtye and art; the relevo so high, that they are almost statues, and doe but seeme to sticke to the ground; some leggs, and other parts, standing holow off, are broken and lost; yet in the whole, it hath a show of rare antiquitye. The third is a Pegasus, with the Nimphs or Muses; one representing the founteyne Pirenne powring out water. These figures are many, but less than halfe the life. . . . The last is a Satyre, sckipping betweene an Hercules, or a wild man, and a woman . . . the one hath a whip in his hand, the other a pot of water behind her, and may signifye a rescue from ravishment; these are above the life, and rather great and stately than delightfull.64

Roe did his best to bribe the captain and soldiers who manned the castle, but to no avail. “The soldiers,” he continued, “cannot steal them, beeing 30 foot and 40 foot high, made fast to the wall with iron pinns; and must bee let downe with scaffolds, and the help of at least 50 men.” The following year, in April 1626, Roe was still hopeful. The reliefs were promised and the money deposited, but a month later the whole project collapsed. No less a person than the Sultan’s Grand Treasurer had agreed to remove the reliefs, but when Roe arrived on the scene, the Castellano and the people mutinied, claiming that the statues were enchanted and that some calamity would befall the city if they were taken down. “Though I could not gett the stones,” concluded Roe, “yet I allmost raised an insurrection in that part of the citty.” 65 Since it is unpublished, it may also be worth quoting the 1670s account of Dr. John Covel, although he does not appear to have inspected the reliefs at close quarters. Covel wrote: In the wall on either side this Gate are severall Bassi Relievi, which have been very good work; we have a particular account of them in P. Gyllius which we must be satisfyed withall, for to attempt to draw them without leave, may be the hazard of a mans head; the Turkes, as is said, being extremely jealous of men making any remarkes [notes], especially about this place. I have often mention’s this matter to his Excellency, the Marquesse Nointel, embassadore from France, who was very curious and had two painters by him for such purposes; he resolved then to get leave for them to designe this whole work, but whether he did it after my departure or no, I know not. It is my present opinion that these several Tables were not thus placed together at first, neither do they seem all of the same Hand, or Order, or Work, or Designe. Travellers do fancy there, Endymion, Phaeton, Hercules, Pegasus, several Cupids, Venus, Adonis, Apollo, Cadmus’ sons of the Earth fighting with clubs, men with baskets of Grapes. Yet I cannot imagine what relation these can have to one another. Therefore I conjecture they might have 64 The Negotiations of Sir Thomas Roe in His Embassy to the Ottoman Porte (London, 1740), 386–87. On this incident, cf. A. Michaelis, Ancient Marbles in Great Britain, trans. C. A. M. Fennell (Cambridge, 1882), 11 ff. 65 Negotiations of Sir Thomas Roe, 434, 444, 495, 497, 512.

CYRIL MANGO

185

been brought of old from other places and here placed to adorne this Gate when it was open, it having been of Great Note formerly.66

Covel fails to say that in his time there were only five or six reliefs remaining, as we learn from Dr. Spon and Sir George Wheler, who met him at Constantinople. These two travelers, who did see the reliefs, mention specifically the Fall of Phaethon, Hercules leading Cerberus, and a sleeping Adonis being approached by Venus. They also say that the Marquis de Nointel was promising to have them sketched.67 Alas, this does not appear to have happened. Gradually, the number of reliefs was reduced to two,68 then to one,69 then to none.70 Although we possess no pictorial record of them, we can attempt to visualize their appearance in the 1620s. Sir Thomas Roe’s account contains two noteworthy statements. The first is that the reliefs were 30 or 40 feet above the ground and that some of them were over the gate. Since he was not a casual observer but concerned with the practicalities of taking them down, it is not likely that he grossly misjudged his figures. Today the bottom tier of frames, which are 2.30 meters tall, is at approximately eye level. The second tier rises another two meters or so. Nor has the ground level of the walkway risen to any appreciable extent. The explanation for this apparent puzzle is provided by a view of the Castle of the Seven Towers by Francesco Scarella of ca. 1685 (Figs. 8, 9).71 It shows the outer gate twice its present height, flanked by four rows of frames.72 The twelve reliefs seen by Gyllius and Roe, among others, would probably have been in the two uppermost tiers. Those in the lower tiers, if they ever existed, would have been destroyed at an earlier date. Possibly they included the Punishment of Prometheus, mentioned only by Chrysoloras. The second interesting statement made by Roe is that some of the figures—he specifically mentions the relief with Hercules, a satyr and a woman (probably Drunken Hercules)73 —were “above the life,” which accords with the height of the two bottom tiers of frames (2.30 and approximately 2 m, respectively). These are unusually large dimensions for mythological or genre reliefs.74 We are not dealing, therefore, with, say, cut-up Roman British Library, Add. MS 22912, fol. 83. J. Spon and G. Wheler, Voyage d’Italie, de Dalmatie, de Grece et du Levant, fait aux anne´es 1675 et 1676 (Lyon, 1678), 1:262–63. 68 So G.-J. Grelot, Relation nouvelle d’un voyage de Constantinople (Paris, 1680), 79, and (copying Grelot?) Cornelis de Bruyn (visit 1678–80), A Voyage to the Levant (London, 1702), 51–52. 69 So Choiseul-Gouffier (ambassador to the Porte from 1784 until 1789), Voyage pittoresque dans l’Empire ottoman (Paris, 1842), 4:477. Not shown in his view of the Golden Gate (pl. 89). 70 Cosimo Comidas de Carbognano, Descrizione topografica dello stato presente di Costantinopoli (Bassano, 1794; repr. Rome, 1992), 35. No reliefs were seen in 1795 by J. Dallaway, “An Account of the Walls of Constantinople,” Archaeologia 14 (1803): 241–42 and pl. XLIV. 71 Austrian National Library, cod. 8627, fol. 5. See F. Babinger, “Francesco Scarella e i suoi disegni di Costantinopoli (circa 1685),” Rivista d’arte 35 (1960): 153–67, esp. 155–56 and fig. 7. In spite of the fact that Babinger described our drawing as a “schizzo singolare,” it has attracted little attention. 72 The drawing appears to be in error in showing the arch of the gate at the level of the fourth rather than the second row of frames. Another interesting feature of the drawing is the superstructure of the two marble pylons. 73 Cf. Lexicon iconographicum mythologiae classicae (LIMC), s.v. Herakles, 3257 ff. 74 Prof. R. R. R. Smith kindly calls my attention to the relatively large reliefs representing the Labors of Hercules from the Roman villa of Chiragan near Toulouse, but these are no more than 1.50 m high: E. Espe´randieu, Recueil ge´ne´ral des bas-reliefs de la Gaule romaine (Paris, 1908), 2: no. 899. 66 67

186 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE sarcophagi but with pieces that must have been obtained from some sumptuous villa or public building, possibly outside Constantinople. These were combined with smallerscale carvings, both antique and Byzantine, to produce a huge montage. A comparable confection, which I discussed on another occasion,75 decorated the maritime facade of the imperial palace and was made up of non-figural sculpture with a few sitting lions thrown in. It was assembled, I believe, in the ninth or tenth century. The Golden Gate ensemble, which I would be inclined to place in roughly the same period, was an altogether grander affair. Some 40 feet high, it would have looked from a distance not unlike the face of a Roman triumphal arch. Only at closer quarters would its makeshift character have been apparent. Covel was, of course, right in saying that the reliefs had been collected from different places, were at different scales, and had no connection with one another. The imprecision of the travelers’ accounts, not all of them based on autopsy, does not, unfortunately, allow us to identify securely more than a few of the compositions: Endymion with Selene (represented by extant fragments), Hercules leading Cerberus,76 Pegasus tended by nymphs,77 probably the Drunken Hercules, perhaps Achilles pursuing Troilus.78 Conceivably, one could read some kind of message into this assemblage—perhaps victory or the blessings of peace following the toils of war—but I would be reluctant to do so. In a passage I have already quoted, the Patria speaks of a female statue holding a crown, i.e., a Victory. According to some manuscripts it stood “outside,” so possibly was over the outer Golden Gate and was about to crown a victor coming from the west. If that statue, said to have been of bronze, was part of the same ensemble, it would have made the whole all the more impressive. We have tried to identify a triumphal monument of the middle Byzantine period, and I believe we have found one. Which particular occasion it commemorated remains, of course, a matter of speculation. One is tempted to think of the triumph of Nicephorus Phocas in 965, celebrated after his victories in Cilicia, for on that occasion he brought with him the bronze gates of conquered Tarsus and Mopsuestia, installing the former at the Acropolis walls, the latter at the Golden Gate.79 One can imagine the ensemble of ancient sculpture being created to set off the Arab spoils. Such an initiative, in any case, would have been meaningful while the Golden Gate was still used for triumphal entries, that is, before the Komnenian period. Exeter College, Oxford 75 “Ancient Spolia in the Great Palace of Constantinople,” Byzantine East, Latin West. Art-Historical Studies in Honor of Kurt Weitzmann (Princeton, 1995), 645–49. 76 Cf. LIMC, s.v. “Herakles,” 2643 (tomb of Nasonii on Via Flaminia). Gyllius’ statement that Hercules was leading “dogs” may be explained by Cerberus’ three heads, probably held by three leashes. 77 Cf. LIMC, s.v. “Pegasos,” C3, D (again, tomb of Nasonii). This is a very rare subject, not represented in stone sculpture. 78 With reference to Roe’s naked young man running by a horse, reaching to pull another off. The identification was suggested by Strzygowski, “Das Goldene Thor,” 33 ff. 79 Skylitzes, ed. I. Thurn (Berlin, 1973), 270; Zonaras, ed. Th. Bu ¨ ttner-Wobst (Bonn, 1897), 3:503.

Appendix The Situation of the Strategion

Current opinion places the Strategion at Sirkeci, that is, in the area of the railroad terminal80 rather than near the Sublime Porte, as had been held earlier. Is it possible to be more precise? We have seen that a Theodosian forum was installed in the Strategion and adorned with an Egyptian obelisk. The expression Strategium, in quo est forum Theodosiacum81 suggests that the forum was not coextensive with the Strategion, but was only a portion of it. The patriographers, for their part, speak of a “great” and a “small” Strategion, designations that have remained somewhat unclear. The obelisk, still standing in the tenth century, was in the great Strategion as were other monumental adornments, including a statue allegedly of Alexander and a tripod. The small Strategion was said to have had a statue of Leo I82 and an inscription in lead (?) letters.83 It can be provisionally suggested that we have here a civic forum flanked by a smaller marketplace, an arrangement that occurs elsewhere, for example, at Palmyra. According to the author of the Patria, the “monolith” at the Strategion was a fragment of the obelisk in the Hippodrome and had been brought from Athens [sic] by the patrician Proklos at the time of Theodosius II.84 At first sight this statement appears acceptable, or nearly so. The Hippodrome obelisk is lacking about two-fifths of its original height, perhaps as much as 12 meters. The fracture may have occurred at Constantinople, and it is quite conceivable that the lower section, after being tapered at the top, was erected at the Strategion by the same Proculus (urban prefect, 388–392), whose name appears in the Hippodrome inscription, along with that of Emperor Theodosius. The patrician Proklos is not mentioned elsewhere in the Patria and is unlikely to have occurred to the semi-educated author through confusion with some other Proklos (e.g., the Neoplatonist philosopher) who lived during the reign of Theodosius II. It is more probable that the author read the name in an inscription. If this scenario is true, the monolith of the Strategion would have been rather squat and thick and would yield no topographical indication. There exists, however, another piece of evidence that may deserve more credence. Gyllius 80 See my De´veloppement urbain, 19. My suggestion that it may have been on the site of the main post office (ibid., 70) is probably incorrect in the light of the evidence presented here. Cf. also Berger, Untersuchungen, 408. 81 Notitia, ed. Seeck, 233, Reg. V. 82 Patria, 184.9. 83 Parastaseis in Scriptores, ed. Preger, 34.8. The wording is very obscure: ejn tv' mikrv' Strathgi´v mo´ libdo" polu` " crhmati´zei, h‘ aujto` " oJ mo´ libdo" h‘ moli´bdou dia´ qesi" e“ggra␸o", etc. I take crhmati´zei (act.) to mean “is, exists.” Cf. Lampe s.v., 5. The meaning cannot be “large amounts of lead are exchanged” as rendered in A. Cameron and J. Herrin, eds., Constantinople in the Early Eighth Century (Leiden, 1984), 87. 84 Scriptores, ed. Preger, 183, no. 60.

188 TRIUMPHAL WAY OF CONSTANTINOPLE AND THE GOLDEN GATE reports that when he first came to Constantinople (1544) he saw an obelisk of Egyptian granite (ex lapide Thebaico) standing within the Seraglio enclosure (intra claustrum regium) on the side of the first hill facing north, next to the workshops of the Sultan’s glaziers, or glassmakers (iuxta domum vitreariarum officinarum regiarum). A little later he saw it moved out of the Seraglio and lying on the ground. It measured 35 feet in height and, if memory served him right, 6 feet to the side, so that its lower perimeter was 24 feet. It was bought by the Venetian nobleman Antonius Priolus (Priuli), who was intending to export it and set it up in the piazza of St. Stephen, presumably the present Campo Morosini or Campo S. Angelo,85 but it does not seem to have ever reached Venice. The measurements given by Gyllius are consistent with the shape of an obelisk and prove that the monument he saw could not have been the bottom part of the Hippodrome obelisk, which would have been considerably wider at the base.86 Nor could it have been the same as the small porphyry obelisk standing today in the garden of the Istanbul Archaeological Museum (width 1.12, preserved height 4.13 m). In the absence of any other recorded obelisk, it is reasonable to suggest that Gyllius saw the one from the Strategion, and since it was standing in 1544, it must have been in situ. Unfortunately, the location of the Seraglio glass workshops does not appear to be recorded, except that they faced the Golden Horn. In view of the smoke generated in the course of smelting, they must have been on the periphery of the palace enclosure. In general, the Ottoman capital imported its glass, mostly from Venice, but there is evidence of considerable local production during the construction of the Su ¨ leymaniye mosque (1550–57).87 Thereafter, the palace workshops probably declined, hence the lack of information about them. It would not be unreasonable to place them just within the Bostancı kapısı (Demirkapı), the northernmost gate of the Seraglio walls. I see no serious objection to locating the Strategion there, that is, some 300 meters east of Sirkeci station, almost directly on the shore of the ancient harbor. There may be a hint at the closeness between water and marketplace in the reference by Themistius (in 350) to “harbors within gates through which the sea flows in and is entwined with those in the midst of the agoras.” 88 De topographia CP, ii. 11. The sides of the extant obelisk range from 2.21 to 2.57 m, but the complete monument must have been more than 1 m wider at the base. 87 See the thorough study by M. Rogers, “Glass in Ottoman Turkey,” IstMitt 33 (1983): 239–67, esp. 242. ¨ O. Ku ¨¸cu ¨ kerman, “Camcılık,” I˙stanbul Ansiklopedisi 2 (Istanbul, 1994):373, mentions the existence in the 1550s of glass workshops “between the Topkapı palace and the Golden Horn,” but cites no reference. 88 Or. 4.60d. 85 86

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1 Istanbul, Cerrahpa¸sa Caddesi, looking east

2

Diagram of ceremonial stations recorded in the Book of Ceremonies

3

4

Istanbul, probable remains of Troadesian porticoes, 1982

Istanbul, outer Golden Gate, south curtain wall

5

Istanbul, outer Golden Gate, north curtain wall

6

Istanbul, outer Golden Gate, south curtain wall, upright frame supports

7

Istanbul, outer Golden Gate, south curtain wall, decorated cornice

8

The Castle of the Seven Towers by Francesco Scarella, ca. 1685 (courtesy of the Österreichische Nationalbibliothek)

9

Detail, Castle of the Seven Towers by Francesco Scarella

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Commercial Map of Constantinople MARLIA MUNDELL MANGO

n an attempt to establish a commercial map of Constantinople across time, three components will be considered: the architectural setting, the geographical setting, and the goods involved.1 The commercial topography of Constantinople was in part determined by the fact that it was a sea-bound city on seven hills, making access from the port to the forum and other commercial premises a key necessity in urban development. In this paper I hope to demonstrate that certain types of structures and certain locations proved convenient for particular types of commerce and that these arrangements endured irrespective of countervailing historical, social, or economic trends. For example, persuasive evidence suggests that the Strategion area, situated by the Prosphorion and Neorion harbors and dating from the time of the early Greek colony, remained an important hub of commercial activity throughout the history of the city, until modern times. Similarly, there is evidence that the porticoed street, the stoa, a traditional base for commerce identified by scholars more with the early than the medieval period, survived at Constantinople until the Fourth Crusade. This study will focus on the siting of “managed commerce,” that is, large-scale enterprise under state supervision, whether destined for state distribution (the annona in the early period) or private, money-based trade. Small-scale commerce (the “neighborhood shop”) will be of secondary interest and only briefly included where relevant. For the purpose of this paper, “commercial activity” will be defined as local production intended for commercial exchange. Limited space precludes comprehensive coverage of all types of commerce or of every topographic feature. The following is presented as an introductory study. Two texts will provide a framework for this investigation of the commercial topography of Constantinople: for the early period, the Notitia urbis Constantinopolitanae, a semiofficial document compiled by a retired civil servant in about 425,2 and, for the middle period, the Book of the Prefect, a legal text of 912.3 The two texts differ in content regarding

I

This paper forms part of a broader investigation of Byzantine commerce, originating in a 1993 exhibition, The Byzantine Bridge between East and West: Trade and Manufacture, AD 300–1453, held in the Ashmolean Museum, Oxford, at the time of the symposium “Constantinople and Its Hinterland”; a publication based on this exhibition, which featured a “Commercial Map of Constantinople,” is in preparation. I should like to thank I. Cartwright, A. A. Wilkins, and R. Wilkins of the Institute of Archaeology, Oxford, for help with the illustrations to this paper. 2 Notitia dignitatum, ed. O. Seeck (Berlin, 1876), 229–43. 3 Das Eparchenbuch Leons des Weisen, ed. J. Koder, CFHB 33 (Vienna, 1991). 1

190

COMMERCIAL MAP OF CONSTANTINOPLE

commercial premises. The Notitia lists numbers of buildings by type, whether in state or private use, while the Book of the Prefect deals only with premises occupied by guilds involved in private commerce regulated by the state. Both documents give geographical and architectural indications, neither document appears to be comprehensive. The geographical setting of managed commerce at Constantinople was linked to architectural requirements. Given the hilly terrain, the size of certain commercial structures helped determine their location. For example, broad, flat ground was needed for fora and horrea. Faced with the paucity of relevant physical evidence of commercial architecture at Constantinople itself, we must look to comparanda in other cities to gain an idea of scale and layout. Examination of the main architectural types found at these sites will provide the basis for much of the discussion of the early period even if, as in the case of stoai/emboloi, they have a direct role in managed commerce only in the medieval period. THE EARLY PERIOD, FIFTH TO SEVENTH CENTURIES The fifth to seventh centuries represent a formative period in the city, when Constantinople was gradually provided with both prestige and utilitarian buildings, and when prospective residents were attracted by the state-supplied food of the annona.4 Aside from bread, it is unclear what else was distributed; if the Abydos tariff can be taken as evidence, the annona included oil, wine, dried goods, and lard.5 It was half of a two-tier system of food distribution in Constantinople in the early period that also included private commerce. Food supply, both state and private, was facilitated by various structures used for its transport (harbors), storage (horrea), and processing and distribution (bakeries, gradus, macella). We know bread was available at the gradus, but how the state distributed other foods remains open to question. Many commercial structures were in place by the time the major topographical document for Constantinople in this period, the Notitia, was compiled ca. 425. The Notitia lists noteworthy buildings within each of the city’s fourteen regions. After citing by name such important regional monuments as Constantine’s column and the Hippodrome, the Notitia enumerates certain types of buildings in each region, including theaters, large and small baths, domus, and cisterns, as well as structures connected with distribution of food and other goods (harbors, fora, horrea, macella, bakeries, gradus). Numbers stated for each region and totals given at the conclusion are sometimes inconsistent. Furthermore, Notitia figures differ dramatically from those in corresponding documents from Rome, the Notitia regionum XIV of 354 and Curiosum urbis regionum XIV of 375.6 Also relevant to late antique commerce generally are the regulations listed in the treatise of Julian of Ascalon, which specify the space required in a portico for stoa shops and the distance between various types of workshops and other buildings, particularly dwellings.7 Emission of smoke, steam, and odors, as well as the threat of fire, were taken 4 C. Mango, Le de´veloppement urbain de Constantinople (IVe–VIIe sie`cles), TM, Monographies 2, 2nd ed. (Paris, 1990), 23–50. 5 J. Durliat and A. Guillou, “Le tarif d’Abydos,” BCH 107 (1984): 581–98. 6 G. Rickman, Roman Granaries and Store Buildings (Cambridge, 1971), appendix 6. 7 The text is preserved in the Geneva manuscript of the Book of the Prefect. C. Saliou, Le traite´ d’urbanisme de Julien d’Ascalon, VIe sie`cle, TM, Monographies 8 (Paris, 1996), 58–60.

MARLIA MUNDELL MANGO

191

into account. Makers of garum and cheese, for example, were considered especially offensive and were relegated to 3.5 stades (655 m) from cities or villages. Other industries had to be separated from houses by specified distances: limekilns up to 46.80 meters, potteries (operating mostly in summer) up to 14 meters, gypsum works by up to 11.23 meters, dyehouses 3.12 meters, bakeries (which heated their ovens at night) 2.80 meters, and forges and oil processors 1.56 meters. Because of the danger of fire, glassmakers (huelourgoi) were to be located outside the city or at some distance from built-up areas within it.8 Architectural comparanda abound of commercial structures, namely fora, horrea, macella, stoai, and other types of commercial structures are known from excavation. The same commercial establishments could, no doubt, have been found in most cities in late antiquity.9 In the city center, for example, at Corinth, we find an agora with shop stoai (beside a second agora area with bouleuterion in the south) and, to the north, a macellum and shops lining the porticoed street that led to the harbor at Lechaion (Fig. 1).10 All these types of commercial premises also once existed at Constantinople. To animate the excavated remains of commercial premises we turn to Libanius’ enthusiastic praise of the delights of urban shopping in his native Antioch. There, markets were open night and day, masonry shops and makeshift booths were everywhere, every quarter of the city was “full of things on sale.” 11 This commercial activity is reflected perhaps in two Antiochene illustrations, the topographical border in the Yakto villa at Daphne (Fig. 2)12 and in the Rossano Gospels (Fig. 3).13 Fora and Agorai Archaeological remains and texts reveal that Constantinople had several agorai or fora in this period (Fig. 4). The traditionally square Greek agora and the rectangular Roman forum served the same functions. One side of the Roman forum was often formed by a basilica, which could house a market or be used for judicial or other civic purposes.14 At Rome, the functions of the Forum Romanum had shifted by the end of the Republican period from commercial to ceremonial use, while commerce moved to the imperial fora and other locations.15 This trend for specialization is seen also in provincial cities with two or more fora, where each served a different purpose. At Palmyra, next to the agora (of ca. 71), which is decorated with two-hundred statues, stood what might well have been a separate commercial agora where marble slabs were engraved Saliou, Traite´, 32–42. D. Claude, Die byzantinische Stadt im 6. Jahrhundert (Munich, 1969), 167–94. 10 R. L. Scranton, Medieval Architecture in the Central Area of Corinth, Corinth 16 (Princeton, 1957), 1–2, 6–26. 11 Libanius, Oration 11 (The Antiochikos). See “Antioche de´crite par Libanius,” trans. A.-J. Festugie`re and comm. R. Martin in A.-J. Festugie`re, in Antioche paı¨enne et chre´tienne: Libanius, Chrysostome et les moines de Syrie (Paris, 1959), 33–35, 56–58. 12 G. Downey, A History of Antioch in Syria from Seleucus to the Arab Conquest (Princeton, 1966), 659–64. 13 Cleansing of the Temple. See W. C. Loerke, “The Rossano Gospels: The Miniatures,” in Codex purpureus rossanensis: Commentarium, ed. G. Cavallo, J. Gribomont, and W. C. Loerke (Rome-Graz, 1985–87), 126–29. 14 ´ cole franc¸aise de Rome 99 (Rome, 1987), 155–85. R. Martin, Architecture et urbanisme, Collection de l’E 15 N. Purcell, “Forum Romanum (the Imperial Period),” in Lexicon Topographicum urbis Romae, ed. M. Steinby (Rome, 1995), 2:336–42. 8 9

192

COMMERCIAL MAP OF CONSTANTINOPLE

with tariff lists (of 137).16 At Ephesus, the lower agora by the harbor (rebuilt in the late 4th or early 5th century) was probably used as a cattle and food market, while the upper agora by the prytaneum and bouleuterion was for civic functions (Fig. 5).17 In smaller cities, a single forum or agora continued to combine both functions. The Notitia of ca. 425 mentions four fora at Constantinople; others (e.g., those of Leo and Marcian) were built later. In addition to the Strategion by the Golden Horn, early fifth-century Constantinople had a line of inland fora: the pre-Justinianic Tetrastoon, the fora of Constantine, Theodosius, and Arcadius, and the Amastrianus (called a forum in the Book of the Prefect) between the latter two. The later fora of Leo and Marcian lay on other arteries. The Sigma of Theodosius II, if a forum,18 is the only one between the two walls, while the Bous on the Mese is of unknown form and function.19 These public spaces are specifically associated by written sources with commerce and/or ceremony, although their functions may have changed, in some cases, between the early and medieval periods. The city’s somewhat unique situation, noted earlier, of being nearly surrounded by water meant that much of its food and other supplies arrived by sea to one of its four harbors. Yet of its several fora, only the Strategion is situated near the coast, that is, by the two harbors on the Golden Horn (Fig. 4). The ancient Strategion was upgraded to a Theodosian forum complete with a Theban obelisk, but it may be surmised that the enigmatic Lesser Strategion, mentioned in the Parastaseis,20 formed a commercial appendage similar to the one described above at Palmyra. Horrea The Notitia mentions only six horrea for Constantinople, compared with threehundred in corresponding documents for Rome. Horrea, or warehouses (as with baths the plural horrea denotes a larger structure than the singular form), are best studied at Ostia and Rome.21 Horrea enclosed a parallel series of oblong storage chambers (Fig. 6). Ostia had more than eleven horrea but as one of Rome’s maritime ports its facilities exceeded those of the average city. At Rome itself, goods arriving in smaller river craft were unloaded and stored in vast warehouses situated in the Emporium region on the Tiber, known from the Severan marble plan (Fig. 7).22 These included, for example, the Horrea Galbana (167 by 146 m) for wine, oil, and other commodities,23 and the Horrea Lolliana. These and most published horrea are earlier in date than the period under consideration here, but two early Byzantine horrea can be cited from Mesopotamia: one built in the new I. Browning, Palmyra (Park Ridge, 1979), 149–59. C. Foss, Ephesus after Antiquity: A Late Antique, Byzantine and Turkish City (Cambridge, 1979), 63 (where it is pointed out that the lower agora was at least occasionally the setting for public meetings) and n. 34, 80–83. 18 Assuming that it is the Theodosian forum mentioned by Marcellinus (Chronicle, ed. T. Mommsen, MGH, AA XI, 79); see Mango, De´veloppement urbain, 50 and n. 81. For a different view see A. Berger, “Tauros e Sigma: Due piazze di Costantinopoli,” in Bisanzio e l’Occidente: Arte, archeologia, storia. Studi in onore di Fernanda de’Maffei, ed. M. Bonfioli, R. Farioli Companati, and A. Garzya (Rome, 1996), 24–28. 19 Mango, De´veloppement urbain, 27–29, 43–46. 20 Parastaseis in Scriptores originum Constantinopolitanarum, ed. T. Preger (Leipzig, 1901–7), 1:34; Patria, ibid., 2:184; Mango, De´veloppement urbain, 43. 21 Rickman, Granaries, 15–121. 22 Ibid., 108–21; C. Mocchegiani Carpano, “Emporium,” in Lexicon Topographicum, 2:221–23. 23 Rickman, Granaries, 97–104, 171; F. Coarelli, “Horrea Galbana,” in Lexicon Topographicum, 2:40. 16 17

MARLIA MUNDELL MANGO

193

city of Dara founded in 508 (Fig. 8) and a smaller structure, identified by its inscription as a horion, that was built in 542 at Constantina, then an important military base (Fig. 9).24 The horrea at Constantinople were situated in only two regions; there were four in Region V and two in Region IX.25 They were all undoubtedly sited on level ground, close to the four harbors on the Golden Horn and Marmara. As their name suggests, the horrea were probably large, like those at Rome (Fig. 7), particularly that known as the Horrea Alexandrina, which accommodated grain shipments from Alexandria.26 Furthermore, one horrea on the Golden Horn was used for oil, the horrea olearia.27 Another horrea in that area may have been used for wine (the Horrea Galbana at Rome stored both oil and wine, Fig. 7). Since wine, oil, and bread were distributed by the annona, the six large warehouses at Constantinople (three with imperial names: the Valentiaca, the Constantiaca, and the Theodosianum) were probably restricted to state use. Undoubtedly there were many more horrea, as at Rome, some possibly small and for private use. Macella Only four, possibly five, macella are listed in the Notitia, two each in Regions V and VIII (Fig. 4).28 Socrates mentions a macellum near the forum of Constantine,29 which may be one of those in Region V or a fifth building. Although related to the forum in basic plan, the macellum was smaller.30 It was composed of a walled area bordered by shops, some of which opened out onto the street (Fig. 10). Sometimes a tholos used as a fountain or as a kiosk stood in the center.31 The macellum was a food market, originally selling a variety of goods including vegetables, but mainly meat and fish,32 until late antiquity, when texts seem to suggest a restriction to meat alone.33 It should be noted, however, that a macellum at Ostia selling both meat and fish (Fig. 11) was restored in 418–420 and therefore, still in use.34 Macella are better documented in the West than in the East, and, although well attested in texts after 300, apparently none of the extant macella studied by de Ruyt was built in our period. However, one can assume that the basic plan known somewhat earlier was retained. Since all fish and most of the livestock sold in the macella at Constantinople would have arrived by sea, at least some of these markets might have been expected near the harbors. This may have been the case with the two macella in Region V, but the Notitia 24 Mango, De´veloppement urbain, 40 and n. 20 (Dara); G. Bell, The Churches and Monasteries of the Tur
Abdin, intro. and notes by M. Mundell Mango (London, 1982), 154 (Constantina/Virans¸ehir). 25 Notitia, ed. Seeck, 233–34, 237. 26 Mango, De´veloppement urbain, 40, where it is suggested that five of the six horrea were granaries. 27 Notitia, ed. Seeck, 233. 28 Ibid., 234, 236; the total number given for the city is five, not four; ibid., 243. 29 Socrates, Historia Ecclesiastica, 1.38.65–67. For the so-called Leomakellon, see most recently A. Berger, “Zur Topographie der Ufergegend am Goldenen Horn in der byzantinischen Zeit,” IstMitt 45 (1995): 152–55. 30 C. de Ruyt, Macellum: Marche´ alimentaire des romains, Publications d’histoire de l’art et d’arche´ologie de l’Universite´ catholique de Louvain 35 (Louvain-la-Neuve, 1983), 284–303, 332–39. 31 Ibid., 295–303. 32 Ibid., 118, 307, 342–45, figs. 44–45, 92. Libanius praises the fish supplied at Antioch to the rich from the sea, to the poor from the lake, and to both groups from the river. Festugie`re and Martin, Antioche, 35, 58. 33 De Ruyt, Macellum, 271–73. 34 The marine mosaic in the taberne dei pescivendoli has been attributed to the mid-3rd century; de Ruyt, Macellum, 124.

194

COMMERCIAL MAP OF CONSTANTINOPLE

says specifically that Region VIII did not extend to the sea.35 This region lay south of the Theodosian forum and west of the street leading to the harbor of Julian, where the ground declines relatively sharply seaward (Fig. 12). Thus, these two macella were probably terraced. Bakeries and Gradus The nineteen public bakeries ( pistrina publica) in Constantinople listed in the Notitia were largely concentrated in three regions: seven in Region V, and four each in Regions I and IX (Fig. 4).36 Regions I and IX also had four gradus each,37 from which the bread was distributed. In addition, there were two public bakeries in Region X and one each in Regions VI and XI.38 Private bakeries ( pistrina priuata) were concentrated particularly in Regions I, VI, VII, and IX.39 With the exception of Region III, there were gradus in every region, their numbers nearly matching those of bakeries in some regions (II, V, VI, and VIII), and being more numerous than bakeries in others (VII, XI, and XII), where the gradus may have served the Mese and the Portico of Domninus (Fig. 4). Stoai and Ergasteria While the foregoing discussion of horrea, macella, and bakeries/gradus was concerned with buildings of a specialized nature, stoai—like fora—were used by a variety of trades. Nevertheless, some commercial grouping (e.g., of silversmiths) within stoai and fora can be noted in the early period. Such grouping of trades became a regular feature in the medieval period (see below), and it is still common in the Islamic world. Once again, architectural comparanda will provide a basis for understanding stoai at Constantinople in both the early period and, perhaps more crucially, in the medieval period, when the eparch exercised control over activities conducted in them. Stoai, or porticoes, lined the streets of many Mediterranean cities and often incorporated commercial premises behind the colonnades.40 In late antiquity and earlier, such streets were identified by inscription as emboloi, for example, at Ephesus (Fig. 5), Sardis (Fig. 13), Gerasa, and Abila.41 Libanius provides us with a glowing description of the stoai of Antioch, when he says that “of all the amenities of a city . . . porticoes are among the most pleasant and the most useful.” Thanks to them, one is no longer a prisoner of the “ . . . nulla maris uicinitate contermina.” Notitia, ed. Seeck, 236. Notitia, ed. Seeck, 230, 234, 237. On the bread supply at Constantinople, see J. Durliat, “L’approvisionnement de Constantinople,” in Constantinople and Its Hinterland, ed. C. Mango and G. Dagron (Aldershot, 1995), 19–33; E. Kislinger, “Pane e demografia: l’approvvigionamento di Costantinopoli,” in Nel nome del pane: Homo edens IV (Trento, 1995), 279–93. 37 Notitia, ed. Seeck, 230, 237. 38 Ibid., 235, 238, 239. 39 Ibid., 230, 235, 236, 237. 40 G. Downey, “The Architectural Significance of the Use of the Words Stoa and Basilike in Classical Literature,” AJA (1937): 194–211; A. Segal, From Function to Monument: Urban Landscapes of Roman Palestine, Syria and Provincia Arabia, Oxbow Monograph 66 (Oxford, 1997), 5–53. 41 Foss, Ephesus, 65 and n. 39; C. Foss, Byzantine and Turkish Sardis, Archaeological Exploration of Sardis Monograph 4 (Cambridge, Mass., 1976), 44–45; C. B. Welles, “The Inscriptions,” in Gerasa, City of the Decapolis, ed. C. H. Kraeling (New Haven, 1938), 471, no. 280; for Abila see P. le Bas and W. H. Waddington, Inscriptions grecques et latines recueillies en Gre`ce et en Asie Mineure (Paris, 1870), 3: no. 1878. 35 36

MARLIA MUNDELL MANGO

195

weather.42 The main porticoes at Antioch incorporated entrances to both private and public buildings as well as shops; the longest street portico may have been supported by 1,400 columns.43 Texts confirm that lighting was provided for porticoes by governors or shopkeepers in Antioch, and in Edessa, Ephesus, Constantinople, and elsewhere.44 Shops known as ergasteria could house both manufacturing and retail operations, sometimes within a stoa. The best-studied late antique stoa shops are at Sardis, where a row of nearly thirty was excavated on the north side of a street running beside a gymnasium, which had been partly converted into a synagogue (Fig. 13).45 The colonnade was single-storied, but the shops had two storeys, usually both with commercial functions. Their depth was constant at about five meters but width varied. The interest of the shops is increased by their recovered contents, which help identify individual functions. These differed, although dyers and glass retailing predominated. Shops labeled E6 and E7 in Figure 13 contained ceramic tubes with dyes, mortars, and steelyards (Figs. 14, 15), while shops E12 and E13 held four-thousand fragments of glass vessels and window panes.46 Five other shops served as taverns.47 A similar mix was observed in the shops lining the alytarch’s stoa behind the Embolos at Ephesus (Fig. 5).48 A more elaborate layout of shops has been uncovered at Scythopolis, where in 507 a semicircular porticoed area (called sigma in two inscriptions) was built into the west side of Palladius Street, an earlier Byzantine colonnaded street lined with shops, which was north of public baths (Fig. 16). The new exedra (13 by 15 m) encompassed three apses and a line of twelve shops revetted in marble and decorated with tessellated pavements. One shop contained glass lamps.49 At Justiniana Prima excavated shops housed metalworkers, tanners, bakers, and potters.50 At some of the double-storied shops in Apamea, built in the second century, lists of goods for sale were painted on their facades, possibly centuries later (Fig. 17).51 From an earlier period, the shops lining the main street at Palmyra had the proprietor’s name and trade inscribed on the lintel over the door.52 Festugie`re and Martin, Antioche, 26–27, 48–49. Ibid., 23, 39–40; see also J. Lassus, Les portiques d’Antioche, Antioch-on-the-Orontes 5 (Princeton, 1972), esp. 125–51; on the number of columns (common, apparently, to both 2nd- and 6th-century phases of the main street), ibid., 146. 44 Festugie`re and Martin, Antioche, 37, 60 (Antioch); The Chronicle of Joshua the Stylite, ed. and trans. W. Wright (Cambridge, 1882), chaps. 29 and 87 (Edessa); Foss, Ephesus, 56–57 and n. 20 (Ephesus and Constantinople). 45 J. Stephens Crawford, The Byzantine Shops at Sardis, Archaeological Exploration of Sardis Monograph 9 (Cambridge, Mass., 1990), 3–11. 46 Crawford, Shops, 60–66, 78–86. 47 Shops E1–E2, W1–W3: ibid., 17–18, 33–49. 48 Foss, Ephesus, 74; on these and other late antique shops at Ephesus see also Crawford, Shops, 108–11; on the alytarch, see Foss, Ephesus, 19. 49 Y. Tsafrir and G. Foerster, “Urbanism at Scythopolis-Bet Shean in the Fourth to Seventh Centuries,” DOP 51 (1997): 121–22, 130, figs. 28–29, 42. The lamps are mentioned by R. Bar-Nathan and G. Mazor in Excavations and Surveys in Israel 11 (1992): 42–44. 50 I. Popovic´ in Caricin Grad, ed. B. Bavant, V. Kondic´, and J.-M. Spieser (Belgrade-Rome 1990), 2:269–306. 51 Crawford, Shops, 119; L. Reekmans, “Fresques des portiques de la grande colonnade,” in Apame´e de Syrie: Bilan des recherches arche´ologiques, 1965–1968, ed. J. Balty (Brussels, 1969), 118, 121. The texts remain unpublished; some are illustrated in J. C. Balty, Guide d’Apame´e (Brussels, 1981), figs. 81, 82. 52 Browning, Palmyra, 138. 42 43

