Journal of Cuneiform Studies (JCS 55) 2003

A VICTORY OVER MARI AND THE FALL OF EBLA* Alfonso Archi and Maria Giovanna Biga Università di Roma “La Sapienza” superior stabat lupus

1) The Supremacy of Mari; 2) {I-daåar of Mari; 3) Relative Chronology of Isar-damu’s Reign (Ministers Ibrium and his Son Ibbi-zikir); 4) The Alliance between Ebla and Mari; 5) The Defeat of Mari; 6) After the Victory; 7) The Peace with Mari and the Triumph of the Minister Ibbi-zikir; 8) The Marriage of Kesdut with the Son of the King of Kis; 9) The Destruction of Ebla

During an initial phase Mari succeeded in conducting an expansionist policy. A letter by the king Enna-Dagan clearly has the aim of intimidating the king of Ebla.2 The tone used is threatening. The letter lists a series of victories won by Mari

1) The Supremacy of Mari For no more than ˜fty years during the twentyfourth century B.C. , the political scene in northern Syria and the region of the Middle Euphrates was dominated by two equally powerful cities: Ebla and Mari. All the other city-states were obliged to submit to the hegemony of one of these two great centers, on occasion passing from one camp to the other.1

conclusion reached by Viganò in “Mari and Ebla: of Times and Rulers,” ibid., 3–24. P. Michalowski, “Mari: The View from Ebla,” in Mari in Retrospect. Fifty Years of Mari and Mari Studies (Winona Lake: Eisenbrauns, 1992), ed. G. D. Young, 243–48, could not use the corrections which A. Archi, “Le synchronisme entre les rois de Mari et les rois d’Ebla au IIIe millénnaire,” MARI 4 (1985) 47–51 (especially p. 47, n. 4), introduced into his previous list of the Mari kings, where the ministers Arennum/Arrukum, Ibrium and Ibbi-zikir were still considered as kings of Ebla. A complete dossier of the relations between Ebla and Mari is in preparation. 2. TM.75.G.2367, preserved in the central archive of the royal palace of Ebla, was originally published by G. Pettinato, “Bollettino militare della campagna di Ebla contro la città di Mari,” OA 19 (1980) 231–45, and ˜rst commented on by B. Kienast, “Der Feldzugsbericht des Ennadagan in literarhistorischer Sicht,” OA 19 (1980) 247–61. It was, however, D. O. Edzard, “Neue Erwägungen zum Brief des Enna-Dagan von Mari (TM.75.G.2367),” SEb 4 (1981) 89–97, who gave the correct interpretation of this letter, which has been reedited by P. Fronzaroli as ARET XIII 4. For an extensive historical and geographical interpretation of this see M. C. Astour, Eblaitica 3 (1992) 26–39; see also J.-W. Meyer, AoF 23 (1996) 156–69.

*Abbreviations used in this article: AAM = Annual Accounts of Deliveries of Precious Metals; MAT = Monthly Accounts of Deliveries of Textiles; D. = DILMUN (determining gín); EDL = Enna-Dagan’s Letter; Ibr. = the minister Ibrium; I.Z. = the minister Ibbi-Zikir. 1. The political and economic relations between Ebla and Mari have been treated by A. Archi, “I rapporti tra Ebla e Mari,” SEb 4 (1981) 129–66.; Id., “Les rapports politiques et économiques entre Ebla et Mari,” MARI 4 (1985) 63–83; F. Pomponio, “Considerazioni sui rapporti tra Mari ed Ebla,” VO 5 (1982) 191–203; Id., “Funzionari di Ebla e di Mari,” in Anatolia Antica. Studi in memoria di F. Imparati (Eothen 11; Firenze: LoGisma, 2002), eds. S. de Martino and F. Pecchioli Daddi, 653– 63. The older economic documents have been ordered chronologically by L. Viganò, “Mari and Ebla: the Archives Reports,” in L. Viganò, On Ebla. An Accounting of Third Millennium Syria (Aula Orientalis Supplementa 12; Sabadell: Editorial AUSA, 1996), 25–51. The present authors cannot accept the

1

JCS 55 (2003)

2

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

in the area of the Middle Euphrates, each of which was swiftly followed by the destruction of the enemy’s city. This expansion was begun by the ˜rst two kings, Anubu and Saåumu.3 Istup-(i)sar reached as far as Emar and annihilated the city. Finally, Iblul-Il ranged repeatedly up and down the valley of the Euphrates, as far as the territory of {a-su-wa-anki, upriver from Karkemis, sowing terror and destruction in his wake.4 In order not to see her own territory invaded, Ebla paid a heavy tribute which was handed over to Mari near Mane (downriver from Emar), the site that continued to act as Ebla’s port on the Euphrates.5 Enna-Dagan does not claim any undertaking other than the completion of a military expedition begun by Iblul-II. It is obvious that the letter, in which Enna-Dagan makes clear his program of continuing Iblul-Il’s political hegemony, was

3. F. Pomponio, AfO 35 (1988) 166, has highlighted that in obv. ii 1 one has to read the PN A-nu-KA. The text has perhaps: A-nu-KA[(x†U)]. A. Alberti, NABU 1990/124, has suggested to identify the An-bu of the Sumerian King List with this Anubu; see also J. S. Cooper, Sumerian and Akkadian Royal Inscriptions I (New Haven: The American Oriental Society, 1986) 4 n. 12. Other identi˜cations with fragmentary names in the List have been proposed by M. J. Geller, Eblaitica 1 (1987) 144–45. The Tell Leilan recension shows now that the six Mari kings of the List are not those attested in the Ebla documents, see C.-A. Vincente, ZA 85 (1995) 241–42, 257–60. 4. Pettinato’s edition has to be corrected here, obv. ix 1–13: in Ne-ra-at ki ù in åÀ-aski (!) {a-su-wa-anki è Ib-lul-il lugal Ma-rí ki ù mu-DU Ib-laki . . . “Against Nerat and åAåas of {asuwan moved Iblul-Il, the king of Mari. The delivery of Ebla he received inside Mane.” The toponym åÀ-aski could be perhaps identi˜ed with åÀ-su/suki, a small center in the territory of Ebla (at least in later times), which is, however, never connected with {asuwan, see A. Archi, P. Piacentini and F. Pomponio, I nomi di luogo dei testi di Ebla (Archivi Reali di Ebla Studi 2; Roma: Missione archeologica italiana in Siria; 1993) (= ARES 2), 134–36. The identi˜cation of {asuwan with later {assum/{assuwa, see ARES 2, 266–67, is not accepted by M. C. Astour, “{assu and {asuwan. A Contribution to North Syrian History and Geography,” UF 29 (1997) 1–66. 5. M. C. Astour, “The Geographical and Political Structure of the Ebla Empire,” in Wirtschaft und Gesellschaft von Ebla (Heidelberger Studien zum Alten Orient 2; Heidelberg: Heidelberger Orientverlag, 1988), eds. H. Waetzold and H. Hauptmann, 146 n. 47, identi˜es this toponym with URUMani-e of the Chaldean Chronicle, see D. J. Wiseman, Chronicles of Chaldean Kings (626–556 B.C.) (London: The Trustees of the British Museum, 1961), 54–55:7: “Manê, Sahiru and Balihu . . . .” See further M. C. Astour, Eblaitica 4 (2002) 111–15.

written soon after he assumed the throne. Events were, however, to take a diˆerent turn as he was soon to die. The document TM.75.G.1953, sums up “the deliveries, tributes,” mu-DU (as they are appropriately de˜ned by Enna-Dagan) paid in gold and silver by Ebla to Mari during the reigns of Iblul-Il, Nizi and Enna-Dagan. The Eblaite administration, for ideological reasons, preferred to call these “gifts,” níg-ba: the obligation to deliver a mu-DU was applied to an individual or a city that had been forced to submit to another authority. The amounts involved are staggering and unparalleled for such an early date (1 ma-na = 470 gr): 6 Silver (ma-na) Iblul-Il (king) 1,164.20 ábba “elders” ,184.18 maskim-e-gi4 ,272.38 “envoys” Nizi (king) ,154.30 ábba “elders” , 49.30 , 77.17 maskim-e-gi4 “envoys” Enna-Dagan (king) ,153.50 ábba “elders” , 53.48 maskim-e-gi4 , 77.50 “envoys” Totals

2,188.01

(kg)

Gold (ma-na)

(kg)

,547.23 , 86.61 ,128.13

89.30 . 0 14.46

42.06

, 72.61 , 23.26 , 36.32

17.27 . 0 . 3

8.19

, 72.29 , 25.28 , 36.57

8.03 . 0 1.40

0.78

1,028.30

134.26

63.15

6.93

1.41 3.78

This document reveals to us the existence of a Mari king by the name of Nizi (Ni-zi), Iblul-Il’s immediate successor, omitted in Enna-Dagan’s letter, perhaps because nothing of great importance happened during his reign. Nizi received the same amount of silver as Enna-Dagan, but ˜ve more kilos of gold. In all probability this tablet was written on the death of Enna-Dagan and marks the start of a new phase, in which Ebla would exchange “gifts,” níg-ba, with Mari on an equal footing within the context of ceremonial exchanges. Certain factors indicate that EnnaDagan only ruled for a few years. The tablet ARET II 4 xv 2–xvi 1 records the death of Iblul-Il: GIBIL-za-il La-da-ad Íl-zi-da-mu 6. Archi, SEb 4 (1981) 132–35; see also Archi, MARI 4 (1985) 64–65 (T. a).

A VICTORY OVER MARI AND THE FALL OF EBLA Ig-na-da-mu su-mu-“tag4” Ib-lul-il lugal DI† mu ug7 ÉxPAP “(gifts which) PN1-PN4 have brought (for) Iblul-Il, the king (of Mari, in) the year of (his) death (for) the funerary ceremony.”7 These four people were among the principal o¯cials (called lugal, “lord,” at Ebla8) during the ˜rst period documented in the archives, corresponding to the reign of Igris-{alab. The death of a king of Mari (together with that of a king of Kakmium) also provides the date for a cattle account, TM.75.G.1574 obv. iii 2–rev. i 4: in DI† mu lugal Ma-rí ki ús ù en [Kak-]mi-umki [ú]s “in the year the king of Mari died and the king of Kakmium died.”9 Darmia, who appears in this document, was the principal “lord,” lugal, during the reign of Igris-{alab. During the last years of Igris-{alab, Arrukum10 was a “lord,” lugal.11 He also appears in ARET II 4 ix 5 on the occasion of a journey to Mari. When Arrukum became minister, Irkab-damu was already king. In the documents of Arrukum the death of Igris-{alab is not mentioned and it is probable that Irkab-damu had only come to the throne a few years earlier. In fact, the main o¯cials mentioned in the mu-DU documents before the period of Arrukum were already in service at the time of king Igris-{alab. Arrukum held the post of minister for perhaps less than ˜ve full years, as he appears in four annual muDU texts, in four Annual Accounts of the distribution of Metals [AAM], and in about sixty Monthly Accounts concerning the distribution of Textiles [MAT].12 He died a few months before 7. On ÉxPAP “funerary ceremony,” see A. Archi, Amurru 1 (1996)17–18. 8. See A. Archi, “Les titres de en et lugal à Ebla et des cadeaux pour le roi de Kish,” MARI 5 (1987) 37–52. 9. SEb 7 (1984) 68–69. 10. This name Ar-EN-LUM was ˜rst read as Ar-en-núm, then as Ar-ru12-lum and now, by the present writers, as Arru12-gúm. 11. The relevant documents for the synchronism between Ebla and the kings of Mari are quoted in Archi, MARI 4 (1985) 48–49. At that time, the ministers Arrukum (Arennum), Ibrium and Ibbi-zikir were erroneously considered kings of Ebla. 12. The four mu-DU texts are listed by A. Archi, “The ‘lords,’ lugal-lugal, of Ebla. A Prosopographic Study,” VO 12 (2000) 19–58 (see pp. 35–36). The monthly documents registering textiles that mention Arrukum will be edited by F. Pomponio.