196

COMMERCIAL MAP OF CONSTANTINOPLE

Trajan’s Market, which borders his forum at Rome, was a different type of commercial complex.53 Irregular in layout and built on several levels, it contrasts with the more regimented plans of other premises (stoai, horrea, macella, etc.). Elsewhere, excavations have uncovered other irregular commercial complexes from late antiquity that evolved casually and were not specifically designed as shops. These are typically ergasteria. Behind a portico of numbered shops recently excavated on the decumanus maximus of Berytus lies a courtyard with basins of undetermined function and a structure with an apsidal room (Fig. 18). These suggest a residential complex with industrial activities.54 The House of Bronzes at Sardis may have been comparable, combining as it did elegant architecture with industrial facilities for bleaching and a storage area (where a collection of bronze items was found). Another large oblong building on the Hypaepa Embolos at Sardis also housed some industrial activity (Fig. 13).55 Excavations at Alexandria uncovered fifth- to seventh-century complexes of artisans’ houses grouped around an interior court where objects of glass, rock crystal, bone, and possibly ivory were produced; in House D, these objects could have been sold in shops that fronted the street.56 Clearly, shops in Constantinople were numerous. Laws of Anastasius and Justinian regulated the revenue of 1,100 shops assigned to the cathedral alone.57 Unfortunately three shops excavated on the Mese behind the so-called House of Lausus were insufficiently published to give an idea of their respective sizes, the nature of any contents discovered, or period of construction or use (Fig. 19).58 These shops may well have been rebuilt following fires in 498 and 532, the latter occurring during the Nika Riot. From the time of Constantine, porticoes had lined the Mese, extending from the Chalke to the Capitolium.59 The Scythopolis exedra (Fig. 16) may furnish a partial model for the plan of the two-storey porticoes of the circular Forum of Constantine.60 Based on the approximate (and relatively large) dimensions of the excavated Mese shops as drawn by Rudolf Naumann (10 m wide by 7 m deep),61 there may have been up to one-hundred shops, fifty to a side, on the Mese between the Milion and the Forum of Constantine. Extending this number to comparable stretches of portico up to the Capitolium yields an estimate of approximately five-hundred shops. Taking the derived module of fifty shops to a side as a minimum for each of the fifty-two porticus listed in the Notitia62 produces a total of 2,600 stoa shops. Such a calculation is, clearly, highly speculative. The lengths of the L. Ungaro, “Mercati di Traiano,” in Lexicon Topographicum, 3:241–45; Martin, Architecture et urbanisme, 173–75. 54 K. Butcher and R. Thorpe, “A Note on Excavations in Central Beirut, 1994–96,” Journal of Roman Archaeology 10 (1997): 297–99. 55 Foss, Sardis, 43–44. 56 M. Rodziewicz, Les habitations romaines tardives d’Alexandrie `a la lumie`re des fouilles polonaises `a Koˆm el-Dikka: Alexandrie (Warsaw, 1984), 3:330–35. 57 Mango, De´veloppement urbain, 48. 58 R. Naumann, “Vorbericht u ¨ ber die Ausgrabungen zwischen Mese und Antiochus-Palast 1964 in Istanbul,” IstMitt 15 (1965): 145–46. 59 Both fires consumed the Mese portico up to the Forum of Constantine; Chronicon Paschale, ed. L. Dindorf, CSHB (Bonn, 1832), 608, 623; see also Chronicon Paschale, 284–628 AD, trans. M. Whitby and M. Whitby (Liverpool, 1989), 99, 120. 60 Mango, De´veloppement urbain, 25–26. 61 Naumann, “Ausgrabungen,” fig. 5. 62 Notitia, ed. Seeck, 230–43. See also M. Mundell Mango, “The Porticoed Streets of Constantinople,” in ˘lu (Leiden, in press). Byzantine Constantinople: Monuments, Topography and Everyday Life, ed. N. Necipog 53

MARLIA MUNDELL MANGO

197

various porticus are not known, not all stoai incorporated only shops, and shop sizes varied even within a single stoa, as demonstrated by those excavated at Sardis, Ephesus, Scythopolis, and elsewhere. Still, Constantinople must have had many thousands of shops, given that the cathedral owned 1,100 and that there was certainly space in the city for many more. Evliya C ¸ elebi gives the number of Ottoman workshops as over 23,000; these later shops were relatively small (mostly 3 m square).63 As mentioned above, concentrations of certain trades outside the horrea/macella/bakeries complexes and within stoai (or fora) can be pinpointed in late antique Constantinople, thanks to various written sources. Figure 20 maps the location of a few of these trades.64 Books were copied and sold in the Law Basilica (composed of four stoai) by the Augustaion.65 To the north of this basilica, coppersmiths are known to have been installed in the area where the church of the Virgin was built in the fifth century.66 Whether they occupied ergasteria in the porticoed streets that ran between the Mese and the Strategion (remains of what might have been a stoa have been excavated on the diagonal modern ˘ı, Fig. 21),67 or whether they filled a forum-type space between the C ¸ atalc¸es¸me Sokag streets is unclear. To the west of this area, silversmiths are known to have operated on the Mese between the Milion and the Forum of Constantine, and in the same general region was located the “basilica of the furriers.” 68 In the seventh century, the Miracles of St. Artemius69 document an alternative to the concentration of specific trades in specialized areas, that is, varied commerce conducted within a single-porticoed street, the Portico of Domninus, the straight north-south artery that could be considered the cardo to the Mese’s decumanus. It linked the Golden Horn and the Marmara, crossing the Mese at the copper-reveted tetrapylon that stood between the fora of Constantine and Theodosius.70 In this portico was found, not the kind of artisanal concentration noted above for copper- and silversmiths, but a mix, typical also of stoai at Sardis and Ephesus,71 of moneychangers, candle makers, and a blacksmith, together with churches, a xenon, and the public baths of Dagistheos. Trends in the Early Period Looking at localized commercial facilities within geographic units of late antique Constantinople, certain general observations can be made (Fig. 4). Food-linked installaR. Mantran, Istanbul dans la seconde moitie´ du XVIIe sie`cle, Bibliothe`que arche´ologique et historique de l’Institut franc¸ais d’arche´ologie d’Istanbul 12 (Paris, 1962), 138, 144. The number of Byzantine shops estimated here would increase if one allowed for smaller dimensions per shop and for an upper storey of shops, as for example at Sardis. 64 See also R. Janin, Constantinople byzantine: De´veloppement urbain et re´pertoire topographique, 2nd ed. (Paris, 1964), 94–99. 65 Vie de Se´ve`re par Zacharie le Scholastique, ed. and trans. M.-A. Kugener, PO 2 (1907), 8. 66 The earliest attestation of the quarter may be that in the Epitome of the Ecclesiastical History of Theodore Lector; see Kirchengeschichte, ed. G. C. Hansen (Berlin, 1971), 102; also Janin, Constantinople, 96, 328. 67 T. Macridy in M. Schede, “Archa¨ologische Funde: Tu ¨ rkei,” AA 44 (1929): 357–58; see now J. Bardill, “The Palace of Lausus and Nearby Monuments in Constantinople,” AJA 101 (1997): 80–83. 68 Chron. Pasch., ed. Dindorf, 623; trans. Whitby and Whitby, 120. 69 In Varia graeca sacra, ed. A. Papadopoulos-Kerameus (St. Petersburg, 1909); The Miracles of St. Artemius: A Collection of Miracle Stories by an Anonymous Author of Seventh-Century Byzantium, trans. V. S. Crisafulli and J. W. Nesbitt, The Medieval Mediterranean 13 (Leiden, 1997). 70 Mango, De´veloppement urbain, 30–32. 71 Miracles of St. Artemius, chaps. 18, 21, 26, 29, 36. 63

198

COMMERCIAL MAP OF CONSTANTINOPLE

tions listed in the Notitia are concentrated in Regions V, VIII, and IX,72 which are comparable to Rome’s Emporium (Fig. 7).73 They are clustered in only one forum, or possibly two: the Strategion in Region V in the northeast and perhaps the Theodosian forum in Region VIII further south. In addition to the Strategion, Region V contained several commercial installations, including the Prosphorion harbor, the scala of Chalcedon, two macella, and four horrea; the Neorion harbor lay just to the west.74 The Strategion itself became what may have been a ceremonial forum (the Theodosian forum) with separate commercial appendage (the Lesser Strategion?). In Region VIII, the main Theodosian forum (which was actually mostly in Region VII) on the Mese may have been close to the macella to the south and was indirectly linked to the Harbor of Julian by the southern extension of the Portico of Domninus.75 Both the Strategion and the Theodosian forum will be seen to occupy central positions on the commercial map of medieval Constantinople. Likewise, the concentration of certain trades within the stoai of a particular area— as was the case of late antique book publishing and silver- and coppersmiths—will emerge as a theme in the Book of the Prefect. THE MEDIEVAL PERIOD, TENTH TO TWELFTH CENTURIES A certain continuity in the types of commercial buildings and their locations within the city can be noted between the early period and the time when the Book of the Prefect was compiled. The intervening economic recession, the “Dark Age” which occurred at Constantinople primarily in the eighth century, and other factors, such as the change in the grain supply and the silting up of the Theodosian harbor, undoubtedly had an impact on the city’s commerce.76 After the early seventh century, the annona ceased, except for military distributions. Nevertheless, by the early tenth century, economic recovery was sufficient to leave few specific traces of the previous depression in the Book of the Prefect, which portrays a level of commercial activity that is buoyant if regulated. Furthermore, commercial establishments documented in the early period can be traced to the medieval period, and at least one architectural type—the stoa—may have still been in use at the end of the twelfth century. As stated above, the Book of the Prefect, a legal text written in 912, provides the fundamental framework for a consideration of commerce in the medieval city. Its regulations cover twenty-two types of activity by guilds, some of which are assigned to designated parts of the city.77 The book mentions fora and stoai by name, and continued use of horrea and macella, while not directly referred to by these terms, can still be inferred or posited. Supplementary information dating from the eighth to tenth centuries is provided mainly Notitia, ed. Seeck, 233–34, 236, 237. See note 22 above. 74 Notitia, ed. Seeck, 233–34. 75 Ibid., 235, 236. 76 ´ tudes sur l’e´volution des strucMango, De´veloppement urbain, 51–62; P. Magdalino, Constantinople me´die´vale: E tures urbaines, TM, Monographies 9 (Paris, 1996), 17–25; idem, “The Grain Supply of Constantinople, Ninth– Twelfth Centuries,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 36), 35–47. 77 Eparchenbuch, ed. Koder. A re´sume´ of topographical aspects of the book (T. Thomov and A. Ilieva, “The Shape of the Market: Mapping the Book of the Eparch,” BMGS 22 [1998]: 105–16) appeared after the present article was submitted for publication and so has not been the subject of detailed commentary here. 72 73

MARLIA MUNDELL MANGO

199

by the Parastaseis78 and Patria.79 Because of the Lycus River that emptied into it, the Theodosian harbor had silted up by this time, when it is also recorded that commercial traffic was transferred from the Neorion harbor to the Julian harbor. Theophanes notes that the Neorion was dredged in 698 and in use by the navy by 715.80 The Prosphorion harbor, however, may have continued in commercial use, as the following discussion of meat markets suggests (Fig. 22). Meat and Fish Markets Among food items dealt with by the Book of the Prefect, livestock and other animals were to be sold in the fora of the Strategion, Theodosius, and Amastrianus,81 spices only in the Portico of Achilles by the Augustaion, and fish in the largest kamarai of the city (Fig. 22).82 By contrast, grocers, taverns selling wine, and (possibly) bakers were located throughout the city,83 on the model praised by Libanius.84 Specifically, non-pork butchers were instructed to travel beyond Nicomedia, as far as the Sangarius River, to obtain better prices directly from the sheep and goat farmers,85 probably arrived there from Paphlagonia and Phrygia. Purchased animals would then be put on ships at Nicomedia or Pylae (Fig. 23). Leo of Synnada saw Pylae thronging with pigs and sheep, as well as asses, oxen, and horses, all bound for Constantinople.86 Sheep probably arrived in the Prosphorion harbor by the Strategion, where they were to be sold, while pigs may have been unloaded at Julian’s harbor and led up to the Theodosian forum, where lambs were also sold at Easter.87 The Patria makes mention of pigs (to be bought at the Theodosian forum?) being herded via the “Artotyrianos” near the Bronze Tetrapylon, whence, as Raymond Janin puts it, “on descendait vers le Kontoscalion” (i.e., south to the Harbor of Julian/ Sophia).88 Animals sold at the Amastrianus, if found unsatisfactory, could be returned to the original owner after six months,89 and so were probably sold as draft animals rather than for slaughter. Some may have been herded into the city from areas outside the walls, to the west. Although this is not stated specifically, it is possible that the live animals led into the city by sheep sellers ( probatemporoi) for sale to butchers (makellarioi) in the Strategion90 Ed. Preger; see note 20 above and also Constantinople in the Early Eighth Century: The Parastaseis Syntomoi Chronikai, ed. A. Cameron and J. Herrin (Leiden, 1984). 79 Ed. Preger. 80 Parastaseis, ed. Preger, 1:67; Patria, ed. Preger, 2:188. Mango, De´veloppement urbain, 55–56; see also Magdalino, Constantinople me´die´vale, 19–20. 81 Eparchenbuch, ed. Koder, 15.1, 15.5, 16.2–3, 21.3. 82 Ibid., 10.1 (myrepsoi); 17.1 (fish). The designated area for the myrepsoi, between the Milion and the Chalke (with its “revered icon of Christ”) was known as the Portico of Achilles; C. Mango, The Brazen House: A Study of the Vestibule of the Imperial Palace of Constantinople (Copenhagen, 1959), 82–84. 83 Eparchenbuch, ed. Koder, 13.1, 18.3 (implied, but see note 104 below), 19.2 (implied). 84 Festugie`re, Antioche, 33. 85 Eparchenbuch, ed. Koder, 15.3. 86 The Correspondence of Leo Metropolitan of Synada and Syncellus, ed. and trans. M. P. Vinson, DOT 8 (Washington, D.C., 1985), letter 54.28. 87 Eparchenbuch, ed. Koder, 15.5, 16.2–3. 88 Patria, ed. Preger, 2:175. Janin, Constantinople, 315–16; on the Chalkoun Tetrapylon, see Mango, De´veloppement urbain, 30–31. 89 Eparchenbuch, ed. Koder, 21.3 and 21.6. 90 Ibid., 15.1–2. 78

200

COMMERCIAL MAP OF CONSTANTINOPLE

were slaughtered within the confines of adjacent buildings, not in the open air. The same could be suggested for the activities of the pork sellers confined to the Theodosian forum,91 although no particular building is specified for the slaughter or sale of meat. At least some of the four or five macella listed in the fifth-century Notitia may have continued to function in the tenth century, particularly since they were probably situated near both the Strategion and the Theodosian forum (Fig. 4), which became the centers of medieval meat-market activity. The sale of fish is said to have been conducted within the largest kamarai of the city.92 Kamara can signify an arch or vaulted covering and could conceivably refer to a vaulted market or macellum (Figs. 10, 11). As noted, in antiquity macella sold both meat and fish (as at Ostia, Fig. 11), and, although texts seem to suggest a narrowing of sales to meat alone by late antiquity,93 it is possible that both continued to be marketed together. The Book of the Prefect specifies that fishmongers (ichthyopratai) acquire their fish from sea beaches or from boats anchored at skalai,94 such as those of Sykae and Chalcedon on the Golden Horn. Until recently, the fish market (balıkpazarı) stood near the former Neorion harbor.95 Bread, Oil, and Wine Markets The information relevant to horrea, bakeries, and bread shops supplied by the Book of the Prefect is less than specific. One passage concerning bakers (18.1) has suggested the existence of a central warehouse for grain. Jules Nicole translates “Qu’ils ache`tent le ble´ dans les magasins de l’assesseur” and E. H. Freshfield likewise, “They are to buy corn in the warehouse of the assessor.” 96 The word warehouse, however, does not appear in the Greek. Indeed, it is difficult to understand what the Greek means: siton exo¯noumenoi ton harmozonta en to¯ sympono¯ touton [touto¯n codd.] ale¯santes. . . . Johannes Koder has a comma after harmozonta and renders it “dieses vor dem (zusta¨ndigen) Assessor (des Eparchen) mahlen,” 97 but en to¯ sympono¯ can no more mean “before the assessor” than it can “in the warehouse of the assessor.” Clearly, the bakers must have bought the grain somewhere, but the text appears corrupt as it stands. In any case, if the text in its original form referred to a warehouse, can it be identified with a horrea named in the Notitia? The seventh-century Miracles of Artemius mentions a grain horrea named Lamia near “the Harbor of Kaisarios”; 98 the Patria refers to the Lamia again in the tenth century,99 while in 899 Philotheus lists the title comes tes Lamias.100 Scholarly opinion identifies the Harbor of Kaisarios with the Theodosian harbor and the Lamia with one of the horrea listed in the Ibid., 16.2. Ibid., 17.1. On fish at Constantinople, see G. Dagron, “Poissons, peˆcheurs, poissonniers de Constantinople,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 36), 57–73, esp. 69–70. 93 De Ruyt, Macellum, 271–73. 94 Eparchenbuch, ed. Koder, 17.3. 95 Notitia, ed. Seeck, 234 (Regions V–VI); see also G. Majeska, Russian Travelers to Constantinople in the Fourteenth and Fifteenth Centuries, DOS 19 (Washington, D.C., 1984), 354. 96 J. Nicole, Le Livre du Pre´fet (Geneva-Basel, 1894), 69; E. H. Freshfield, Ordinances of Leo VI c. 895 from The Book of the Eparch (Cambridge, 1938), 42. The following observations on the phrase “en to ¯ sympono ¯” I owe to Cyril Mango. 97 Eparchenbuch, ed. Koder, 18.1. 98 Miracles of St. Artemius, chap. 16. 99 Patria, ed. Preger, 2:179, 246 (chap. 85). 100 N. Oikonomides, Les listes de pre´se´ance byzantines des IXe et Xe sie`cles (Paris, 1972), 113. 91 92

MARLIA MUNDELL MANGO

201

Notitia, possibly the Alexandrina in Region IX (Fig. 4).101 Attempts have also been made to reconcile the conflicting evidence in the Parastaseis and Patria concerning the location and function of the Modion.102 Bakeries, all now apparently private and combining the services of both pistrinum and gradus, may have been situated throughout the city, as the Notitia indicates they had been in the early period (see above). The Book of the Prefect specifies only that bakers grind their own grain, bake bread, and not locate their ovens under dwellings.103 Although the book does not state where bakers were located, a number of them were probably concentrated in the Artopoleia, or “bakeries quarter,” near the Bronze Tetrapylon on the Mese.104 Oil and wine are only briefly mentioned in the Book of the Prefect. Olive oil was sold by grocers (saldamorioi), who were located throughout the city.105 No specific regulations pertain to processing or importation. John Hayes’s study of the ceramic finds from Sara¸chane demonstrates continuous importation of amphorae that may have carried oil or wine (Fig. 24) into the city after late antiquity.106 No mention is made in the Book of the Prefect, or elsewhere in this period, of the oil horrea, the olearia, that were near the Strategion in the early period (Fig. 4).107 Wine was sold by innkeepers (kapeloi) in fixed measures of thirty and three litrai,108 not just by the glass. Like wax and soap chandlers, grocers sold goods weighed by the steelyard (Fig. 25). The Book of the Prefect lists these steelyard-weighed goods as meat (but distinct from the meat sold by butchers), salted (not fresh) fish, meal, cheese, honey, vegetables, and butter, in addition to olive oil. Besides food, grocers also sold resin, cedar and linseed oils, camphor, gypsum, pottery, and nails.109 They were not permitted to sell items weighed by balance scale (Fig. 26), which were only to be sold by myrepsoi.110 These latter goods included spices and ingredients for perfume, medicines, and dyeing.111 Apothecaries per se are not mentioned by the Book of the Prefect. Grocers, like inns and possibly bakers, were scattered throughout the city, while myrepsoi operated only between the Milion and Chalke (Fig. 23).112 Markets for Non-food Goods Continuity from late antiquity may again be noted in the location of markets for nonfood goods. The sale and/or manufacture of these goods, including textiles, precious metals, saddles, and soap and candles (made by lorotomoi, saponopratai, and keroularioi, respectively) was also regulated by the Book of the Prefect. The locations of saddle and Mango, De´veloppement urbain, 54–55 and nn. 21–24. Magdalino, Constantinople me´die´vale, 21–25. On the Modion see also Cameron and Herrin, Constantinople, 186–89. 103 Eparchenbuch, ed. Koder, 18.1 and 18.3. 104 Janin, Constantinople, 315; Mango, De´veloppement urbain, 55; Magdalino, Constantinople me´dievale, 21–25. 105 Eparchenbuch, ed. Koder, 13.1. 106 J. W. Hayes, Excavations at Sarac¸hane in Istanbul, vol. 2, The Pottery (Princeton, 1992), 61, 71–77, figs. 23–26; types 35–45 (8th century), types 46–54 (10th–11th centuries), types 55–59 (11th century), types 60–67 (12th–13th centuries). 107 Notitia, ed. Seeck, 233. 108 Eparchenbuch, ed. Koder, 19.1. 109 Ibid., 11.9, 12.9, 13.1. 110 Ibid., 10.5, 13.1. 111 Ibid., 10.1. 112 See notes 82–83 above. 101 102

202

COMMERCIAL MAP OF CONSTANTINOPLE

soapmakers are not specified, but the Book of the Prefect does indicate, however, that some candlemakers were attached to Hagia Sophia. The premises of candle- and soapmakers had to be separated from other buildings for safety.113 By contrast, some traders in textiles and precious metals are explicitly grouped in specific areas. Already noted was the concentration of silversmiths in the Mese during the early period; now they are assigned there. Precious metals and garments imported from Syria were to be sold tes Meses or en heni topo¯ tou embolou,114 while raw silk merchants (metaxopratai) were to sell en to phoro, probably the Forum of Constantine (Fig. 27).115 Other merchants involved in the silk trade (vestiopratai, katartarioi, serikarioi) had no designated locations.116 Linen merchants were instructed not to display their goods on counters (abbakia) or in their workshops but to carry them on their shoulders on market days.117 A document preserved on Patmos records the sale in 957 and 959 (that is, 45 years after the publication of the Book of the Prefect in 912) of five shops used, respectively, by linen, silk, goatskin, and Syrian cloth merchants, all perhaps in the same area, which for the linen and Syrian cloth merchants was designated as “the forum [of Constantine].” 118 If in compliance with Book of the Prefect regulations, the Syrian cloth shop ( prandioprates) may have been at the corner of the Mese and forum (Fig. 27). Perhaps this de luxe shopˆme (with its central honorific colping district could be compared with the Place Vendo umn) and the rue de Rivoli in Paris today, where Chanel, Bulgari, and other exclusive shops are located. The locations of glass- and copper-working establishments, not covered in the Book of the Prefect, are indicated by another middle Byzantine text, the Miracles of St. Photeine. There is described a fire that started in a glass smelting workshop (ergasterion huelopsestikon) on the broad uphill street that led from the Strategion to Hagia Sophia and the Milion and that spread to the area of copper workshops around the Chalkoprateia church (Fig. 27).119 While the location of copperworkers in this area is already documented in the early period (see above), glass factories are not otherwise mentioned in the city. The manufacture of glass in medieval Byzantium is a controversial subject among scholars, some of whom suggest glass was imported rather than manufactured within the empire, despite the contrary evidence provided by glass factories at Corinth (Fig. 28)120 and the glass cullet from an eleventh-century ship wrecked at Serc¸e Limanı, which is now considered to have been destined for a Byzantine factory.121 The St. Photeine text conEparchenbuch, ed. Koder, 11.2, 12.3. Ibid., 2.11, 5.2. 115 Ibid., 6.1 and 6.13. 116 Ibid., chaps. 4, 7, 8. 117 Ibid., 9.7. 118 N. Oikonomides, “Quelques boutiques `a Constantinople au Xe s.: Prix, loyers, imposition (Cod. Patmiaeus 171),” DOP 26 (1972): 345–56. 119 Invention of Relics of St. Photeine, ed. F. Halkin, CCSG 21 (1989), chap. 9, 122–24; see also A. Kazhdan and A.-M. Talbot, “The Byzantine Cult of St. Photeine,” in Presence of Byzantium. Studies Presented to Milton V. Anastos in Honor of His Eighty-fifth Birthday, ed. A. Dyck and S. Taracs ⫽ ByzF 20 (1994): 103–12; A.-M. Talbot, “The Posthumous Miracles of St. Photeine,” AB 112 (1994): 85–105; and note 122 below. 120 G. R. Davidson, The Minor Objects, Corinth 12 (Princeton, 1952), 80–122. 121 F. H. von Doorninck, “The Serc¸e Limani Shipwreck: An 11th Century Cargo of Fatimid Glassware Cullet for Byzantine Glassmakers,” in First International Anatolian Glass Symposium, April 26th–27th, 1988 (Istanbul, 1990), 58–63. For general bibliography and summary of this shipwreck, see A. J. Parker, Ancient Shipwrecks of the Mediterranean (Oxford, 1992), 398–99. 113 114

MARLIA MUNDELL MANGO

203

firms glass manufacture within the capital city,122 and the fire it describes proves the wisdom of Julian of Ascalon’s regulation prohibiting glassmakers (huelourgoi) in built-up urban areas.123 Stoai The specific references in the Book of the Prefect and other medieval texts to stoai (that is, “the Embolos” and “under the portico”)124 raise the question of whether the colonnaded street survived as an architectural form at Constantinople into the tenth century or even later. In medieval Corinth, the shops, and apparently their portico, on the west side of the Lechaion road north of the agora were restored and remained in use, while some new shops built within the space of the agora also had a type of porticoed front or porch (Fig. 28).125 Elsewhere, porticoed streets are known to have been built in new or renovated cities of the sixth century, including Dara, Zenobia, and Justiniana Prima.126 They were certainly also restored and maintained in many other places, as attested by excavation and inscriptions (see above). New porticoes with shops were built into the eighth century at Scythopolis (on Silvanus Street in the 720s), Anjar (ca. 715), and Palmyra (45 shops, some illustrated in Fig. 29).127 After the tenth century, emboloi are mentioned at Constantinople by the Patria in ca. 1000 and by Anthony of Novgorod in 1200, as well as in intervening Latin contexts.128 They also figure in accounts of the fire that broke out on 19 August 1203 during the Latin siege of the city, which might, indeed, have brought about their final destruction.129 In words reminiscent of the earlier St. Photeine account of the glass-factory fire, Villehardouin described the path of the fire, which raged eight days, as spreading above “the port” (Golden Horn), penetrating into the city in the direction of Hagia Sophia, and moving on to the Marmara. He watched “les grandes rues marchandes bruler.” 130 That the fire spread along the stoai is suggested by the fact that Niketas Choniates called it “a flowing river (hos eis holkon hena pyroentos synepieto potamou).” He stated that it destroyed the porticoes of Domninos and the covered porticoes 122 See J. Henderson and M. Mundell Mango, “Glass at Medieval Constantinople: Preliminary Scientific Evidence,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 36), 343–46. 123 See above, p. 191, and Saliou, Traite´, 40. 124 Eparchenbuch, ed. Koder, 5.2; for Patmos document, no. 2, see Oikonomides, “Boutiques,” 345–46. 125 Scranton, Medieval Architecture, 37, 48–49, 58–60, 67–68, and 77–78, where it states that the columns on the east side of the Lechaion Road were probably removed at some time between the 10th and the end of the 12th century. 126 Emboloi at Dara: Ioannis Malalae Chronographia, ed. and trans. L. Dindorf, CSHB (1831), 399; stoai (porticoed streets) at Zenobia and Justiniana Prima: Procopius, Buildings, 2.8.25, 4.1.23; these latter two have been excavated: see J. Lauffray, Halabiya-Zenobia: Place forte du limes oriental de la haute Mesopotamie au VIe sie`cle (Paris, 1983), 17, 129–31, and Caricin Grad, 2:56–60, 2:107–10, 2:116–19. 127 Tsafrir and Foerster, “Scythopolis-Bet Shean,” 123–24, 138–40, 144, figs. 55–59; for Anjar’s porticoed streets of ca. 715, see K. A. C. Creswell, Early Muslim Architecture: Umayyads, A.D. 622–750, 2nd ed. (Oxford, 1969), 1.2.478–81; for Palmyra’s 45 shops begun in the Abbasid period, see K. al-As
ad and F. M. Stepniowski, “The Umayyad Suq in Palmyra,” Damaszener Mitteilungen (Deutsches Archa¨ologisches Institut) 4 (1989): 205–23. 128 Patria, ed. Preger, 2:138, 148, 174, 200, 280, 283; Kniga Palomnik, ed. C. M. Loparev, PSb 51 (1899): 23, 29, 30; Magdalino, Constantinople me´die´vale, 78–85. 129 On which see T. F. Madden, “The Fires of the Fourth Crusade in Constantinople, 1203–1204: A Damage Assessment,” BZ 84–85 (1991–92): 72–93. 130 Villehardouin, La Conqueˆte de Constantinople, vol. 1, 1199–1203, ed. and trans. E. Faral (Paris, 1938), 161, 209.

204

COMMERCIAL MAP OF CONSTANTINOPLE

extending from the Milion to the forum of Constantine, and continuing on to the Philadelphion, took in everything in between, down to the Harbor of Sophia (Julian) and the former harbor of Theodosius (Figs. 22, 27).131 This fire may have been the end of the porticoes, although even today one finds in Istanbul, a few descendants, for example, an eighteenth-century portico near the Bazaar. Xenodocheia The Miracles of St. Artemius refer repeatedly to people connected with trade who were on business in Constantinople from Alexandria and elsewhere.132 Foreign merchant colonies were also an important feature of medieval Constantinople. Syrians, who in the early period had been Byzantines themselves, were now foreigners.133 The Russian Varangians bringing furs, slaves, and wax, who are mentioned already at the beginning of the tenth century, were confined to the suburb of St. Mamas up the Bosporos.134 By the mid-tenth century, the first of the Italian colonies had appeared—the Amalfitans, followed by the Venetians, the Pisans, and the Genoese—and the Germans. These groups were restricted to the Golden Horn,135 while Jews were assigned to Pera, perhaps from 1044.136 The Book of the Prefect required foreign merchants to stay in inns, or metata. Those from Syria could remain for three months; their imports were to be deposited in nearby warehouses for eventual sale on the Mese.137 Also in the tenth century, Romanus I established a xenodocheion in the Portico of Domninos (also called Embolos tou Maurianou) for those staying at Constantinople on business, pursuing lawsuits, and so on. It provided them with accommodation as well as stabling, food, and clothing.138 This is the very area where Ottoman hans were to be concentrated (see below). Trends in the Middle Period Two key commercial areas in the middle period can be identified, thanks to the Book of the Prefect. The first was an outer band of animal/food markets situated at the Strategion to the north and in the Theodosian and Amastriana fora to the west. These were mostly served by the Prosphorion and Julian/Sophia harbors, respectively. The second key area was an inner band of luxury markets in the Forum of Constantine and in stoai extending west along the Mese and the Portico of Achilles, which led to the Chalke of the palace. The two-tier circulation of food that had been in force in the early period had ceased with the disappearance of the annona in the early seventh century. Thereafter food sales were private transactions regulated by the state. Evidence suggests that at least one early Nicetae Choniatae Historia, ed. J. L. van Dieten (Berlin–New York, 1975), 554–55. Libanius had referred to Antioch’s porticoed streets as themselves being like rivers; Festugie`re, Antioche, 24. 132 Miracles of St. Artemius, chaps. 32, 35. 133 Eparchenbuch, ed. Koder, chap. 5. 134 J. Shepard, “Constantinople—Gateway to the North: The Russians,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 36), 243–60, esp. 251–53. 135 Magdalino, Constantinople me´die´vale, 78–90, with references. 136 D. Jacoby, “The Jews of Constantinople and Their Demographic Hinterland,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 36), 221–32, esp. 225. 137 Eparchenbuch, ed. Koder, 5.2 and 5.5. 138 Theophanes Continuatus, ed. I. Bekker, CSHB (Bonn, 1838), 430; Janin, Constantinople, 91, 386–87. 131

MARLIA MUNDELL MANGO

205

grain horrea remained in use, and, as argued above, the early macella may have served the animal markets at the Strategion and the Theodosian forum. The distribution of bakeries is unclear; stoai, however, apparently continued to house commerce and may have survived until the end of the twelfth century. THE LATE PERIOD, THIRTEENTH TO FIFTEENTH CENTURIES This final section makes no attempt to provide a survey or detailed discussion of commerce in Palaiologan139 and early Ottoman140 Constantinople. Instead, it offers only a few observations that bear on the foregoing discussions of the early and middle periods, in particular the commercial activities based in the regions of the Strategion and the Theodosian/Constantinian fora. While the stoai that had lined many streets of Constantinople since the days of Constantine may have finally disappeared in the fire of 1203, other types of commercial premises and settings known from earlier centuries apparently persisted, although probably changed in form or appearance. A rare graphic glimpse of this world is afforded by a mid-thirteenth-century fresco at Arta that may illustrate commerce in a medieval forum at Constantinople, where the heavy icon belonging to the Hodegon monastery is seen circulating,141 as described by Pero Tafur (1430s)142 and Russian pilgrims.143 The public square has an arcaded portico with upper storey windows, and the fresco’s inscriptions identify itinerant grocers selling vegetables (lachana), fruit (apples, pears, etc.), caviar (ho chazaris poulon to chabiari), and drinks in small pots ( phokadia) (Fig. 30). Elsewhere, texts provide a verbal picture, revealing both continuity and change. We begin with the Strategion region. Constantinople’s north shore continued to thrive thanks to alterations in trading patterns within the city that developed after the days when the Book of the Prefect was written. The growing role of Latin merchants in Constantinople accelerated a demographic shift in the city’s population, which had begun before the Fourth Crusade (see above) and increased in the Palaiologan period. This resulted in a greater concentration of commerce on the Golden Horn, where the Pisans, Genoese, and Venetians chose to settle.144 Commerce also moved over to Galata. In this late period, the Prosphorion was apparently no longer an enclosed harbor. Rather, Crusaders and others make repeated references to the Golden Horn itself as a harbor,145 served by skalai. There, in the late fourteenth century, the anonymous Russian pilgrim refers to “the large Basilike market near the wharves and ferry crossing to Galata.” 146 The buildings men139 On which see, e.g., N. Oikonomides, Hommes d’affaires grecs et latins `a Constantinople, XIIIe–XVe sie`cles (Montreal, 1979); E. Kislinger, “Gewerbe im spa¨ten Byzanz,” in Handwerk und Sachkultur im Spa¨tmittelalter, Internationaler Kongress Krems an der Donau, 7. bis. 10. Oktober 1986 (Vienna, 1988), 103–26. 140 See the discussion in H. Inalcik, “The Policy of Mehmed II toward the Greek Population of Istanbul and the Byzantine Buildings of the City,” DOP 23–24 (1969–79): 248 and nn. 84–85; Mantran, Istanbul, 179–487. 141 M. Achimastou-Potamianou, “The Byzantine Wall Paintings of Vlachernae Monastery (Area of Arta),” in Actes du XVe Congre`s international d’e´tudes byzantine, Athe`nes 1976, vol. 2, Art et arche´ologie. Communications (Athens, 1981), 1–14. 142 Pero Tafur, Travels and Adventure, 1435–1439, trans. and ed. M. Letts (London, 1926), 139–41. 143 Majeska, Russian Travelers, 362–66. 144 Magdalino, Constantinople me´die´vale, 78–90, with references. 145 E.g., Villehardouin, Conqueˆte, 209. 146 Majeska, Russian Travelers, 140, 353–54.

206

COMMERCIAL MAP OF CONSTANTINOPLE

tioned in the vicinity of the market allow us to place this market in the general vicinity of the present Spice, or Egyptian, Bazaar—in other words, by the old Neorion harbor and the Strategion (Fig. 31). The market may derive its name, not from the nearby imperial (i.e., Basilike) Gate,147 but from a market building, that is, a basilica, dating back to the time of the Notitia or even earlier. The Basilike Gate, the present Zindankapı, would then have been named after the market basilica, not vice versa. In the fifteenth century, Clavijo described this area as “the trading quarter of the city.” 148 We turn now to consider the Theodosian/Constantinian fora region. From “the large Basilike market” and Basilike Gate on the shore of the Golden Horn, the Portico of Domninos (or, as the Russian pilgrims called it in the 14th century and later, the Maurianos Embolos) led south to the area of the Constantinian and Theodosian fora. That these latter had also remained commercial centers in the Palaiologan period may be inferred from circumstantial evidence of the mid-fifteenth century, that is, from the earliest Ottoman period. Mehmet the Conqueror built two key structures in this area. One became known as the Covered Bazaar (Kapalıc¸arsı). It was situated in an area just to the north between the Constantinian and Theodosian fora, straddling the Portico of Domninus, an area that had been a hub of Byzantine commerce. The Old Bedesten had four gates, shops without doors, and seventy watchmen. Goldsmiths, operating also as bankers, occupied this inner building. The Sandal Bedesten, or Brocade Market, lies to the east.149 Near the two bedestens stood Mehmet’s other new building, his first palace (the Eski Saray).150 Viewed together his palace and bedestens may represent a repetition of the Byzantine past, if not unbroken continuity with it (Fig. 32). In the middle Byzantine period, luxury goods (perfume, precious metals, silks) had been sold in the stoai leading from the imperial palace to the Forum of Constantine. The palace itself had fallen out of use by the Palaiologan period, after the court moved to Blachernai. Now, Mehmet created a new prestige area in the center of the city, revitalizing the area of the Theodosian and Constantinian fora with his palace and adjacent luxury market (precious metals, silks). Mehmet also built the Sarac¸hane, another specialized market for leather goods.151 The commercial grouping of retail operations by type of merchandise, which was observed in the early period and became a salient feature of the middle period, is thus taken a step further in the early Ottoman period, when the groups are separately housed. By function, the covered markets recall the Roman commercial basilicas. Another feature of the early Ottoman period that can be linked with earlier Byzantine commercial topography concerns the location of the hans used by itinerant merchants. Indeed, the hans occupy an area beside the Portico of Domninos, which links the two regions of greatest interest in this paper, the Strategion and the area of the Theodosian/ Constantinian fora. The portico was where Romanus I built his xenodocheion in the tenth century (Fig. 31). It is unsurprising that these hans were concentrated here in the area between the Golden Horn, where ships were unloaded, and the Covered Bazaar, where As postulated by Berger, “Ufergegend,” 152–55. Gonzalez de Clavijo, Embajada a Tamorla´n, ed. F. Lopez (Madrid, 1943), 57; quoted by Majeska, Russian Travelers, 354. 149 Mantran, Istanbul, 463–67, map 14. 150 A. Berger, “Zur sogenannten Stadtansicht des Vavassore,” IstMitt 44 (1994): 342–44. 151 Mantran, Istanbul, 415. 147 148

MARLIA MUNDELL MANGO

207

goods were sold. Furthermore, the hans with their rooms laid out around a court (Fig. 33),152 recall in their architectural form such late antique inns as the xeneon at Umm alHalahil in Syria (Fig. 34).153 CONCLUSION In sum, in the early period food circulated by a two-tier system partly operated by the state (the annona) and partly by private enterprise. A systematic grouping of commercial premises by type was observed in the case of food-linked wholesale activities (at least some of which were operated by the state), whose establishments were often close to harbors (Fig. 4), and in certain crafts, such as the copperworkers opposite Hagia Sophia and silversmiths on the Mese (Fig. 20). By contrast, variety characterized the types of commerce in the Portico of Domninus. The areas of the Strategion (Region V) and of the Theodosian forum (Regions VII–VIII) were important centers of wholesale food enterprise. In the medieval period, the annona had ceased (except for military uses), while commercial grouping continued and was regulated by the city prefect. Myrepsoi, who dealt in spices and unguents, were located in the Portico of Achilles between the Milion and Chalke, silversmiths were still on the Mese, and textiles were sold in the forum of Constantine (Fig. 27). Wholesale meat, fish, and grain continued to be traded in specific locations, some of which had apparently been in use in the early period; the Strategion and the area of the Theodosian forum remained important (Fig. 22). Grocers, bakers, and inns selling wine operated everywhere. Stoai continued to house commerce and may have stood until 1203. Visiting merchants were controlled and housed in metata. In the Palaiologan period, the Latins—who were increasingly important in the city’s trade—occupied the area of the Golden Horn, which had itself become a “harbor.” The general area of the Strategion remained a focus of commerce, stretching west at least to the Basilike Gate at the end of the Portico of Domninos. In the early Ottoman period, the area of the latter street, where Romanus I had built an inn, became filled with hans (Fig. 31). Just to the south, the area of the Constantinian and Theodosian fora became the site of Mehmet the Conqueror’s new palace and luxury market, the two bedestens of the Bazaar. Thus, luxury commodities move from the vicinity of one imperial palace (Byzantine) to that of another (Ottoman) (Fig. 33). The commercial groupings of retail merchandise noted in the early and middle periods are now given separate housing by Mehmet the Conqueror in the bedestens and Sarac¸hane. So, it becomes clear—as stated at the start of this paper—that certain parts of the city proved to be convenient for commercial enterprise and thus remained in use throughout the centuries. Institute of Archaeology, Oxford 152 153

Mu ¨ ller-Wiener, Bildlexikon, 342–57. J. Lassus, Inventaire arche´ologique de la re´gion au nordest de Hama (Paris, 1935), 67.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Corinth, plan of the central area in late antiquity (based on R. L. Scranton, Medieval Architecture in the Central Area of Corinth, Corinth 16 [Princeton, 1957], plan IV)

Fig. 2.1

Fig. 2.2–2.3

Fig. 2.4 2

Daphne, Yakto villa, pavement, topographical border showing details of merchants selling or preparing (1) sweets?; (2) chopped meat; (3) oil; and (4) fish (after Antioch-on-the Orontes, vol. 1, Excavations of 1932, ed. G. W. Elkerkin [Princeton, 1934], figs. 13.13, 15.20, 16)

3

Rossano Gospels, fol. 2r, Cleansing the Temple ( after D. V. Ainalov, The Hellenistic Origins of Byzantine Art [New Brunswick, N. J., 1961], fig. 59)

4

Constantinople, plan of the city within the Constantinian walls, showing commercial installations in Regions I–XII, as listed in the Notitia of Constantinople, ca. 425 (drawing by A. Wilkins)

5

Ephesus, plan of agora areas in late antiquity (based on C. Foss, Ephesus after Antiquity: A Late Antique, Byzantine and Turkish City [Cambridge, 1979], f ig. 12)

6

Ostia, showing (1) Horrea ( Reg. I.Is.viii.2) and (2) Horrea Epagathniana et Epaphroditiana (after G. Rickman, Roman Granaries and Store Buildings [Cambridge, 1971], f ig. 3)

7

Rome, plan of the Emporium region showing the Horrea Galbana, Horrea Lolliana, and other horrea (after Rickman, Roman Granaries and Store Buildings, f ig. 24)

8

9

Dara, horreum of ca. 508 ( photo: C. Mango)

ˇ cenko) Constantina (modern Viransehir), horion of 542 (photo: I. Sevˇ

10

Perge, plan of the macellum (2nd/3rd century) and the south gate of the city (after C. de Ruyt, Macellum: Marché alimentaire des romains, Publications d’histoire de l’art et d’archéologie de l’Université catholique de Louvain 35 [Louvain-la-Neuve, 1983], f ig. 49)

11

Ostia, plan of the macellum with f ish merchant (1st/2nd century, restored 418–420) (after de Ruyt, Macellum, f ig. 44)

12

Istanbul, view from the area of the tetrapylon on the Mese down the street leading to the harbor of Julian ( photo: C. Mango)

13

Sardis, plan of Byzantine stoa shops, the House of Bronzes, and Building R on Hypaepa Embolos (based on J. Stephens Crawford, The Byzantine Shops at Sardis, Archaeological Exploration of Sardis Monograph 9 [Cambridge, Mass., 1990], f ig. 4)

14

Sardis, plan of lower storey of Byzantine stoa shops E6 ( left) and E7 (right) showing location of finds (after Crawford, Byzantine Shops at Sardis, f ig. 267)

15

Sardis, drawing of hypothetical section of Byzantine stoa shop E6 (after Crawford, Byzantine Shops at Sardis, f ig. 32)

16

Scythopolis, plan of Palladius Street showing Sigma portico and shops built in 507 (after R. Bar-Nathan and G. Mazor, in Excavations and Surveys in Israel 11 [1992]: f ig. 60)

17

Apamea, signs painted on facade of cardo shops (after J. C. Balty, Guide d’Apamée [Brussels, 1981], f ig. 81)

18

19

Berytus, plan of decumanus maximus with portico shops and residential complex (with apsidal building) incorporating industrial installations (based on K. Butcher and R. Thorpe, “A Note on Excavations in Central Beirut, 1994–96,” Journal of Roman Archaeology 10 [1997]: fig. 7)

Constantinople, plan showing stoa shops excavated on the Mese behind the so-called House of Lausus (after R. Naumann, “Vorbericht über die Ausgrabungen zwischen Mese und Antiochus-Palast 1964 in Istanbul,” IstMitt 15 [1965]: f ig. 5)

20

Constantinople, plan of Constantinian city showing locations of some non-food commerce (drawing by A. Wilkins)

21

Constantinople, drawing and section of stoa (?) remains (including entablature, above) excavated on Çatalçesme Soka˘g ı (after T. Macridy in M. Schede, “Archäologische Funde: Türkei,” AA 44 [1929]: f igs. 20, 21)

22

Constantinople, plan of Constantinian city showing localized food commerce in the medieval period, based largely on the Book of the Prefect of 912 (drawing by A. Wilkins)

23

Constantinople and its hinterland (after J. Lefort, “Les communications entre Constantinople et la Bithynie,” in Constantinople and Its Hinterland, ed. C. Mango and G. Dagron [Aldershot, 1995], 208)

24

Amphorae excavated at St. Polyeuktos: (A) Type 35 (8th century); (B) Type 54A (10th–11th centuries); (C) Type 56 (11th century); ( D-E) Types 61, 67, respectively (12th–13th centuries) (after J. W. Hayes, Excavations at Saraçhane in Istanbul, vol. 2 [Princeton, 1992], f igs. 23–26)

25

Athens, Benaki Museum, drawing of 7th-century steelyard (after Greece and the Sea, ed. A. Delivorrias [Athens, 1987], no. 159)

26

London, University College, Petrie Museum, wooden box housing balance scale and flat weights (after W. M. Flinders Petrie, Glass Stamps and Weights: Ancient Weights and Measures [Warminster-Encino, 1974], pl. XVI)

27

28

Constantinople, plan of Constantinian city showing localized non-food commerce in the medieval period, based on the Book of the Prefect of 912 and other texts (drawing by A. Wilkins)

Corinth, plan of central area in the medieval period, showing glass and pottery factories and their porticoed showrooms (based on R. L. Scranton, Medieval Architecture in the Central Area of Corinth, Corinth XVI [Princeton, N. J., 1957], plan IV)

29

Palmyra, plan of Umayyad stoa shops built within the earlier Roman stoas on the cardo (after K. al-As‘ad and F. M. Stepniowski, “The Umayyad Suq in Palmyra,” Damaszener Mitteilungen 4 [1989]: f ig. 1)

30

Arta, church, drawing of fresco showing procession of Hodegetria icon, possibly at Constantinople (after M. Achimastou-Potamianou, “The Byzantine Wall Paintings of Vlachernae Monastery (Area of Arta),” in Actes du XVe Congrès international d’études byzantine, Athènes 1976, vol. 2, Art et archéologie. Communications [Athens, 1981], f ig. 14)

31

32

Constantinople, plan of central area showing Ottoman hans near the Portico of Domninos (Maurianos Embolos), the Basilike Gate (Zindankapi), the Spice and Covered Bazaars, and the fora of Theodosius and Constantine (drawing by A. Wilkins based on Müller-Wiener, Bildlexikon zur Topographie Istanbuls, fig. 393)

Constantinople, plan showing localized non-food commerce in the Byzantine and Ottoman periods (drawing by A. Wilkins)

33

34

Constantinople, plan of Kürkçü Hani, 16th century (after Müller-Wiener, Bildlexikon zur Topographie Istanbuls, f ig. 404)

Umm al-Halahil, Syria, plan and inscription of xeneon, 6th century (after J. Lassus, Inventaire archéologique de la région au nordest de Hama [Paris, 1935], f igs. 70, 71)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Maritime Neighborhoods of Constantinople: Commercial and Residential Functions, Sixth to Twelfth Centuries PAUL MAGDALINO

onstantinople, like New York, is a city not only by the sea, but also, to a large extent, in the sea. The effect of the sea on the fabric of the city is strongly pervasive, and it makes sense to start from the sea when investigating urban neighborhoods. By far the best evidence for the texture of urban neighborhoods comes from twelfth-century documents concerning the real estate conceded to the Italian maritime republics of Venice, Pisa, and Genoa—real estate that lay close to the shores of the Golden Horn. Since the sea is not far from any part of the city or its suburbs, and is indeed visible from almost anywhere within the Theodosian walls, it may well be asked what is meant by a maritime neighborhood. What distinguishes it from an inland neighborhood? Where does the one end and the other begin? Eleven of the twelve urban regions of the fifth-century Notitia urbis Constantinopolitanae bordered on the sea, but only Regions I and IX had a long coastline.1 Most of the other regions extended from a narrow stretch of coast to a narrow bloc of the city center. However, the regions were administrative rather than social or economic units. If we take into account the topography, the layout of public spaces, and the location of public monuments, we can draw a broad working distinction between those parts of the city that looked primarily toward the sea and those orientated toward the central avenue (Mese), the fora, and the great public buildings. Only in rare cases was a focal point such as the Strategion or the Leomakellon situated so close to the sea as to constitute a rival attraction.2 In this paper, I shall be concerned with those neighborhoods whose proximity to the sea may be assumed to have been decisive, whether directly or indirectly, for the location of houses and businesses. The assumption is that a seaside location was desirable, first, for the loading, unloading, storage, and marketing of seaborne merchandise, and second, for the recreation afforded by a view of the sea. The commercial importance of proximity to the sea is self-evident, although, given the low status of commerce in Byzantine society and culture, it is almost never

C

1 In Notitia dignitatum, ed. O. Seeck (Berlin, 1876), 229–43. For German translation, commentary, map, and earlier bibliography, see A. Berger, “Regionen und Straßen im fru ¨ hen Konstantinopel,” IstMitt 47 (1997): 349–414. 2 On these locations, see A. Berger, Untersuchungen zu den Patria Konstantinupoleos, Poiki´la Buzantina´ 8 (Bonn, 1988), 406–11, 515–56, and below.

210

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

articulated by the sources and must be inferred from the geographical incidence of harbors and landing-stages. Byzantine sources are rather more eloquent on the recreational value of a sea view. Imperial legislation relating to Constantinople in the fifth and sixth centuries concurs with the treatise on urban planning emanating from sixth-century Palestine under the name of Julian of Ascalon in insisting that new buildings should not obstruct a neighbor’s view of the sea.3 The legislation proved difficult to enforce, and the eleventh-century judge Eustathios Romaios ruled that it did not apply outside Constantinople. However, the grounds on which he justified his decision show that the sight of the sea had lost none of its appeal: For where a man can go and walk on the shore, what need is there to urge a neighbor to keep the regulation distance? For here we are enclosed by the walls, and it is not possible for us to leave our homes and spend the night on the shore. But outside, people are not shut off from the sea by walls, and there is nothing to prevent them from spending as long as they like on the seashore.4

The site of Constantinople is bordered by three expanses of sea: the Sea of Marmara, the Bosporos, and the Golden Horn. I shall have little to say about the area beside the Bosporos, the area consisting of the eastern slope of the ridge that terminates in the acropolis of ancient Byzantium. This was certainly a residential area: there were private houses next to the Great Palace in the ninth century, and several individuals are recorded as living on the Acropolis.5 But there is no evidence for commercial premises along this stretch of coast, where the only proper harbor was that of the Great Palace,6 and the few landing-stages that are attested were used either for ferries across the Bosporos7 or for servicing the great religious and imperial houses that took up most of the space between the Great Palace and the Acropolis point.8 Partly because of the proximity of the Great Palace and partly because this part of the city had developed out of the sacred and recreational area of ancient Byzantium, it tended to be dominated by a few large religious complexes—the orphanage, the Mangana, the Monastery of the Hodegoi, the churches of the Archangels and St. Menas—that did not really add up to an urban neighborhood. I shall therefore concentrate on the parts of the city beside the Golden Horn and the Sea of Marmara, and in particular on those coastal districts within the Constantinian wall that were important because they were close both to the commercial axis of the Mese and to the harbors that handled most of the city’s maritime traffic during the height of its Laws of Zeno and Justinian: CIC, CI 8.10.12.2–4; CIC, Nov 63; Julian of Ascalon, 52.2, ed. and trans. C. Saliou, Le traite´ d’urbanisme de Julien d’Ascalon, TM, Monographies 8 (Paris, 1996), 72–73. Cf. in general S. N. Troianos and K. G. Pitsakis, Fusiko` kai` domhme´ no periba'llon sti`" buzantine` " nomike` " phge´ " (Athens, 1998). 4 Peira, 18.5; ed. P. Zepos and I. Zepos, Jus graecoromanum (Athens, 1931; repr. Aalen, 1962), 4:68–69. 5 Theophanes Continuatus, ed. I. Bekker, CSHB (Bonn, 1838), 328.14–15; ibid., 382, 838 (Georgius Monachus Continuatus); John Skylitzes, Synopsis historiarum, ed. I. Thurn, CFHB 5 (Berlin–New York, 1973), 500; Life of St. Luke the Stylite, ed. and trans. P. Vanderstuyf, “Vie de Saint Luc le Stylite (879–979),” PO 11 (1915), 246–47. 6 See C. Mango, “The Palace of the Boukoleon,” CahArch 45 (1997): 47. 7 Vanderstuyf, “Saint Luc”; Notitia, 5.15, ed. Seeck, p. 233. 8 A landing-stage at the Mangana is mentioned by Michael Attaleiates, Historia, ed. I. Bekker, CSHB (Bonn, 1853), 73. The ship that brought Pope Martin I to Constantinople in 653 docked “near the Arkadianai” (PG 92:392), which must refer to the deep-water anchorage near the Baths of Arcadius: Procopius, De aedificiis, 1.11.2–4; R. Janin, Constantinople byzantine: De´veloppement urbain et re´pertoire topographique, 2nd ed. (Paris, 1964), 311–12. 3

PAUL MAGDALINO

211

development in the fifth and sixth centuries. I shall not, however, be dealing primarily with this period but with its aftermath, with the question of how the port neighborhoods of Constantinople evolved between the reign of Justinian I and 1204. I shall be revisiting problems that Cyril Mango posed in the mid-1980s9 and solutions that I suggested in the mid-1990s.10 I take the opportunity to develop, refine, and modify my ideas, particularly in the light of recent work by Albrecht Berger.11 In conclusion, I shall attempt to convey something of the texture of the urban fabric that is documented in the twelfth-century evidence for the Italian quarters beside the Golden Horn (see Fig. 1). One of the great natural advantages of the site of Constantinople, as Byzantines from Procopius to Pachymeres were well aware, is the presence of a large sheltered anchorage in the shape of the Golden Horn.12 The commercial port and the naval dockyard of ancient Byzantium, the Prosphorion and Neorion, were on the north coast of the city, and in Ottoman times all important shipping, apart from a small fleet of war galleys, used the Golden Horn. In Byzantine times, however, things were much less one-sided, at least until the shock of the Fourth Crusade, which prepared the way for the development of the Ottoman city. From the fifth to the thirteenth century, business was more evenly distributed between the north and south coasts of the city, and for a time, from ca. 550 to ca. 1050, the south coast was probably busier. This was entirely due to the construction, by the emperors Julian and Theodosius I, of two large artificial harbors that gave the Marmara coast a port capacity at least as great as that of the Golden Horn. It is evident that both harbors were constructed as part of the infrastructure supporting the rapid growth of population and built-up area in the century following the foundation of the city.13 The work coincided with the extension of the city’s water supply and is explained, in part, by the need to cater for an increase in the food supply. The Notitia of Theodosius II lists two granaries, the Horrea Alexandrina and the Horreum Theodosianum, in the ninth region of the city, which stretched between the harbors of Julian and Theodosius.14 These harbors thus handled part of the grain shipments coming from Egypt. However, the storage facilities on this side of the city were not equal to those near the ports of the Golden Horn. In Region V, which contained the Strategion and the Prosphorion harbor, the Notitia lists three granaries in addition to an oil storage depot, the Horrea Olearia;15 one of the granaries, the Horrea Valentiaca, was evidently constructed by the emperor Valens (364–378), so after the Harbor of Julian (361–363). Such was the concentration of food supply infrastructure in this region that the Notitia describes it as containing the “essential buildings of the city” (necessaria civitatis aedificia)— something not said of Region IX. The greater importance, or at least the higher profile, of the northern complex is suggested by a fifth- or sixth-century text preserved in the 9 C. Mango, Le de´veloppement urbain de Constantinople (IVe–VIIe sie`cles), TM, Monographies 2, 2nd ed. (Paris, 1990). 10 ´ tudes sur l’e´volution des structures urbaines, TM, Monographies 9 P. Magdalino, Constantinople me´die´vale: E (Paris, 1996). 11 See below, p. 221. 12 Procopius, De aedificiis, 1.5; Leo the Deacon, Historia, ed. C. B. Hase, CSHB (Bonn, 1828), 129; George Pachymeres, Relations historiques, ed. A. Failler (Paris, 1984), 2:469. 13 Mango, De´veloppement urbain, 37 ff. 14 Notitia, 10.6, 9, ed. Seeck, p. 237; Berger, “Regionen,” 369. 15 Notitia, 5.13, 15–17, ed. Seeck, pp. 233–34; Berger, “Regionen,” 364.

212

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

Book of Ceremonies that details the procedure to be followed when the emperor goes to inspect the granaries of the Strategion16 —it makes no mention of those on the south coast. The conclusion would seem to be that the food storage facilities on this side were not commensurate with the capacity of the vast new artificial harbors and that these were not constructed solely in order to receive grain imports but to handle other traffic. The commodity that comes most readily to mind is building material—the timber, the bricks, and, above all, the Proconnesian marble needed for the great building programs of the late fourth and early fifth centuries.17 In the main, the most massive constructions of this period, the palaces and the public monuments of the Theodosian dynasty, were closer to the south than to the north coast.18 Given the continuing importance of the Golden Horn for the urban food supply during Constantinople’s first two centuries, the subsequent decline of the port areas on the north coast is all the more striking. The most important single piece of evidence for this development is the statement of the Parastaseis, repeated by the Patria, that under Justinian the wholesale import market (aiJ ajgorai` tw'n qalassi´wn ejmporenma´ twn) was transferred from the Neorion to the Harbor of Julian.19 The historical information of these texts is suspect, but there is no need to doubt that it always reflects the material reality of the times when they were written. In other words, during the eighth and tenth centuries, the wholesale business of the port of Constantinople was concentrated beside the Harbor of Julian. There is, moreover, an accumulation of circumstantial evidence to prove not only that the port of Julian and the adjacent neighborhoods along the Marmara coast were flourishing at the expense of the old harbors and urban neighborhoods at the lower end of the Golden Horn, but also that the shift dated from the middle of the sixth century. 1. By the beginning of the ninth century, at the latest, the Harbor of Julian was alternatively known as the Harbor of Sophia.20 This supports the information, contained in the Patria and certain chronicles, that the harbor was dredged by Justin II (565–574), embellished with statues, and renamed in honor of De cerimoniis aulae byzantinae, 2.51, ed. J. J. Reiske, CSHB (Bonn, 1829), 1:699–701; cf. M. McCormick, “Bateaux de vie, bateaux de mort: Maladie, commerce, transports annonaires et le passage ´economique du Bas-Empire au Moyen Age,” Morfologie sociali e culturali in Europa fra tarda antichita` e alto medioevo, Settimane, 45 (1996): 37–40. 17 See N. Asgari, “The Proconnesian Production of Architectural Elements in Late Antiquity, Based on Evidence from the Marble Quarries,” in Constantinople and Its Hinterland, ed. C. Mango and G. Dagron (Aldershot, 1995), 263–88. 18 This is obvious in the cases of the fora of Theodosius and Arcadius (see Mango, De´veloppement urbain 43–45) and the domus of Pulcheria and Arcadia in the third and ninth regions (Notitia, 4.8, 10.7, ed. Seeck, pp. 232, 237; see below, p. 216). The domus of Placidia, Eudocia, and Arcadia in the tenth region (Notitia, 11.11–13, ed. Seeck, pp. 237–38) are probably to be sought in the area just to the north of the Capitol, where a number of Theodosian family mansions were situated: cf. Magdalino, Constantinople me´die´vale, 47 n. 170. I will deal further with the topography of these residences in a forthcoming study. 19 Scriptores originum Constantinopolitanarum, ed. T. Preger, 2 vols. (Leipzig, 1901–7; repr. 1989), 67, 188; Constantinople in the Early Eighth Century: The Parastaseis Syntomoi Chronikai, ed. and trans. A. Cameron and J. Herrin (Leiden, 1984), 152–53, 267. 20 Theophanes, Chronographia, ed. C. de Boor (Leipzig, 1883; repr. Hildesheim, 1980), 1:184. 16

PAUL MAGDALINO

213

his wife Sophia.21 Although the Neorion on the Golden Horn was dredged in 698, this was almost certainly to restore the harbor to use as a naval dockyard, and the dredging was, significantly, connected with an outbreak of the plague.22 2. Justin II and Sophia are credited with three buildings in the neighborhood of the harbor: their own palace, which they inhabited before Justin became emperor,23 and two churches, dedicated to the Archangel Michael and St. Thekla, which they built or restored.24 Another important complex of buildings near the harbor, comprising a hospital and two adjacent churches, is attributed to one of their ministers, the praipositos Narses.25 It is to be noted that this foundation provided the southern port area with a symmetrical counterpart to the fifth-century hospital of Markianos and the church of St. Eirene at Perama on the north coast.26 It is also to be noted that the northern port neighborhoods saw no comparable development in the late sixth century, the church of the Theotokos ta Protasiou, attributed by the Patria to Justin II, being the only construction datable to this period.27 3. The churches and monasteries which the iconoclast emperor Constantine V (741–775) is said to have converted to secular use were all on the south coast, and most were in the vicinity of the Harbor of Julian.28 4. By the tenth century, the only public granary still in use was the one known as the Lamia, in the vicinity of the port of Theodosius; this is probably to be identified with the Horrea Alexandrina or the Horreum Theodosianum of the fifth-century Notitia.29 5. According to the tenth-century Synaxarion of Constantinople, the feast of St. Thekla was celebrated at her church “in the barley market” (ejn toi'" Kriqopwlei´oi").30 If, as seems likely, this church was identical with the one restored by Justin I or Justin II, it may be deduced that the market for barley and other bulk foodstuffs was situated close to the Harbor of Julian. In this connection, it is worth noting that of the few urban fires that tenth-century 21 Scriptores, ed. Preger, 184, 229–31; George Kedrenos, Historiarum compendium, CSHB (Bonn, 1838), 1:685. 22 Theophanes, Chronographia, ed. de Boor, 1:370; cf. Mango, De´veloppement urbain, 56, and see below, pp. 218–19. 23 Corippus, In laudem Iustini Augusti minoris, 1.97–114, ed. and trans. A. Cameron (London, 1976), 39, 89, 132–33; ed. and trans. S. Ante`s (Paris, 1981), 20–21; see below, p. 216. 24 Scriptores, ed. Preger, 228–29; Berger, Untersuchungen, 563–66, 578–80. Procopius, however (De aedificiis, 1.4.28), attributes St. Thekla to Justin I. 25 Scriptores, ed. Preger, 248–49; Berger, Untersuchungen, 591–96. 26 Ibid., 447–49; for the hospitals, see T. S. Miller, The Birth of the Hospital in the Byzantine Empire, 2nd ed. (Baltimore, 1997), 91–92. 27 Scriptores, ed. Preger, 220; Berger, Untersuchungen, 403–4. 28 Theophanes, Chronographia, ed. de Boor, 1:439 (St. Euphemia), 443 (monasteries of Kallistratos, Dios, Maximinos); Nikephoros, Antirrhetika, 3, PG 100:493D (monasteries of Phloros and Kallistratos); Scriptores, ed. Preger, 148, 217, 240, 258 (churches of St. John the Baptist ta Probou, St. Euphemia, St. Julian the Myrelaion, St. Andrew at Boukinon, all near the Harbor of Julian). 29 Mango, De´veloppement urbain, 54–55. 30 Synaxarium CP, col. 75.

214

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE chroniclers considered worth recording, two ravaged the quarter of ta Amantiou, near the Harbor of Julian.31 From later evidence, it is clear that neighborhoods where merchandise was stored were particularly at risk.32 6. When sources of the seventh to tenth centuries mention a precise embarkation or disembarkation point for sea travelers, this is mostly the Harbor of Julian or Sophia, and never the Golden Horn. It was at the Harbor of Sophia that Heraclius landed in 610,33 that Eustratios, abbot of the monastery of Agauros on Bithynian Olympus, disembarked in the mid-ninth century,34 and that Leo of Synada set sail on his diplomatic mission to Rome in 996.35 7. We know of several people who lived on the south coast of the city from the eighth to eleventh centuries. There was a cluster of aristocratic residences near the former Harbor of Theodosius.36 A succession of illustrious persons lived at the Harbor of Sophia, possibly in the palace of Justin and Sophia,37 and in the mid-eleventh century, an imperial secretary, Nicholas, had a “not very fine house” at Bykinon, between the harbor and the Hippodrome.38 Most significantly, we hear of three tradesmen (ejrgasthriakoi´), one of them said to be very rich, who lived near the Harbor of Sophia.39 The same period yields only one reference, of which I am aware, to a resident of the area near the old harbors on the Golden Horn. This was Antony, a patrikios under Michael III, who owned a fine house, complete with church and bathhouse, near the dockyard of the Neorion.40

In every respect, the evidence for the development and prosperity of the south coast, particularly in and around the Harbor of Julian, is as striking as the almost complete lack of evidence for business and residential activity in the area beside the Golden Horn, which had been the main hub of the city’s economy in the fourth to sixth centuries. The picture changes slightly if one takes into account the marketing of livestock, which, according to the Book of the Prefect, was divided between the Strategion and the Forum Tauri. This distribution may be much older than is generally supposed, because it corresponds to the location of the main meat markets, the macella, listed by the Notitia of Theophanes Continuatus, ed. Bekker, 354, 462. See the poem of Constantine Stilbes on the fire of 1197, ed. J. Diethart, “Der Rhetor und Didaskalos Konstantinos Stilbes” (doctoral diss., University of Vienna, 1971), lines 165–73; George Pachymeres, De Andronico Paleologo, ed. I. Bekker, CSHB (Bonn, 1835), 2:227, 582 (see also the fuller text of this second passage published by A. Failler in REB 36 [1978]: 157–58). 33 Theophanes, Chronographia, ed. de Boor, 1:299. 34 Ed. A. Papadopoulos-Kerameus, Ana j ´ lekta JIerosolumitikh'" Stacuologi´a" (St. Petersburg, 1897), 4:391. 35 The Correspondence of Leo Metropolitan of Synada and Syncellus, ed. and trans. M. P. Vinson, CFHB 23 (Washington, D.C., 1985), 14–15. 36 Life of St. Basil the Younger, ed. A. N. Veselovsky, in Sbornik otdela russkogo jazyka i slovestnosti imperatorskoj akademii nauk 26 (1889): 6, supplement, 57, 72; Mango, De´veloppement urbain, 59. 37 See below, pp. 216–17. 38 Michael Psellos, Orationes forenses et acta, ed. G. Dennis (Leipzig-Stuttgart, 1994), 172. 39 Life of St. Basil the Younger, ed. Veselovsky, 51, 54; Pseudo-Symeon, in Theophanes Continuatus, ed. Bekker, 674. 40 Synaxarium CP, col. 935. 31 32

PAUL MAGDALINO

215

Theodosius II: two in Region V and two in Region VIII.41 But here again, the trend seems to have been toward a greater concentration in the south; horses were sold in the Amastrianon, and if we can believe the Patria, Constantine V moved the cattle market to the Forum of Theodosius (Tauri) from the area of the Prosphorion harbor.42 What brought about this apparent gravitational shift in the city’s maritime economy? By docking on the Marmara coast, ships avoided the strong currents and headwinds of the Bosporos; yet these hazards were known long before the foundation of Constantinople. The silting up of the old harbors on the north coast was certainly a problem; it may have encouraged business to move not only to the south coast but also, as we shall see, further up the Golden Horn, where ships could moor closer to the sea walls.43 But as we shall also see, silting did not prevent the eventual re-use of the old harbor area, and it affected the Marmara coast just as badly: the Harbor of Julian had to be dredged two centuries after its construction, and the Harbor of Theodosius was allowed to silt up almost completely.44 The question is, why of all the four artificial harbors that had served Constantinople in the fifth century was the Harbor of Julian kept open? The shrinking of the urban population is surely part of the answer, as Cyril Mango has suggested.45 Another part lies, no doubt, in the changing nature of supply. The cessation of grain shipments from Egypt in the seventh century meant, presumably, that basic commodities were imported, to a greater extent, over smaller distances in lighter loads and smaller craft that did not draw much water and did not need elaborate docking facilities. The trend toward smaller ships may have begun under Justinian with the construction of a large granary at Tenedos, where the large grain transports from Alexandria could leave their cargoes to be carried on by local vessels.46 Certainly, the picture we get from Attaleiates in the eleventh century is one of boats unloading at jetties all along the coast rather than in specially localized harbors.47 However, none of this properly explains why the depopulated Constantinople of the early Middle Ages gravitated toward the south coast instead of concentrating around the original commercial center beside the Golden Horn, where the “necessary buildings” of the city had always been situated. The explanation may have more to do with the residential than with the commercial attractions of the south-facing Marmara coast—this was the place to be in during the harsh winter.48 The Harbor of Julian was also conveniently close to the Great Palace. The coastal district to the west of the harbor was known by 425 as Kainopolis, or New City, and this has led Cyril Mango to identify it with the large built-up area that Zosimus says 41 Notitia, 6.27, 9.17, ed. Seeck, pp. 234, 236; Leonis Sapientis Liber Praefecti, 5.1, 5, and 16.2, ed. and trans. J. Koder, Das Eparchenbuch Leons des Weisen, CFHB 33 (Vienna, 1991), 122–25. 42 Ibid., 21.3, 8, pp. 136–39; Scriptores, ed. Preger, 263–64; cf. Berger, Untersuchungen, 425. 43 Magdalino, Constantinople me´die´vale, 89; see below, pp. 221–23. 44 See A. Berger, “Der Langa Bostanı in Istanbul,” IstMitt 43 (1993): 467–77. 45 Mango, De´veloppement urbain, 53 ff. 46 Procopius, De aedificiis, 5.1.7–16; note that Procopius refers to the harbors of Constantinople in the plural (ej" tou` " Buzanti´ou" lime´ na"). On the capacity of the grain fleet, cf. McCormick, “Bateaux de vie,” 103–7. 47 Attaleiates, Historia, ed. Bekker, 277–78; cf. P. Magdalino, “The Grain Supply of Constantinople, Ninth– Twelfth Centuries,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 17), 41–43. 48 E.g., the winters of 716–717 and 762–763: Theophanes, Chronographia, ed. de Boor, 1:396–97, 434–35.

216

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

was reclaimed from the sea.49 The identification finds some support in the Patria, which records a tradition that a harbor on this stretch of coast, the Harbor of Eleutherios, was filled in with construction debris from the Forum of Theodosius.50 What is interesting here is the evident demand, in the first phase of the city’s expansion, for building land along the south coast in preference to other parts of the city, where land did not have to be reclaimed from the sea. The desirability of residing on the south coast is demonstrated by the presence here of large aristocratic residences that seem to have had almost no equivalent on the Golden Horn, at least within the Constantinian wall. The earliest mentions of these are found in the Notitia of Theodosius II, which records a palace in Region III belonging to the Augusta Pulcheria, then the most powerful person in Constantinople, and one belonging to the Nobilissima Arcadia in Region IX; both women were important enough to have second homes elsewhere in the city.51 We do not know what happened to these palaces after the fifth century, but to judge from what we know of other Theodosian buildings in Constantinople, they must have been magnificent and solid constructions of a kind that would have been easier to re-use than demolish.52 Is it coincidence, then, that we find later references to large and important princely residences in similar if not identical locations? At the end of the eighth century, the favorite residence of the empress Eirene, her palace at ta Eleutheriou, was in the vicinity of the palace of the Nobilissima Arcadia.53 The palace of Justin II and Sophia overlooked the Harbor of Julian/Sophia on one side and the open sea on the other;54 it is therefore likely to have stood on the east side of the harbor, somewhere between the Hippodrome and the sea. This is exactly the part of Region III where we should look for the palace of Pulcheria.55 Three centuries later, the “enormous house” Nikephoros Phokas the Elder received from Basil I and passed on to his son Bardas stood “near the church of St. Thekla,” according to one version of the Logothete Chronicle; according to Leo the Deacon, it was “on the descent of the street 49 Notitia, 10.5, ed. Seeck, p. 237; Zosimus, Historia nova, ed. F. Paschoud (Paris, 1971), 1.108, 2.35; Mango, ´ Developpement urbain, 17–18, 45. 50 Scriptores, ed. Preger, 184–85. 51 Notitia, 4.8, 10.7, ed. Seeck, pp. 232, 237, and see 11.12, 12.9 for the domus of Arcadia and Pulcheria in Regions X and XI, respectively. Region III included the Hippodrome and the whole area to the south of the Mese between the Hippodrome and the Forum of Constantine. This has led Berger, “Regionen,” 361, to suggest that the domus Pulcheriae of the Notitia was originally the palace of Antiochus, the remains of which have been excavated on the north-west side of the Hippodrome. The suggestion is plausible in that the building existed, and had become imperial property, by the time the Notitia was composed. Against it, however, is the fact that the house of Antiochus was subsequently known by the name of its original owner. It is also to be noted that the domus of other Theodosian princesses (Placidia, Marina, Arcadia) in the southern part of the city were evidently close to the sea. On Pulcheria, see K. Holum, Theodosian Empresses (Berkeley– Los Angeles, 1982). 52 The main evidence comes from the excavations of the palace complexes to the west of the Hippodrome: R. Naumann and H. Belting, Die Euphemia-Kirche am Hippodrom zu Istanbul, IstForsch 25 (Berlin, 1966); J. Bardill, “The Palace of Lausus and Nearby Monuments in Constantinople: A Topographical Study,” AJA 101 (1997): 67–95. Cf. also C. L. Striker, The Myrelaion (Bodrum Camii) in Istanbul (Princeton, 1981); C. Mango, “The Palace of Marina, the Poet Palladas and the Bath of Leo VI,” in Eujfro´ sunon. Afie j ´ rwma sto´ n Mano´ lh Catzhda´ kh, ed. E. Kypraiou, 2 vols. (Athens, 1991), 1:321–30; Magdalino, Constantinople me´die´vale, 42–43. 53 Berger, Untersuchungen, 588–89. 54 See the translation of the relevant passage of Corippus, below, p. 217. 55 See above, note 52.