3

his master Irkab-damu.13 Of the four o¯cials who carried funerary gifts to Mari for Iblul-Il, only GIBIL-za-Il was still active under minister Arrukum.14 Iblul-Il must therefore have died either during the very last years of Igris-{alab or in the ˜rst year of Irkab-damu’s reign, if we attribute to this king about eight to ten years on the throne. Baba, the spouse of Iblul-Il, is mentioned in four documents.15 The death of Nizi is recorded in TM.75.G.1299 (SEb 4, 137–39: T. 3) rev. iv 3–4: in ud Ni-zi ús-ús DI† mu “the year in which Nizi died.” In this document Arrukum ˜gures among those sent by Ebla to deliver tribute to the king of Mari (obv. iii 3). Some of the people mentioned here, however, were only active in the period preceding that 13. Archi, Amurru 1 (1996) 23–26. 14. The lists of the lugals were published by Archi, VO 12 (2000) 48–55. 15. For the identi˜cation of this queen see F. Pomponio and M. G. Biga, NABU 1989/114, who interpreted correctly l. 4 of the Mari inscription MP 12 of I. J. Gelb and B. Kienast, Die altakkadischen Königsinschriften des Dritten Jahrtausends v. Chr. (FAOS 7; Stuttgart: Franz Steiner Verlag, 1990) 12: [Ib-lul]-il lugal Ma-rí ki Pa4-ba4 nin “[Iblul]-Il, king of Mari (and) Paba, the queen.” She is also attested in ARET VII 3 obv. v 5, rev. iv 7; 7 rev. i 3; 12 obv. iii 1. The prosopographic data con˜rm the dating of these three documents to the last years of king Igris-{alab. Another document, TM.75.G.1559 (= T. 9 in SEb 4 [1981] 155–62), from the time of Enna-Dagan, covers four years and mentions Enna-Il, the steward of Pa4-ba4, in obv. I 2. The title of the queen was nin, see also TM.75.G.2460 (to be dated to the minister Ibbi-Zikir) obv. v 14–vi 1: lugal Ma-rí ki . . . nin Ma-rí ki. The spouse of {I-daåar, the last king of Mari known from the Ebla archives had the same name, ARET VIII 533 viii 13–16: Ba-ba4 dam {I-da-ar Ma-rí ki “(gifts for) Baba, the spouse of {I-daåar, (the king) of Mari” (this document has to be dated to the very last years of the central archive). Pa-a-ba4, ARET I 44 obv. i 7, rev. ii 3, must be another writing of the same name. W. Sallaberger, Subartu 4,2 (1998) 36–37, has suggested that Pa4-ba4, attested to in: F. Ismail, W. Sallaberger, Ph. Talon and K. van Lerberghe, Administrative Documents from Tell Beydar (Seasons 1993–1995). Subartu 2 (Turnhout: Brepols, 1996) (= Subartu 2), 23 i 1, is to be identi˜ed with the spouse of Iblul-Il of Mari. This suggestion could hold true if Paba was a princess from Nagar given in marriage to Iblul-Il. It is also possible that this Paba was the spouse of king {I-daåar, who visited Ebla about two years before the destruction of the city, see ARET VIII 533 viii 13– 21: Ba-ba4 dam {i-da-ar Ma-rí ki lú gi4 mi-nu Ib-laki si-in Ma-rí ki “Baba, the spouse of {I-daåar (king) of Mari who returned from Ebla to Mari” (see below, at the end of § 8). In this case, the chronology of the tablets from Tell Beydar suggested by Sallaberger should be lowered (see also note 68).

4

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

which saw Arrukum at the head of the administration, such as Ladat “the judge,” di-kud (obv. iii 6–7). Ladat is also attested in TM.75.G.1987 (SEb 4, 139: T. 4) and in ARET VII 1 + TM.75.G.12683, obv. ii 1 (!). The two “judges” of the four yearly mu-DU documents of the minister Arrukum were Enna-I(l) and Ibdur-isar.16 The latter had taken the place of Ladat, who was “judge” together with Enna-I(l) in the years immediately preceding. Arrukum, who appears regularly in the documents relating to Nizi, is the head of the delegation to Mari in TM.75.G.1866 (SEb 4, 139–140: T. 5). Nizi reigned for at least three years: ARET VII 1 + TM.75.G.12683 rev. x 4–7: dub-gar Ni-zi lugal fimu-flDU 3 mu “document (concerning) Nizi, the king: delivery (for) three years.” ARET VII 16 (41): dub-gar lú su-ba4-ti in ud Ni-zi lugal 3 mu “document of what was received (in Mari) when Nizi (was) king (in his) third year.” ARET VII 16 (19) mentions Ma-ra-AN en Na-gàr ki “M. king of Nagar.” This king is possibly to be identi˜ed with AMAR-AN dumu UR-dUTU.†A [en?] Nagarki, who dedicated a statue of Iblul-Il to Inanna-ZA.ZA in Mari.17 Enna-Dagan, who succeeded him, already appears in some of his predecessor’s documents as the receiver of gifts.18 ARET VII 17 relates to the ˜rst year of Enna-Dagan’s reign, rev. vi 1–3: ás-du En-na-dDa-gan lugal 1 mu “since EnnaDagan has been king: one year.” Of the two o¯cials mentioned in obv. ii 4–6, Iba-zinu and Isgi-daåar, only the latter was still active when Arrukum became minister. Some other documents of king Enna-Dagan must, for prosopographical reasons, be dated to the period immediately before Arrukum was minister, such as MEE II 13 obv. i 3–ii 7:19 mu-DU Ig-na-da-mu ul-ki . . . mu16. Archi, VO 12 (2000) 35–36. 17. Gelb and Kienast, Königsinschriften, 12, no 12 (= MAM 3, 318–19 no 11). See Sallaberger, Subartu 4,2 (1998) 35 with n. 59. 18. TM.75.G.1299 (SEb 4, 137–38: T. 3) rev. II 2–3; ARET VII 1 obv. iii 6, rev. iii 1, vii 2; 16 obv. vi 8, viii 1, rev. i 4. An Enna-Dagan from Mari was active for several years as a messenger after the death of the king of the same name, see TM.75.G.1923 obv. xii 8; 2507 rev. xix 8. 19. Rev. v 3–5 + vi 3 reads: ás-du En-na-dDa-gan lugal 1 mu. This document, therefore, comes from the same year of ARET VII 17.

DU Ti-ir è níg-ba En-na-dDa-gan lugal “delivery of Igna-Damu, the ulki; . . . delivery of Tiåir: expenditure (as) gift (for) Enna-Dagan, the king (of Mari).” Tiåir, one of the most important “lords” before Arrukum, remained in the administration under Arrukum; Igna-damu, instead, withdrew from service. TM.75.G.1233 (SEb 4, 145–54: T. 8) must be of the same date as MEE II 13. The text MEE II 35 rev. x 3–5 shows that Enna-Dagan was king for more than two years: dub-gar níg-ba En-na-[d]Da-gan [1/2+]2 mu “document of the gifts (for) Enna-Dagan (for) 3/4 years.” This text, which refers to a number of years, must have been drawn up immediately after the death of king Irkab-damu, rev. viii 9–XI 1: in ud níg-ba en su-mu-“tag4” ÉxPAP “(silver as gift for PN) on the occasion of the gift (which) the king (of Ebla) has brought (to his own?) funeral ceremony.”20 Apart from the famous letter which allows us to reconstruct the expansion of the Mari state, the archives of Ebla also include another document from Enna-Dagan, TM.75.G.1913+, which deals with events of his reign.21 When Nizi died, certain o¯cials were still in service who would then disappear under minister Arrukum. The latter died a few months before his king, Irkab-damu, and had held his post for four years and several months.22 The period of EnnaDagan’s reign covers those few years in which Arrukum was minister, and began at least one year earlier. During the reign of Irkab-damu, Mari suˆered a defeat, possibly at the hands of Ebla. ARET VII 115 rev. i 1–ii 6: sikil dåÀ-da Ìr-kabda-mu sikil in mu Ma-rí ki àga-kár ás-ti åÀ-ti-ni ki “puri˜cation of the god Hadda, Irkab-damu has puri˜ed (?) in the year (in which) Mari was vanquished by åAtini.”23 This must have been a minor con˘ict seeing as it is mentioned nowhere else. 20. The expression PN su-mu-“tag4” ÉxPAP has been discussed by A. Archi, ZA 92 (2002) 184. Minister Arrukum, who died several months before king Irkab-Damu, is mentioned in obv. viii 5. 21. This document has been published by P. Fronzaroli as ARET XIII 1. 22. See note 12. 23. On the sikil ceremony, see A. Archi, NABU 2000/65.