PAUL MAGDALINO

217

leading down to the sea where the Harbor of Sophia opens out.” 56 These two pieces of information combined again point to a location on the east side of the harbor.57 In the twelfth century, the brother of one of the Komnenian emperors lived in a grand house whose location is described by Choniates in terms very similar to those used by Leo the Deacon.58 It is tempting, and I think not unreasonable, to see all these references as pertaining to a single house occupied, over the centuries, by close relatives or favored associates of the emperor. What does seem certain is that the re-investment in the Harbor of Julian and its neighborhood made by Justin II and the praipositos Narses was not unconnected with a preference for this location, which Justin II clearly shared with eminent people before and after him. The charms of the spot are evoked by Corippus in his poem in praise of Justin: One side looks out over the wide sea, the other backwards over the harbour—the harbour formed by the embrace of the arms of the two banks, with walls on top; they make it defy the swift winds, and render the open sea quiet inside the anchorage. They break the waves of the sea with their marble barrier and keep away the waters as they flow back with their narrow neck. The royal pair loved this place; from it they used to watch the waves in the strait and the curving ships carrying all the trade of two worlds.59

The description is too precise to be merely an ekphrastic topos, and it is reminiscent of what Julian of Ascalon has to say about the importance of a sea view: “If a man can see a harbor or the shore, or even just look at ships at anchor in the case of a town or village which does not have a proper harbor, his view of them should in no way be impaired or removed, for they are a source of recreation to those who behold them.” 60 But it was one thing to invest in renewing the Harbor of Julian and another thing to do so at the expense of the existing economic hub of the city by the Golden Horn. If there was a major relocation in the mid-sixth century, what caused it? I think we have to focus attention on the greatest catastrophe to hit the Mediterranean world at this time, the bubonic plague. There is still no systematic study of the sixth-century plague in all its aspects, and its significance has been debated.61 It has been cogently argued, by Jean Durliat and Mark Whittow, that the long-term demographic and economic effects of the plague were negligible compared with the wars and the territorial losses of the seventh century.62 The evidence for building programs and problems of overcrowding in late sixth-century Constantinople tends to support this view.63 However, plague mortality in the short term was undoubtedly devastating and shocking. The horrific eyewitness acH. Gre´goire, “La carrie`re du premier Nice´phore Phocas,” in Prosfora` eij" Sti´lpwna Kuriaki´dhn (Thessalonike, 1953), 2:250; Leo the Deacon, ed. Hase, 83–84. 57 This is where Berger, Untersuchungen, 566, places the church of St. Thekla. 58 Nicetae Choniatae Historia, ed. J. L. van Dieten, CFHB (Berlin–New York, 1975), 1:445; cf. Magdalino, Constantinople me´die´vale, 47, 91. 59 Corippus, In laudem Iustini, trans. Cameron, 89. 60 Saliou, Traite´, 72–73. 61 For a recent discussion, with some new insights, see McCormick, “Bateaux de vie,” 48 n. 20, 52–65. 62 J. Durliat, “La peste du VIe sie`cle: pour un nouvel examen des sources byzantines,” in Hommes et richesses dans l’Empire byzantin (Paris, 1989), 1:107–19; M. Whittow, The Making of Orthodox Byzantium, 600–1025 (London, 1996), 66–68. 63 Magdalino, Constantinople me´die´vale, 58–59. 56

218

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

counts by Procopius and John of Ephesus of the initial outbreak in 542 are not to be dismissed.64 Later outbreaks were less severe, but they ensured constant concern about the likelihood, and measures for the prevention, of future recurrences. The official and the popular explanation was that the plague was a scourge sent by God, and this undoubtedly quickened the pace of investment in pious and charitable foundations that is so marked during the age of Justinian and his successors. At the same time, the rational and natural explanations offered by ancient medical theory were not entirely discredited, especially as it became evident that the plague struck the righteous and the unrighteous with equal ferocity. Not only did the medical profession adhere to the wisdom of Hippocrates and Galen, which held that the body was predisposed to plague infection by bad air,65 but the seventh-century theologian known for convenience as Anastasius of Sinai decided that the credibility of Divine Providence was better preserved by attributing the plague to natural causes. Witness one of his erotapokriseis: Q. Whence do plagues arise, and why do they not occur in certain desert lands of the nations, but mostly in densely inhabited, crowded and filthy cities? A. Fatal diseases often arise from corrupt air, and dust, and the stench of dead bodies, summer rains, and exhalations of land and sea.66

It seems to me that considerations of this kind were bound, eventually, to affect residential patterns in plague-stricken cities. In the case of Constantinople, it may also be relevant that stagnant waters and those polluted by the effluent from large settlements were believed to be sources of noxious exhalations.67 It is clear from the Notitia of Theodosius II that in pre-plague Constantinople, the highest concentration of ordinary housing, and therefore the greatest source of human waste, lay in the area beside the Golden Horn, to the west of the Neorion.68 The Golden Horn is not flushed out by currents or waves. And, as we have seen, the dredging of the Neorion in 698 was associated with a bad outbreak of plague; the wording of Theophanes, our source, leaves no doubt that the association was perceived to be causal as well as temporal.69 John of Ephesus says that in the plague of 542 many bodies were dumped in the sea; any dumped in the Golden Horn would not have been washed away. Both Procopius and John of Ephesus tell us that when the authorities got around to disposing of bodies in a more organized way, the 64 Procopius, De bello Persico, 2.22–23; John of Ephesus, in Pseudo-Dionysius of Tel-Mahre, Chronicle, Known Also as the Chronicle of Zuqnin: Part III, trans. W. Witakowski, Translated Texts for Historians 22 (Liverpool, 1996), 74–98. 65 See, e.g., Aetius of Amida (early 6th century), Libri medicinales, 5.95, ed. A. Olivieri, Corpus medicorum graecorum 8.2 (Berlin, 1950), 80–82; Stephen the Philosopher (6th–7th century), A Commentary on the Prognosticon of Hippocrates, 1.17, ed. and trans. J. M. Duffy, Corpus medicorum graecorum 11.1, 2 (Berlin, 1983), 56, 62. 66 PG 28:661; see also PG 89:744–45, 748, 765–68. On the author, see J. F. Haldon, “The Works of Anastasius of Sinai: A Key Source for the History of Seventh-Century East Mediterranean Society and Belief,” in The Byzantine and Early Islamic Near East, vol. 1, Problems in the Literary Source Material (Princeton, 1992), 107–47. 67 See, e.g., Galen, De sanitate tuenda, 1.11, 15 ff, ed. H. Koch et al., Corpus medicorum graecorum 5.4, 2 (Leipzig-Berlin, 1923), 27: healthy air is oJ mht∆ ejk limnw'n h‘ ejlw'n ajnaqumia´ sew" ejpiqolou´ meno" . . . ou”tw de` kai` o”sti" e“k tino` " ojcetou' tw'n kaqairo´ ntwn h‘ mega´ lhn tina` po´ lin h‘ polua´ nqrwton strato´ pedon ejpiqolou'tai, mocqhro` " iJkanw'" ejsti. 68 Notitia, 7.13, 8.19, ed. Seeck, 234–35, 238; Berger, “Regionen,” 382–83. 69 Theophanes, Chronographia, ed. de Boor, 1:370 (trans. Mango and Scott, 517); cf. McCormick, “Bateaux de vie,” 64.

PAUL MAGDALINO

219

disposal took place at Sykai, the suburb north of the inlet. According to Procopius, bodies were piled inside the towers of the fortification; according to John of Ephesus, vast pits were dug to receive the corpses oozing pus and putrefaction. Either way, rot would have been rapid in the summer heat and the stench carried by the prevailing north wind to the south shore of the Golden Horn overpowering. Altogether, there is reason to suppose that the plague was a strong incentive to move business and residence to the Marmara coast. In this connection, it is worth noting that the worst outbreaks of plague recorded in the sources, the first in 542 and the last in 747, occurred under the same emperors, Justinian I and Constantine V, who are said by the Patria to have transferred commercial facilities away from the Golden Horn.70 Whatever it was that caused the shift to the Marmara coast, when and why did the Golden Horn re-emerge as the city’s main commercial artery? It is clear that a decisive moment was the establishment of the Italian trading quarters in an area corresponding roughly to what had been the commercial harbor in late antiquity.71 We know about these trading quarters partly from Greek and Latin narrative histories but most importantly from documentation in the archives of Venice, Genoa, and Pisa concerning the real estate grants the city communes received from a succession of Byzantine emperors.72 The pattern of grants was set by Alexios I Komnenos in 1084 when he granted Venice a wharf, a church, a mall (embolos), and houses close to Perama, the embarkation point for the ferry to the northern suburb of Pera. Pisa and, much later, Genoa followed with similar acquisitions further to the east, near the ancient ports of Neorion and Prosphorion. In the course of the twelfth century, each city requested and received additional grants of property, extending their original enclaves both inland and along the shore—though not, it must be emphasised, so far that their properties adjoined. A document of 1192 also reveals the existence of an Amalfitan presence closely associated with that of Pisa: a wharf that became included within the Pisan section of the waterfront and a quarter inland from the Pisan enclave.73 The events of 1204 and 1261 led to further changes beyond the scope of this paper. There can be no doubt that the establishment of the Italians increased the commercial importance of the Golden Horn. But would the Italians have asked for concessions in this area if it had not been fairly important already to their business interests? The 70 See above, pp. 212, 215; for the plague of 747, see Theophanes, Chronographia, ed. de Boor, 1:422–43 (trans. Mango and Scott, 586–87), and Nikephoros, Patriarch of Constantinople, Short History, ed. and trans. C. Mango, CFHB 13 (Washington, D.C., 1990), 138–41. 71 For what follows, see Magdalino, Constantinople me´die´vale, 78–90, and the important study of A. Berger, “Zur Topographie der Ufergegend am Goldenen Horn in der byzantinischen Zeit,” IstMitt 45 (1995): 149–65, which appeared after my book had gone to press. 72 Venice: Urkunden zur ¨alteren Handels und Staatsgeschichte der Republik Venedig, ed. T. L. F. Tafel and G. M. Thomas, vol. 1 (Vienna, 1856); new edition of the early charters by M. Pozza and G. Ravegnani, I trattati con Bisanzio, 992–1198 (Venice, 1993); a private document of 1184 of topographical interest is published by S. Borsari, Venezia e Bisanzio nel XII secolo: I rapporti economici (Venice, 1998), 154–56. Pisa: Documenti sulle relazioni delle citta` toscane coll’oriente cristiano e coi Turchi fino all’anno MDXXXI, ed. G. Mu ¨ ller (Florence, 1879; repr. Rome, 1966), 46–49; the Greek version also in MM 3:16–23. Genoa: A. Sanguineti and G. Bertolotto, “Nuova serie di documenti sulle relazioni di Genova coll’Impero bizantino,” Atti della Societa` ligure di storia patria 28 (1896–98): 337–573. 73 Mu ¨ ller, Documenti, 47–49, 56–57; Magdalino, Constantinople me´die´vale, 86; the same conclusion was reached independently by Berger, “Ufergegend,” 161.

220

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

Venetians and Amalfitans had been active in the empire’s trade since the tenth century, and possibly earlier, so it is likely that the neighborhoods where their quarters were located in the twelfth century were those that they had always frequented. There are two indirect indications that the area’s recovery was well under way by the time Alexios I formalized the grant of the Venetian quarter in 1084. One is the information that in 1056 the emperor Michael VI proposed to renovate the Strategion, perhaps indicating that this square was now more than the pig- and sheep-market that figures in the Book of the Prefect.74 The other, more useful, indication lies in the twelfth-century documents concerning the Italian quarters. These documents give details of property ownership in the area at the time when the Italians were granted their premises. As I have attempted to show elsewhere, the pre-existing proprietors—that is, those whose premises either adjoined or were acquired by the Italians—consisted overwhelmingly of religious institutions founded or refounded in the tenth and early eleventh centuries.75 The almost complete absence of earlier foundations is striking. Assuming that the endowments were made at the time of the foundations, it is reasonable to suppose that they reflect the profitability and availability of commercial and residential real estate in lower Golden Horn neighborhoods in the period 900–1050. In other words, this was an area where rent-producing property was still available for endowment purposes in the tenth and eleventh centuries, because prior to then it had not been sufficiently valuable. If this reading of the evidence is correct, it means that the beginnings of the revival of the lower Golden Horn area coincided generally with the arrival of the Venetians and Amalfitans on the trading market of Constantinople. Yet, as we have seen, in the tenth century the market was still oriented primarily toward the Harbor of Julian and the Sea of Marmara. So why did the Italians apparently not operate in this area? What were the incentives, or the constraints, that made them base their operations on the north coast? The answer is to be sought, I believe, in a consideration of the independent evidence for several foci of commercial and maritime activity beside the Golden Horn. The settlement at Pera (the ancient Sykai, to the north of the inlet) created, at the very least, a demand for ferry services to and from Perama; from the mid-eleventh century, if not earlier, the Jewish quarter, with its tanneries, was situated at Pera.76 East of Perama, on the south coast, was the Neorion, the naval dockyard, which generated business building, servicing, and supplying the imperial fleet; in this context, we should note that Venetians and Amalfitans were, according to Liudprand of Cremona, engaged in the empire’s armed forces in the 960s.77 At Perama itself, an important focus was the mitaton of the Saracens at Perama, near the ancient church of St. Eirene and the hospital of Markianos. This mitaton is first attested in 1203 as the site of a mosque.78 Although this mosque may well have been the one inaugurated—or Skylitzes, ed. Thurn, 482; Liber praefecti, ed. Koder, 122–25; Magdalino, Constantinople me´die´vale, 51–52. Ibid., 83. 76 D. Jacoby, “Les quartiers juifs de Constantinople `a l’e´poque byzantine,” Byzantion 37 (1967): 168–73; repr. in idem, Socie´te´ et de´mographie `a Byzance et en Romanie latine (London, 1975); see also idem, “The Jews of Constantinople and Their Demographic Hinterland,” in Mango and Dagron, Constantinople and Its Hinterland (as above, note 17), 224–25. Jacoby does not, in my opinion, offer convincing proof that the Jews had not been confined to Pera prior to 1044. 77 Liudprand of Cremona, Relatio de legatione Constantinopolitana, 45, ed. and trans. B. Scott (Bristol, 1993), 17, 46; Magdalino, Constantinople me´die´vale, 83. 78 Choniates, ed. van Dieten, 1:553–54; Magdalino, Constantinople me´die´vale, 88. 74 75

PAUL MAGDALINO

221

renewed?—in 1188–89 at the insistence of Saladin, there is no reason to assume that the hostel and entrepot for visiting Arab traders with which it was associated, as the name mitaton clearly indicates, were of such recent creation.79 In other words, the mitaton of the Saracens at Perama can plausibly be identified as one of the mitata used by the Syrian merchants mentioned in the Book of the Prefect,80 and is likely to have dated from the very beginnings of Muslim trade with Constantinople. The Venetians and Amalfitans first developed their international trade by exporting slaves to Muslim North Africa. By the time they became active in Constantinople, they had a history of trading with the Arab world that would have linked their business interests with those of visiting Arab traders. The proximity of the Saracen mitaton to the Venetian quarter cannot be coincidental. A final key point on the Golden Horn was the Leomakellon, whose location and significance have recently been highlighted by Albrecht Berger.81 Whether or not the Leomakellon had anything to do with the emperor Leo I, whose name became attached to it in urban folklore, the second part of the word indicates the existence of a macellum— a market for meat and possibly other products—at a coastal site in the area of modern Unkapanı.82 This market can plausibly be identified with an agora mentioned by Procopius near the church of St. Akakios,83 and, perhaps, with the market called Basilike mentioned by a Russian traveler in 1390.84 Whether or not this second identification is correct, there are other indications to confirm that an important retail market was located beside the Golden Horn throughout the Middle Ages, well to the west of the area where the economic hub of early Constantinople had been and where the Italians were based. First, the main food markets of the city in Palaiologan times were located along the Golden Horn between Blachernai and Perama.85 Second, John Tzetzes alludes to a perfumer/druggist workshop at the Leomakellon market.86 Third, the stretch of coast near St. Akakios was known, from at least the tenth century, as the Heptaskalon, meaning “seven skalai,” which suggests that this was a particularly active port area.87 Finally, and perhaps most important, it can be inferred from the Genoese documentation that the most sought-after stretch of waterfront was to the west of the Venetian concession. When the Genoese were negotiating the terms of their treaty with Manuel I, their ambassador was instructed as follows: You will ask for and strive by all means to obtain an embolos and skalai in Constantinople between the embolos of the Venetians and the palace of the Despot Angelos. And if you don’t manage this, then in “Perforo” [Prosphorion]. And if not there, then in some other The assumption is made by S. V. Reinert, “The Muslim Presence in Constantinople, 9th–15th Centuries: Some Preliminary Observations,” in Studies on the Internal Diaspora of the Byzantine Empire, ed. H. Ahrweiler and A. Laiou (Washington, D.C., 1998), 140–43. 80 Liber praefecti, 5, ed. Koder, 94–97; Reinert, “Muslim Presence,” 131 ff. 81 Berger, “Ufergegend,” 152–55. 82 It was thus situated at the center of the most densely populated area of the 5th-century city: see above, note 68. 83 Procopius, De aedificiis, 1.4, 26. 84 G. Majeska, Russian Travelers to Constantinople in the Fourteenth and Fifteenth Centuries, DOS 19 (Washington, D.C., 1984), 150. 85 N. Oikonomides, Hommes d’affaires grecs et latins `a Constantinople, XIIIe–XVe sie`cles (Montreal, 1979), 97– 100, 106. 86 Ioannis Tzetzae epistulae, ed. P. A. M. Leone (Leipzig, 1972), 85–86. 87 Berger, Untersuchungen, 464–68. 79

222

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE convenient place within the city of Constantinople. If indeed you are in no way able to obtain an embolos and skalai within the city you should strive to obtain them in Pera.88

This passage offers a unique insight into the commercial preferences of the Italian merchants. As we might expect, they wanted to be in the city rather than the suburbs. Less expected is the revelation that the location where the Genoese eventually received their trading quarter, the old harbor area of Prosphorion,89 was not their first but their second choice. Their first choice was for a location between the Venetian quarter and the palace of Constantine Angelos.90 This building is not otherwise known, but it is clear from the context that it must have been a prominent landmark near the sea walls. If it had been east of the Venetians, it would have been between them and the Pisans; the distance was not great, and neither Venetian nor Pisan documents refer to such a palace. It is therefore likely to have been west of the Venetian quarter. This likelihood is strengthened by the consideration that there is no evidence for large aristocratic palaces on the coast in the lower Golden Horn area: the other palaces mentioned in the Italian documents, the palace of a sebastokrator and that of Botaneiates, both lay inland, on the hills to the south of the original Venetian and Genoese quarters, respectively. On the other hand, we do know of one aristocratic oikos in the neighborhood of the Leomakellon—the one occupied in 1056 by the proedros Theodosios, cousin of the late Constantine IX Monomachos.91 Theodosios’s senior title, and the fact that he considered himself to have a right to the throne, suggest that his house was a fairly grand affair—grand enough, perhaps, to have been a fitting residence, a century later, for the son-in-law of Alexios I. It is now clear that the decline of the Golden Horn area in the early Middle Ages was by no means absolute, and that the picture presented earlier of a concentration of maritime traffic and wholesale business at the Harbor of Julian must be qualified by the evidence for continuing foci of commercial activity on the north coast of the city. The Italians sought concessions in this area because it was good for business, and they were already doing business there well before the late eleventh century. As their presence increased, it undoubtedly made the Golden Horn busier than the Harbor of Julian. I think it is revealing that in Ptochoprodromos’s satire on the hegoumenoi, the poor novice who is sent to go shopping for the senior monks goes to the forum, to the Milion, to the Venetians, and to ta Eugeniou, a location on the lower Golden Horn.92 This does not mean, however, that the Harbor of Julian was deprived of business overnight, and we should not forget that when the Venetians and Amalfitans started trading at Constantinople in the tenth century, it remained, so far as we can tell, the most important port area. If I am right in thinking that the Italians developed their business interests at Constantinople through association with the Arabs and their mitaton at Perama, they gravitated to the Golden Horn because this was the place for foreigners to trade. Furthermore, the Genoese ambassador’s instructions show that there was a ranking of commercial locations Ed. Sanguineti and Bertolotto, “Documenti,” 346. This location is confirmed by the 1170 description of the quarter as being “in positione locorum Onorii,” a toponym clearly deriving from the Thermae Honorianae of the 5th century: Notitia, 6.7, ed. Seeck, p. 233. 90 See Magdalino, Constantinople me´die´vale, 80–83. 91 Skylitzes, ed. Thurn, 481. 92 Ptochoprodromos, ed. and trans. H. Eideneier, Neograeca Medii Aevi 5 (Cologne, 1991), no. IV, lines 120–21, 450, 456, 571. 88 89

PAUL MAGDALINO

223

along the lower Golden Horn, rising in value from east to west. It is no surprise that of the three major Italian trading communes in the twelfth century, it was Venice, the longest established and most privileged, that enjoyed the best position. But this does not mean that the Venetians were the most highly favored of all the rentiers along the coast. We should remember that the Italians obtained their concessions at the expense of former property-holders, who were dispossessed in their favor. Thus, what they received depended not only on what they wanted but also on what vested interests would be affected by satisfying them. This may be why we encounter no major eleventh- or twelfthcentury foundations, and no members of the imperial family, among the owners whose properties were granted to the Italians—we tend to see proprietors whom the emperor could afford to offend. Sometimes the properties in question needed major investment, like the burnt-out houses granted to the Pisans and Genoese in 1192.93 The one occasion on which the Italians received a politically sensitive grant of highly lucrative real estate was in 1189, when Isaac II, under pressure from his Venetian patriarch Dositheos, and desperate to secure Venice’s support against any threat from the Third Crusade, granted Venice’s request to be given the emboloi and skalai in the possession of the French and Germans.94 We do not know where these properties were situated, but it was probably to the west of the existing Venetian quarter. Nor do we know when and why the French and Germans acquired these lucrative concessions, but we can guess that it had something to do with the importance of France and Germany in the Crusades and in the international diplomacy of the Komnenian emperors.95 Indeed, when we consider the care that John II and Manuel I had put into trying to form marriage alliances with the French and German royal dynasties, we may suppose that the properties in question were at least as valuable as those originally granted to Venice. What we can conclude with confidence, I think, is that the twelfth-century documents concerning the Italian concessions on the Golden Horn present a picture of urban neighborhoods that had revived after a long period of depression but that, although prosperous, were not yet the most prosperous parts of the city. With this in mind, let us now consider the texture of these neighborhoods. The first point to note is that the holdings of all three Italian cities lay on either side of the sea wall. From 1148, the intra- and extramural sections of the Venetian quarter seem to have been treated as a continuum, but the Pisan and Genoese documents maintain a clear distinction between properties pertaining to the embolos inside the wall and the skalai, or landing-stages, outside the wall. By 1192 the Pisans acquired properties up to the wall on either side, but in 1201 the Genoese were still seeking to join up their separate blocs, and although they received further concessions from Alexios III in 1202,96 he evidently declined their ambassador’s request for the monastery between their embolos and their skalai or the church that separated their embolos on the southern side from the aristocratic palace granted to them in 1192.97 In general, it seems that trading quarters were not granted en bloc but as composEd. Mu ¨ ller, Documenti, 47 (Greek text), 56 (Latin text); Sanguineti and Bertolotto, “Documenti,” 443. I trattati, ed. Pozza and Ravegnani, 105–10; for Dositheos, see Choniates, ed. van Dieten, 1:404–5. Cf. D. M. Nicol, Byzantium and Venice (Cambridge, 1988), 115–16. 95 See P. Magdalino, The Empire of Manuel I Komnenos, 1143–1180 (Cambridge, 1993), chap. 1. 96 Sanguineti and Bertolotto, “Documenti,” 475–76 (Greek text), 483–84 (Latin text). 97 Ibid., 470. 93 94

224

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

ite packages of individually negotiated units. Of these units, the most complex were the skalai. A skala comprised not only a quayside, formed of an earth embankment fronted by a wall of wooden piles, but also a fenced rectangular terrain built over with houses, workshops, and the booths of money-changers. Otherwise, the standard unit of transfer was the single house (oi“khma, rendered in Latin as habitaculum) or cluster of houses. Thus, when the documents refer to a commune’s requesting or possessing an embolos, this is shorthand for acquiring a section of frontage on a street that itself belonged to the state or, in the case of a small unroofed passage in the Genoese quarter, to a neighboring monastery.98 Apart from the city wall, the main lines of division within the neighborhoods were the water courses and drainage channels coming down to the sea, and the streets. Between the walls and the shore in the Pisan and Genoese quarters, and, no doubt, in the Venetian area as well, ran a public road (dhmosi´a oJdo` ") that bisected the strips of land belonging to the skalai. Inside the wall, the main spatial feature of each quarter was the embolos, also on an east-west axis. This was evidently a covered portico: the earliest description of the Genoese quarter specifies that another embolos, belonging to a local monastery, was “unroofed.” 99 The description of the Venetian quarter in 1148 also mentions a transverse embolos, and the Pisan and Genoese documents mention alleys (rJumi´de") running in both directions. The existence of east-west emboloi in each of the three Italian quarters raises the question whether these were not, in fact, sections of a single covered portico running parallel to the coast. The question is well founded, in view of the following passage in the account in the Patria of Constantine’s foundation of the city: Also, he built four emboloi with masonry vaults from the palace as far as the land walls. One went by the Tzykanisterion and the Magnaura and the Acropolis and ta Eugeniou and extended as far as St. Antony’s; the other went by way of the Daphne and the Sophiae as far as Rabdos; the other two emboloi went by way of the Chalke and the Milion and the Forum to the Tauros, the Ox, and the Exakionion.100

Again, this seems to be a case of the Patria making fanciful sense of a visible reality; here, a series of emboloi aligned with the sea walls on both coasts and looking as if they were meant to form a continuum. We find a trace of the southern line of emboloi in the Book of Ceremonies, in the mention of an embolos in front of the church of St. Panteleemon beside the Harbor of Julian.101 This may well have been the curved portico, built by Julian, that gave access to the harbor.102 Although the Pisan and Genoese quarters must have been close to the Strategion, this square is not mentioned in any of the documents. Indeed, apart from the city wall, they mention only one local landmark known from other sources, the hospital of St. Markianos. All the well-known churches and monasteries that owned property in these neighborhoods were themselves located in other parts of the city. Churches and monasteries actually located in these neighborhoods are not otherwise known, confirming the Ibid., 364–65: prescriptus absque tecto parvulus embolus pertinet et idem monasterio tu Apologothetu. Ibid. 100 Scriptores, ed. Preger, 148. 101 De cerim. 2.13, ed. Reiske, 1:561. 102 Zosimus, Historia nova, 3.11.3, ed. F. Paschoud, vol. 2.1 (Paris, 1979), 25; cf. Berger, “Regionen,” 361. 98 99

PAUL MAGDALINO

225

suspicion that the majority of modest religious foundations in Constantinople have gone unrecorded.103 A few open spaces are recorded in the Pisan quarter, but on the whole the neighborhoods the Italians moved into were dense concentrations of oikemata/habitacula. These varied considerably in use and in architectural form. Some were purely residential, while others were partly or wholly occupied by workshops. Many houses had specific luxury features familiar from descriptions of imperial or aristocratic residences: a reception/ dining hall (triklina´ rion), a solarium (hJliako` n), a chamber (kou´ bouklon). It is notable, however, that no building was more than two storeys high, in a city where a house of three or more storeys was a recognized mark of social distinction. It is also notable that interior courtyards are rarely mentioned, in interesting contrast to the evidence for Thessalonike, where the urban properties described in the documents of Athonite monasteries were generally grouped around courtyards.104 Where supporting columns are mentioned, these were invariably wooden. In general, the impression is of fairly modest buildings of recent construction. Considering that these were trading quarters, the number and variety of businesses mentioned is surprisingly low. Most numerous were the booths of money changers (numulerarii/katalla´ ktai): eight in the intramural part of the Venetian quarter in 1148, four on the Pisan skalai in 1192, and one on the Genoese part of the waterfront in the same year. There were bakers in all three quarters, a butcher and a tavern in the Genoese section, and three candlemakers among the Venetians—but these were providers of basic everyday necessities, and their equivalents were no doubt found in every urban neighborhood. The only businesses that dealt in a specialized product were the workshops of the oarmakers, which gave the area colonized by the Genoese its name, Koparia. It is not clear whether these workshops supplied commercial shipping or, as I rather suspect, the galleys of the imperial fleet. What is clear, however, is that the Italian emboloi were not important markets or manufacturing areas. Things may have been different on the landing-stages, where several ergasteria are listed, but unfortunately the documents do not say what they produced. As for the residential functions of the neighborhoods, we can only assume as a probability that the oikemata of the Italian quarters were actually occupied by Italians. There was a Latin baker named Walter in the Pisan quarter, but all the lay tenants and the lay neighbors of the properties acquired by Genoa were Greeks: Kaparina, the widow of the exarch Alexios, John Pastos, Leo Strobiliates, John Rapsommates, Makrogenes, the Opsikianos brothers, the widow Eudokia, the head of the (palace?) goldsmiths (a“rcwn tw'n crusoco´ wn) Kyriakos, Eudokios.105 The documents present a picture of Byzantine urban neighborhoods into which Italian traders were moving or had moved recently. The properties they describe in detail are those that the Italians had just acquired. The documents do not, therefore, illustrate the extent to which the neighborhoods were being transformed by the Italian presence, Cf. Magdalino, Constantinople me´die´vale, 63. E.g., Actes de Docheiariou, ed. N. Oikonomides, Archives de l’Athos 13 (Paris, 1984), no. 4. See in general E. S. Papagianni, Morfe` " oijkodomw'n kata` th` n u”sterh Buzantinh` peri´odo: plhrofori´e" ajpo` nomika` e“ggrafa (Athens, 1995), 40 ff. 105 Sanguineti and Bertolotto, “Documenti,” 475 ff. 103 104

226

MARITIME NEIGHBORHOODS OF CONSTANTINOPLE

and we should certainly resist the temptation to visualize them in terms of the more plentiful evidence for the Venetian and Genoese colonies of later centuries. The description of the Pisan quarter in 1192 provides, however, one revealing detail: it mentions two big churches, one of St. Peter and one of St. Nicholas, that the Pisans had built since becoming established in 1112.106 These churches, which had not existed before, may well have been built in Tuscan Romanesque style. That the Byzantine imperial chancery described them as large suggests that they loomed over the neighborhood in a way that the average middle Byzantine church would not have done. They were, therefore, highly visible symbols of the wealth and power of the Latin West that was imposing itself on the other cultures of the Mediterranean world. The sight of them might have done much to inflame the mob that massacred the Latins of Constantinople in 1182, “that race of Latins who, in accordance with ancient custom, were set apart on the shore of the Horn of Byzantion, in the area of Phosphorion.” 107 University of St. Andrews, Scotland Mu ¨ ller, Documenti, 47–48. Both churches had existed—though not necessarily in the same form—since 1162, and that of St. Nicholas is mentioned in 1141: ibid., 10.4. 107 Eustathios of Thessalonike, La espugnazione di Tessalonica, ed. S. Kyriakides, trans. S. Rotolo (Palermo, 1961), text repr. with same pagination and English trans. by J. R. Melville-Jones (Canberra, 1988), 34–37. “Phosphorion is evidently a variant of Bosporion,” a name sometimes applied to the Prosphorion harbor: Berger, Untersuchungen, 424. 106

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Medieval Constantinople. The approximate locations of Italian trading quarters are indicated by V ( Venetians), A (Amalf itans), P (Pisans), and G (Genoese).

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Recent Work on the Land Walls of Istanbul: Tower 2 to Tower 5 METIN AHUNBAY AND ZEYNEP AHUNBAY

uilt during the reign of Theodosius II (408–450), the seven-kilometer land walls of Constantinople are one of the remarkable accomplishments of antiquity. The fortification was deemed invincible because of its superior design, the strength of its building materials, and details of its construction. The imagination and amount of manpower involved in the design and execution of this magnificent work is stupendous. Even in their ruinous state, the wall sections and towers bear testimony to the grandeur of fifthcentury military architecture. Today, the Theodosian walls stand as a landmark delineating the western border of the ancient city. The inhabitants of Istanbul are privileged to have a cultural heritage of outstanding importance and beauty, but preservation of such a gigantic monument is a difficult task, requiring substantial financial resources and expertise. The municipality of Istanbul is the authority responsible for the land and sea walls surrounding the historic city. As early as 1939, with the first urban development plan for Istanbul, the land walls and their environs were designated as a conservation area. A large, green belt bordering the historic city on the west was envisaged in the master plan. After UNESCO’s designation of the historic quarters and monuments of Istanbul as a World Heritage Site in 1985, the municipality initiated a project for the preservation of the land walls and their environs.1 The first step was to free them from accretions. This included the transfer of the tanneries near the southern end of the land walls to another site and the demolition of dreary buildings on or near the walls. After pursuing several lawsuits, the municipality of Istanbul succeeded in removing industrial plants from the land wall conservation area and demolished the tanneries. Thus in 1991, the blighted area near the shore of the Sea of Marmara was cleared, revealing a 200 meter section of the wall that had been concealed for nearly a century. Major changes had taken place around this part of the land walls since the nineteenth century (Fig. 1). About 1870, the railroad was introduced into the historic city, breaking

B

1 During the first campaign, which took place between 1987 and 1989, the Belgrade, Silivri, and Mevlevihane Gates were restored according to projects developed by the Foundation for the Preservation of Turkey’s Monumental, Environmental, and Touristic Assets. The second campaign started in 1991, and several teams worked along the land walls, Golden Horn, and sea walls until 1994. The project presented here is one of the six.

228

LAND WALLS OF CONSTANTINOPLE

through the wall between Towers 6 and 8 (T6 and T8). At the beginning of the twentieth century, Kazlıc¸es¸me,2 an industrial district outside the city walls, started to grow toward the east. Factories were built very close to the wall, and some parts of the outer wall were damaged or totally lost during this process. Further changes were introduced in the late 1950s, during construction of the motorway along the Marmara coast of the historic center. The new road passed between the Marble Tower (the first tower of the sea wall) and the first tower of the land walls. Due to this construction, the coastal character of Tower 1 and its connection to the Marble Tower were lost. A comprehensive survey of the land walls was carried out in the 1920s and 1930s and published in 1943 by B. Meyer-Plath and A. M. Schneider,3 but since the tanneries prevented access to the wall between Tower 1 and Tower 5 (T1 and T5), there was almost no documentation of this sector in their work. After clearance of the area in 1991, it was possible to examine the state of the walls and towers. Since it had been restored in 1988, T1 was in a good state of preservation. Towers 2, 3, 5, and 6 had serious structural problems. The main wall was preserved almost to its full height between T1 and T6, but the outer wall had suffered severely. Nothing from the moat was visible above ground, except some merlons near T5a (Fig. 2). The wall surfaces had been stained by dyes or chemicals used in tanning; several cavities had been sunk in the wall surfaces and flues had been dug into the wall mass. A thick layer of earth was deposited above the rampart walk and towers, providing sufficient soil to support fully grown trees. The rundown appearance of the site was an eyesore at one of the main entrances to the historic city; its repair had priority for archaeological and cityscape reasons. The work started in 1991.4 Vegetation and the earth deposits above the rampart walk were carefully removed and 1:50 scale surveys (plans, cross-sections, and elevations) produced. The rampart floors were examined and documented with 1:20 scale plans and cross sections. On the eastern side of the wall, access to the towers was barely possible, because earth and debris had accumulated to a height of about three meters. Excavation was conducted near the towers until the original ground level was reached. To provide complete documentation and understanding of their structure, research continued within the towers—revealing interesting details about their history. 2 Kazlıc¸es¸me is an extramural settlement founded in the 15th century. It takes its name from a fountain with a marble plate depicting a bird (or goose?). 3 A. van Millingen, Byzantine Constantinople: The Walls of the City and Adjoining Historical Sites (London, 1899), is one of the important early works on the subject. B. Meyer-Plath and A. M. Schneider, Die Landmauer von Konstantinopel (Berlin, 1943), furnishes short descriptions, historical and epigraphical data, and some architectural drawings. A concise presentation, including a list of publications after 1943, is found in W. Mu ¨ llerWiener, Bildlexikon zur Topographie Istanbuls (Tu ¨ bingen, 1977), 286–300. C. Foss and D. Winfield, Byzantine Fortifications: An Introduction (Pretoria, 1986), 41–77, contributes a substantial analysis of a new chronology for the land walls of Constantinople. 4 The Technical University of Istanbul (ITU) was appointed to prepare a restoration project for the wall section T1–T6. The present authors, M. Ahunbay (architect and archaeologist) and Z. Ahunbay (conservation architect) from the Faculty of Architecture directed the team, which comprised experts from several ¨ rmeci and O. Mu ˘lu from the Faculty of disciplines. Photogrammetric surveys were prepared by C. O ¨ ftu ¨ og Civil Engineering (I˙ TU). The structural engineering team consisted of M. Yorulmaz and F. C ¸ ılı. N. Toydemir and E. Gu ¨ rdal provided expertise in materials science. A large team of architect-restorers (D. Sarı, D. Maz˘an, M. Demirer), architects, and students of architecture contributed to the lum, A. Kangu ¨ lec¸, H. Erdog preparation of the architectural surveys and restoration proposals.

METIN AHUNBAY AND ZEYNEP AHUNBAY

229

CONSERVATION METHODOLOGY The main objective of the restoration work was to strengthen the existing fabric of the walls, so the towers would suffer less from climatic factors and future earthquakes. The selection and use of materials compatible with the originals were an important aspect of the project, requiring research into the original mortars, brick sizes, and types of stones used, as well as repair phases. Towers could not resist earthquakes; tremors caused serious vertical or diagonal cracks or led to total failure. After earthquakes some towers were totally renewed, others restored, preserving remaining wall sections. Different phases of construction and repair can be distinguished by the building materials and techniques used, and they provide a means for deriving a relative chronology. Inscriptions on some of the towers help in dating the repairs. Conservation of old repairs along with the original fifth-century fabric posed problems that had to be handled with care.5 MATERIALS AND REPAIR TECHNIQUES The main wall is ca. 4.85 meters thick and was fortified by towers placed approximately 50 meters apart.6 The curtain wall rises to a maximum height of about 12 meters. It consists of a rubble core, confined between two shells built of squared blocks, 30–50 centimeters deep. Facing stones in the fifth-century masonry exhibit fine quality, nearly ashlar, with very close joints of 0–10 millimeters. On the city side, the stone blocks are dressed and cut with less care, and the joints are generally wider than 1 centimeter. Brick bands consisting of five courses were laid at intervals of between seven and eleven courses of dressed stone. They run through the entire thickness of the wall, binding the structure firmly at different levels. The thickness of the brick bands ranges between 38 and 40 centimeters. The facing blocks of the fifth-century structures were made of a cream-colored sandstone quarried from the area outside the southern end of the land wall. These quarries are no longer in operation;7 they were exhausted in Byzantine times by continuous use. Therefore, a matching sandstone had to be provided from another source. In accordance with the results of laboratory analysis of several samples, sandstone from Kandıra (Bithynia) was selected for use in repairs. Bricks measuring 37/38 by 37/38 by 4.5–5 centimeters had been used in the fifthcentury construction.8 However, within the wall core bricks 35/36 centimeter square were 5 The key to understanding the phases of repairs has been worked out by Meyer-Plath and Schneider, Landmauer, 22–26, and by Foss and Winfield, Byzantine Fortifications, 52–70, with however, different datings for some phases, cf. note 22. 6 All the measurements given in the text refer strictly to the section studied for the restoration project. 7 ¨rkiye Mermerleri ve I˙ns¸aat Tas¸ları (Istanbul, For these stones and quarries, see M. Sayar and K. Erguvanlı, Tu 1962), 21–27. 8 Early buildings of Constantinople down to Hagia Sophia frequently use large-sized bricks; cf. J. B. WardPerkins, “Notes on the Structure and Building Methods of Early Byzantine Architecture,” in The Great Palace of the Byzantine Emperors, Second Report, ed. D. Talbot Rice (Edinburg, 1958), 76. E. Reusche, “Polychromes Sichtmauerwerk byzantinischer und von Byzanz beeinflußter Bauten Su ¨ dosteuropas” (diss., University of Cologne, 1971), 40. The physical properties of Byzantine bricks in Istanbul, including samples from the land walls, were investigated by Y. Kahya, “Physikalische und mechanische Eigenschaften der Mauerziegel byzantinischer Bauten in I˙ stanbul,” in Werkstoffwissenschaften und Bausanierung, Materials Science and Resto¨blingen, 1993), 3.3:1878–95. ration (Ehingen bei Bo

230

LAND WALLS OF CONSTANTINOPLE

in abundance; they made for easier and quicker work. Smaller and/or thinner bricks were employed during Byzantine repairs and renovations. The original mortar was basically a mixture of lime, crushed brick, and brick powder, the latter providing a hydraulic quality that simple lime mortar lacks.9 During the later Middle Ages, lime-based mortars were used as well. For restoration the composition of mortar mixes was investigated and new mixes with matching color and composition were developed by laboratory research. CURTAIN WALL BETWEEN TOWER 1 AND TOWER 2 Restoration work started in 1992 with the repair of the section between T1 and T2. Tower 1 and the eastern (city) side of T1–T2 had been repaired in 1988 by the metropolitan municipality, but the presence of tanneries had made it impossible to work on the western (field) side. We had to work on the western side and the rampart level. The fifth-century fabric of the lower part of the western elevation was preserved, but the upper sections had been subject to alterations in the Middle Ages. We repaired the cavities that had been dug into the lower section of the wall, using a technique similar to the original work. Each missing stone course was carefully rebuilt with squared blocks and backed with rubble and mortar. The repair work in the upper part of the wall was more complex, since the masonry fabric was not homogeneous due to repairs at various dates. Missing or decayed parts of the facade were rebuilt using their original construction techniques, in an effort to make the small patches conform. The efficacy of chemical and mechanical cleaning techniques was tested on stained areas. However, nondestructive methods did not prove very successful, due to the deep penetration of dyes in the stone, and, finally, no attempt was made to clean the surfaces of walls and towers. During the restoration in 1988, a crenellated parapet had been added by contractors to the top of the wall in the section next to T1, although there is not enough evidence to prove the presence of such a battlement at this point. Our investigations in the area close to T2 disclosed the traces of two battlements, which were similar in size and construction to the battlements between T3 and T8. With the help of information obtained from the battlements between T3 and T4 and between T4 and T5, the traces of these two battlements around T2 were consolidated and partially restored, making their position and form legible on the rampart walk. By studying the better-preserved sections of the rampart walk, it was also possible to restore the brick-tiled floor, giving it an inclination toward the city side. TOWER 2 The second tower (T2) presented serious structural problems (Fig. 3). Its octagonal form had been damaged by earthquakes and mutilation by the tanneries. A passageway 9 The traditional Turkish term applied to this type of mortar is “khorasan,” see S. Akman, A. Gu ¨ ner, and I. H. Aksoy, “The History and Properties of Khorasan Mortar and Concrete,” in Turkish and Islamic Science and Technology in the Sixteenth Century, Proceedings of Second International Congress on the History of Turkish and Islamic Science and Technology, 28 April–2 May 1986 (Istanbul, 1986), 1:101–11; A. Gu ¨ lec¸ and A. Ersen, “Characterization of Ancient Mortars: Evaluation of Simple and Sophisticated Methods,” Journal of Architectural Conservation 4.1 (1998): 56–67.

METIN AHUNBAY AND ZEYNEP AHUNBAY

231

had been bored through the west side to provide for easy access into the tower from the tanneries. Having lost its outer facing and weakened by cracks, the masonry on the south side had become loose and dangerous. More than a quarter of the dome covering the ground floor had been lost. At the northern side, the upper half of the tower was partially intact, with scanty remains of an upper dome. The upper parts of the southeastern, southern, and southwestern walls were totally lost. The tower stood as an impressive ruin that merited preservation as it was. Yet, its structural instability necessitated measures against further losses. In order not to change the behavior of the historic structure, insertion of foreign elements during the strengthening operation was restricted to a minimum. The weak parts of the wall were stabilized and the missing portion of the lower dome was restored, returning the tower’s structural integrity. Vertical cracks on the walls were stitched, and a reinforced concrete beam was inserted on top of the northern side of the tower, to hold the tall structure together. Entrances to the towers are generally from the east, through vestibules crowned by brick arches, resting on attached piers (Figs. 4, 6). As observed in many of the towers along the main circuit of the land walls, the collapse of the vestibule arch had resulted in dangerously cantilevered wall sections above the entrance to T2. Considering the difficulty of preserving the stone blocks in position, the vestibule arch was reconstructed using bricks produced especially for this restoration. On its northern side, the tower had a loophole that had been unrecognizably mutilated by the tannery. Cleaning and excavation within the window niche provided evidence about its original form and size (Figs. 5, 7). It measured 29.5 to 30 centimeters wide at the facade, and the lateral walls splayed out into a wide niche in the wall thickness. Vestiges of a similar window were found on the west elevation of the tower. During the restoration of the western facade, the outer facing was raised to the base of the dome, and the western window was partially reconstructed, using the northern one as a model. Apparently the lower dome of T2 was a later addition, for it blocked the loopholes (Fig. 8). The present dome shell rests on a setback of the inner cylinder, above which the tower has a wider interior diameter. Originally, the setback must have supported a timber floor flush with the sill of the window niche (Fig. 9). Thus the loopholes did not serve as light sources for the ground floor but were apertures for archers or arrow- or boltshooting machines positioned in the upper room. At a higher level on the northwest facade, enough had remained to prove the presence of a window, filled in at a later date with rubble stone (Fig. 10). Its brick vault is preserved on the inner side of the dome. It probably accommodated heavier artillery, such as a large catapult. Sockets for timber beams just below the base of the upper dome and at the same level as the window are indications of a second room with a timber floor. The position of the window at the northwest side of the tower shows that two or, at most, three artillery machines were employed in this room, which was confined within the hemisphere of the original dome. Access to this floor is from the rampart walk, through a passage leading down to the chamber (Figs. 8, 9). The passage was covered by an inclined vault; remains of its brickwork are visible in the masonry. On the ground floor of the tower, a large number of limestone and marble blocks were found in a pile that rose to 1.5 meters. Some of the pieces were Corinthian capitals and impost blocks with crosses on their sides. The pile appears to have resulted from a

232

LAND WALLS OF CONSTANTINOPLE

repair operation, during which the marble blocks were arranged in such a way that they surrounded three sides of a small rectangular area (Figs. 6, 8). The fact that the original entrance level from the vestibule was maintained during this operation suggests that the purpose was not to raise the floor level but to strengthen the structure. The exterior facing of T2 is in harmony with the fine masonry of the fifth-century land walls. The interior masonry, however, is not uniform, and the upper parts are of inferior workmanship. Vestiges of fine masonry were preserved only in the lowest part of the interior wall, close to the entrance. The variation in the masonry is probably an indication that T2 was repaired after a disaster, possibly the earthquake of 447. A need for rapid rebuilding must be responsible for the crude masonry of the interior wall surfaces. A new series of earthquakes (862, 864, or 989) must have produced the large vertical cracks through the lower-level artillery slits on each of the tower’s cardinal sides (Fig. 3).10 The piling up of material on the ground level might have been a means of stiffening the structure after the second catastrophe. The timber construction of the first and second floors was demolished during the renovation of the tower. The ground level was vaulted over by a brick dome, and a new window was introduced above the tower entrance. The vaulting of the vestibule was also changed, as the semidome over the vestibule was replaced by a flat ceiling supported by timber beams. The history of the tower after this intervention remains obscure; there is no information about its later history, except in the very late period, when the floor level was raised in accordance with its immediate surroundings. Excavation within the tower revealed that after the collapse of the lower dome, its debris filled the rectangular area in the center. This collapse might be dated roughly to the end of the nineteenth century, when the tower became a total ruin and was used occasionally by the people growing vegetables in the gardens next to it (Fig. 1). CURTAIN WALL BETWEEN TOWER 2 AND TOWER 3 The curtain wall between T2 and T3 had been seriously damaged on its eastern side. Only in the middle section of the wall was its original facing preserved from ground level to the top; the rest suffered from stone removal and serious damage caused by root action. A peculiar intrusion in the form of a trapezoidal opening disrupts the regular course of the curtain wall at a point close to T2 (Fig. 11). This gap is visible from the east, while a thin wall screens it from the field side. The breach was probably a later defensive measure intended to isolate a strategically important section.11 The slanting walls of the opening Recorded earthquakes in Constantinople up to 1000 A.D. are listed with sources and modern bibliography in E. Guidoboni, Catalogue of Ancient Earthquakes in the Mediterranean Area up to the Tenth Century (Rome, 1984); for a geophysical interpretation of earthquakes, N. N. Ambraseys and C. F. Finkel, “Long-Term Seismicity of Istanbul and of the Marmara Sea Region,” Terra Nova 3 (1991): 527–39, with a list of earthquakes from 32 A.D. to 1912 and bibliography. The tectonic structure of Istanbul as a source for frequent tremors ˘ini Olus¸turan Tektonik recorded in the historical times is assessed by A. Barka, “I˙ stanbul
un Depremsellig ¨lgelendirme Denemesi,” in I˙stanbul ve Deprem Sempozyumu, 4 Mayıs 1991 Yapılar ve I˙ stanbul ic¸in bir Mikrobo (TMMOB I˙ ns¸aat Mu ¨ hendisleri Odası I˙ stanbul S ¸ubesi) (Istanbul, n.d.), 35–56; according to his estimations the most devastating earthquakes in Istanbul recur every 200 to 250 years, less hazardous ones in 100-year cycles. It is possible to deduce that the tremors of 447, 740, and 989 were major ones. 11 As suggested by Robert Ousterhout. 10

METIN AHUNBAY AND ZEYNEP AHUNBAY

233

exhibit a different masonry technique from that used elsewhere. Bands of three brick courses alternate regularly with five courses of squared blocks, sheathing the exposed rubble core of the curtain wall on either side of the interruption. This new rhythm in the refacing is extended to the field side of the curtain wall toward T2. A very hard, pink mortar pointing was used in the joints around the squared blocks, enclosing them in rectangular frames. The regular alternation of brick and squared block courses suggest Palaiologan style. A similar fabric was used in the upper level of Mermer Kule (the Marble Tower), which has been tentatively identified with a Palaiologan princely residence.12 Consequently, the wall section up to T2 can be considered as an extension of Mermer Kule.13 During examination of the wall walk between T2 and the V-shaped opening, cavities left by deteriorated timber beams and their missing traverses were partially uncovered under the much-repaired and disturbed floor tiles. They are probably remains of a palisade built to protect the rampart walk in front of T2 against attack from the field and from the northern side of the V-shaped gap. This might be a further indication of the defensive nature of the intrusion. Excavation on the ground level along the eastern side of the main wall revealed the remains of double stairs leading to the rampart walk above T3 (Figs. 12, 14). This is an element not documented in Meyer-Plath and Schneider’s survey. Only the lower parts of the stair and the springings of the vaults supporting the stairway were preserved in situ. They provided information about the original ground level at this side of the wall, as well as details about the stair construction. The masonry of the stairs was not bonded with the main wall, perhaps explaining the loss of the upper parts. The vault supporting the stairs was made of brick, while several large stone blocks were laid out to form the steps. No attempt was made to reconstruct the whole structure; only the missing blocks of the steps were replaced, and the ruin was stabilized. An interesting detail was observed at the point where the south wall of T3 joined the curtain wall. One of the brick bands, third from the top, has eight brick courses (Fig. 13). The unusual number of brick courses at this point is explained by the meeting of brick bands from two adjacent structures, curtain wall and tower. The lower three courses of the band belong to the south side of T3; the two courses directly above are shared by both structures. This awkward transition from the tower to the curtain wall and the abrupt interruption of some brick courses after they join the curtain wall seems to have been intended to achieve a strong bond between the tower and the curtain wall—contrary to the advice of the military theoretician Philo of Byzantium,14 who advised against bonding two elements of a fortification in order to avoid possible damage resulting from differences in settling. Brick bands act as both stiffeners along the walls and bonding agents between the facing blocks and the wall’s rubble core. Inserting brick courses, even parU. Peschlow, “Mermerkule—Ein spa¨tbyzantinischer Palast in Konstantinopel,” in Studien zur byzantin¨r Horst Hallensleben zum 65. Geburtstag (Amsterdam, 1995), 93–97. ischen Kunstgeschichte. Festschrift fu 13 A similar V-shaped gap interrupts the course of the main wall close to the Yedikule fortification and apparently isolated the Golden Gate, which was reorganized as a defensive citadel in the Palaiologan period, C. Mango, “The Date of the Anonymous Russian Description of Constantinople,” BZ 45 (1952): 380–85, esp. 383 f; also, Peschlow, “Mermerkule,” 96; Meyer-Plath and Schneider, Landmauer, 41 f. 14 For Philon’s Poliorketika, see Y. Garlan, Recherches de poliorce´tique grecque (Paris, 1974), 298, 362 (62a–63); cf. translation and comments of A. C. Lawrence, Greek Aims in Fortification (Oxford, 1979), 84, 85. 12

234

LAND WALLS OF CONSTANTINOPLE

tially, into the masonry of an adjoining structure established a strong bond. Whether it was adequate is less certain. TOWER 3 Tower 3, which is rectangular in plan, was in a very poor state of preservation. Its southern wall, half of its west wall, and the barrel vault over the ground level were completely lost. The north wall had a dangerous cavity running through its entire thickness (Fig. 15). Excavation within the tower revealed an unexpected feature: the lower parts of three piers within the fifth-century walls. The piers were probably part of a medieval renovation program, during which the original barrel vault was not rebuilt but, instead, replaced with a new vault resting on the piers. The wall above the gateway leading from the vestibule to the interior of the tower had largely collapsed. The gateway arch had to be rebuilt as part of the reconstruction of the eastern wall. A reinforced concrete beam was inserted above this arch to relieve the load of the masonry above. The original barrel vault over the ground floor had left its mark on the eastern wall; during restoration, its contour was retraced (Fig. 17). Although the northern wall of the tower had large areas dating from the Middle Ages, the evidence indicated that the large cavity on this elevation was limited to a portion from the fifth century (Fig. 15). During the repair, the original masonry techniques were used to rebuild the missing courses (Fig. 16). Only half of the western wall was preserved (Figs. 14, 18). Its stone and recessed brick construction indicated a repair from the eleventh and twelfth centuries. By excavation, it was possible to reveal the southwest corner of the tower and a small portion of the south wall, which had fifth-century fabric intact. During restoration, the southwest corner of T3 was raised enough to project above the ground, so that the full perimeter of the tower would be visible, revealing its rectangular plan (Fig. 19). CURTAIN WALL BETWEEN TOWER 3 AND TOWER 4 The city side of this wall section had a nearly horizontal band of damage within the fifth-century fabric (Fig. 20). Near the entrances to T3 and T4, the damaged surface widened, due to the collapse of the crowning arches of the vestibules and loss of their supporting piers. The wide band of missing courses, where the depth of the lost masonry was almost one meter, was repaired using material and workmanship similar to the originals. The upper levels of the wall had late Palaiologan repairs, which are distinguished by their roughly hewn stone blocks and insertion of some brick fragments into horizontal joints of the brick bands. This wall section was repaired using similar block sizes, surface texture, and pointing technique as the original. The stone surfaces of the western elevation of T3–T4, were relatively well preserved, except in some parts damaged by smoke and tar from tannery chimneys. The deteriorated blocks were replaced with finely cut and chiselled new stone blocks of equal size. The upper part of the northern half of the curtain wall, a hastily repaired section, where irregular stone material had been employed, bulged dangerously. By grouting, the detached area was bonded to the intact core. On the rampart walk of this section, all of the

METIN AHUNBAY AND ZEYNEP AHUNBAY

235

battlements except one were relatively well preserved. This offered the chance to study the battlements in more detail. Two types of battlements have been identified on the main line of the land walls; one with a T-shaped plan, the other, a crenellated parapet. The latter, which is obviously of a later date, generally crowns the repaired sectors of the towers between T1 and T6. The former has been claimed to represent the earliest form of the battlements on the wall.15 The following paragraphs detail the main features (with some variations) of the T-type battlements observed between T3 and T6. The floor of the rampart walk is the uppermost course of a brick band, and this is reflected in the elevations as the last course of the topmost brick band. When viewed from the field side, one sees ashlar courses rising on top of this brick band (Fig. 22). They represent the exterior face of a parapet wall with a thickness of 55–60 centimeters and a height of approximately 80 centimeters. The brick merlons on top of the parapet are, in fact, the heads of T-shaped battlements. The leg of the T is constructed of a mixed masonry composed of brick and stone, and buttresses the merlon with its full height. The arms of the T rest partially on the floor of the wall-walk. An embrasure of 80–110 centimeters is reserved between two juxtaposed battlements for fighters. Decorated marble spolia used within the masonry of the battlements provide evidence that the T-type battlement was not part of the original construction. At points between T4 and T5 and T3 and T4, where the T-type battlements were severely disturbed, the stone parapet yielded information about the earlier construction. The parapet wall has inward projections, each measuring ca. 60–62 centimeters wide and 50 centimeters deep. The fine finish of the blocks indicates that the whole inner face of the parapet was exposed to view. Though the projections were preserved only randomly, it was possible to measure the distance between two successive pilasters, connected by an unbroken line of parapet construction, as 2.50 meters (Fig. 21). The role of the attached piers can be interpreted in two ways. Originally, the battlements could have been composed of a simple crenellated parapet with piers as traverses (the present parapet would then be the part remaining after the crenellations had been demolished). Each traverse would buttress a merlon 1.70–1.90 meters wide, with an embrasure 0.60–0.80 meters between merlons. If, on the other hand, the wall had a closed battlement, the extant parapet would represent the lower part of a high screen wall, which would have protected the walk and the fighters. Between two traverses, the screen would have had an embrasure. The traverses or wall piers could have equally buttressed the screen and joined with free-standing piers on the rear side of the walk or with the neighboring traverses to support a vault. The main wall would then have created a closed battlement and a vaulted gallery affording a second storey for defense. This reconstruction yields a battlement level with features common with the outer wall (Fig. 24).16 The walk level of the outer wall has a closed battlement and is roofed by a vault (Fig. 32). The lowest part of the screen wall consists of dressed sandstone blocks, surmounted by brick masonry with embrasures, which were at a later date converted to loopholes. The screen is buttressed by wall piers, set 1.50 meters apart. Corresponding to the but15 16

Meyer-Plath and Schneider, Landmauer, 28, 82 with notes on curtain wall T6–T9. Ibid., 33 f and 85, fig. 21 with variations.

236

LAND WALLS OF CONSTANTINOPLE

tresses are free piers along the rear edge of the walk. A second level of defense was supported by these vaults. The plan, taken at the base of the screen, below the brick masonry, shows the construction to be similar to the older parapet of the main wall not only in design but in some dimensions as well. In addition to the analogy with the outer wall, further evidence from the wall walk supports the existence of a covered battlement on top of the main wall. Within the brick floor of the rampart walk, were uncovered thick flagstones with rectangular sockets (Fig. 23). They were observed mainly between T2 and T3, but several others have been detected and documented elsewhere, up to T6. These were placed very close to the parapet, but their positions were not quite in line with the embrasures between the T-type battlements. It can, however, be inferred from the matching spans of the parapet projections and stone slabs that the flagstones were evenly positioned between the parapet piers and were part of the earlier battlement system. The perforations in the flagstones extended as far as the five-coursed brick floor, that is, approximately 40 centimeters vertically. The sockets and perforations probably served to fix the vertical supports of catapults employed on the walk level. Although it is not certain which type of catapult engine was used, these weapons could have been heavier than those employed in the outer wall.17 The evidence from the rampart walk leads to the conclusion that up to eighteen artillery engines were positioned between the traverses of the parapet. Based on the model of the outer wall, it can be further assumed that the guns and their shooters (ballistrarii) were secured against hostile missiles by a screen wall with apertures for sighting and shooting—that is by a closed battlement (Fig. 24). Since it was highly desirable to keep the torsion engines roofed, the preference would have been to protect the artillery on top of the Theodosian fortification with a vault supported by the attached wall piers.18 Since they were constructed directly on the floor tiles, the wall piers at the rear of the wall walk have disappeared without trace. Abandoning the slender structure with two shooting stories can be attributed to disuse of artillery in later centuries or simply to the structure’s vulnerability to earthquakes.19 A difficulty arises, however, in the discrepancy between the heights of T2 and the two-storey battlement. The estimated height of the two-storey battlement reaches the level of the crenellated platform of T2 (Fig. 24). A similar height for both features can be considered, but this deprives T2 of a tower’s normal dominating effect. Bearing in mind the fact that T1 has two storeys with domes, one wonders whether T2 originally According to E. W. Marsden, Greek and Roman Artillery: Historical Development (Oxford, 1969), 195, and idem, Greek and Roman Artillery: Technical Treatises (Oxford, 1971), 234 ff, only arrow and bolt-shooting torsion machines were used by the late Roman army, and the stone throwers were one-armed torsion engines, i.e., onagers. 18 Lawrence, Greek Aims, 375. The height of T2 was the dictating element in the reconstruction proposal. 19 Intact galleried curtains are still preserved in the Theodosian walls, however, with different functions, construction, and dimensions: for Pteron, the northern end of the defense line, see Meyer-Plath and Schneider, Landmauer, 102, pl. 42, for the sea wall near the Boukoleon Palace, Mu ¨ ller-Wiener, Bildlexikon, 314, fig. 360. For some other late Roman walls with a gallery in their curtains, see I. A. Richmond, The City of Wall of Imperial Rome (Oxford, 1930), fig. 3 and p. 67 f, for the Maxentian phase. The fortification of Amida (modern Diyarbakır) has a non-continuous gallery with shooting apertures: A. Gabriel, Voyages arche´ologiques dans la Turquie orientale (Paris, 1940), 1:107, fig. 78; see also M. van Berchem and J. Strzygowski, Amida (Heidelberg, 1910), 283, fig. 226; so probably Mayyafaˆrikıˆn (modern Silvan), Gabriel, Voyages, 107 nn. 2, 3. 17

METIN AHUNBAY AND ZEYNEP AHUNBAY

237

also had another room above its upper dome.20 No other evidence offers proof for this proposed two-storied wall top.21 Further research on other parts of the land walls, during the course of restorations, may provide new finds pertaining to the problem presented here. TOWER 4 Tower 4 is one of the best preserved along the main wall circuit, having been almost totally reconstructed after an earthquake. The inscription band near the top of the tower bears the name of Emperor Romanus, dating it to the second half of the tenth or the first half of the eleventh century.22 After clearing the entrance to the tower, excavation continued inside to the level of the original floor (Fig. 25). A postern connecting the tower to the peribolos (the inner terrace between the main and outer walls) was revealed. At ground level, interior of the tower is circular in plan, while the upper chamber is octagonal, with deep recesses on each side (Fig. 26). Nothing remains from the original floor covering of the upper chamber. Curiously, there is a circular hole in the center of the floor. The most serious damage to the structure is a diagonal crack that starts above the window level on the western facade and runs down about 6 meters after turning 45 degrees from the horizontal. The external stairway leading to the platform level of T4 was partially preserved; in fact it is one of the rare well-preserved sections in the main circuit of the land walls. The floor had enough of its stone paving in situ to give an idea of the original design. The central part of the floor was higher; the slabs were inclined toward the sides of the octagon (Figs. 27, 28). Small waterspouts placed under the parapets helped to keep the crenellated platform dry. CURTAIN WALL BETWEEN TOWER 4 AND TOWER 5 Between Towers 4 and 5 the western elevation of the wall was relatively well preserved, except for the upper part, which had been disturbed by the growth of trees. The fifth-century fabric could be followed up to the top of the wall. The eastern elevation, however, had been seriously damaged. Near the rampart level, areas repaired in the later Byzantine period had been detached from the wall. Sections about to collapse were carefully surveyed, and their individual stones and bricks numbered. Then they were dismantled and reassembled, using the original material. Meyer-Plath and Schneider, Landmauer, 71 f, notes on T1. Some features of Hellenistic fortifications, such as Philon describes, seem to have been reevaluated and reorganized into an efficient system in the Theodosian walls, see M. Spieser, “Philon de Byzance et les fortifications pale´ochre´tiennes,” in La fortification dans l’histoire du monde grec, Actes du Colloque International. La fortification et sa place dans l’histoire, Valbonne, December 1982 (Paris, 1986), 363–68. Whether a galleried artillery level on the curtains is also among Philon’s recommendations is a disputed issue, cf. Garlan, Poliorce´tique, 347 (A 17b), 348 (A 19), 359 (A 46a); and Lawrence, Greek Aims, 77 (1.17), 370. 22 Romanus in the inscription had been identified as Romanus III Argyrus (1028–37) by Meyer-Plath and Schneider, Landmauer, 72, 124 no. 4a, but as Romanus II (959–963) by C. Foss, “Anomalous Imperial Inscriptions of the Walls of Constantinople,” in Studies Presented to Sterling Dow, Greek, Roman and Byzantine Monographs 10 (Durham, 1984), 77–87, esp. 80 f. 20 21

238

LAND WALLS OF CONSTANTINOPLE

In the northern half of the wall the battlements were preserved almost to their full height (Fig. 29). Also significant was the inclined rampart walk, which was partially preserved. Starting at a point near T4, the wall walk had an upward inclination toward the north, as the curtain ascended a mild slope. The floor was not uniform throughout T4– T5, but had level and inclined sections, following changes in the slope. Observations revealed that the inclined planes in the walk were constructed by a rubble fill over the last brick band of the wall. A single layer of brick over the fill served as paving. Some of the T-type battlements between T4 and T5 stood on the inclined floor of the walk. Their bottom courses were adapted to the rise in the wall walk; in the construction of the first two courses, stone blocks with different heights were used to correct for irregularity in the bedding courses caused by the slope. TOWER 5 Rectangular Tower 5, which had lost its floors and roof, stood as an empty shell. The interior was filled with earth and debris up to half its height. In the course of restoration, the vestibule was cleaned and the arch over the entrance repaired. It is unfortunate that the work had to stop before the survey could be completed. The inscription fragment in T523 should not be considered as a secure base for dating, since it is used as building material in a repair section. Two other fragments, not belonging to the former inscription, are incorporated in the same repair work. Another inscription block with letters AUTOKPATOP (Fig. 30) was unearthed during our work in 1992 in the peribolos between T4 and T5 (closer to the latter and now in the Archaeological Museums of Istanbul). The block is probably not from T4; it might have fallen from another tower between T2 and T6 or have been brought from elsewhere. WORK ON THE OUTER WALL The outer wall is positioned about 16 meters to the west of the main wall.24 Its towers are placed so that they stand midway between the main towers, thus supporting the defense from the main towers at a lower level. Nothing was visible of the outer wall near T1 and T2. Because of the reluctance of the municipal administration to permit excavation within the new park, it proved impossible to trace the remains of the undocumented south end of the outer wall. The outer wall becomes visible above ground starting at a point in front of T3 in the main wall. Tower 3a in the outer wall does not exist today, but it must have stood until the 1940s, since it is marked on Jacques Pervititch’s map from 1939 (Fig. 1). By examining the outer wall closely and taking measurements it was possible to detect traces where the lateral walls of T3a bonded with the outer wall. Excavation was conducted to find the foundations of T3a, but this attempt was in vain because at a depth of about 1.5 meters the reinforced concrete foundations of the former tanneries were encountered. Meyer-Plath and Schneider, Landmauer, 124, no. 6. A new date for the construction of the outer wall, in 447, was provided by a recently recovered inscription, H. Kalkan and S. S ¸ahin, “Ein neues Bauepigramm der theodosischen Landmauer von Konstantinupolis aus dem Jahr 447,” Epigraphica Anatolica 23 (1994): 145–56; W. D. Lebek, “Die Landmauer von Konstantinopel und ein neues Bauepigramm,” Epigraphica Anatolica 25 (1995): 107–54. 23 24

METIN AHUNBAY AND ZEYNEP AHUNBAY

239

From T3a to T4a, the outer wall is only partially preserved (Fig. 31). The extant parts include the piers and the screen wall (with windows) as well as a latrine next to T4a.25 Tower 4a is rectangular and has lost its floor and platform. There is a postern leading down to the lower terrace on the tower’s north wall. A survey of the outer wall and Tower 4a was conducted, and a conservation plan developed (Fig. 32) but it could not be implemented, as work came to a close at the end of 1993. As part of the World Cultural Heritage, the Theodosian wall (Fig. 33) deserves treatment with utmost care. We hope that, in the future, it will be possible to continue our research and preservation work up to the line of the railroad, as was foreseen at the start of the project. Istanbul Technical University 25 Similar corbel stones for a latrine are preserved from the walls of Resafa: Karnapp, Stadtmauer von Resafa, 11.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Map showing the southern end of the land wall and the industrial development around Kazlıçe¸s m e in 1939. Archive, Istanbul Technical University, Faculty of Architecture ( J. Pervititch, 1939)

2

Istanbul, general plan of the land walls work site, Tower 1 to Tower 6, 1993

3

Istanbul, tower 2 from the west, 1991

5

4

Istanbul, entrance to Tower 2, 1991

Istanbul, remains of the loophole at the north wall of Tower 2, 1992

6

Istanbul, Tower 2, ground-level plan, 1992

7

Istanbul, Tower 2, plan of upper chamber, 1992

8

Istanbul, Tower 2, east-west section, looking north, 1992

9

Istanbul, Tower 2, east-west section, looking north, reconstruction drawing

10

Istanbul, Tower 2 from the northwest, after restoration, 1993

11

Istanbul, trapezoidal opening on the eastern side of the wall between Tower 2 and Tower 3, before restoration, 1992

12

Istanbul, remains of double stairs on the eastern side of the wall between Tower 2 and Tower 3, 1992

13

Istanbul, brick band with eight courses at the juncture of Tower 3 with the curtain wall, 1992

14

Istanbul, plan of Tower 3, before restoration, 1993

15

Istanbul, north wall of Tower 3, before restoration, 1992

16

Istanbul, north wall of Tower 3, after restoration, 1993

17

Istanbul, detail from the restoration of Tower 3, 1994

18

Istanbul, Tower 3 from the south, before restoration, 1991

19

Istanbul, Tower 3 from the south, after restoration, 1994

20

Istanbul, east side of wall between Tower 3 and Tower 4, before restoration, 1992

21

Istanbul, plan of the wall walk between Tower 4 and Tower 5

22

Istanbul, the west facade of the curtain between Tower 6 and Tower 7, 1987

23

Istanbul, a flagstone and remains of a battlement on the wall walk between Tower 2 and Tower 3, 1995

24

Istanbul, reconstruction drawing of the battlement level in the Theodosian wall

25

Istanbul, Tower 4, ground-level plan, 1993

26

Istanbul, Tower 4, plan of upper chamber, 1993

27

Istanbul, Tower 4, plan of platform level, 1993

28

Istanbul, Tower 4, east-west section, looking south, 1993

29 Istanbul, restoration project for the eastern facade of the curtain between Tower 4 and Tower 5 (implemented in 1993)

30

Istanbul, block inscribed with AUTOKPATOP, found between Tower 4 and Tower 5, 1993

31

Istanbul, outer wall and Tower 4, with the remains of stairs leading to Tower 3a, 1993

32 Istanbul, detail from the restoration project for the outer wall between Tower 3a and Tower 4a

33 Istanbul, general view of Tower 1 to Tower 4, after restoration (photo: A. Neftçi, 1994)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Contextualizing the Later Churches of Constantinople: Suggested Methodologies and a Few Examples ROBERT OUSTERHOUT

n the introduction to his handbook on Byzantine architecture, Cyril Mango warns us that Constantinople “had in the course of the Middle Ages more than five hundred churches and monasteries. Of these about thirty have survived in varying degrees of ruination, that is, less than ten percent.” 1 That meager ten percent, however, marks a considerably better rate of survival than that for other types of buildings in the Byzantine capital, most of which have vanished without a trace. What is more, we have virtually nothing resembling urban archaeology to assist us in visualizing the historic city. Obviously, we architectural historians are not playing with a full deck, and, for the uninitiated, the buildings that have vanished completely often seem more interesting than those that remain. The surviving churches have typically been studied in isolation. Although we may be able to reconstruct the complicated movements of the liturgy within the interior, we are completely lost when we step outside the front door. The churches must have been constructed with some regard to existing buildings, the street system, and the urban population—that is, they must have responded to the existing fabric of the city. But is it possible for modern scholars to somehow reintegrate these isolated structures into the context of the society that created them? In the following pages, I offer three possible approaches to the later monuments of Constantinople, which may be pursued with varying degrees of success. The first approach might be called topographical. Something similar has been employed by textbased scholars concerned with historical geography in an attempt to position the hundreds of unsecured placenames into the ever-changing jigsaw puzzle of the Byzantine city. A second approach takes history as its context. Accordingly, the evolution of the forms of an individual building over a period of time might be understood as a reflection of larger urban concerns. A third method of analysis concentrates on construction: a careful study of building fabric tells us about the masons and their working methods, and this might also help provide a human context for the churches of Constantinople. In developing these three approaches to the Byzantine monuments of Constantinople, the following paper is divided into three unequal sections. Although the discussion is

I

1

C. Mango, Byzantine Architecture (New York, 1976), 11.

242

LATER CHURCHES OF CONSTANTINOPLE

limited to a select number of examples, the methodologies proposed here could be effectively applied to numerous other buildings. TOPOGRAPHY As an architectural historian, my approach to the historical geography of Constantinople is slightly different from that of the text-based scholar. Rather than privileging the text, I begin with the building, with the physical clues preserved on the site. Fragmentary foundations, terrace walls, orientations of buildings, positions of portals, and other details can be reassembled to suggest the immediate context of the monument in question. I shall here limit myself to a single example, the site I know best, the Kariye Camii, the Byzantine monastery of the Chora (Fig. 1).2 It may be possible through a careful study of the evidence to fix some of the monastic boundaries and perhaps also the locations of other monastic buildings, as well as to set the monastery into a larger urban context. The main entrance to the Chora monastery was to the south of the church, probably directly south of the belfry, near the present parking lot for tour buses. In 1979, the remains of a cistern were visible in this area, about 60 meters from the building.3 The extrados of a domical groin vault constructed of brick was exposed before the porch of a neighborhood house, its concrete steps partly laid over and partly cut into the vault. A small trap door was apparently originally intended for the removal of water from the area below the vault. A marble column shaft lay nearby. Local tradition alleges the cistern was a tunnel leading to Hagia Sophia. Before it could be properly recorded, however, the cistern was bulldozed, in the early 1980s, to create the parking lot. Yet, it was not completely destroyed: another part of it was removed in 1996 when the parking lot was widened, and the last remaining portion is now concealed behind a retaining wall. This hapless ruin is significant because when Stephan Gerlach visited the Chora monastery in 1578, he saw in the same area a dry cistern in which the Jews were spinning and preparing silk.4 It was located by a fallen wall and the “noble gate” to the monastery. The site plan by Alexander Ru ¨ dell, drawn at the beginning of the twentieth century, shows a Byzantine wall in this area.5 A street from this point must have connected with the north Mese near the cistern of Aetius. This route of access would explain why Gerlach misidentified the Chora as the “Monastery of Aetius,” and why the troops of Fatih Mehmet were able to find their way to the Chora so easily immediately upon entering the city in 1453.6 In some unpublished drawings that Cyril Mango brought to my attention, Ernest Mamboury recorded the remains of Byzantine walls in the same area as Ru ¨ dell’s wall (Figs. 2–5). About 25 meters east of this point, Mamboury also recorded the remains of what he called the hagiasma of John the Baptist. The first drawing shows the position of the remains in relationship to the southwest corner of the Kariye, where the base of the R. Ousterhout, The Architecture of the Kariye Camii in Istanbul, DOS 25 (Washington, D.C., 1987), esp. 4–6. R. Ousterhout, “A Sixteenth-Century Visitor to the Chora,” DOP 39 (1985): 117–24, esp. 120 and fig. 3. 4 Stephan Gerlachs dess Aeltern Tage-Buch (Frankfurt, 1674), 455–56; discussed by Ousterhout, “Visitor to the Chora,” 120. 5 A. Ru ¨ dell, Die Kahrie-Dschamisi in Constantinopel: Ein Kleinod byzantinischer Kunst (Berlin, 1908), fig. 1. 6 Ducas, Historia Turco-Byzantina, 39.13–15, trans. H. J. Magoulias, Decline and Fall of Byzantium to the Ottoman Turks (Detroit, 1975), 224–25. 2 3

ROBERT OUSTERHOUT

243

belfry projects (Fig. 2, left half and bottom). Another drawing shows the details of the plan of the hagiasma, with a note concerning a marble head of St. John that had been stolen (Fig. 3). Yet another shows the transverse section, with the main access by stair to the north; the depth of the hagiasma is given as 1.90 meters below the floor (Fig. 4). The longitudinal section includes a blocked vaulted extension to the west (Fig. 5). The small, apsed chamber was covered by a domical vault and looks suspiciously like the substructure for the diakonikon of a church. In fact, this may have also been a part of the monastery, but unfortunately these ruins have disappeared with the urban development of the area. The situation of the main entrance to the Chora monastery may also help to explain certain features in the design of the church. The belfry, decorated with the monograms of the founder, seems to have been positioned in direct visual relationship to the entrance (Fig. 1), and this route of access might also explain the amount of detail devoted to the south facade and the significant position of the south portal. To the east of the Kariye Camii is a retaining wall, below the present garden wall (Fig. 6).7 It is lined with niches flanked by cut-stone corbels, and with broken masonry indicating projecting spur walls spaced at about 2.6-meter intervals. For a brief moment when I first saw this wall, I thought I had discovered the remnants of the Byzantine monastic cells, with each niche forming a fireplace within the individual cell. On the other hand, if this was an important residential area of the monastery, it would have been facing the wrong way—that is, away from the church and the functional center of the monastery. Moreover, the construction does not match anything else at the site, and it may well be later in date. Although the niches are lined with brick, the wall is of a rough stone construction, broken by the lines of reinforcing beams. Nevertheless, this terrace wall may mark the position of the eastern limit of the monastery, at the point where the ground falls sharply away. Other buildings lay immediately to the north of the Chora church, both in the Byzantine and Ottoman periods. On the north facade of the narthexes, several details are preserved: brackets above the north exonarthex door suggest that it connected to something, perhaps a portico of light construction.8 On the north facade of the inner narthex, several metal pins of the type used to anchor marble revetments are set into the wall. These only appear at this point on the exterior, and they suggest that this might have been an internal surface—that is, where another building abutted the church. To the ¨rbe, recorded to be the location of the former imaret, or northeast, lies a small Ottoman tu soupkitchen, of the Kariye mosque. As the Ottomans normally adapted Christian buildings to related Muslim functions, this may have been the location of the trapeza and kitchen of the Chora monastery.9 To the west of the church was a broad courtyard, still evident in Gerlach’s day (Fig. 7). The west wall of the monastery must have been somewhere before the land wall, for there was probably a street along the inside of the fortification. This street would have led from Adrianople Gate to the Tekfur Saray and the Blachernai palace further to the See the plan in W. Mu ¨ ller-Wiener, Bildlexikon zur Topographie Istanbuls (Tu ¨ bingen, 1977), fig. 159. Ousterhout, Architecture, 86–87 and fig. 147. 9 As suggested by C ¨ lersoy, “Kariye,” Arkeoloji ve Sanat 1.2 (1978): 18–22. ¸ . Gu 7 8