A VICTORY OVER MARI AND THE FALL OF EBLA

The locality åAtini appears to occur only in this document.24 It is uncertain if the news of a defeat of Mari registered in the MAT TM.75.G.1405 (to be dated to the king Irkab-damu, because the minister is Arrukum) refers to the same event, obv. vi 12–vii 1: (1 cloth) Bù-da-ma-lik lú En-na-ì dikud níg-mul Ma-rí ki til “(1 cloth for) Buda-malik, the man (i.e. son ?) of Enna-i(l), the judge, (for) the news (that) Mari was defeated.” Iku(n)-isar, Enna-Dagan’s successor, is mentioned in an AAM from the ˜rst year of minister Ibrium (in the ˜rst months of the same year in which king Irkab-damu died): TM.75.G.1705 (Ibr. 1.b) rev. vi 4–10: tar kù-gi I-ku-i-sar en Ma-rí ki Du-tum Ma-rí ki su-mu-“tag4” “30 (shekels) of gold for Iku-isar, king of Mari, Dutum of Mari has brought.” The only other reference to this king is in ARET I 11 (17): I-ku-sar en Ma-rí ki I-ga-is-ru12 su-mu-“tag4” “(garments for) Iku(n)-(i)sar, king of Mari, Iga-isru has brought.” This MAT registers in section (41) the gifts for Tahir-malik, daughter of minister Ibrium, on the occasion of her marriage. It has to be dated to the same year as the AAM Ibr. 1.b, which mentions “the marriage of a Ibrium’s daughter” in rev. XI 18–21: in ud nígmu-sá dumu-mí Ib-rí-um and has other parallel sections. All these elements suggest that Enna-Dagan of Mari and Irkab-damu of Ebla died during the same year. MEE II 35 is, therefore, the summary of gifts sent to Mari during the reign of EnnaDagan. 2) {I-daåar of Mari The successor of Enna-Dagan of Mari, Ikunisar, is only mentioned in the ˜rst AAM of Ibrium (Ibr. 1.b) and in one MAT from the same year, which was also his ˜rst year (see above). The name of {I-daåar, who succeeded Ikun-isar, never appears in the AAMs, but only in some MATs from the ˜nal years of Ebla, that is about thirty years later (see Table 1). 24. If this åAtini is to be identi˜ed with A-ti-in/ni/nu/núm ki, a small center belonging to Ebla (see ARES 2, 120), this unfavorable event for Mari had to be just a clash.

5

1) TM.75.G.1249 rev. v 4–6: 2 maskim {I-da-ar Ma-rí ki (in rev. v 21–26, marriage of the princess Tagris-Damu with the crown prince of Nagar;25 last two/three years of Ebla). 2) ARET VII 64 (2): NI-ba-ku-tu dumu-nita {Ida-ar (in section [1]: Ni-zi Na-gàr ki: about the same year as the preceding document). 3) TM.75.G.2241 obv. i 2–4: {I-da-ar lugal Marí ki (in obv. vii 4–8: death of Zugalum, queen of {arran. The Eblaite princess Zugalum married the king of {arran in I.Z. 1; she is mentioned in I.Z. 3 and I.Z. 6. Her death does not appear in the preserved section of the AAMs. It has to be occurred, therefore, in the last three years of Ibbi-zikir, whose AAMs are extremely fragmentary). 4) TM.75.G.2270 obv. XI 13–XII 1: †um-da-ar dumu-nita {I-da-ar lú Ma-rí ki (in obv. vii: gifts of the queen for Za-åà-se, daughter of Ibbi-zikir, who had married the crown prince Ir-åà-ag-da-mu; last years of Ibbi-zikir). 5) TM.75.G.2278 rev. ix 5–8: †um-da-ar dumunita {I-da-ar Ma-rí ki (year of the peace after the victory of Ebla on Mari). 6) TM.75.G.2328 rev. iii 19–22: 3 dumu-nita tur maskim dumu-nita {I-da-ar. 7) TM.75.G.2334 obv. vii 5–8: A-ha-ar-sè dumunita {I-da-ar Ma-rí ki (last year of Ebla; death of the priestess Tarib-Damu, who is still alive in I.Z. 15) 8) TM.75.G.2353 obv. xi 16–18: maskim-e-gi4 {I-da-ar lugal Ma-rí ki (in obv. ii 1–2: expedition against Za-bur-rúmki; last years of Ibbi-zikir). 9) ARET VIII 521 vi 14–18: Ab-ba-ì-lum I-tinu maskim {I-da-ar (last two/three years of Ebla). 10) ARET VIII 522 xv 5–8: {I-da-ar Ma-rí ki . . . ábba-sù (last two/three years of Ebla). 11) ARET VIII 525 xx 16–19: †u-ma-dAs-dar Puzur4-ra-BE 2 maskim {I-da-ar (last two/ three years of Ebla). 12) ARET VIII 529 xxii 1–4: A-hu-DÙG maskim {I-da-ar Ma-rí ki (last two/three years of Ebla). 25. See M. G. Biga, Subartu 4,2 (1998) 18–19.

6

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

Table 1 Kings of Ebla

Ministers years

Kings of Mari years

Abur-lim Agur-lim Ibbi-damu Baga-damu Enar-damu Isar-malik Kun-damu Adub-damu Igris-{alabd Irkab-damu

Synchronisms with Babyloniaa

Ikun-Marib Ikun-†amaganb Ikun-†amasb Anubuc Saåumc Istup-isarc Iblul-Ilc d 1 (?) 2 3

Isar-damu

years

NIzi

1

(three years?)

2

Arrukum

3

4

2

Enna-Dagan

1

5

3

(four years?)