244

LATER CHURCHES OF CONSTANTINOPLE

north. There may have been a monastic gate in the northerly direction as well, considering the well-documented connections between the Chora monastery and the Blachernai. The connection of the Chora to the Blachernai and to the land walls had important symbolic overtones as well. Before the fourteenth century, the monastery had been supported by the imperial family, and its new founder, Theodore Metochites, sought to link himself as closely as possible to it. The monastery was dedicated to the Virgin, and Metochites referred to both monastery and Virgin as his protection and refuge. These ideas are expressed in the well-known image of the Virgin positioned above the main entrance to the church (Fig. 8).10 She is of the Blachernitissa type, repeating the significant features of a miraculous protective icon kept at the imperial church of the Blachernai, which also housed the robe of the Virgin, the sacred palladium of the city. Both the robe and the icon were paraded around the walls when the city was under siege. In the Chora image, the Virgin’s cascading robe once framed a view westward toward the land walls, emphasizing this association. The Virgin was thus the imperial protector of both Theodore Metochites and his monastery—and she was also the protector of the city and its walls. The mosaic helps us situate the monastery, at least symbolically, into a broader urban context. Like the Zoodochos Pege and Blachernai churches, the Chora offered another spiritual outpost for the defense of the city and its walls. It was, significantly, at the Chora that the protective icon of the Virgin was housed during the final siege of the city.11 HISTORY In our discussion of the Chora, our attempt to establish its urban context took into consideration outlying remains from a variety of periods that may have been associated physically or symbolically with the church. In fact, older remains survive directly under the building as well. Substructures dated to the sixth and ninth centuries determined the position of the present building, whose superstructure may be dated to rebuildings of the eleventh, twelfth, and fourteenth centuries.12 The site was of such significance that it was reemployed repeatedly in spite of its instability, which resulted in collapses and fractures, and prompted the construction of the prominent yet ineffectual flying buttress on the east facade (Fig. 1).13 The significance of its site is not something unique to the Chora, and this fact serves to introduce a second method of contextualizing the churches of Constantinople that might be called historical. Few monuments in the capital were limited to a single phase of construction, and the process of growth and change within an individual building often reflects larger themes in the evolution of the city. For example, the site of the Myrelaion church evidences any number of urban transformations. In its first phase, it supported a late Antique rotunda of uncertain designation, probably a palace vestibule.14 R. Ousterhout, “The Virgin of the Chora,” in The Sacred Image East and West, ed. R. Ousterhout and L. Brubaker (Urbana-Chicago, 1995), 91–109. 11 Ducas, Historia, 39.13–15. 12 Ousterhout, Architecture, 12–15 and pls. 41–48. 13 Ibid., 85. 14 C. L. Striker, The Myrelaion (Bodrum Camii) in Istanbul (Princeton, 1981), 13–16 and pl. 26; R. Janin, La ge´ographie eccle´siastique de l’Empire byzantin, vol. 1, Le sie`ge de Constantinople et le patriarcat œcuminique, pt. 3, Les ´eglises et les monaste`res, 2nd ed. (Paris, 1969), 351–54. 10

ROBERT OUSTERHOUT

245

This was filled with a colonnaded cistern to form the substructure for a palatial residence in the early tenth century, and an additional substructure was added to support the adjoining private chapel. The difference in scale between the late antique palace and middle Byzantine one is clearly evident—with the whole of the latter sitting atop the vestibule of the former. The private chapel became the setting for the tombs of Romanos Lekapenos and his family, marking a change in imperial burial practices. And the whole oikos was converted to a nunnery after Romanos’ accession to the throne; the conversion of secular properties was a common occurrence in the middle Byzantine era. In the Palaiologan period, the substructure of the chapel, which previously had been unused, was outfitted as a mortuary chapel, evidence of the late Byzantine concern for privileged burials.15 The church of St. Euphemia at the Hippodrome underwent a similar evolution, reflecting in turn the urban transformations around it.16 The area of the Kalenderhane Camii, carefully studied by Striker and Kuban, had an equally long and complex history.17 The reuse and redevelopment of numerous late antique sites and establishments represent important aspects of urban continuity into the medieval period, as Paul Magdalino has recently emphasized.18 Because it was raised on tall substructures, the Myrelaion church has remained standing in splendid isolation, while almost all other churches in the capital were expanded with annexes. In the Palaiologan period, the most important constructions were additions to existing churches. This has led to the assumption that late Byzantine architectural production in the capital was restricted by the poverty of its citizens, who could no longer afford to build complete churches. Should we assume that the inhabitants of Thessalonike, Arta, Mistra, Trebizond, Messembria, and elsewhere could afford to do what the citizens of Constantinople could not? We must recall that of the more than five hundred ecclesiastical structures known from the texts, the vast majority were in existence before the Latin occupation. Although many needed refurbishing in the final centuries, the city certainly did not need more churches. Thus, we must find a different explanation for the development of the Palaiologan church complexes. Clearly, all of them provided prominent spaces for prestige burials, but I offer also some other historical explanations. The first is the significance of the Pantokrator monastery, the imperial mortuary church of the Komnenoi, that set the standards for Palaiologan architecture (Fig. 9). The church complex grew by accretion, with its three churches, serving three different functions, built in rapid succession in three phases between 1118 and 1136 (as will be discussed in additional detail below).19 In its first phase, the church consisted of a monumental Striker, Myrelaion, 29–31. R. Naumann and H. Belting, Die Euphemia-Kirche am Hippodrom in Istanbul und ihre Fresken, IstForsch 25 (Berlin, 1966). 17 C. L. Striker and Y. D. Kuban, Kalenderhane in Istanbul: The Buildings (Mainz, 1997), esp. 7–16, by A. Berger. 18 ´ tudes sur l’evolution des structures urbaines, TM, Monographies 9 P. Magdalino, Constantinople me´die´vale: E (Paris, 1996), esp. 40–48. 19 A. H. S. Megaw, “Notes on Recent Work of the Byzantine Institute in Istanbul,” DOP 17 (1963): 333–64; ´ glises, 515–23; see also R. Ousterhout, Z. Ahunbay, and M. Ahunbay, “Study and Restoration of the Janin, E Zeyrek Camii in Istanbul: First Report, 1997–98,” in this volume. 15 16

246

LATER CHURCHES OF CONSTANTINOPLE

block, measuring approximately 100 Byzantine feet on each side (Fig. 10). This contrasts dramatically with the familiar jumble into which it very quickly grew. In its final form, the significance of the Pantokrator was expressed by its complexity; separate functional spaces are clearly distinguished on the exterior, identified by their distinctive apses and domes, which were prominent features on the urban skyline. In effect, as the Pantokrator grew, complexity replaced monumentality as the primary visual expression. We might compare the Pantokrator to the contemporaneous descriptions of the mansions of the wealthy, the great oikoi of the period, which the texts claim to have been like “cities within the city” or “resembling cities in magnitude and not at all unlike imperial palaces in splendor.” 20 Palaiologan architecture followed suit, as the Palaiologoi associated themselves with the Komnenoi. The Lips monastery was expanded by the dowager empress Theodora toward the end of the thirteenth century to become their dynastic mausoleum. A south church and an ambulatory were added in two closely related phases.21 Following the model of the Pantokrator, there is little attempt at visual integration. The building complex is distinguished by an irregular row of apses along its east facade and is topped by an asymmetrical array of domes. The parts read individually, with a marked contrast between the middle and late Byzantine forms. The same contrasts are evident at the Chora, and I would argue that the lack of integration, the contrast and abrupt juxtaposition of old and new, was intentional (Fig. 1). In all of the Palaiologan complexes, the new portions may be understood as a response to history, an attempt to “build” a symbolic relationship with the past. The new additions never obscure the older edifice but are joined to it in a way that seems to respect its character. At the Chora, for example, in spite of stylistic differences, the domes over the naos and parekklesion are aligned, and the detailing of the naos apse is reflected in that of the parekklesion. Moreover, the masons seem to have been inspired by the difficulties of adding to an older building. The self-conscious, “mannerist” style of these late Byzantine complexes may result, in part, from the attempt to design around the historical core of the monastery while maintaining its integrity. In these examples, the masons would appear to be addressing not just new functional considerations but also the symbolic significance of the historical setting. It may be useful to recall Theodore Metochites’ concern with the past and with his own position in history.22 The considerations we see reflected in the architecture are the same we find expressed in the decoration of the Chora. For example, the Deesis mosaic spells out his ktetoric lineage, perhaps most obviously in the “family resemblance” between Isaak Komnenos in the Deesis mosaic and Metochites’ dedicatory image in the adjacent panel.23 These two portraits establish a sort of visual dialogue with the past that corresponds to the architectural relationships. For the oikoi, see P. Magdalino, “The Byzantine Aristocratic Oikos,” in The Byzantine Aristocracy, IX–XIII Centuries, ed. M. Angold, BAR International Series 221 (Oxford, 1984), 92–111. In fact, the organizational structure of an oikos and a monastery was similar, as Magdalino stresses (pp. 102–5). 21 T. Macridy et al., “The Monastery of Lips (Fenari Isa Camii) at Istanbul,” DOP 18 (1964): 251–315. 22 ˇevcˇenko, “Theodore Metochites, the Chora, and the Intellectual Trends See the excellent essay by I. S of His Time,” in The Kariye Djami, ed. P. A. Underwood, vol. 4 (Princeton, 1975), 15–55. 23 P. A. Underwood, The Kariye Djami, 3 vols. (New York, 1966), 1:42–43, 45–48; 2: pls. 26–29, 36–41; Ousterhout, Architecture, 96–100. 20

ROBERT OUSTERHOUT

247

The architectural sculpture in these late Byzantine churches may similarly reflect a concern with the past. Among the sculpture, we find both the prominent display of spolia and the imitation of older forms. At the Vefa Kilise Camii, for example, a mixed collection of reused closure panels, columns, and capitals decorates the facade.24 At the Lips monastery, the pattern of the tenth-century cornice is replicated in the late-thirteenth-century addition.25 Reuse and imitation are already in evidence at an earlier date at the Pantokrator, where, for example, the distinctive bird cornice of the north church follows an Early Christian model, and many of the liturgical furnishings of the south church were spolia.26 CONSTRUCTION The search for a historical context has gradually narrowed our focus from the city and its neighborhoods to the details of the individual churches. This leads to the third method of analysis, which might be called constructional. A close examination of construction techniques can help us reconstruct the architectural practices of the Byzantine period, for which little textual evidence survives. This, in turn, may place the churches of Constantinople in a human context. Because of the relatively conservative nature of Byzantine architectural design in the centuries after Iconoclasm, it seems likely that architectural drawings were not used in most construction projects, and that a minimum of pre-planning was done on site.27 This means that, unlike architecture today, design and construction were part of the same process, with the details of the elevation determined as the building rose. Indeed, this is the impression we get from Michael Psellos’ descriptions of imperial building projects of the eleventh century. At the Theotokos Peribleptos, for example, built by Romanos III, “One on top of another new parts were added, and at the same time another part would be pulled down. Often, too, the work would cease and then suddenly rise up afresh, slightly bigger or with some more elaborate variety.” 28 At St. George of the Mangana, according to Psellos, Constantine IX had the design altered and expanded three times during its construction.29 Although the tone of Psellos’s remarks reflects badly on the patrons, his descriptions may represent actual architectural practices. In fact, the final forms of many middle Byzantine buildings are the result of growth and transformation. The development of the Pantokrator monastery may best be understood in the context of a process of construction similar to that described by Psellos. Although the buildOusterhout, Architecture, pl. 162. C. Mango and E. J. W. Hawkins, “Additional Notes,” in Macridy et al., “Monastery of Lips,” 309–19 and fig. 39. 26 T. F. Mathews, The Byzantine Churches of Istanbul: A Photographic Survey (University Park, Pa., 1976), 10–41 and pls. 10–40, for which a significant Early Christian parallel was found near the Ru ¨ stem Pas¸a Camii and is now in the Archaeological Museum; Megaw, “Recent Work,” 345–46 and figs. 7–9. The pieces from St. Polyeuktos may have only been reused in the 13th century or later. 27 See discussion in R. Ousterhout, “Building Medieval Constantinople,” Proceedings of the Patristic, Medieval, and Renaissance Conference 19/20 (Villanova, 1994–96), 35–67, esp. 44–45; expanded in idem, Master Builders of Byzantium (Princeton, 1999), esp. 58–85. 28 Psellos, Chronographia, ou, Histoire d’un sie`cle de Byzance (976–1077), ed. E. Renauld (Paris, 1967), 41–46; Psellos, Fourteen Byzantine Rulers, trans. E. Sewter (Harmondsworth, 1966), 71–75. See Ousterhout, Master Builders, 86–127, for a fuller development of this theme. 29 Chronographia, ed. Renauld, 143–51; also Sewter, Fourteen Byzantine Rulers, 250–52. 24 25

248

LATER CHURCHES OF CONSTANTINOPLE

ing was completely irregular in its final form, the structure was too important for us to dismiss its form as simply the unfortunate product of an inept master mason (Figs. 9–11). The master mason, by the way, is named in an addition to the Synaxarion as a certain Nikephoros, said to be the synergates of the empress and “a new Bezalel.” 30 Thus, in spite of what we might think of his abilities, he was held in high regard during Byzantine times. Following the construction of the south church, the Pantokrator underwent a more or less continuous process of expansion and modification for more than a decade, as our recent analysis testifies. I have been working in conjunction with Metin and Zeynep Ahunbay to document the building for the replacement of the roof, and more details about the building’s present state are given in our first report.31 In the process of restoration, we have been able to make some useful observations about the building’s construction history. The south church seems to have been brought to completion before the additions were begun. It was plastered on the exterior, and significant areas of pink plaster are preserved beneath the level of the Ottoman roof. The later Byzantine additions abut this plastered surface, which was left intact (Figs. 12, 13). Evidence of the external plaster is visible where the narthex galleries of the south and north churches join—that is, between Phase 1 and Phase 2, and where the dome of the Phase 3 middle church abuts the Phase 1 south narthex gallery. Although the construction may be conveniently divided into three distinct phases, there are also indications of sub-phases, that is, of changes that were effected during the construction of each of the churches. For example, the central, funerary church has a unique twin-domed design. The two domes undoubtedly related to the double function of the middle church, divided between a liturgical space to the east and the burial area to the west. The typikon suggests as much: the east dome is called the “dome of the Incorporeal,” whereas the western area is the “heroon of the outside,” being outside the bema.32 An analysis of the construction indicates that the west dome was completed first, and then the east dome was built against it, with some unfinished surfaces where the two join (Fig. 14). The forms of the two domes are actually quite different: the west dome has a ribbed inner surface, whereas the east dome is a pumpkin dome. Where they join, windows in the drums open between them, but their sills and crowns occur at different heights. We may hypothesize that the middle church was begun as a single-domed space, but that its design was modified during the process of construction. Because the plan had already been determined, the east dome had to be constructed above an oblong bay, resulting in its unprecedented oval form. Another change of design was effected during the construction of the exonarthex. The stepped pilasters and arches of the facade were simplified as the outer narthex was joined to it. The setbacks were cut away to create an angled and slightly concave surface within each arch (Fig. 15). Moreover, sometime during the construction of the exonarthex, the masons decided to increase the height of its vaulting. On the lateral walls, remnants of lower arches are still evident, framing triple-light windows that were subsequently blocked. The outer narthex was apparently intended to have a lower roof level Megaw, “Recent Work,” 342 n. 24; G. Moravscik, La´zlo´ Iea´nya ´es a Biza´nci Pantokrator-monosotor (Budapest, 1923), 43–47. 31 Ousterhout, Ahunbay, and Ahunbay, “Study and Restoration.” 32 P. Gautier, “Le Typikon du Christ Sauveur Pantocrator,” REB 32 (1974): 81, lines 867–68. 30

ROBERT OUSTERHOUT

249

that would have corresponded with the height of the inner narthex vaults. It is unclear why this particular change was made, because it has resulted in a space that is lofty but dark, with windows positioned only in the lower walls. Moreover, this change also apparently motivated several other alterations. The most important of these was the addition of the gallery dome of the south church. The evidence indicates that this was not part of the first phase, but that it was added only after the outer narthex was constructed. Above the present roof, the arch of the central bay rises above the others, with its springing at roof level (Fig. 11). The pilasters to either side are quite broad, and the profiles of the central arcade are considerably simpler than the others, consisting of two setbacks where the others have four. Within the attic below the present concrete roof, the details of the central arcade are considerably different (Fig. 16). There, the pilasters have multiple setbacks and are set further apart; between them, a broad arch springs from a lower point. Within the arch are the setbacks for the original window openings. Just above the extrados of the exonarthex vaults, a line of pink plaster is still preserved, which extends into the window reveals. Built against this are masonry additions that correspond to the window mullions visible above the roof. When did the alteration of the central arcade occur? In the attic zone, the mullions stop above a level of unfinished masonry. This indicates that the modification only happened after the exonarthex was added, that is, in Phase 3, when the unfinished area was already covered by the exonarthex roof. The details preserved in the attic, combined with a careful analysis of the heavily reworked pilasters of the lower facade, allow us to propose a reconstruction of the original facade of the Phase 1 church (Fig. 17). On both levels, it had a flattened arcade over the broad central bay, the height of which corresponded to the other arcades, and broad windows within each bay. Intriguingly, the tall, blocky facade with arcades of the same height recalls twelfth-century Russian churches, such as the Cathedral of Vladimir. The raising and narrowing of the central arcade bay correspond to the raising of arches to the north and south—that is, to the creation of a square bay for the addition of the dome. The dome blocked the west lunette of the naos, whose painted marble mullions are still visible on the interior. But what motivated the construction of yet another dome in an already complicated building? It was undoubtedly not added in response to ceremonial or liturgical concerns. Rather, the construction of the exonarthex had blocked almost all natural light from the inner narthex on the ground floor. Apparently to remedy this gloomy situation, the dome was added on the gallery level and the floor of the gallery immediately beneath removed, creating a lightwell to illuminate the main entrance to the naos (Fig. 18).33 A similar lightwell was constructed later in the twelfth century at the Kalenderhane Camii, perhaps inspired by this example.34 At the Pantokrator, however, the lightwell was clearly the result of a step-by-step process of modification. The observations on the Pantokrator are preliminary and, doubtless, will require some nuancing after we are able to examine other areas of the building fabric. Nevertheless, they give a good indication of the number of modifications and changes in design 33 34

Megaw, “Recent Work,” 343. Striker and Kuban, Kalenderhane, 94.

250

LATER CHURCHES OF CONSTANTINOPLE

that could occur in a single construction project. If this is a reliable indication of architectural practices at the highest level of patronage, the implications are significant—for both architecture and the Byzantine rhetorical tradition. As Henry Maguire has argued in relationship to Byzantine gardens, Michael Psellos might actually be telling the truth about Byzantine construction practices, in spite of his ironic tone.35 In more general terms, the detailed technical examination of buildings may help us reconstruct workshops and workshop practices, and may assist us in tracking the dissemination of architectural ideas in a more concrete way than do the elusive “influences” and “appropriations” based on stylistic analysis that are the stock-in-trade of art historians. An often-cited example is the so-called recessed brick technique, a hallmark of Constantinopolitan architecture of the eleventh and twelfth centuries, in which alternating courses of brick are set back from the wall surface and covered by what appear to be exceedingly wide mortarbeds.36 Probably too much has been said about this technique already, but it is important because its distinctive details were invisible in a completed building, and it could have only been learned through direct participation in the construction process. These and similar details may help us identify individual workshops, a subject I have developed more fully in a recent study.37 The appearance of the recessed brick technique in such far-flung locations as Kiev and Jerusalem, for example, indicates the presence of workshops from the capital.38 Other seemingly insignificant details, such as the pattern of putlog holes left by scaffoldings built into the walls, the impressions left by formwork, and incisions in the mortarbeds may also be useful in the identification of workshops.39 To sum up, what we are able to say about a Byzantine building is often limited by the questions we ask. If our analysis of a Byzantine church is restricted to typological analysis, or to the liturgical organization, or if it is simply presented as a backdrop for historical events or in the context of patronage, we have not learned all we can from it. We should be able to read a building, just as we read a text, as a historical document, for it can tell us much about the society that produced it. The details preserved in buildings and their sites can help reconstruct a context for the Byzantine churches of Constantinople— whether that context is the surrounding neighborhood, the site’s history, or the lives of the artisans who formed a significant element of the urban workforce. All were a part of the fabric of the city. University of Illinois, Urbana-Champaign H. Maguire, “Gardens and Parks in Constantinople,” in this volume. ¨ B 28 P. L. Vocotopoulos, “The Concealed Course Technique: Further Examples and a Few Remarks,” JO (1979): 247–60, with additional bibliography. 37 Ousterhout, Master Builders, 169–84. 38 For Kiev, see H. Scha¨fer, “Architekturhistorische Beziehungen zwischen Byzanz und der Kiewer Rus im 10. und 11. Jahrhundert,” IstMitt 23/24 (1973–74): 197–224; for Jerusalem, R. Ousterhout, “Rebuilding the Temple: Constantine Monomachus and the Holy Sepulchre,” Journal of the Society of Architectural Historians 48 (1989): 66–78. 39 Ousterhout, Master Builders, 181–92; see also G. Velenis, Ermenia tou Exoterikou Diakosmou ste Byzantine Architektonike (Thessalonike, 1984), and A. Pasadaios, Ho keramoplastikos diakosmos ton byzantinon kterion tes Konstantinopoleos (Athens, 1973). 35 36

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Istanbul, Chora (Kariye Camii) seen from the southeast

2

Sketch by E. Mamboury showing the hagiasma of St. John (courtesy of Cyril Mango)

3

Sketch by E. Mamboury showing the hagiasma of St. John (courtesy of Cyril Mango)

4

Sketch by E. Mamboury showing the hagiasma of St. John (courtesy of Cyril Mango)

5

Sketch by E. Mamboury showing the hagiasma of St. John (courtesy of Cyril Mango)

6

Retaining wall to the east of the Kariye Camii

7

Plan of the Chora area of Istanbul, showing positions of monastic components (redrawn after W. MüllerWiener, Bildlexikon zur Topographie Istanbuls [Tübingen, 1977], fig. 159)

8

Chora ( Kariye Camii), mosaic of the Virgin Blachernitissa

9

Istanbul, Pantokrator (Zeyrek Camii), plan showing three phases of development, ca. 1118–36

10

11

Pantokrator, hypothetical plan of the first phase, begun ca. 1118

Pantokrator seen from the minaret, with the south church in the foreground

12

13

Pantokrator, detail of vaulting extrados showing the joint between the galleries of the south church on the left (Phase 1) with plaster surface preserved and the north church on the right (Phase 2)

Pantokrator, detail showing the joint between the gallery facade of the south church ( left, Phase 1, with plaster surface preserved) and the dome of the middle church (right, Phase 3)

14

15

Pantokrator, interior of exonarthex, looking north

Pantokrator, view of the domes of the middle church, from the southeast

16

Pantokrator, detail drawing of the central arcade of the south narthex gallery, beneath the modern roof, showing Phase 1 stepped pilasters and Phase 3 modifications in plan and elevation

17

18

Pantokrator, hypothetical reconstruction of the west facade of the south church (Phase 1) ( based on J. Ebersolt and A. Thiers, Les églises de Constantinople [Paris, 1913], pl. XLVIII)

Pantokrator, longitudinal section through south church, looking north, showing narthex dome and light well (Ebersolt and Thiers, Les églises de Constantinople, pl. XLIII)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Gardens and Parks in Constantinople HENRY MAGUIRE

he many gardens and parks of Constantinople are primarily known to historians today through descriptions in texts.1 The actual remains of gardens are scant, though not, as we shall see, nonexistent.2 As for the texts, they take many forms, from brief mentions in histories and chronicles, to longer ekphrastic descriptions in prose or in verse. On the face of it, the ekphraseis of the green spaces of the city should be our most useful source for their understanding and reconstruction, but they present us with problems of interpretation. At one time, scholars used to value the ekphraseis as repositories of factual information about lost works of art and architecture, but they were frequently frustrated by the perceived distortions and inaccuracies produced by the rhetorical genre of ekphrasis and its conventions.3 Now, a newer generation of scholars sees the ekphraseis in a more positive, but not positivist, light, as expressions of the Byzantine imagination, more relevant to mentality and ideology than to concrete reality.4 In either case, the Byzantine texts have been viewed as suspect guides to the physical reconstruction of lost monuments. My intention in this paper is to use the topic of gardens and parks in Constantinople to propose a middle way of reading the ekphraseis. I will argue

T

This paper owes much to the assistance of Robert Ousterhout and Engin Akyu ¨ rek, which is here gratefully acknowledged. 1 The modern literature on Byzantine gardens is now extensive. For a survey of the bibliography, see A. Littlewood, “The Scholarship of Byzantine Gardens,” in Byzantine Garden Culture, ed. A. Littlewood, H. Maguire, and J. Wolschke-Bulmahn (Washington, D.C., 2001). On the gardens of Constantinople, see especially A. R. Littlewood, “Gardens of the Palaces,” in Byzantine Court Culture from 829 to 1204, ed. H. Maˇevcˇenko, “Wild Animals in the Byzantine Park,” in Littleguire (Washington, D.C., 1997), 13–38; N. P. S wood, Maguire, and Wolschke-Bulmahn, Byzantine Garden Culture. On later developments, see G. Neci˘lu, “The Suburban Landscape of Sixteenth-Century Istanbul as a Mirror of Classical Ottoman Garden pog Culture,” in Gardens in the Time of the Great Muslim Empires: Theory and Design, ed. A. Petruccioli (Leiden, 1997), 32–71. 2 See C. Barber, “Reading the Garden in Byzantium: Nature and Sexuality,” BMGS 16 (1992): 1–19, esp. 1: “The subject of the Byzantine garden is primarily to be tackled from texts, and is very much a subject of its texts.” 3 See, e.g., J. Beckwith, Early Christian and Byzantine Art (Harmondsworth, 1970), 161: “Many of the poems and epigrams referring to works of art may be dismissed as cliche´s by which the scholar displayed his learning.” On the role of rhetoric, see C. Mango, “Antique Statuary and the Byzantine Beholder,” DOP 17 (1963): 55–75, esp. 64–68. 4 See, especially, L. James and R. Webb, “ ‘To Understand Ultimate Things and Enter Secret Places’: Ekphrasis and Art in Byzantium,” Art History 14 (1991): 1–17; Barber, “Reading the Garden in Byzantium,” esp. 1–5; L. James, “ ‘Pray not to Fall into Temptation and Be on Your Guard’: Pagan Statues in Christian Constantinople,” Gesta 35.1 (1996): 12–20, esp. 14.

252

GARDENS AND PARKS IN CONSTANTINOPLE

that these texts were not completely divorced from the works of art that they describe. In spite of modern depredations, enough survives in Istanbul of some Byzantine gardens to show that the descriptions of those spaces by Byzantine writers were at least partly accurate and supportive of a literal reading. At the same time, it cannot be denied that the ekphraseis were not simply objective descriptions but constructions that reflected the cultures of their composers and audiences. The second aim of my paper, therefore, will be briefly to illustrate a change in Byzantine attitudes toward gardens, and indeed landscape in general, between late antiquity and the Middle Ages. This change of outlook structured the Byzantine perceptions of garden spaces, whatever their actual form. This paper will concentrate on four parks and gardens where texts and topography can be studied together. I will begin with two suburban parks, the Philopation and the Aretai, and then move inside the city to consider the Mesokepion and the Mangana, both in the area of the Great Palace. The Philopation is the most frequently mentioned of the parks of the city.5 The earliest record of it is from the ninth century, when the future emperor Basil I killed a large wolf that had startled his hunting party by rushing out at them from some woods “in the place called Philopation.” 6 The Philopation seems to have remained in use as an imperial hunting ground until the fall of the city to the Latins at the end of the twelfth century. The best description of the area is the one provided by Odo of Deuil, the chaplain of Louis VII, who saw the park in 1147, during the Second Crusade. Although Odo’s description is well known, it is worth quoting in full: Before the city stood a spacious and impressive ring of walls enclosing various kinds of game and including canals and ponds. Also, inside were certain hollows and caves which, in lieu of forests, furnished lairs for the animals. In that lovely place certain palaces which the emperors had built as their springtime retreat are conspicuous for their splendor.7

Some of the details in this description are confirmed by Byzantine writers. George the Monk, for example, speaks of the “ravines and hollows” of the Philopation, in which Jacovitzes, the henchman of Basil I, had a fatal accident while hunting.8 Kinnamos refers to the Philopation as “an imperial dwelling place,” which “is overgrown with leaves and produces rich grass” and “bears everywhere a green appearance.” 9 The use of the suburban palace of the Philopation for imperial receptions is evoked by a passage in Choniates that speaks of Alexios Angelos’s holding court there while sitting freshly bathed on a couch spangled with gold.10 On the basis of these descriptions we should imagine the Philopation as an enclosed ˇevcˇenko, “Wild On the Philopation, see, most recently Littlewood, “Gardens of the Palaces,” 37–38, and S Animals,” who give the earlier literature. 6 eij" to` lego´ menon Filopa´ tion. Vita Basilii, 14, ed. I. Bekker, CSHB (Bonn, 1838), 231–32. 7 De profectione Ludovici VII in orientem, 3, ed. and trans. V. G. Berry (New York, 1965), 48. 8 Fa´ ragga" kai` boqu´ nou". Imperium Basilii Macedonis, 2, in Theophanes Continuatus, ed. I. Bekker, CSHB (Bonn, 1838), 839. 9 tv' katantikru` teice´ wn basilikv' gegonw` " ejndiaith´ mati . . . th` n fu´ lloi" komw'san po´ an te dayilh' ajniei'san (ajmfilafh` " ga` r oJ cw'ro" kai` ejpi´cloon aJpantach' fe´ rei to` pro´ swpon). Historiae, 2.14, ed. A. Meineke, CSHB (Bonn, 1836), 74–75; translation from C. M. Brand, Deeds of John and Manuel Comnenus by John Kinnamos (New York, 1976), 63. 10 Nicetae Choniatae Historia, ed. J.-L. van Dieten (Berlin–New York, 1975), 1:457. 5

HENRY MAGUIRE

253

park, containing not only a natural landscape for hunting but also constructed features such as canals, pools, and pavilions. Of the latter nothing survives today, but at least we know the approximate location of the park, which lay immediately outside the walls of the city. In the map accompanying Raymond Janin’s book on the topography of Constantinople, the Philopation straddles the valley of the Lycus, just outside the land walls of the city, about two kilometers from their northern end (Fig. 1).11 Other authors, however, have placed the estate further to the north, nearer to the Golden Horn.12 There is no doubt that the Philopation was close to the land walls. Kinnamos says that when Conrad reached Byzantium in 1147, he was able to survey the walls of the city from the Philopation, and even to see that the people standing on the outworks were unarmed. According to Kinnamos, Conrad decided immediately that the city was impregnable, and, being unable to take it, swiftly crossed the bridge across the Golden Horn.13 Another passage, in Niketas Choniates’ Historia suggests that the plains above the Philopation were visible from the Blachernai palace. Speaking of Isaak Angelos, Choniates wrote: Often, he [Isaak] held newly wrought arrows in his hands and said that he would have them sharpened in order to pierce the hearts of the Germans. Then he would point to a window of the palace of the Blachernai through which were visible the plains outside the battlements that were suitable for horsemanship and that sloped down to the Philopatia, and he would say that the missiles would be shot through it, to strike down the Germans, thus moving his listeners to laughter.14

From this passage it is evident that the higher ground above the Philopation could be seen from the palace at the Blachernai, at the northern end of the land walls. Presumably, the sloping plains referred to by Choniates are the ones that descend toward the north, to the flat ground alongside the Golden Horn; from the Blachernai it would have been impossible to see the slopes on the southern side, those that descend to the Lycus, as they are over the brow of the hill. The modern view from the Blachernai palace takes in a multilane expressway, but we should imagine that the view was once pleasant enough for the construction of a belvedere adjoining the palace, on top of the so-called Tower of Isaak Angelos (Fig. 2). It is debatable whether this is the very tower that Niketas Choniates says was built at the Blachernai as a residence by Isaak Angelos from the spoils of demolished churches.15 However, it is presently provided with a spacious room at the top with generously proportioned arched windows. These opened, on the south and west sides, onto balconies con11 R. Janin, Constantinople byzantine: De´veloppement urbain et re´pertoire topographique, 2nd ed. (Paris, 1964), 144–45 and plan I. 12 C. Emereau, “Le Philopation, le ‘Vincennes’ de Byzance,” EO 21 (1922): 181–85; A. M. Schneider, Byzanz: Vorarbeiten zur Topographie und Archa¨ologie der Stadt, IstForsch 8 (Berlin, 1936), 81. 13 Historiae, 2.14, Bonn ed., 74–75. 14 kai` be´ lh neoca´ lkeuta fe´ rwn polla´ ki" ejn tai'n ceroi'n taujt∆ e“fasken ajkona'n kata` kardi´a" paghso´ mena tw'n Alamanw' j n⭈ deiknu´ wn de` kai` para´ quron ejn toi'" basilei´oi" tw'n Blacernw'n, di∆ h»" oJrata` h«n ta` iJpph´ lata pedi´a ta` e“xwqen tw'n ejpa´ lxewn ejn toi'" Filopati´oi" kaquptia´ zonta, ejk tau´ th" e“lege me´ llein ajfie´ nai ta` be´ lemna, ba´ llein ´ ", w”ste kai` ejn ge´ lwti h“geto toi'" ajkou´ ousi ta` lego´ mena. Choniates, Historia, te kai` kataba´ llein tou` " Alamanou j ed. van Dieten, 1:404; adapted from H. J. Magoulias, trans. O City of Byzantium, Annals of Niketas Choniates (Detroit, 1984), 222. 15 A. van Millingen, Byzantine Constantinople. The Walls of the City and Adjoining Historical Sites (London, 1899), 143–45.

254

GARDENS AND PARKS IN CONSTANTINOPLE

spicuously supported on reused columns projecting from the walls. This mirador must have provided a fine view of the landscape of the Philopation. Further clues concerning the location of the Philopation can be found in the ekphrasis by Nikolaos Mesarites of the church of the Holy Apostles, which he composed around the year 1200. Mesarites said that someone who climbed up onto “the parapets and towers” of the Holy Apostles could see the Philopation, together with its activities of hunting and the marshaling of troops. “See,” declared Mesarites, “the ruler has gone out for the salvation of his people, and he is staying in the emperor’s tents (skenas), which stand opposite the emperor’s palace. . . . ”16 Presumably the palace opposite which the tents were pitched was the Blachernai, unless Mesarites was referring to a palace within the park of the Philopation itself. Another suburban park that we know from the written sources is the Aretai,17 which, in the words of Anna Komnena was located “near the city.” About the Aretai, she wrote: “It is elevated above the plain, and, to those standing and looking on it from below, gives the appearance of rising into a ridge, inclining one of its flanks towards the sea, and the other towards Byzantium, and the two others to the north and to the west, being exposed to all the winds.” 18 She goes on to say that the place was delightful, temperate, and continually provided with pure water, although in her day it had been denuded of trees. It had buildings, which Anna describes as splendid and worthy of emperors, who would briefly take their leisure there. She tells us that this palace had been built as a retreat by Romanos IV Diogenes, who reigned 1068–71.19 However, it seems that some kind of palace already existed on the site of the Aretai before the time of this emperor, for the poet Christopher of Mytilene, who is otherwise known to have been active in the first half of the eleventh century, devoted an epigram to a stone statue of Hercules, which he saw “standing in the palace of the Aretai.” 20 Therefore, Anna, who was writing some eighty years after the death of Romanos IV may have been mistaken, or else Romanos restored an existing palace. The palace and park of the Aretai are probably described in an 85line ekphrastic poem by John Geometres, who was active in the second half of the tenth century.21 Although the poem does not have a lemma specifically identifying the subject as the Aretai, the third line appears to have a punning allusion to the site, for it reads, 16 o”ra moi toi'nun⭈ ejxh'lqen eij" swthri´an laou' oJ kratw'n kai` parame´ nei ejpi ta` " kate´ nanti me` n tw'n ajnakto´ rwn ajpodiistame´ na" . . . Basilei´ou Skhna´ ". . . . kai` oJrato` " thnikau'ta meta` su´ mpanto" tou' strateu´ mato" tv' para` ta` " ejpalxei" iJstame´ nv tou' new' te kai` ta` purgw´ mata. Description of the Church of the Holy Apostles, 5.1–6, ed. and trans. G. Downey, TAPS, n.s., 47.6 (1957): 864, 898. 17 H. Maguire, “A Description of the Aretai Palace and Its Garden,” Journal of Garden History 10.4 (1990): 209–13. 18 to´ po" de` ou»to" ajgcou' th'" po´ lew" diakei´meno", uJperkei´meno" me` n th'" pedi´ado" kai` toi'" ka´ twqen iJstame´ noi" kai` pro` " tou'ton oJrw'sin eij" lofia` n ajnateino´ meno" kai` th` n eJte´ ran me` n pleura` n pro` " qa´ lattan ajponeu´ wn, kata` de` th` n eJte´ ran pro` " to` Buza´ ntion, tai'" de´ ge loipai'" dusi` pro` " a“rkton kai` du´ sin, panti` ajne´ mv katapneo´ meno". Alexiad, 2.8.5, ed. B. Leib (Paris, 1937), 1:90. 19 Ibid. 20 [Eij" to` n ajndria´ nta tou' JHr] akle´ w", to` n iJsta´ menon ejn tv' palati´v tw'n Aretw' j n. E. Kurtz, Die Gedichte des Christophoros Mitylenaios (Leipzig, 1903), 99, no. 143. On the dated poems of Christopher of Mytilene, see E. Follieri, “Le poesie di Cristoforo Mitileneo come fonte storica,” ZRVI 8.2 (1964): 133–48. 21 Ed. J. A. Cramer, Anecdota graeca e codd. manuscriptis bibliothecae regiae parisiensis (Oxford, 1841; repr. Hildesheim, 1967), 4:276–78. On the identification of the subject, see Maguire, “Aretai Palace.” On the career of John Geometres, see M. Lauxtermann, The Byzantine Epigram in the Ninth and Tenth Centuries (Amsterdam, 1994), 150–69. See also idem, “John Geometres–Poet and Scholar,” Byzantion 68 (1998): 356–80, esp. 376–78.