2

6

4

3

7

5

4

1

1

1

2

Ibrium

2

2

3

3

3

4

4

4+x

5

5

6

6

7

7

8

8

First defeat of Kis

9

9

Conquest of Aksak

10

10

11

11

Conquest of Kis by Ensakusana

12

12

13

13

14

14

15

15

16

16

17

17

18

18

19

Ibbi-zikir

1

20

2

21

3

22

4

A VICTORY OVER MARI AND THE FALL OF EBLA

7

Table 1 Kings of Ebla

Ministers years

Kings of Mari years

23

5

24

6

25

7

26

8

27

9

28

10

29

11

30

12

31

13

32

14

33

15

34

16

35

17

years

Synchronisms with Babyloniaa Conquest of Adab by Lugalzagesi 1–7<

{idaåar

x+1

Lugalzagesi +/-15

Fall of Ebla

?

Isgi-Mari

1 (?) 2 3 4 5 6 7

Lugalzagesi +/-25

8 9 10 (?)

Fall of Mari (?)

a

Synchronism according to W. Sallaberger Chronological sequence uncertain c Kings of Mari mentioned in the Ebla documents d Synchronism on the basis of o¯cials of the period of Igris-{alab who are active in documents which mention Iblul-Il b

13) ARET VIII 533 viii 13–21: Ba-ba4 dam {Ida-ar Ma-rí ki lú gi4 mi-nu Ib-laki si-in Ma-rí ki “(1 cloth 2 toggle pins for) Baba, the spouse of {I-daåar (king) of Mari who returned from Ebla to Mari.” 14) ARET VIII 541 iii 14–15: {I-da-ar Ma-rí ki. 15) The following document, from archive L. 2712, has to be dated to the last two years of Ebla, ARET IX 102: (very large amount

of victuals for) åÀ-si-hu dumu-nita {I-da-ar; (large amount of victuals for) {I-da-ar. {I-daåar is mentioned in other MATs, all from the period of Ibbi-zikir, but these cannot be dated even approximately at the current time: 16) ARET IV 21 obv. iv 1–2: 100 íb+III-TÚGgùn {I-da-ar Ma-rí ki su-ba4-ti lú DU.DU si-in

8

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

níg-kas4 “100 kilts: {I-daåar of Mari has received for a military expedition;”26 17) TM.75.G.2640+ obv. V 6–8: Sá-mu maskim {I-da-ar; 18) TM.75.G.10081 obv. III 5–7: dumu-nita {Ida-ar Ma-rí ki; 19) TM.75.G.10156 obv. III 14–19: Ba-ba Ì-lum-a-bí Puzur4-ra-dAs-dar maskim Ma-rí ki; 20) TM.75.G.10248 obv. V 2–4: [PN] lú {I-da-ar Ma-rí ki. Documents nos. 1–13 demonstrate that {I-daåar was the Mari king who suˆered defeat at the hands of Ebla, and then lived to witness the ˜nal destruction of the city. We still do not know, however, when Ikun-isar was succeeded by {I-daåar. The former could have reigned for the entire period of minister Ibrium’s mandate. {I-daåar is already present in documents from the time of Nizi. ARET VII 16, from Nizi’s third year, records in sections (18) and (31) the presence at Ebla of one of his agents: (10/5 shekels of silver for) Puzur4-As-dar maskim-e-gi4 {I-da-ar. ARET VII 1+, again dated to the period of Nizi, records one mina of silver and three pieces of clothing for {I-daåar, and ten shekels of silver for his agent Puzur-Asdar; an amount of gold was, furthermore, taken to Mari for {I-daåar by PuzurAsdar (sections [18]–[20]). TM.75.G.1233+ (SEb 4, 145–54: T. 8) is dated to the same time, and registers 20 shekels of silver for {I-daåar. {I-daåar receives a notably more substantial gift according to TM.75.G.2236 (= MEE XII 25) obv. x 12–xi 2: 5 ma-na babbar:kù 1 ma-na kù-gi {I-da-ar “˜ve minas of silver (and) 1 mina of gold (for) {I-

26. According to the Ebla documents, kaskal means “travel, commercial expedition,” níg-kaskal “travel provision,” níg-kas4 “military expedition.” See A. Archi and M. G. Biga, “Wars at Ebla” (in preparation). For the meaning of níg-kas4 see the passage TM.75.G.2277 rev. xiv 1–11, quoted below, section 5 § 8, with the sequence: níg-kas4 GN1 . . . GN1 til “military expedition against GN1 . . . GN1 was destroyed.” The term níg-kas4 occurs in several passages of the chancellery documents published by P. Fronzaroli in ARET XIII. He translates: “travel; military expedition” in the index at p. 289; the single passages, however, request “military expedition,” as this author has seen. See e.g., p. 101, no. 9 (47): [DU.DU] si-in níg-kas4 GN1 GN2 GN£ su-ba4-ti “[when he went] to the military expedition (against) GN1, GN2 conquered GN£;” p. 193, no. 19 (18)–(19) “you did not give your good men to Ebla; you gave your bad men (for) the military expedition with (níg-kas4 ás-da) Ebla.”

Two Lines Long

daåar.” This document, like ARET VII 1+ and 16, mentions “Gulla messenger of Irba,” Gul-la u5 Ìrbaxki. This is the time when Ebla received “the news (that) Mari had vanquished the town of Terqa,” obv. X 4–8: níg-mul Ma-rí ki Ter5-gaki àgakár. TM.75.G.1222 has also to be dated to the time of Nizi, because Enna-Dagan (who was not yet king) is mentioned several times; {I-daåar appears in obv. x 1 and rev. ix 6. TM.03.G.1000 registers 1081 jars of oil (la-ha ì-gis), about 32,430 liters, sent as a tribute to Mari “in the second year (of Nizi) (2 mu).” Also this document mentions EnnaDagan and {I-daåar. If {I-daåar was about twenty years old when Nizi was king, he would have been sixty when Ebla was destroyed. 3) The Relative Chronology of Isar-damu’s Reign (Ministers Ibrium and his Son Ibbi-zikir) The chronological framework that enables us to order the events recorded in the Ebla tablets is provided by the Annual Accounts of Delivery of Metals (AAMs), arranged on a prosopographic basis with the assistance of the Monthly Accounts of Deliveries of Textiles (MATs). The ˜rst four AAMs date back to the minister Arrukum. These documents, like the older ones of Ibrium, contain relatively little data, unlike the AAMs of the later period, when the greater riches pouring into Ebla allowed for a larger number of gold and silver gifts. A few synchronisms with the annual mu-DU documents con˜rm the following chronological sequence of the AAMs:27 a) Minister Ibrium (Ibr.) 1.a: TM.76.G.672+; 1.b: TM.75.G.1705+ 2.a: TM.75.G.10077+; 2.b: TM.75.G.2362+ 3.a: TM.75.G.2464; 3.b: TM.75.G.2333 (?) 4: TM.75.G.1771 (= MEE VII 47) 5: TM.75.G.3052+ 6: TM.75.G.3481+ 27. A ˜rst study of the AAMs, with a provisional chronological order, was provided by A. Archi, “Les comptes rendus annuels des métaux (CAM),” Amurru 1 (1996) 73–99. The list of the AAMs given here is the result of further work on the fragmentary tablets; it should be considered as ˜nal. For the sequence of annual mu-DU documents see Archi, “The ‘lords’,” VO 12 (2000) 19–58.