HENRY MAGUIRE

255

“The virtues [aretai] not only of earth but of all creations came together here.” 22 Like the park described by Anna Komnena, the Aretai of the poem occupied an elevated position, “rising . . . a small height from the ground.” 23 The poem describes the virtues of the site in extravagant terms, as boasting beautifully decorated buildings, sculptures more accomplished than those of Praxiteles, Phidias, Lysippos, and Polykleitos, and waterworks, including both streams pouring from the rock and formal fountains. According to Geometres, all this was set within a natural landscape that included pastures, peaks, ravines, hollows, and wooded vales. The park rejoiced in the songs of birds and cicadas, and in the leaping of game, specifically hares and roe-deer. Since the park has now vanished, a twelfth-century ceramic bowl from Corinth may serve to illustrate its charms; here can be seen the characteristic creatures of the game park, the deer, the hares, and the birds against a background of foliage (Fig. 3).24 Where, then, was the park of the Aretai located? According to A. M. Schneider it may have been situated in the area of Topc¸ular, facing the northern portion of the land walls, at the position numbered 1 on the satellite view shown in Figure 4.25 But, as Janin pointed out, such a site does not correspond well with Anna Komnena’s description, for she says that one flank of the park faced the sea, which is some distance from this point— about 4.5 kilometers. Instead, Janin proposed that the Aretai should be placed some 3 kilometers north of the ancient Hebdomon, a little more than 3 kilometers west of the southern portion of the land walls, at the position marked 2 on Figure 4.26 The matter can be resolved, I believe, by reference to another of the poems of John Geometres, which describes a fortification tower.27 In this case, we know exactly which area he is describing, because the tower of his poem can be identified as Tower 1 on the inner line of the Theodosian land walls. This tower stands at the extreme southern end of the land walls, at the point where they adjoin the sea wall.28 Geometres describes the tower as standing at the junction of land and sea.29 He also specifies that it is hexagonal in its upper parts;30 Tower 1 of the land walls is pentagonal below, where it abuts the curtain wall, but hexagonal above, where it rises clear of the wall (Fig. 5). The occasion for the composition of Geometres’ poem may have been the restoration of the land walls by Basil II and Constantine VIII after the earthquake of 989, for the poem appears to refer to this tremor,31 which, according to Leo the Deacon, brought down “the towers of Byzan22 oujk ajretai` gh'", ajlla` pa´ ntwn ktisma´ twn sunh'lqon w»de. Ed. Cramer, 276, lines 5–6. 23 kai` mikro` n ejk gh'" w”sper u”yo" hjrme´ nh. Ed. Cramer, 276, line 14. 24 H. C. Evans and W. D. Wixom, eds., The Glory of Byzantium, exh. cat., Metropolitan Museum of Art (New York, 1997), 268, no. 190. 25 Schneider, Byzanz, 81. 26 Janin, Constantinople, 443 and plan VIII. 27 Ed. Cramer, 278–80. On the identification of the subject, see H. Maguire, “The Beauty of Castles: A Tenth-Century Description of a Tower at Constantinople,” Delt.Crist.
Arc.
Et. 17 (1993–94): 21–24. 28 On the architecture of this tower, see F. Krischen, Die Landmauer von Konstantinopel (Berlin, 1938), 1: pls. 40–41; B. Meyer-Plath and A. M. Schneider, Die Landmauer von Konstantinopel. Zweiter Teil (Berlin, 1943), pls. 1, 4, 12, 25; W. Mu ¨ ller-Wiener, Bildlexikon zur Topographie Istanbuls (Tu ¨ bingen, 1977), figs. 328–29. 29 Ed. Cramer, 279, line 3. 30 Ed. Cramer, 279, line 12. 31 Ed. Cramer, 278, line 28.

256

GARDENS AND PARKS IN CONSTANTINOPLE

tium.” 32 The hexagonal upper stage of Tower 1 still bears an inscription naming the emperors Basil and Constantine.33 For our purposes, John Geometres’ poem about the tower is significant for its last two lines. They read, “Do you wish to know the whole prospect of the tower? Raise your eyes straight, look from nearby at the virtues (aretas) of the earth.” 34 If this is taken as another reference to the park of the Aretai, then the park should lie near the southern end of the land walls, near Tower 1, and within sight of it. There is, indeed, a wellwatered hill that rises into a ridge opposite the land walls at this point, in the area marked 3 on the satellite view in Figure 4. Like the site described by Anna Komnena, this hill has views of the sea and of the city on its south and east sides, respectively. The location proposed by Janin, north of Hebdomon (at 2 on the satellite view in Fig. 4), seems to be too far to the west; because of the intervening ridges, it would not be easy to see Janin’s site from the land walls. Also, it cannot be described as “nearby” to the fortifications. In conclusion, it may be proposed that the park of the Aretai occupied a position analogous to the Philopation, near the walls, but at their southern rather than northern end. It is not possible to determine when the park was first laid out, but John Geometres referred to the features he described as a “new creation”;35 this statement suggests that they should be ascribed to the second half of the tenth century. There is one other intriguing piece of evidence, kindly brought to my attention by Robert Ousterhout—namely a fragmentary relief portraying a goose, which is now incorporated into an Ottoman fountain at the base of the hill that I am identifying here as the Aretai (Figs. 6–8). According to its inscription, the Ottoman fountain, known as Kazlı C ¸ es¸me, was constructed in 1537.36 The fountain and its sculpture were both recorded by the seventeenth-century traveler Evliya C ¸ elebi in his section on the suburbs of Istanbul. Describing the pungent, but well-watered, quarter of the tanners, outside the Yedikule, Evliya C ¸ elebi remarks, “Outside this suburb is a fountain, where on a square piece of marble is engraved a goose, of admirable workmanship.” 37 This sculpture, as we see it now, is carved on two large interlocking blocks of marble that are incorporated into the back of the fountain (Fig. 6). The upper block is carved with a plump goose, standing in front of a leaf on a curved stem (Fig. 7). The surface of the lower, and larger, piece is now smooth, except for the lower part of the leaf stem, which is carved at its upper left, and the arched frame for the water outlet, the profile of which is Ottoman (Fig. 6). Even though the body of the goose has been damaged and restored, enough remains of it to show that it was related to Byzantine sculptures of the tenth and eleventh centuries. In the carving on our fountain we find a distinctive combination of naturalistic and unnatuta´ te purgw´ mata tou' Buzanti´ou pro` " gh'n kateri´pwse. Historia, ed. C. B. Hase, CSHB (Bonn, 1828), 175–76. Meyer-Plath and Schneider, Landmauer, 123. 34 Qe´ lei" to` pu´ rgou pa'n pro´ swpon manqa´ nein… Ai«re pro` " ojrqo` n o“mma, ble´ yon ejggu´ qen ta` " ajreta` " gh'". Ed. Cramer, 279, line 33–280, line 2. 35 Ed. Cramer, 276, line 3. 36 H. Yelmen, “Kazlic¸es¸me,” in Istanbul Ansiklopedisi (Istanbul, 1994), 4:512–14. 37 Evliya Efendi, Narrative of Travels in Europe, Asia and Africa in the Seventeenth Century, trans. J. von Hammer (London, 1850), 1.2:29–30. I am indebted to Antony Greenwood for this reference. 32 33

HENRY MAGUIRE

257

ralistic forms. The goose has a plump well-rounded body, but the wing feathers are rendered by means of a shallow surface decoration of stylized dartlike forms, which are terminated on the right by a curved border of drilled holes framed by buttonlike discs (Fig. 8). The combination of these forms is characteristic of Byzantine rather than Ottoman carvings. Among dated Byzantine works, for example, they can be found in the early-tenth-century sculptures from the church of Constantine Lips, the Fenari Isa Camii in Constantinople.38 Here, we can compare a plump bird with the characteristic dartlike decoration of the upper wing feathers (Fig. 9), a peacock with a similar treatment of the wing feathers and with the buttonlike discs framing the drilled holes that represent the “eyes” of the tail (Fig. 10),39 and an eagle with a curved band of drilled holes defining the upper segment of the wing, as in the carving of the goose (Fig. 11).40 Such similarities with the sculptures from the church of Constantine Lips, and also with other Byzantine carvings of the eleventh century, show that the marble relief now incorporated into the fountain is certainly not Turkish.41 But what was its original purpose? The height of the two interlocking panels makes them unsuitable to have served as parapet slabs. Indeed, it is possible that they were created to serve as part of a fountain in the first place, since, from the Roman period onward, sculptures of geese and ducks had been associated with gardens and with fountains. Examples include a pair of bronzes that decorated the peristyle garden of the House of the Vettii at Pompeii. Each bronze depicted a boy holding a large duck, from whose beak a stream of water jetted into a basin set between the two statues.42 From the late antique period there is a bronze statue of a goose, probably dating from the third or fourth centuries A.D. that was said to have been found in the Hippodrome at Constantinople and is now in the British Museum (Fig. 12). To judge from the holes pierced in the end of the bird’s beak and under its tail, this sculpture also was used as part of a fountain.43 Where did the marble goose of the Kazlı C ¸ es¸me originally come from? The simplest suggestion is from the park of the Aretai, on this very site. We know from the poem of John Geometres that the park of the Aretai was graced with “fountains, pools, and their myriad devices,” and with sculptures.44 But, of course, it is also possible that these marbles, large and heavy though they are, could have been moved here from elsewhere, 38 On these sculptures, see A. Grabar, Sculptures byzantines de Constantinople (IVe–Xe sie`cle) (Paris, 1963), 100– 122, pls. 47–56; T. Macridy, “The Monastery of Lips and the Burials of the Palaeologi,” DOP 18 (1964): 253–77, esp. 262–65, figs. 17–21, 31–83; C. Mango and E. J. W. Hawkins, “Additional Notes,” DOP 18 (1964): 304–9, figs. 8–39. 39 Mango and Hawkins “Additional Notes,” 307, fig. 13. 40 Macridy, “The Monastery of Lips,” 259, fig. 18. 41 For similar plant forms, with thick rounded stems and a raised ridge at the center of each lobe of the five-lobed leaves, see the relief of an eagle seizing a deer now in the Staatliche Museen, Berlin (no. 3250); A. Grabar, Sculptures byzantines du moyen ˆage, vol. 2, XIe–XIVe sie`cle (Paris, 1976), 95–96, no. 80, pl. 64b. For the dartlike wing feathers, compare a relief of an eagle seizing a rabbit in the Byzantine Museum, Athens (no. 157); ibid., 67, no. 56, pl. 35a. 42 W. K. Jashemski, The Gardens of Pompeii, Herculaneum and the Villas Destroyed by Vesuvius (New York, 1979), 35–36, fig. 54. 43 D. Buckton, ed., Byzantium: Treasures of Byzantine Art and Culture, exh. cat., British Museum (London, 1994), 44, no. 22. 44 krh´ nai, li´gnai, te´ cnai de` tou´ twn muri´ai. Ed. Cramer, 277, lines 5, 21–25.

258

GARDENS AND PARKS IN CONSTANTINOPLE

and the evidence of coincidences should not be pushed too far. At the least, it can be proposed that the marbles of the Ottoman fountain probably did once decorate a Byzantine fountain in or near Constantinople. The second pair of imperial gardens that I will consider, the Mesokepion and the Mangana, were inside the walls of the city. The Mesokepion was incorporated into the Great Palace. As is well known, the Great Palace was constructed on a series of terraces that descended the hill in steps from the level of the Hippodrome and Hagia Sophia, about 31 meters above sea level, down to the level of the shore. Some of the retaining walls of these terraces still survive among the houses of the modern city. The recent investigations of Eugenia Bolognesi have distinguished a total of six levels of terraces, at approximately 5-meter intervals.45 The Mesokepion was evidently built on one of the lower levels, in the eastern portion of the palace, although its exact location has not yet been determined.46 Unusually precise descriptions of the Mesokepion survive in the text of Theophanes Continuatus. While relating the building activities of Theophilos on the eastern side of the Great Palace, this text describes a terrace that he set up, “facing the north,” from which one could see the old polo grounds (tsoukanisterion). The text adds that the old polo ground “occupied the area upon which the glorious emperor Basil erected the New church, the two fountains, and the enclosed garden,” that is, the Mesokepion.47 These constructions of Basil I are described in a lengthy passage of the Vita Basilii, which takes us through the atrium of the Nea, with its two resplendent fountains of porphyry and Sangarian marble,48 through the church itself, and finally through the porticoes that extended eastward from the church on its north and south sides as far as a new pologround that Basil constructed to replace the old one. The text describes one of the two porticoes, which was reached through the north door of the church, as a peripatos, or covered walk, but the other portico, which was reached through the church’s southern, sea-facing door, is called a diaulos, or double course or promenade. The new polo field was evidently on the edge of the palace complex, because the text specifies that the emperor had to expropriate and demolish houses in order to build it.49 According to the Vita Basilii, the emperor turned the space enclosed between the two porticoes linking the east of the Nea with the new pologround into a garden, “which, on account of its position, we are wont to call Mesokepion.” The text describes the garden conventionally as “abounding in every kind of plant and irrigated with abundant water.” 50 E. Bolognesi, “The Great Palace Survey: The First Season,” Aras¸tırma Sonuc¸ları Toplantisı 11 (1993): 19–21. On the position of the Nea and its complex, which included the Mesokepion, see, most recently, C. Mango, “The Palace of Marina, the Poet Palladas and the Bath of Leo VI,” in Eujfro´ sunon. Afie j ´ rwma sto´ n Mano´ lh Catzhda´ kh, ed. E. Kypraiou, 2 vols. (Athens, 1991), 1:321–30, esp. 323. A general survey of the texts can be found in Janin, Constantinople, 118–19, 390; idem, La ge´ographie eccle´siastique de l’Empire byzantin, vol. 1, Le sie`ge de Constantinople et le patriarcat oecume´nique, pt. 3, Les ´eglises et les monaste`res, 2nd ed. (Paris, 1969), 361–64. 47 to` . . . hJliako` n uJp∆ ejkei´nou katabeblhme´ non kata` pro´ swpon tou' borjrJa', ejx ou» dh` wJra'to kai` to` palaio` n tzoukanisth´ rion, ejkei'se thnikau'ta tugca´ non, e“nqa kai` hJ Ne´ a e“ktistai jEkklhsi´a kai` aiJ du´ o fia´ lai eijsi`n kai` to` mesokh´ pion uJpo` Basilei´ou tou' ajoidi´mou geno´ mena basile´ w". Bonn ed., 144, trans. C. Mango, The Art of the Byzantine Empire, 312–1453 (Englewood Cliffs, N.J., 1972), 163. 48 On these fountains, see the discussion by Littlewood, “Gardens of the Palaces,” 31–32. 49 Bonn ed., 327–29, trans. Mango, Art of the Byzantine Empire, 194–95. 50 para´ deison . . . pantoi´oi" komw'nta futoi'" kai` u”dasin ajfqo´ noi" o“nta kata´ rjrJuton. Bonn ed., 329; trans. Mango, Art of the Byzantine Empire, 195. 45 46

HENRY MAGUIRE

259

From these passages it is evident that the Mesokepion was entirely enclosed, on the north and the south sides by porticoes, on the west by the church, and presumably on the east by the back wall of the pologround, since the gardeners would not have wanted the polo balls among the plantings. In addition, since it was situated in the area formerly occupied by the old pologround, it was overlooked on its south side by the terrace built by Theophilos. This may be the explanation for the term diaulos used for the southern portico, the one on the side toward the sea, for this portico may have had two walkways— not side by side, but on two levels, that of the upper terrace and that of the garden itself. Some confirmation for this layout is provided by the instructions contained in the Book of Ceremonies, despite the usual difficulties in interpretation arising when this text is used to reconstruct topography. It is clear from several passages that the Nea was situated on one of the terraces at a lower level. The instructions for the feasts of St. Elias and of the dedication of the Nea, for example, speak of the imperial procession’s descending from the terrace of the church of the Pharos by means of two flights of stairs to reach the narthex of the Nea. After a circuit of the church, the sovereigns arrive in the “narthex on the side of the sea,” that is, in the southern narthex. From there the sovereigns, in the words of the text, “pass through that narthex and the secret stairs that are there, and by way of the uppermost terrace of that narthex they enter the palace.” 51 This passage seems to confirm that there was indeed an upper-level terrace on the south side of the Nea, which was accessible by stairs from the lower storey of the church. In summary, the Mesokepion was a small, secluded garden, an enclosed paradise that must, to some extent, have had the appearance of a sunken garden, especially when viewed from the southern terrace. It was, therefore, in the tradition of such Roman imperial gardens as the long sunken garden on the east side of the Palatine in Rome, which was surrounded by porticoes on two levels (Fig. 13).52 The whole situation of the Mesokepion was vividly evoked by John Geometres in a prose ekphrasis of his own small garden in Constantinople. John Geometres praises his private paradise because it was easily accessible from his house, and not sunk into the ground, like many other gardens: “It is not necessary to wind through many labyrinths, nor to go up and down many stairs, in order to discover a garden hidden, or rather buried, somewhere under the ground, as I know is the fate of many gardens, so that a toiling [visitor] would sink with exhaustion before reaching those plants withering in their sick-beds.” 53 It is tempting to see this passage as a reference to the Mesokepion itself, with its tortuous access through covered corridors and twisting staircases, far below the upper levels of the palace. The next garden to be considered, the expansive landscape that surrounded the ejkbai´nonte" eij" to` n pro` " th` n qa´ lassan na´ rqhka . . . oiJ despo´ tai die´ rcontai mustikw'" dia` tou' aujtou' na´ rqhko" kai` tou' ejkei'se mustikou' ajnabasi´ou, kai` dia` tou' ajnwta´ tou hJliakou' tou' aujtou' na´ rqhko" ejise´ rcontai oiJ despo´ tai eij" to` pala´ tion. Ed. A. Vogt, Le Livre des Ce´re´monies (Paris, 1935), 1: chap. 28, pp. 108–9; chap. 29, pp. 111–12. 52 J. B. Ward-Perkins, Roman Imperial Architecture (Harmondsworth, 1981), 83, fig. 37. 53 oJ de` probe´ blhtai me` n kai` ouj dei' pollou` " me` n laburi´nqou" dielicqe´ nta, polla` " d∆ ajnelqo´ nta h‘ katelqo´ nta kli´maka" ou”tw pou ejn katagai´oi" katakekrumme´ non, ma'llon de` katorwrugme´ non, to` n kh'pon euJrei'n —o’ pollou` " oi«da tw'n kh´ pwn ejgw` pa´ sconta"—w”ste ponou'nta´ tina pro´ teron ajpeipei'n h‘ toi'" futoi'" ejkei´noi" oi»a katakeklime´ noi" kai` tetariceume´ noi" sw´ masin ejntucei'n. A. R. Littlewood, ed., The Progymnasmata of Ioannes Geometres (Amsterdam, 1972), 11, no. 3, with commentary on p. 55. See also A. R. Littlewood, “Gardens of Byzantium,” Journal of Garden History 12.2 (1992): 126–53, esp. 143. 51

260

GARDENS AND PARKS IN CONSTANTINOPLE

monastery of St. George of the Mangana, could not have been more different from its secluded neighbor, the Mesokepion. The monastery of St. George of the Mangana was a foundation of Constantine IX Monomachos, who constructed an extensive complex that also included a large palace and a hospital.54 These buildings were set in expansive grounds, which extended from north to south at least 800 meters between the ramparts of the Seraglio, on the west, and the sea wall on the east (Fig. 14). Unlike the Aretai and the Philopation, the Mangana has preserved its landscaped aspect to the present day. According to Clavijo, who visited the monastery in 1402, at that time the church was still surrounded by gardens.55 The Turkish sultans replaced the Byzantine structures with belvedere pavilions, and they built a zoo on the western side of the garden, under the southern end of the Topkapı complex.56 However, they kept the terracing of the Byzantine gardens, as seen in an early-nineteenth-century plan (Fig. 15).57 Today the site is a well-tended military base. In his chronography, Psellos provides a lengthy description of the gardens laid out by Constantine Monomachos at the Mangana, and also mocks the emperor’s passion for landscaping, including the transplanting of mature fruitbearing trees and of turf for lawns.58 According to Psellos, the whole complex of the Mangana was enclosed by a wall. Inside, there were buildings with covered galleries for walking, meadows filled with flowers, both on the perimeter of the estate and in its center, and groves of trees, some suspended in the air, others set on the flat ground below. There were extensive waterworks, including canals irrigating the lawns, basins, and graceful baths filled with water. The whole area, according to Psellos, was suitable for horseback riding; the eye could not see the boundaries of it, nor take in all of the meadows that were enclosed.59 Psellos’s account has an ironic, or even satirical, tone, mocking the extravagances of an emperor who tried to create a rus in urbe. One is reminded of the descriptions by Suetonius and Tacitus of the Golden House of Nero, with its long galleries and pools, its fields and woods, its open spaces and vistas, all within the city of Rome.60 Thus we might suspect that Psellos’s portrait of Constantine IX, like the Vita Basilii’s portrait of Michael III, is making a rhetorical comparison between the emperor and the excesses of Nero.61 Consequently, we may be tempted to dismiss Psellos’s account of the gardens of the Mangana as overstated and inaccurate. The suspicion that Psellos was indulging in literary exaggeration is heightened by the denouement of his story of Constantine IX. According to Psellos, the emperor caught a fatal chill while bathing in a pool dug in the middle of a meadow in one of his gardens, and thus, by a poetic justice, Constantine Monomachos 54 ´ glises, 70–76. On the excavations, see R. Demangel and For a survey of the literary sources, see Janin, E E. Mamboury, Le Quartier des Manganes (Paris, 1939). See also Mu ¨ ller-Wiener, Bildlexikon, 136–38 and plan on p. 497; H. Tezcan, Topkapı Sarayı ve C ¸ evresinin Bizans Devri Arkeolojisi (Istanbul, n.d.). 55 ´ glises, 74. Janin, E 56 ˘lu, Architecture, Ceremonial, and Power: The Topkapı Palace in the Fifteenth and Sixteenth Centuries G. Necipog (Cambridge, Mass., 1991), 201–4. 57 The plan is found in A. Melling, Voyage pittoresque de Constantinople (Paris, 1819). 58 Chronographia, 6.173–75, 186–87, 201, ed. E. Renauld (Paris, 1926), 2:56–57, 62–63, 70. 59 Chronographia, 6.186–87, ed. Renauld, 2:62–63. 60 Suetonius, Nero, 31; Tacitus, Annals, 15.42. 61 On Michael III and Nero in the Vita Basilii, see R. J. H. Jenkins, “Constantine VII’s Portrait of Michael III,” Bulletin de la Classe des Lettres et des Sciences Morales et Politiques, Acade´mie Royale de Belgique, 5th ser., 34 (1948): 71–77.

HENRY MAGUIRE

261

became the only Byzantine emperor to die as a result of his obsession with landscape architecture.62 In the case of the Mangana, however, we can check some of Psellos’ statements against the existing evidence of the site, and we discover that his overall characterization of the gardens was no more than the truth. The site has two main levels. The lower level is about 250 meters wide at the southern end, but narrows to a point toward the north (Fig. 14). This area is bordered by the sea wall on the east and by a high terrace wall on the west. The monastery and the palace were built on this lower level, close to the sea wall. The terrace wall on the west is massive and contains recessed brickwork at several points in its construction. The recessed brick technique, in which alternate courses of bricks were set back and covered with mortar, is generally dated between the late tenth and the mid-thirteenth centuries.63 It is used extensively in the substructures of the church of St. George and the palace. Therefore, the portions of the terrace wall that are built in this technique probably date to the time of Constantine Monomachos. Figure 16, for example, is a view of the southern end of the terrace, corresponding to the portion marked A on the plan in Figure 14. Here, the top one-third to one-half of the wall has extensive areas of recessed brick masonry, together with later repairs in stone. This wall extended more than 600 meters to the north, where there are still traces of recessed brick masonry, as can be seen in Figure 17 (B on the plan in Fig. 14). There are remains of two substantial cisterns built against the terrace wall toward its southern end (C and D in Fig. 14). Although these cisterns appear to be of early Byzantine date, the southern one (C) shows traces of medieval repairs. The upper terrace that is created by this wall measures about 190 meters in width at its southern end, and narrows to a point at the north. It is enclosed on the west side by the wall that supports the Topkapı Palace and its forecourt (E in Fig. 14); this wall, also, shows traces of recessed brickwork, behind the arcades of the Ottoman zoo, which was later built in front of the wall, at its southern end. It also served as the western boundary of the Mangana gardens. Even this brief and incomplete survey of what can still be seen on the site of the Mangana shows that the description by Psellos was essentially accurate. The gardens were indeed laid out on two levels, with the upper terrace still providing space for groves of trees that appear to be “suspended in the air,” as Psellos put it (Fig. 16). There are the large cisterns, which provided water for the canals and the baths. On the lower level, there is a flat area suitable for horseback riding along the sea walls (Fig. 18). This area was still used for equestrian sports during the Ottoman period.64 And most of all, the whole site, as defined by the lower terrace wall with its recessed brickwork, was extremely long. The site has the shape of an elephant’s tusk, curving to the west along the line of the cliff at its northern end (Fig. 14). Hence, it is indeed impossible to see the whole extent of the garden from either of its ends; Psellos did not exaggerate on this point either. It is evident that the garden of the Mangana presented a striking contrast with the Chronographia, 6.201–2, ed. Renauld, 2:70–71. ´ urcˇic´, Early Christian and Byzantine Architecture, 4th ed. (Harmondsworth, 1986), R. Krautheimer and S. C 354, 504–5 n. 3. 64 ˘lu, Topkapı Palace, 206. Necipog 62 63

262

GARDENS AND PARKS IN CONSTANTINOPLE

Mesokepion inside the Great Palace. Whereas the Mesokepion was small and confined, the Mangana was expansive and open. The Mesokepion was situated below the main part of the palace, so that one looked down into it; the gardens of the Mangana, on the other hand, rose on a high terrace above the palace, so that one looked up at them, hanging in the air. From this comparison we can get a sense of the variety of designs that were possible in Byzantine landscape architecture. We have seen, then, that there are sufficient points of convergence between the literary descriptions and the physical fabric of the city to demonstrate that the texts were anchored in reality; they were not purely figments of their authors’ imaginations. The relationship of the park of the Aretai to the hexagonal first tower of the land walls, the secret confinement of the Mesokepion, the openness of the terraced gardens of the Mangana—all of these features, which are described in the texts, conform to what can still be seen today in the modern city. However, as I proposed at the beginning, the descriptions reflected not only the gardens themselves, but also the changing attitudes of the observers. I would like to conclude this paper, therefore, with two brief examples, one late antique and one medieval, of how each observer’s state of mind could color his response to a similar view. My first witness is the fourth-century description by the emperor Julian of a small estate and garden that he owned in Bithynia.65 This property, which he had been given by his grandmother, was situated near Constantinople, about three kilometers from the sea. In a letter, Julian describes it as a delightful place, with excellent springs of water, a charming bath, a garden, and trees. The estate had a hill that offered fine prospects of the sea, the Propontis, the islands, and the city of Constantinople. Presumably, therefore, the estate was situated somewhere on the rising ground above the modern district of Maltepe, overlooking Princes’ Islands to the south and the city to the west. Julian recalls with nostalgia the pleasant experience of looking upon these views: “Very peaceful it is to lie down there and glance into some book, and then, while resting one’s eyes, it is very agreeable to gaze at the ships and the sea. When I was still hardly more than a boy, I thought that this was the most delightful summer place. . . . ”66 Julian’s delight in the views to be obtained from his country estate is a topos of Roman and late Roman garden descriptions. From Cicero to Basil the Great, a standard element in the praise of a villa or an estate was the pleasure to be obtained from gazing at its surrounding scenery, whether it was the sea, a river, or the countryside.67 Over eight centuries later, on the eve of the Fourth Crusade, a Byzantine churchman, Nikolaos Mesarites, also gazed upon the city of Constantinople and the sea, but his response to these views, in spite of some superficial similarities, was fundamentally different from that of the fourth-century emperor. Mesarites begins his long ekphrasis of the church of the Holy Apostles by praising its situation, which is in the heart of the city. Letter 25, trans. W. C. Wright, The Works of the Emperor Julian, 3 vols. (London–New York, 1923–53), 3:76–80. 66 hJsuci´a de` pollh` kataklinome´ nv kai` ei“" ti bibli´on ajforw'nti, ei«ta dianapau´ onti th` n o“yin h”diston ajpidei'n eij" ta` " nau'" kai` th` n qa´ lattan. tou'to ejmoi` meiraki´v komidh' ne´ v qeri´dion ejdo´ kei fi´ltaton. Ibid., 78. 67 A. R. Littlewood, “Ancient Literary Evidence for the Pleasure Gardens of Roman Country Villas,” in Ancient Roman Villa Gardens, ed. E. B. MacDougall (Washington, D.C., 1987), 9; H. Maguire, “Paradise Withdrawn,” in Littlewood, Maguire, and Wolschke-Bulmahn, Byzantine Garden Culture. 65

HENRY MAGUIRE

263

Mesarites stresses, however, that the Holy Apostles is not “in the middle of confusion,” and “jostled by the mob” like “the rest of the churches,” but rather lies in the peaceful environs of well-watered gardens.68 There is still an echo of this setting today, in the formal gardens that currently surround the Fatih, the mosque that replaced the Holy Apostles. Mesarites describes the gardens around the church with copious borrowings from Libanius’ encomium of Antioch, but he concludes his praises with a significant passage that is not lifted from the fourth-century author.69 Here, Mesarites explains that the people living near the Holy Apostles can grow enough food in their gardens to make them independent of supplies from the outside: “People from whom this church is far distant can see from afar the wheat being brought in; for those who live near it, the wheat alone which grows in the land about their houses is sufficient for their nourishment, and they need have no care for invasions of barbarians, for the mighty waves of the sea, [or] for the dangers from pirates. . . . ”70 Immediately after this passage comes Mesarites’ description of the view to be obtained from the upper levels of the church, to which I have already referred in discussing the Philopation.71 Mesarites’ reactions to the scenery are expressed in a kind of collage, made up of quotations from Libanius and passages of his own devising. The view of the medieval writer is distinctly ambivalent. At first Mesarites, like Julian, seems to delight in the prospect of the sea that is spread before his eyes. “A man who walks about in the upper galleries of the church,” he says, “enjoys a varied pleasure as he gazes over the backs of the northern and southern seas. For one can see from there the sea, which itself lies there tranquilly and on its back bears freight ships before a fair breeze, a sweet sight to all men and a source of rejoicing and pleasure.” 72 But then comes a note of fear, as Mesarites continues, “And neither is he [the viewer] frightened by the floods which the sea is wont to spew up, now here, now there, because he stands at a fitting distance from it, nor does he hear the groans of the sailors or the cries of drowning men.” 73 Although many parts of this description of the waters surrounding Constantinople were taken from Libanius, Mesarites chose and framed his quotations so as to emphasize the theme of the dangers coming from the sea, which had colored his previous paragraph. Certainly, Mesarites followed the conventions of ekphrasis by describing the views before him in terms of oppositions and by borrowing heavily from ancient writers. At the same time, he expressed the concerns of his own epoch, the heightened threats of barbarDescription, 3.1–4.3, ed. and trans. Downey, 862–63, 897–98. The borrowings from Libanius have been identified by Downey, 862, note III, 1. 70 toi'" me` n ga` r a“lloi", o”soi" oJ new` " ou»to" ajfesthkw´ " ejsti, mh´ koqen e“stin ijdei'n kai` to` n si'ton eijskomizo´ menon, toi'" d∆ e“ggista tou´ tou mo´ no" oJ ejk tou' tw'n dwma´ twn pericw´ rou oi«den ajrkei'n pro` " diatrofh´ n, kai` ou“t∆ ejqnw'n tou´ toi" me´ lei ejpidromh´ , ouj trikumi´ai qala´ ssh", ouj ki´ndunoi peiratw'n. Description, 4.2, ed. and trans. Downey, 863, 898. 71 See above, note 16. 72 karpou'tai de` kai` poiki´lhn tina` carmonh` n oJ peri` ta` tou' naou' di´wrofa ejmfilocwrw'n kai` pro` " qala´ tth" borinh'" kai` noti´ou nw'ta ejpentrani´zwn ⭈ e“sti ga` r ejkei'qen ijdei'n qa´ lassa´ n te galhniw'san aujth` n kai` pro` " nw'ta tau´ th" ferome´ nhn oJlka´ da di∆ oujri´ou tou' pneu´ mato", hJdu` de` tou'to toi'" a”pasi qe´ ama kai` qumhdi´a kai` terpwlh´ . Description, 5.1, ed. and trans. Downey, 864, 898. 73 kai` ou“te pro` " ta` " ejpiklu´ sei", oJpo´ sa´ " per a“llote a“llw" oi«den ajpereu´ gesqai qa´ lassa, seso´ bhtai´ pote tv' diesta´ nai tau´ th" summe´ trw", ou“te ta` " tw'n nautw'n oijmwga` " ejpa?ei kai` tw'n katapontizome´ nwn tou` " kwkutou´ ". Description, 5.2, ed. and trans. Downey, 864, 898. 68 69

264

GARDENS AND PARKS IN CONSTANTINOPLE

ian attack, the resurgence of piracy after the middle of the twelfth century, the uncertainty of food supplies.74 Mesarites’ view of the gardens around the Holy Apostles and of their setting was tinged by the circumstances of his time; for him, the crowning feature of the gardens was their self-sufficiency. His view of Constantinople and its maritime setting was more anxious, inward looking, and closed to the outer world than that of Julian. In conclusion, the Byzantine descriptions of gardens, parks, and scenery may be read on two levels. They can still serve as useful factual records of the lost physical landscapes of Byzantium; at the same time, they reveal the changing mental landscapes of the Byzantine beholder. Johns Hopkins University 74

On the recent pirate attacks, see Downey, 863, note IV, 2.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Constantinople (after R. Janin, Constantinople byzantine, développement urbain et répertoire topographique, 2nd ed. [ Paris, 1964], plan I)

2

Belvedere, “Tower of Isaak Angelos”

3

Ceramic bowl decorated with creatures of the hunt, ancient Corinth, Archaeological Museum

4

Istanbul, satellite view (after D. Kuban, Istanbul, an Urban History: Byzantion, Constantinopolis, Istanbul [ Istanbul, 1996])

5

Inner land wall of Constantinople, view of Tower 1 from the southwest, ca. 1870 (after B. Meyer-Plath and A. M. Schneider, Die Landmauer von Konstantinopel [Berlin, 1943], 2: pl. 25)

6

7

Istanbul, Kazlı Çe¸sm e, Byzantine sculpture, detail, upper block

Istanbul, Kazlı Çe¸sme, Byzantine sculpture

8

Istanbul, Kazlı Çe¸sm e, Byzantine sculpture, upper block, detail, wing feathers of goose

9

Carving from Fenari Isa Camii (church of Constantine Lips), Istanbul, Archaeological Museum

10

11

Carving from Fenari Isa Camii (church of Constantine Lips), Istanbul, Archaeological Museum

Detail, dome cornice, Fenari Isa Camii (church of Constantine Lips), Istanbul (after T. Macridy, “The Monastery of Lips and the Burials of the Palaeologi,” DOP 18 [1964]: f ig. 18)

12

Bronze goose, London, British Museum ( photo: courtesy of the museum)

13

Flavian Palace, reconstruction, Rome, Palatine (after J. B. Ward-Perkins, Roman Imperial Architecture [Harmondsworth, 1981], f ig. 37)

14

Istanbul, area around the Mangana (after W. Müller-Wiener, Bildlexikon zur Topographie Istanbuls [Tübingen, 1977], 497)

15

Plan of the Topkapı complex (after A. Melling, Voyage pittoresque de Constantinople [Paris, 1819])

16

Istanbul, Mangana gardens, terrace wall at southern end (labeled A in Fig. 14)

17

Istanbul, Mangana gardens, terrace wall at northern end (labeled B in Fig. 14)

18

Istanbul, Mangana gardens, view to the lower level from the north

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Istanbul, Zeyrek Camii, plan showing phases of construction, ca. 1118–36

2

Zeyrek Camii, general view from the minaret, looking north, during the restoration, 1998

3

Zeyrek Camii, Middle Building, stone-by-stone plan at floor level, 1997

4

Zeyrek Camii, Middle Building, detailed elevation of east facade, 1998

5

Zeyrek Camii, the complex plan of at roof level, prior to restoration, 1997

6

Zeyrek Camii, South Building, detailed drawing of cornice blocks, 1997

7

Zeyrek Camii, North Building, cleaning of crack in north crossarm, 1998

8

Zeyrek Camii, view from South Building, looking north, showing roofing in progress and eastern dome of Middle Building during repair, 1998

9

Zeyrek Camii, North Building, view toward the apse vault during cleaning, showing Byzantine amphorae in the fill behind the conch and Ottoman cornice blocks joined with iron ties, 1997

10

Zeyrek Camii, North Building, vaulting of the prothesis, showing Byzantine roofing tiles above the conch, 1998

11

12

Zeyrek Camii, brickstamp found in the f ill of the vaulting of the South Building

Zeyrek Camii, brickstamp found in the fill of the vaulting of the South Building

13

Amphora removed from the apse of the North Building, Zeyrek Camii, Istanbul (drawing by A. Özügu¯ l, 1988)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

Excavations of an Early Christian Pilgrimage Complex at Bir Ftouha (Carthage) SUSAN T. STEVENS The Bir Ftouha project, begun in 1994, was designed to meet threats to the site from the expansion of modern suburbs through excavation of the latest phases of the early Christian basilica complex. The 1999 season, partially funded by Dumbarton Oaks, was the last of five summer campaigns, during which an open-area excavation of about 1,600 square meters revealed parts of nine attached structures of a major Christian pilgrimage complex now tentatively dated to the sixth century. Although several years of study remain to be done, sufficient data has been gathered to answer questions regarding the plan, phasing, dating, and decorative program of the complex’s monumental center. The Bir Ftouha complex is important historically because it is the only known suburban basilica at Carthage for which it will be possible to write a reasonably detailed history from stratigraphic excavation. Although the east end of the site was excavated by Paul Gauckler in 1897 and the west end by Alfred-Louis Delattre in 1928–29, most of the basilica lay between their excavations.1 Damaged since antiquity by the depredations of stone robbers, tomb robbers, and cultivators, all the colonnades and wall lines of the complex, save two small fragments of elevation wall, are proposed on the basis of surviving foundations. From field observation and preliminary study, one major phase of construction can be recognized in the monumental center. The Bir Ftouha church appears to have been an ambu1 L. Ennabli, Carthage: une me´tropole chre´tienne du IVe `a ´ tudes d’antiquite´s africaines (Paris, la fin du VIIe sie`cle, E 1997), 111–41.

latory basilica of modest size, divided into three aisles and eight bays, with an enclosed apse in the east and narthex in the west. The ambulatory around the outside of the apse connected the church to a large annex building that included a baptistery marking the eastern extent of the complex.2 A nine-sided polygonal entrance structure was attached to the narthex at the west end of the complex. Pottery dates the construction of the complex to the early sixth century. A second phase, in which minor modifications were made to some parts of the complex, appears to date to the mid-sixth century.3 This preliminary report, intended as an update of previously published reports, focuses on the baptistery at the east end of the basilica complex and the entrance structure at the west end, the major discoveries of the 1998 and 1999 seasons.4 These two structures illustrate the features that effectively distinguish Bir Ftouha from the other four suburban churches of Carthage.