A VICTORY OVER MARI AND THE FALL OF EBLA

7: TM.75.G.1904+ 8: TM.75.G.10148+ 9: TM.75.G.2359 10: TM.75.G.10143+ 11: TM.75.G.10144 12: TM.75.G.1464+ 13: TM.75.G.2365+ 14: TM.75.G.1730+ (= MEE VII 34) 15: TM.75.G.2502 16: TM.75.G.2465 17: TM.75.G.10210 18: TM.75.G.1923+ b) Minister Ibbi-zikir (= I.Z.)28 1: TM.75.G.1860 (= MEE X 20) 2: TM.75.G.2462 3: TM.75.G.10088+ 4: TM.75.G.10201 5: TM.75.G.1918 (= MEE X 29) 6: TM.75.G.10074+ 7: TM.75.G.2622+ 8: TM.75.G.2428 9: TM.75.G.2508 (= MEE XII 37) 10: TM.75.G.2429 (= MEE XII 36) 11: TM.75.G.2507 12: TM.76.G.534+ (= ARET VIII 534) 13.a: TM.75.10202+ (?); 13.b: TM.75.G.2426 14: TM.75.G.12450 15: TM.75.G.12278(+)15772 / TM.75.G.12373 16: TM.75.G.12373 / TM.75.G.12278(+)15772 Table 2. The kings of Mari and their titles according to the Ebla tablets 1) 2) 3) 4)

Anubu en: EDL obv. ii 1 Saåumu en: EDL obv. iii 1 Istub-(i)sar lugal: EDL obv. v 4 Iblul-Il en: EDL obv. vi 6, vii 3. Iblul-Il lugal: EDL obv. viii 10, ix 9, rev. I 10, ii 13, iv 13; TM.75.G.1953 obv. ii 3, v 2; ARET II 4 xv 7. 5) Nizi lugal: TM.75.G.1953 rev. i 4; TM.75.G.1987 obv. i 3; TM.75.G.1866 obv. i 3; ARET VII 1+ obv. i 7, rev. x 6; 16 rev. x 7.

28. TM.75.G.2271, which concerns gold deliveries (mu-DU) under the charge of Ibbi-Zikir, ascribes to him seventeen years as minister, see Archi, VO 12 (2000) 20–21. The administrative years of the mu-DU documents and those of the AAMs could have begun, however, in diˆerent months.

9

6) Enna-Dagan en: EDL obv. i 3, rev. xiii 12. Enna-Dagan lugal: ARET VII 17 rev. vi 2; MEE II 13 obv. ii 7. 7) Iku(n)-isar en: ARET I 11 obv. vi 5; TM.75.G.1705. rev. vi 6. 8) Hidaåar lugal: TM.75.G.2241 obv. i 2; TM.75.G.2270 obv. x 17; TM.75.G.2353 obv. xi 15. Commentary The order of Mari kings according to the EnnaDagan Letter (TM.75.G.2367 = EDL), was: Anubu, Saåumu, Istub-(i)sar, Iblul-Il and Enna-Dagan, see D. O. Edzard, SEb 4 (1981) 89–97. Enna-Dagan is de˜ned as “en,” according to the terminology in use at Ebla, where this Sumerian term corresponds to mal(i)kum “king.” An unpublished seal of Ikun-†amas shows that this title was once common at Mari (see below, § 9). In EDL Istub-(i)sar (who is mentioned only once) bears the title of lugal = sarrum, whilst Iblul-Il is twice de˜ned as en and ˜ve times as lugal, see A. Archi, MARI 5 (1987) 37–52. Not only the royal inscriptions from Mari, but now also the unpublished seals of {Idaåar and Isgi-Mari, prove that lugal was the o¯cial title of the kings of Mari towards the end of the Early Dynastic period. The administrative documents of Ebla use the title lugal for the kings of Mari Iblul-Il, Nizi, Enna-Dagan and {I-daåar. Only Iku(n)-isar has the title en. 4) The Alliance between Ebla and Mari The brief reign of the two kings of Mari, Nizi and Enna-Dagan, marked the slow decline of Mari in the third millennium, ˜rst defeated by Ebla in a direct confrontation and then conquered by Akkad. In approximately the year of EnnaDagan’s death, Isar-damu, still a minor, ascended the throne of Ebla. The succession was imposed by his mother Dusigu, Irkab-damu’s favorite. The “queen,” maliktum, of Irkab-damu had died a few months after her marriage, possibly in childbirth.29

29. The ritual for the marriage of Irkab-Damu, preserved in ARET XI 1, was celebrated when Arrukum was already minister (he is mentioned in rev. vii 4, [13], viii 1).

10

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

For more than twenty years Dusigu was the most important woman at court. Isar-damu married his own maliktum only fourteen years after ascending to the throne. For seven more years Dusigu preceded the new queen in the administrative lists, until her death in the third year of the minister Ibbi-zikir.30 It was therefore Ibrium, who had been appointed minister by Irkab-damu only a few months before his death, who carried on the task of strengthening Ebla, a trend that had begun some time earlier. The treaty in which Ebla dictated conditions to Abarsal (= Tell Chuera ?) should be attributed to the time of Igris-{alab, or to the very ˜rst years of Irkab-damu. It was Irkabdamu who brought Emar into Ebla’s sphere of in˘uence;31 Tuttul on the Balih remained under the control of Mari. The border between Ebla’s and Mari’s spheres of in˘uence ran between Emar and Tuttul.32 For roughly thirty years the two states exchanged messengers and ceremonial gifts on an equal footing. Already during the time of Enna-Dagan, the most important delegations from Mari were lead by a “steward,” sagi.33 Among the messengers who traveled most frequently to Ebla during the mandates of Ibrium and Ibbi-zikir (sometimes more than once a year) we ˜nd Tesna (Tés-na), who was active during the reign of Enna-Dagan, as well as Warutum. Contacts between Ebla and Mari were already becoming frequent in the initial years of the