2 The annex building was the source of the mosaic with bird medallions and scenes of the doe and stag drinking from the four rivers of Paradise lifted by Gauckler and currently displayed in the Bardo (Tunis) and Louvre; M. Yacoub, Muse´e du Bardo (Tunis, 1996), 39, 43, nos. 203–209; F. Baratte and N. Duval, Mosaı¨ques romaines et pale´ochre´tiennes du Muse´e du Louvre (Paris, 1978), 79–80, no. 39. 3 The ceramic finds from stratified contexts are few and poor in finewares, so the dates furnished by pottery provide rather broad termini post quos. It should be possible to refine the dating after the coin evidence becomes available. Clearly, the phasing will be refined in the course of writing the stratigraphic report, now in process. 4 S. Stevens, A. Kalinowski, and H. vanderLeest, “The Early Christian Pilgrimage Complex at Bir Ftouha, Carthage: Interim Report,” Journal of Roman Archaeology 12 (1998): 371–83.

272

EARLY CHRISTIAN PILGRIM COMPLEX AT BIR FTOUHA THE BAPTISTERY

The round baptistery (see Figs. 1A, 1B) at Bir Ftouha lay on the main east-west axis of the basilica, behind the apse.5 It was about 16 meters in diameter on the outside, with a little over half of its circumference enveloped by the annex building. The baptistery consisted of an ambulatory about 4 meters wide and an inner circular room about 6 meters in diameter. Only the packed clay foundations of the ambulatory floor survived, but the abundance of marble floor tiles in the fills of adjacent robber trenches suggest it was once paved with marble tiles. The original paving of the baptistery must have been at current ground level. The solid concrete foundations between the ambulatory and the central room probably supported a double colonnade. They were some 2 meters wide, approximately double the width of the foundations of the outer wall of the baptistery, they may have supported a considerable weight of vaulting. The only known entrance into the baptistery from the annex building was through a colonnade of five columns in the northwestern part of the baptistery’s outer wall. The inner room was probably once covered in a mosaic now represented by a layer of crushed mortar and loose tesserae around the font in the center of the room. The font, once revetted in Carian and cipollino marble, was extensively looted; deducing its exact dimensions and shape from the remaining third of its rubble core is difficult. It appears to have had three steps: the top level was quatrefoil or cruciform with the arms aligned on the main eastwest axis of the church, the middle step was probably oval, and the lowest, at a maximum depth of 1 meter, was circular.6 The ciborium 5 Round baptisteries, although not common, are attested in North Africa at Djemila and Henchir Harat/ Segermes; S. Ristow, Fru ¨christliche Baptisterien, JbAC Erganzungsband 27 (Mu ¨ nster, 1998), 117, no. 75; 257, no. 714. For 5th–6th century African baptisteries located behind the apse, see Henchir Deheb, Ksar es Zit/Siagu, Maktar (Hildeguns), Sbeitla (Vitalis), and Sidi Habich/Enfida, among others. Ibid., 117, no. 77; 260, no. 729; 261–62, no. 734; 263, no. 741; 264, no. 746. 6 The closest African parallels may be the fonts that use the same elements, though not necessarily in the same order, at Kelibia, Carthage, Sayda, Oumcetren/Sidi Daoud, also Henchir Deheb; A. Khatchatrian, Les baptiste`res pale´ochre´tiens (Paris, 1962), 38, nos. 279, 275, 278, and 30, no. 23.

most likely consisted of four columns, probably placed between the arms of the cross, and it included some gilded stucco elements. A drain from the font ran down into a small tank (with a capacity of approximately 4 cu. m) built between the foundations of the annex building and the curving outside wall of the baptistery. Its west and south walls (the foundations of the annex and baptistery) were lined with plaster. By contrast, its east and north sides were built of courses of cobble masonry pierced by numerous drainage holes. The bottom of the tank was also water permeable. This feature is likely to have been a soakaway chamber, designed to take water out of the font and leech it gradually into the clayey subsoils away from the buildings.7 The prominence of the baptistery in the design of the Bir Ftouha complex is perhaps its most remarkable feature. The baptistery was the culmination of the complex, lying at the eastern end of the main east-west axis of the complex, probably with a high clerestory and floor surfaces built up 40–50 centimeters above the mosaic floor of the annex building. This is the most prominently placed of all the baptisteries in Carthage, and its location is especially challenging to explain in a suburban church. Like the baptisteries added to Roman pilgrimage churches, it may reflect a response to the popular belief in the particular efficacy of baptism in the presence of saints.8 THE ENNEAGON Balancing the baptistery at the opposite, western end of the complex was a unique, nine-sided polygonal entrance structure (Figs. 2A, 2B), one side of which was attached to the narthex of the basilica. Like the baptistery, the enneagon was about 16 meters across on the outside and consisted inside of a round central room and surrounding ambulatory. The ambulatory was about 3 meters wide, separated from the central room, which was approximately 9 meters in diameter, by a colonnade of nine columns, one opposite each corner of the polygon. In the center of the enneagon was a striking mosaic medallion of lilies 7

Thanks to Andrew Wilson for this suggestion. R. Krautheimer, Rome, Profile of a City, 312–1308 (Princeton, 1980), 56. 8

SUSAN T. STEVENS in a laurel wreath surrounded by a border of tendrils. Radiating from it was a pattern of interlocking peltas with pendant grape leaves gradually increasing in size. Entrance into the enneagon at the west may have been through two doorways on either side of the corner of the building that marked the main east-west axis of the basilica.9 Opposite this corner a single doorway led from the enneagon up one step into the narthex. The enneagon, lacking any obvious built feature at its center, seems to have been designed to funnel visitors into the basilica. Its most unusual decorative element was a roof of fine, decorated, red-slipped tile, which was found in great abundance in destruction layers. At least four different molded motifs were recovered, the most common of which was the facing heads of Mauritania and Africa familiar from late Roman African Red Slip plates (Hayes form 46) usually dated 400–420.10 After the enneagon was constructed, at least eleven tombs were sunk into its floor, almost all in the ambulatory. Subsequent robbing of the tombs removed virtually all ancient layers, leaving behind a few tenuous islands of mosaic in the ambulatory, including two fragments of tomb mosaics cut into the pelta mosaic. Adjacent to the enneagon toward the southeast was a burial chamber containing another four tombs. The arrangement of all the tombs suggests that they were individual burials of various dates, carefully placed to avoid disturbing earlier ones. The deep shafts, generous size, and remains, albeit scanty, of the elaborate tomb structures suggest that they were all high-status burials. An analogous but smaller group of late tombs was cut into the mosaic of the annex building at the east end of the complex; six tomb mosaics were found by Gauckler and one during the current excavations.11 9 A similar arrangement is found at San Vitale (Ravenna), where two doorways lead from the narthex into the church, one on either side of a corner of the octagon. 10 Cf. J. W. Salomonson, “Late Roman Earthenware with Relief Decoration Found in Northern Africa and Egypt,” Oudheidkundige Mededelingen uit Rijksmuseum van Oudheden te Leiden 43 (1962): pl. xxxii.1–2; J. Garbsch, “Spa¨tantike Sigillata-Tabletts,” Spa¨tantike Keramik aus Nordafrika in der Pra¨historische Sammlung, Acta rei cretariae romane fautores, suppl. vol. 5 (August 1980), 31, fig. 23. The Bir Ftouha rooftiles are being studied by Susan Pringle. 11 P. Gauckler, Inventaire des mosaı¨ques de la Gaule et de l’Afrique, vol. 2, Afrique Proconsulaire (Tunisie) (Paris, 1910),

273

A twelfth tomb, among the worst preserved in the enneagon, may have been the most privileged in terms of its location. The only known tomb in the western third of the building, it lies in the corner of the enneagon between the hypothetical doorways, oriented east-west on the main axis of the basilica. Although no differently constructed than the others, could this tomb have been the impetus for building the enneagon and subsequently attracted later burials? Might the nonagon have been built as a mausoleum for a single privileged individual, who was buried not in the center but at the structure’s apex? The removal of ancient strata by tomb robbing makes a definitive answer impossible. The tomb’s placement against the narthex of an ambulatory basilica seems to be an architectural reference (differences in scale and function notwithstanding) to the mausoleum of Helena attached to the martyr church of Sts. Peter and Marcellinus, a fourth-century ambulatory church on the outskirts of Rome.12 Most of the current excavation took place inside the monumental center of the complex and produced few of the disarticulated human remains and fragments of tomb inscriptions that ordinarily signal disturbed tombs. Trenches on the edge of the complex uncovered one tomb against the exterior of the north wall of the annex building and two against the exterior of the south wall of the basilica. Less privileged individuals may have been consigned to rest outside the monumental center of the complex. Except for the focal tombs in the apse and choir around which the basilica was built, burials seem never to have been permitted in the church proper.13 The tombs in the enneagon and annex building demonstrate that even privileged burials were restricted to the outer buildings of the complex and arrived relatively late in the history of the church.14 The careful control of burials makes 264–66, nos., 787–91, 793; and S. Stevens, “Excavations at Bir Ftouha in 1994 and 1996: Preliminary Report,” CEDAC Carthage Bulletin 18 (1998): 9. 12 J. Guyon, Le cimetie`re aux deux lauriers, BEFAR 264 (Rome, 1987), 207–35, 250–55. 13 The focal tombs were those original to the construction phases of the basilica as currently understood: the three tombs in the apse of the basilica (first phase) and the one in the center of the nave (second phase). 14 This pattern may even hold for the Christian buildings near the modern wellhouse. Delattre was astonished that although he found ten sarcophagi in a trefoil funerary

274

EARLY CHRISTIAN PILGRIM COMPLEX AT BIR FTOUHA

Bir Ftouha different from the other suburban basilicas at Carthage, funerary basilicas which often have a high concentration of tombs, especially inside the church. While the ambulatory attests that the original focus of the Bir Ftouha church was the group of special tombs in the apse and the choir, these did not attract other subsequent burials. Rather, later burials appear to have been excluded from the church proper. They were permitted in less important areas, inside the enneagon and annex building and outside the buildings of the complex. With its quadratum populi of 17 by 27 meters, the Bir Ftouha basilica is modest in size by comparison with the other suburban basilicas at Carthage.15 In fact, it was substantially smaller than its annex building, measuring 20 by 40 meters (on the outside) (Fig. 3). Despite its size, however, the annex building enhanced rather than detracted from the church: it added a north-south axis and, at the same time, extended the basilica’s original east-west axis with the baptistery. Although the hourglass shape of its interior is unparalleled, the annex building is an elaborate chevet, a feature that regularly enclosed the apses of North African basilicas. The ambulatory design of the Bir Ftouha annex building is related, in spirit if not in plan, to the chevet of the basilica at Siagu.16 The purpose of the Bir Ftouha annex seems to have been to circulate visitors through various parts of the complex, providchapel, in all the dirt he moved in and around that structure no other tomb inscriptions and only a few other burials were found, Delattre, CRAI (1929): 29. 15 Damous el Karita is 45 ⫻ 65 m (9 aisles, 11 bays), Basilica Maiorum is 45 ⫻ 61 m (9 aisles, 13 bays), St. Monique is 35.5 ⫻ 71.3 m (7 aisles, 14 bays), Bir el Knissia is 25.6 ⫻ 46.6 m (3 aisles, 11 bays). 16 N. Duval, “Le choeur de l’e´glise de Siagu (Tunisie),” Felix Ravenna, ser. 4, 127–30 (1984–85): 159–99.

ing limited access to the tombs in the apse. In this, Bir Ftouha resembles the pilgrimage churches of St. Peter and St. Paul in Rome, whose transepts performed this function.17 Since the excavation has just been completed, the stratigraphic report is in progress, and the thorough analysis of the finds has just begun, any conclusions about the site would be prematurely drawn. Nevertheless, the remarkably unified plan that has emerged during five campaigns of excavation deserves some preliminary comment. The basilica at the core of the Bir Ftouha complex is unusual for its ambulatory, but more remarkable for the extraordinary embellishments to the basilica plan. The enneagon is without parallel in early Christian architecture, and it must have been designed to meet the specific needs of this complex. It is also a clever variation on the octagon or circle. The balancing of the basilica’s east-west axis, prolonged by the baptistery and the enneagon, with the transverse axis of the annex gave the Bir Ftouha complex the shape of a Latin cross overall. This cruciform plan, an unusual hybrid of longitudinal and centralized forms, is unique in Africa. The innovative and sophisticated design of the Bir Ftouha complex will provide a counterbalance to a long-held opinion that Christian architecture in North Africa was conservative.18 If current dating of the phases is sustained and places the development of Bir Ftouha late in the Vandal period, the complex will stand as another small corrective to the commonplace that Vandal Carthage was a city of stagnation and decline. 17 F. Tolotti, “Il S. Sepolcro di Gerusalemme e le coeve basiliche di Roma,” Ro¨mische Mitteilungen 93 (1986): 498–512. 18 R. Krautheimer, Early Christian and Byzantine Architecture (New York, 1975), 199–206 passim.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1A View of the Bir Ftouha baptistery from the south. The soakaway is just visible in the upper right corner. (photo: N. Rough)

1B Baptistery in the working reconstruction plan superimposed on the 1998 Bir Ftouha stateplan (drawing: B. Dayhoff )

2A View of the Bir Ftouha nonagon from the south. The narthex of the basilica is visible in the upper right. (photo: N. Rough)

2B Enneagon in the working reconstruction plan superimposed on the 1999 Bir Ftouha stateplan. The privileged tomb is located at the top left. (drawing: B. Dayhoff )

3

Annex building in the working reconstruction plan superimposed on the 1999 Bir Ftouha stateplan. The ambulatory basilica is on the left. (drawing: B. Dayhoff)

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

The Byzantine Basilica at Priene STEPHAN WESTPHALEN The basilica at Priene was investigated during two campaigns in 1995 and 1996.1 The church had already been excavated by a German mission at the end of the nineteenth century and was published by Theodor Wiegand.2 Since then few studies have been devoted to the church, and none of these presented significant new information.3 The ruins are almost completely excavated and comparatively well preserved. In addition, much of the liturgical furniture is still extant, including the ambo and particularly, the posts and screens of different chancels. In the course of our investigation it was decided that no new excavations be undertaken, but rather that we document the architecture by producing large-scale plans (scale 1:25) and the liturgical furniture with drawings and photographs. Defining the history of the building and its reconstruction were our main objectives.4 GENERAL DESCRIPTION The basilica is situated on the upper terrace of the Hellenistic town above the agora, in the This paper was translated by Katja Lembke and Amy Papalexandrou. 1 The campaigns were supported by the German Archaeological Institute and conducted by W. Koenigs. For the 1996 campaign, Dumbarton Oaks provided a project grant. 2 T. Wiegand and H. Schrader, Priene (Berlin, 1904), 475–85. 3 V. Schultze, Altchristliche Sta¨dte und Landschaften, Kleinasien (Gu ¨ tersloh, 1926); P. Verzone, “La ‘Cattedrale’ di Priene e le sue scolture,” Felix Ravenna (1970): 261–75; M. Schede, Die Ruinen von Priene, 2nd ed. (Berlin, 1964), 108–11. 4 For a detailed study, see S. Westphalen, “Die Basilika von Priene: Architektur und liturgische Ausstattung,” IstMitt 48 (1998): 279–340, pls. 30–43. See also the short reports of S. Westphalen and W. Koenigs, “Priene 1995,” XIV. Aras¸tırma Sonuc¸ları Toplantısı (Ankara, 1997), 71–76, figs. 1–9; eidem, “Priene 1996,” XV. Aras¸tırma Sonuc¸ları Toplantısı (Ankara, 1998), 308–11, figs. 2, 3.

general area inhabited during the Byzantine period.5 The building defines the north side of an older insula. The Theaterstrasse lies immediately to the north. To the east, the church is built against a Roman bath. The area west of the church is not excavated, so that it is as yet unknown whether the church was entered there from an open space, an atrium, or a narrow street. The building is divided into three parts: the apse, the three-aisled naos, and the narthex (Figs. 1, 2). The latter is a simple rectangular hall with two portals in the western wall. Three further portals connect the narthex with the central nave and the aisles. The south wall of the south aisle is punctuated by two doors. Ten pairs of columns originally divided the nave from the aisles. The lower drums of the southern row of columns are still in situ, whereas only the drums of the first, third, and tenth columns (counted from the east) are extant in the northern row. Rectangular piers were placed between the columns during a rebuilding phase. The apse is not visible from the outside, since it was built directly into a niche of the thick wall of the Roman bath. HISTORY OF THE BUILDING In his study of the site, Wiegand had already distinguished two phases of the building.6 In the earlier phase, the church was a simple basilica with columns. Later, when the church was rebuilt, the columns were replaced with strong piers. During our investigation we were able to distinguish a phase earlier than Wiegand’s first phase. It is visible at the wall that divides the aisles from the narthex. This wall 5 W. Mu ¨ ller-Wiener, “Mittelalterliche Befestigungen im su ¨ dlichen Ionien,” IstMitt 11 (1961): 47, fig. 10. 6 Wiegand, Priene, passim.

276

BYZANTINE BASILICA AT PRIENE

abuts the side walls and covers older plaster. Furthermore, it is built on top of the marble floor slabs. This wall was therefore set subsequently into the building, as were two wall piers at the end of the two rows of columns. The construction of the columns thus belongs to the same phase of rebuilding. Phase 1 The side walls belong to the earliest building phase of the basilica. The church is 13.80 meters wide in the east, 13.55 meters wide in the west. These dimensions were maintained up to the latest phase. We know very little about the area and its surroundings in pre-Byzantine times. It is bounded on the east by the Roman bath. The western wall of the bath is very thick (ca. 3.80 m) and was punctuated by a rectangular niche. The niche was opened to the outside during the construction of the basilica and rebuilt as the apse of the church. The northern wall of the church encroaches on the Theaterstrasse, thereby reducing it to a narrow street. Perhaps other buildings standing to the south necessitated the church’s placement further north on the street. Older constructions were demolished to prepare the area for the church. Below the floor, the wall of an older building crosses the basilica. Unfortunately, it was not possible to discern its ground-plan. In any case, this earliest division of the area was later abandoned. In general, we can conclude that preparations for the church were extensive, since an opening was made to the bath, the Theaterstrasse reduced, and an older building pulled down. To the oldest parts of the church also belong the marble floor slabs, which are of spolia of varying sizes and upon which the wall dividing the three aisles from the narthex was built. In this area, the floor is laid regularly, whereas about two meters east of the dividing wall it is haphazard. Here a strip of irregularly laid slabs crosses the church, presumably marking the placement of the western wall in phase 1, before it was demolished in phase 2 and covered with the floor slabs. The earliest nave was, therefore, slightly shorter than the extant nave, which measures 24.45 meters. The placement of the columns cannot be reconstructed with certainty, but it does not seem too daring to propose eight

pairs of columns with intercolumniations of 2.20 meters, for a total length of about 24 meters. Phase 2 The ground-plan of the second phase is clearly visible in the preserved state of the ruins. The nave is about 26.75 meters long and 6.80 meters wide. Ten columns with intercolumniations of around 1.80 meters divide the nave from the aisles. The columns are spolia of the Hellenistic period that did not come from the nearby gymnasion, as Wiegand thought, but mainly from the stoa south of the temple of Athena.7 The Doric columns measure some 4.50 meters in height. They are dated to the second or first century B.C. and were, therefore, already five to six hundred years old (and perhaps had fallen down) by the time of their reuse in the basilica. In the Byzantine construction, the columns were not as tall as they had been in the Hellenistic structure. They consisted of two or three drums and were only about 3.10 meters high, including the capitals. The columns seem short in comparison to their diameter (62 cm) and stood in narrow rows. Besides the Hellenistic marble spolia, we also found fragments of the arcades above the columns. The stones of the arcades are socalled breccia, that is, a porous local stone. The mason clearly did not place much emphasis on polished surfaces or a precise cut. It is likely that the arcades were covered with plaster, as were the walls, where parts of the painted plaster are still preserved. Phase 3 The most striking feature from this phase are the large piers, the massive stumps of which are preserved between the aforementioned columns. Perhaps structural problems were the reason for this rebuilding activity, as suggested by the fact that deeper foundations were laid out. The piers, about 2 meters in length and 1.30 meters in width, were inserted into every third intercolumniation. They were partly set into the nave, thereby narrowing that space to a width of 5.20 meters. Wiegand interpreted the piers as an indication that the ear7

Ibid., 481.

STEPHAN WESTPHALEN lier, wooden-roofed basilica was vaulted at this time.8 It is possible to reconstruct arches between the piers, which defined three square bays of considerable regularity. While such vaulting or a configuration of three domes cannot be excluded, there are no further indications that this design was adapted. The earlier excavators make no mention of large heaps of rubble, which would have pointed to collapsed vaults; nor is there any significant quantity extant of brick, the building material generally used for vaulting. THE ARCHITECTURE The architecture is difficult to date, because it corresponds to a type used over centuries. The walls contain no architectural sculpture from the Byzantine period, because only spolia were used in their construction. The basilica of phase 2 can be compared with other churches in the area. Similarities can be noted between our basilica and that in the adyton of the temple of Apollo at Didyma, the ruins of which were pulled down during the excavation of the 1910s. This three-aisled basilica with galleries was built in the late fifth or early sixth century A.D.9 The churches at Didyma and Priene are nearly identical in terms of their dimensions, proportions, and building techniques. At Priene the naos covered a space of 353 square meters (26.75 ⫻ 13.20 m), whereas at Didyma the church was just slightly larger, with an interior space of 387 square meters (25.83 ⫻ 15 m). The nave at Priene is about 6.80 meters wide, while that at Didyma is 6.72 meters wide according to the published ground plan. The primary difference between the two churches concerns the placement of the columns. The basilica at Priene had ten columns with intercolumniations of about 1.80 meters. At Didyma, on the other hand, the basilica had only eight columns with wider intercolumniations of some 2.35 meters. As at Priene, older material was reused at Didyma, and little care was taken in its placement. At Didyma the columns of the ground floor were assembled from spolia. For the galleries, by contrast, new capitals

were made. They are similar to those used in the Bischofskirche at Miletus and were perhaps sculpted in the same workshop.10 The Bischofskirche is the older of two churches excavated at Miletus. Its construction was completed before that of the Justinianic city walls and is therefore dated to the early sixth century A.D. at the latest.11 It has the form of a basilica with galleries and includes an atrium to the west and additional rooms, including a baptistery and a chapel of central plan, at the sides. The decoration was quite rich; columns and capitals were carved, and the floors were covered with figural mosaics. Most striking, however, is the small size of the Bischofskirche, the nave of which covered a space of only 257 square meters (18 ⫻ 14.30 m) and was divided from the side aisles by only six columns. These comparisons suggest that the basilica at Priene was probably also of the type with galleries above the side aisles, even though this is not definitively proven by the preserved material on site. Access to the galleries was probably obtained by a stair ramp adjoining the building, since there is no trace of stairs inside the church. The similarity of our basilica with the buildings at Didyma and Miletus is striking and suggests that the church at Priene (phase 2) must also have been built in the late fifth or early sixth century A.D. In any case, basilicas with wooden roof, columns, and galleries were a widespread phenomenon and render a precise dating impossible.12 Perhaps of more interest than the standard building type employed are the reasons for the selection of the site (phase 1). At Didyma, the church was built in the middle of the ancient temple. Its construction, therefore, signals the replacement of pagan worship by Christianity, since the members of a pagan community would not have been allowed to enter a holy enclosure reserved for priests. Presumably, reoccupation of the traditional religious space was more important at the famous oracle of Didyma than at Priene, where a similar attitude guiding the selection of the site is not appar10

Peschlow, “Byzantinische Plastik,” 211. W. Mu ¨ ller-Wiener, “Die ‘Grosse Kirche’ (sog. Bischofskirche) in Milet,” IstMitt 23/24 (1973/74): 131–34, plan 3; Peschlow, “Byzantinische Plastik,” 211. 12 R. Krautheimer, Early Christian and Byzantine Architecture, 4th ed. (Harmondsworth, 1986), 106 ff. 11

8

Ibid., 484. H. Knackfuß and T. Wiegand, Didyma (Berlin, 1941), 29–37; U. Peschlow, “Byzantinische Plastik in Didyma,” IstMitt 25 (1975): 211. 9

277

278

BYZANTINE BASILICA AT PRIENE

ent. Here, the church is situated well outside the temple precinct, near the bath and theater. The presence of revised building material, like the columns from the stoa on the terrace of Athena, does not indicate that a certain edifice was pulled down because of religious reasons. Secular reasons likely played a role in the selection of the site, for example, the availability of an inexpensive piece of land. Precisely when the site was leveled for the church is not known, but it cannot be excluded that construction of the basilica had begun already in the first half of the fifth century A.D., that is, when the first bishops from Priene are recorded.13 The wooden-roofed basilica was altered by the construction of piers (phase 3), which divided the nave into three bays. It is likely that these bays were covered by arches, vaults, or even domes. Similar rebuilding programs characterized by the addition or even replacement of older columns with piers are known from several other basilicas. In Binbirkilise, church no. 1 was restored during the ninth century A.D., after the town had been deserted for a long period. Rectangular piers were placed in front of the supports of the early Byzantine period in order to carry a new vault over the nave.14 At Binbirkilise it is believed that this change was effected to achieve a different conception of space (Raumgedanke) rather than for structural reasons. Generally, piers in the churches of Asia Minor probably date to the ninth century A.D.15 We may argue, however, that the alternation of columns and piers, as well as the domed or vaulted form of basilica, was not unknown in the sixth century A.D. The church at Priene can be compared, for example, with a church from the nearby island of Samos. Here, at Pythagorion in 1979, a church was excavated that had been incorporated into an older building, the so-called Dontia.16 The three-aisled basilica has an externally polygonal apse, a narthex, and an atrium. The 13

Wiegand, Priene, 475 f. M. Restle, “Binbirkilisse,” RBK 1:699. 15 W. Nausch, “Untersuchungen zur Pfeilerbasilika in Kleinasien” (Ph.D. diss., Ludwig-Maximilians-Universita¨t, 1977), 55, 58, 163. 16 Short report by K. Tsakos, A⌬ 34 (1979): Cronika´ 355–57, pl. 162b-d; N. Poulou-Papademetriou, “Samos pale´ochre´tienne: l’apport du mate´riel arche´ologique” (Ph.D. diss., Sorbonne, 1985), 73–78, pls. 42–45. 14

naos is 330 square meters (22 ⫻ 15 m), the same area as the basilica at Priene, and the nave is also 6.80 meters wide. During a later restoration, the church was outfitted with piers alternating with two columns. The piers (1.25 m long, 1 m wide) are thinner than those at Priene, but colossal piers at the outer walls indicate that the space was vaulted. The rebuilding can be dated to Justinianic times by its architectural decoration. Finally, we should recall the great buildings of the sixth century, such as the church of St. John in Ephesos17 or the basilica at the nymphaeum in Hierapolis in Phrygia,18 whose impressive masonry vaults may well have influenced the rebuilding of formerly wooden-roofed basilicas elsewhere. THE LITURGICAL FURNITURE The liturgical furniture of the Priene basilica is relatively well preserved. The synthronon in the apse and the foundations of the altar and ambo in the nave are still partly in situ. In addition, there are approximately fifty pieces of chancel screens and posts, preserved either completely or as fragments in the church.19 At the synthronon, two different states can be distinguished. At first it featured four steps and was covered by plaster colorfully painted in imitation of small marble slabs. Later, probably during phase 3 of the building, the synthronon was rebuilt in brick and reduced in size to just three steps.20 The most important furnishing to survive is the ambo (Figs. 3, 4), which can be reconstructed in most of its parts.21 In the nave lies an octagonal plinth for the small columns that carried the platform, now lost. West of this stands a staircase on a socle, which is worked 17 H. Ho ¨rmann, Die Johanneskirche, Forschungen in Ephesos 4.3 (Vienna, 1951), 157 ff, pl. 68 f; M. Bu ¨ yu ¨ kkolanci, “Zur Bauchronologie der justinianischen Johanneskirche in Ephesos,” Akten des 12. Internationalen Kongresses fu ¨r Christliche Archa¨ologie, Bonn 22.–28. September 1991, JbAC Erga¨nzungsband 20.1 (Mu ¨ nster, 1995), 598–602. 18 P. Verzone, “Hierapolis,” RBK 2:1215–20. 19 For a detailed catalogue, see Westphalen, “Basilika von Priene,” 330–39, nos. 62–117. 20 For details see ibid., 303–5, fig. 13, pl. 36.1. 21 Verzone, “La ‘Cattedrale’ di Priene,” 272 ff, figs. 9–13; O. Feld, “Christliche Denkma¨ler aus Milet,” IstMitt 25 (1975): 198 f, pl. 34.2; idem, Spa¨tantike und fru ¨hes Christentum, Propyla¨en Kunstgeschichte, suppl. 1, ed. B. Brenk (Frankfurt, 1977), 168, no. 133; Peschlow, “Byzantinische Plastik,” 238, fig. 2, pl. 49.4.

STEPHAN WESTPHALEN

279

as a monolith from a square block of marble. It has four risers, the upper one of which is half as high as the three below. The top of the staircase is flat. The sides are divided by semicircular niches and decorated with reliefs of vine scrolls and peacocks. Only the central section of the ambo remains in situ. A fragment belonging to a second staircase of this ambo was also found. Perhaps more important was the discovery of another block, also decorated with vine scrolls and containing two risers, the lower one of which is half as high as the upper one. This block fits perfectly with the remains in situ, and together they form the complete staircase of the ambo. It consists of five risers leading to the now lost platform that lay ca. 1.90 meters above the floor level. The type of ambo with two flanking staircases is widespread.22 The construction of staircases out of massive blocks, together with its decoration of vine scrolls and peacocks, is limited to the region of Caria and Ionia.23 The ambo of Priene is, thus, a representative example of this regional type. The fragments from the chancel screens are particularly numerous and indicate a variety of liturgical functions. Most are screens that served to divide the public from the altar. There are, however, clear signs of another screen that separated the nave from the aisles. Vertical grooves in the columns (Fig. 5a) were carved to receive closure slabs, of which no parts remain. The size and shape of these slabs can, however, be easily deduced from the form of the sockets. The slabs were placed around 60 centimeters above the floor, most likely on a socle of stone or bricks. They were between 90 and 100 centimeters high and capped with a cornice. The sockets are still to be seen on eighteen drums, which belong to phase 1 of the church. Such high screens placed between the nave and aisles were perhaps used to separate men and women, as described by John Chrysostom in the churches of Antioch.24 These

divisions of the aisles of the church were abandoned by the time of its second construction phase. The chancel screens dividing the altar from the nave (the templon) were rearranged several times. To phase 1 belong the screens and the piers that, according to an inscription, were donated by a certain Iulianos, an eparchikos (Figs. 6, 11e, f). The screens are decorated on the front side with a lattice pattern, on the back with a cross.25 This screen arrangement was presumably demolished, at the latest, during the rebuilding of the church. In phase 2, the chancel was placed between the third pair of columns. The southern column is still in situ (Fig. 5b). On its northern side is a groove intended to hold a screen. The western side of the column is marked by crosses, parts of which are carefully carved and were originally filled with metal, parts of which are scratched, much like graffitto. In phase 3, the chancel was placed even further to the west. Of its screens only their impressions made in the plaster of the large piers are preserved. These changes in the chancel screens reflect the various rebuildings of the church. In every phase, the liturgical furniture was restored or, to a certain extent, replaced. The church contained other liturgical furniture as well, fragments of which are preserved, but cannot be easily reconstructed. To this group belong posts of different dimensions and bearing different patterns. The posts of the chancel screen of phase 1 have a simple soffit decoration (Fig. 11e) on the front side (101 cm high). From these a post of the same size can be distinguished, decorated with a vine scroll (Fig. 11b). Some larger posts (135 cm high) were provided with simple flutes or with soffit decoration (Figs. 7, 11a). The screens likewise, are heterogeneous. The slabs decorated with a lattice pattern have already been mentioned (Figs. 6, 11f). There were also large screens (131 cm high) bearing a simple cross

22 P. Jakobs, Die fru ¨hchristlichen Ambone Griechenlands (Bonn, 1987), 44–50. 23 Feld, “Christliche Denkma¨ler,” pls. 34, 35; Peschlow, “Byzantinische Plastik,” pl. 49. M. Falla-Castelfranchi, “Amboni carii: un caso desueto,” in Bild und Formensprache der spa¨tantiken Kunst. Hugo Brandenburg zum 65. Geburtstag [Boreas 17] (Mu ¨ nster, 1994), 49–52. 24 F. van de Paverd, Zur Geschichte der Messliturgie in Antiochea und Konstantinopel gegen Ende des vierten Jahrhunderts:

Analyse der Quellen bei Johannes Chrysostomos, OCA 187 (Rome, 1970), 25–27. 25 Wiegand, Priene, 482–84, figs. 590, 591, 596, 597; T. Ulbert, “Studien zur dekorativen Reliefplastik des o ¨stlichen Mittelmeerraumes (Schrankenplatten des 4.–10. Jahrhunderts)” (Ph.D. diss., Ludwig-Maximilians-Universita¨t, 1968), cat. 29, no. 85, text 16; Verzone, “La ‘Cattedrale’ di Priene,” 266, fig. 6.7; Schultze, Altchristliche Sta¨dte, 140 f, fig. 50.

280

BYZANTINE BASILICA AT PRIENE

(Figs. 10, 11c) and lower screens (67 cm high) with rhomboid decoration (Figs. 9, 11d). The most richly decorated screen (Fig. 8) bears a cross and two confronted peacocks (79 cm high). The decorated sculpture exhibits several common characteristics. The motifs are limited to a few common symbols and patterns. The relief is simply carved. The marble was mainly reused spolia. All features indicate that the liturgical furniture was made by workshops in the region and was not imported from Constantinople or elsewhere. Beyond these similarities, however, the pieces have obvious differences. As one example, we can compare the ambo (Figs. 3, 4) with the screen decorated with peacocks (Fig. 8). In each the tail feathers are rendered differently: on the screen, the eyes of the feathers appear like pearls on a string, whereas on the ambo the eyes are carefully differentiated from the hatched barbs of the feathers. Differences can also be recognized between the ambo and the post with the vine scroll (Fig. 11b). On the post, the leaves are formed as stylized hearts, whereas the vine leaves of the ambo are rendered in a more naturalistic manner with four or five leaf tips. Stylistic differences are likewise apparent in the carving technique. The motifs are incised, as in the slabs containing the rhomboid decoration (Figs. 9, 11d), or are worked in high relief, as in the large screens with a cross surrounded by a circle (Figs. 10, 11c). How can we explain these differences? The material might not have been worked at the same time. As we have shown, the church underwent fundamental reconstruction, and the chancel screens were altered several times. The screens and posts might have been restored and reworked again and again. In any case, the sculpted material is not suitable for showing temporal differences or for presenting chronological variations. Moreover, it is highly questionable that workshops existed in early Byzantine Priene. Perhaps the liturgical furniture was ordered from foreign shops, or perhaps itinerant craftsmen traveled to Priene when required. Presumably

they were Ionian or Carian, but they cannot be localized with any certainty. Only rarely can good parallels for the Priene carvings be found. The slabs with the rhomboid motif (Figs. 9, 11d), for example, can be compared with an ambo in the Bischofskirche in Miletus. As in Priene, the motifs there are shallowly carved with a round groove.26 The dating of the Bischofskirche to the late fifth or early sixth century A.D. can thus suggest a similar date for the slabs in Priene. The church at Didyma was built at the same time. To this edifice belong posts decorated with vine scrolls that are markedly similar to those on the post in Priene (Fig. 11b). In both cases, two leaves and a fruit grow out of each tendril. These examples reflect well-known motifs taken up by the workshops in the Proconnesos and subsequently spread all over the Byzantine world. Closely related motifs can be found in the chancel screen sculpted in the 530s A.D. and reused in the Romanesque Schola Cantorum of San Clemente in Rome.27 To conclude, we note that the relief sculpture of Priene demonstrates the same dichotomy that Friedrich Wilhelm Deichmann observed at Ephesos: while the motifs are similar to those found at Constantinople and surroundings, the workmanship is different. The situation at Ephesos is paradigmatic. Local workshops were still active everywhere in Asia Minor during late antiquity, although on a more modest scale. Deichmann’s judgment concerning the workshops of Ephesos, that is, that they were still producing some worthy, albeit simple, works, is valid for Priene as well.28 26 O. Feld, “Milet 1972,” IstMitt 23/24 (1973/74): 136 f, pl. 38, 3.4. 27 F. Guidobaldi, C. Barsanti, and A. G. Guidobaldi, San Clemente: la scultura del VI secolo, San Clemente Miscellany 4.2 (Rome, 1992), 210 f, figs. 136–97. See also U. Peschlow, ` " Timhtiko´ " “Zum Templon in Konstantinopel,” in Armo J to´ mo" sto` n kaqhghth` N. K. Moutso´ poulo (Thessalonike, 1991), 3:1449–75. 28 F. W. Deichmann, “Zur spa¨tantiken Bauplastik von Ephesos,” Mansel’e Armag˘an, Me´langes Mansel (Ankara, 1974), 549–70; reprinted in Rom, Ravenna, Konstantinopel, Naher Osten (Wiesbaden, 1982), 560 f, 570 f.

This is an extract from:

Dumbarton Oaks Papers, No. 54 Editor: Alice-Mary Talbot Published by

Dumbarton Oaks Research Library and Collection Washington, D.C. Issue year 2000 © 2000 Dumbarton Oaks Trustees for Harvard University Washington, D.C. Printed in the United States of America

www.doaks.org/etexts.html

1

Plan, basilica of Priene

2

Priene basilica, general view, looking east ( photo: German Archaeological Institute, Istanbul)

3

Priene basilica, the ambo, south side of the staircase ( photo: German Archaeological Institute, Istanbul)

4

Priene basilica, the ambo, reconstructed

a

b

5

Priene basilica, columns with grooves for chancel screens: (a) phase 1; (b) phase 2

6

Priene basilica, screen with lattice pattern ( phase 1) (photo: German Archaeological Institute, Istanbul)

7

Priene basilica, large post with soffit decoration (photo: German Archaeological Institute, Istanbul)

8

9

Priene basilica, screen decorated with peacocks (photo: German excavations at Miletus, University of Bochum)

Priene basilica, screen with rhomboid decoration (photo: German Archaeological Institute, Istanbul)

10

Priene basilica, screen with a cross ( photo: German Archaeological Institute, Istanbul)

c

b a

d

e

11

Priene basilica, various screens and posts (drawings by M. Pfister and Stephan Westphalen)

f