reigns of Isar-damu and Ikun-isar. TM.75.G.10077 (Ibr. 2.a) rev. X 12–17 records the news—undoubtedly received with satisfaction at Ebla—of a defeat suˆered by Mari (by whom, it does not say): níg-mul àga-kár Ma-rí ki wa TIL Wa-ru12-tum. TIL cannot refer to Warutum, the Mari messenger, unless by mistake, because he is mentioned also in later documents.34 In the same year, Tesna received a small gift for having taken part (?) in an expedition against Ga/uraman, TM.75.G.2362 rev. V 2–7: níg-ba Tés-na Ma-rí ki in níg-kas 4 Gú-rá-ma-anki. Two new kings had recently come to the throne, in Ebla as well as in Mari. The leadership of Mari was weakened, and neither city was capable of conducting an oˆensive policy. This naturally indicated the need to seal the peace, to which the Mari representatives swore in the temple of Kura, the main god of Ebla. A similar ceremony would have also been performed at Mari. The following passage, which registers the gifts for the delegation that came from Mari, indicates that the tablet bearing the treaty was deposited in the temple of the city. TM.75.G.2464 (Ibr. 3a) obv. II 16–III 22: 5 ma-na babbar:kù níg-ba A-ha-ar-si 3 ma-na babbar:kù níg-ba åÀ-sa 2 ma-na babbar:kù níg-ba Íl-a 1 ma-na babbar:kù níg-ba gurus-gurus maskim-sù Ma-rí ki in ud nídba ì-gis Ma-rí ki 10 gín D. babbar:kù níg-ba Du-tum Ma-rí ki 50 gín D. babbar:kù nu11-za 1 dub lú nam-kud Ib-laki wa Ma-rí ki é dKu-ra “5 minas of silver: gift (for) Ahaarse; 3 minas of silver: gift (for) åAsa; 2 minas of

30. See the lists of “the women of the king,” dam en, in Eblaite Personal Names and Semitic Name-Giving (Roma: Missione Archeologica Italiana in Siria, 1988), ed. A. Archi (= ARES I), 245–59; M. V. Tonietti, “Le liste delle dam en: cronologia interna,” in Miscellanea Eblaitica 2 (Firenze: Dipartimento di Linguistica, 1989) ed. P. Fronzaroli, 79–115. The AAM registering the marriage of Isar-damu is TM.75.G.1730+ (MEE VII 34) (Ibr. 14); the ritual for this marriage is ARET XI 2. The AAM mentioning the death of Dusigu is TM.75.G.10088+ (I.Z. 3). 31. On Abarsal and Emar, see ARES 2, pp. 87–91, 286–90. 32. On Tuttul in the Ebla texts, see A. Archi, “Tuttul-surBalih à l’âge d’Ebla,” in: De la Babylonie à la Syrie, en passant par Mari. Mél. J.-R. Kupper (Liège: Université de Liege, 1990), 197–207. 33. See A. Archi, “The Steward and his Jar,” Iraq 61 (1999) 147–58.

34. The EDL has: PN àga-kár GN “PN vanquished GN,” see Edzard, SEb 4 (1981) 91, which is the expression used in Old Akkadian inscriptions, see B. Kienast, W. Sommerfeld, Glossar zu den altakkadischen Königsinschriften (FAOS 8. Stuttgart: Franz Steiner Verlag, 1994), 162: TÙN.KÁRA-sè; AHw III, 1118: saåaru(m); CAD, † I, 2. The sequence àga-kár GN should be translated: “victory over / plunder of GN,” as the following two passages show, TM.75.G.10210 obv. vii 1–3: . . .] níg-mul Du-liki àga-kár, xiii 15–21: 1 ma-na babbar: kù níg-ba Tés-na Ma-rí ki níg-mul Du-liki su-ba4-ti “(x silver: gift for PN)] who brought the news that Duli was vanquished; . . . . 1 mina of silver: gift (for) Tesna of Mari, who brought the news (that) Duli was conquered.” TM.75.G.1630 (= MEE VII 12) presents the sequence: udu nídba – udu ug7 – udu àga-kár “sacri˜ced, killed, plundered (?) sheep.” See ARET X 100 obv. xiv 22–26: níg-kas4 Ib-al6ki àga-kár udu-udu “expedition (against) Ibal (for) plundering the sheep;” TM.76.G.534+ obv. 1u–4u: . . .] àga-kár udu-udu Wa-za-ru12ki lú Ib-al6ki.

A VICTORY OVER MARI AND THE FALL OF EBLA

silver: gift (for) Ilåa; 1 mina of silver: gift (for) the men (who were) their representatives, (all) of Mari, on the occasion of the oil oˆering (of) Mari. 10 shekels of silver (for) Dutum of Mari. 50 shekels of silver: a sheet (for covering) one tablet: that of the oath of Ebla and Mari (for) the temple of Kura.” In the following year, the peace treaty notwithstanding, Ebla apparently supported a rebellion of Tuttul against Mari, TM.75.G.1771 (= MEE VII 47; Ibr. 4) obv. v 1–9: . . .] sub si-in kinx-ak-sù lú ìna-sum SA.ZAxki in Du-du-luki in níg-kas4 Ma-rí ki “(weapons ?)] . . . which the Palace has given to Tuttul for the expedition (against) Mari.” The following passage may also refer to this event; this comes from a MAT dated to the period of Ibrium, TM.75.G.1889 rev. I 1–6: En-na-ni-il níg-mul Ma-rí ki til ás-ti Du-du-luki “(a gift for) Ennani-Il (who) brought the news that Mari was defeated by Tuttul.” In the same year, according to the AAM TM.75.G.1771 rev. xii 21–xiii 6, Tesna announces the defeat of Kis: 1 ma-na babbar:kù [níg-ba] Tés-na Ma-rí ki níg-mul Kiski til “1 mina of silver: [gift] (for) Tesna of Mari (who) brought the news (that) Kis was defeated.” No mention is made of precisely who brought about the defeat of the capital of the Babylonian region or whether Mari participated in the event. A few years later Mari turned towards its northern borders, defeating Nagar (Tell Brak), the city that controlled the northern plain of the {abur.35 The news was soon brought to Ebla, TM.75.G.1904 (Ibr. 7) obv. xvi 4–10: 1 ma-na babbar:kù níg-ba Ib-gí-tum Ma-rí ki níg-mul Na-gàr ki [t]il “1 mina of silver: gift (for) Ibgitum of Mari, (who) brought the news (that) Nagar was defeated.” The campaign was taken up again the following year, resulting in another victory, TM.75.G.10148+ (Ibr. 8) obv. vii 5–9: 20 gín D. babbar:kù níg-ba níg-mul Na-gàr ki til “20 shekel of silver: gift (for whom) brought the news (that) Nagar was defeated.” Mari had to confront Nagar so as to prevent this city’s in˘uence from extending over the entire area of the southern {abur as far as the Euphrates, and it was therefore in it’s interests to main35. On Nagar, see A. Archi, “The Regional State of Nagar According to the Texts of Ebla,” Subartu 4,2 (1988) 1–15.

11

tain friendly relations with Ebla. The eastern sector must have also given rise to worries before Kis came under attack, most probably from an enemy to its south. Mari, therefore, had to bow to the inevitable, that is, that Ebla have access to the Euphrates, by means of the city of Mane.36 Mari, however, maintained its control over Tuttul and the southern valley of the Balih, while Ebla concentrated on consolidating its power in northern Syria. Six years after the defeat reported in the AAM of Ibr. 4, Kis suˆered another attack, probably carried out by Ensakusana of Uruk (see below, § 9). This time we can assume that Mari took part in the war. TM.75.G.10143+ (Ibr. 10) rev. vi 14– vii 1: Na-zu-mu in ud níg-kas4 si-in Kiski 1 ma-na babbar:kù [. . .] [. . .] [. . .] Kiski “(a dagger decorated with gold for) Nazumu on the occasion of the expedition against Kis; 1 mina of silver (for) [NP . . . to] Kis.” Another two campaigns were necessary before a de˜nitive victory over Kis could be claimed. TM.75.G.10144 (Ibr. 11) rev. v 10–16: 1 ma-na bab[bar:]k[ù] [níg-ba] Wa-ru12-tum Ma-rí ki níg-mul Kiski til-til “1 mina of silver: [gift] (for) Warutum of Mari (who) brought the news (that) Kis was completely (?) defeated.” The following year (the third) we have the decisive campaign. Various pieces of news arrived in Ebla regarding this con˘ict. As usual, it is Tesna who announced an initial defeat of Kis, TM.75.G.1464 (Ibr. 12) obv. ii 29–iii 5: 30 gín D. babbar:kù [níg-ba] Tésna Ma-rí ki níg-mul Kiski til. As the campaign continued, Kis’s adversary succeeded in conquering some towns. Mari now intensi˜ed its commercial relations with Ebla; two men from Mari, acting as agents for a commercial group from Tuttul, arrived in Ebla, obv. vii 12–viii 4: 30 babbar:kù níg-ba Wa-ru12-tum Ma-rí ki níg-mul in uruki-uruki Kiski su-ba4-ti 6 gín D. babbar:kù níg-ba Is-dub-Il wa Dab6-ti Ma-rí ki maskim es8:ga Du-du-luki “30 (shekels) of silver: gift (for) Warutum of Mari (who) brought the news concerning the towns of Kis (which) were conquered. 6 shekels of silver: gift (for) Istub-Il and Dabti of Mari, agents of the merchant(s) of Tuttul.” 36. On Mane, see above note 5.

12

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

It would appear that Warutum made a second journey to Mari “traveling from Kis” together with Nazumu, who had participated in the earlier expedition against Kis in the year Ibr. 10, TM.75.G.10143+ obv. xvii 4–10: 20 gín D. babbar:kù níg-ba Wa-ru12-tum Ma-rí ki e11 ás-ti Na-zu-mu mi-nu Kiski “20 shekels of silver: gift (for) Warutum of Mari (who) came up from Kis with Nazumu.” Finally, Tesna returned to Ebla to announce an even greater victory over Kis, rev. x 19–25: 1 ma-na babbar:kù níg-ba Tés-na Ma-rí ki níg-mul Kiski til-til “1 mina of silver: gift (for) Tesna of Mari (who) has brought the news (that) Kis was completely (?) defeated.” This does not mean that Mari itself did not go through some di¯cult moments: defeats were obviously not communicated to allies such as Ebla, which were, in reality, rivals. However, six years later, the two cities renewed their alliance. Ebla found herself in di¯culties: the minister Ibrium, who had led the state for eighteen years, was ailing and died a few months after the agreement with Mari. The following fragmentary passage does not say who led the delegation from Mari, which also included a member with the same name as the former king, Enna-Dagan, TM.75.G.1923 (Ibr. 18) obv. xii 1–17: . . .] 30 babbar:kù níg-ba Ar-rúm 20 gín D. babbar:kù níg-ba Mi-su-wa-du 20 gín D. babbar:kù En-nada-gan 1 ma-na tar babbar:kù níg-ba ábba-ábba Ma-rí ki in ud nídba ì-gis Ib-laki wa Ma-rí ki “. . .] 30 (shekels) of silver (for) Arrum; 20 shekels of silver (for) Misuwadu; 20 shekels of silver (for) Enna-Dagan; 1.30 mina of silver: gift (for) the elders of Mari on the occasion of the oil oˆering of Ebla and Mari.” This is immediately followed by the mention of clothing acquired “for a military expedition of (/ against) Mari,” xiii 1–7: . . .] 14 aktum-TÚG 48 åà-da-um-TÚG-I lú KA-ba-an níg-sa10 in níg-kas4 Ma-rí ki. This sequence is con˜rmed by the mu-DU text of the same year, marked by the death of Ibrium. First of all there is the record of the gift of garments by the king of Mari, MEE II 1 (TM.75.G.1261) rev. ii 2–9: nígki-za lugal Ma-rí ki in ud nídba ì-gis Ib-laki wa Ma-rí ki “(1+1+1+2 garments:) competence of the king of Mari on the occasion of the oil oˆering of

Ebla and Mari.” This is a less valuable gift than the one that Ebla gave to the delegates from Mari and this exchange of ceremonial gifts seems imbalanced in favor of Mari. It is probable, however, that a delegation from Ebla travelled to Mari and received gifts in turn. There then follows an amount of gold sent to Ebla from Kablul, a city that had been under Ebla’s hegemony for at least twenty years,37 rev. ii 13–18: 45 (shekels) of gold: delivery of Kablul for the expeditions of (/ against) Mari,” in níg-kas4 Ma-rí ki. This passage, as TM.75.G.1923 obv. xiii 1–7 (mentioned above) is ambiguous insofar as the preposition si-in “against” is lacking. Ebla provides clothes for (in) an expedition against or in support of Mari. Kablul sends gold to Ebla for the same purpose. There can be no doubt, however, that military action was taken against Mari during that year and the mu-DU text reveals that this expedition had a positive outcome, rev. vi 13–vii 5: Bùda-ma-lik in †a-ra-bí-igki su-ba4-ti níg-mul Marí ki til “(gifts for) Buda-malik (which) he has received in †arabig (because) he brought the news (that) Mari was defeated.” We cannot rule out the possibility that these events are not recorded in chronological order and that the peace with Mari mentioned in the ˜rst part of the document resulted from the defeat suˆered by the city. When Ibbi-zikir became minister, following in his father’s footsteps, diplomatic relations between Ebla and Mari intensi˜ed and the exchange of gifts increased. The steward †ugadu and his delegation visited Ebla, often more than once a year, and Tesna and Warutum continued to act as messengers, along with others.38 In the ˜rst ten years Ibbi-zikir consolidated Ebla’s power over northern Syria, leading numerous military campaigns. Of particular note is the one conducted against Ibal (AAM I.Z. 8), an important center in the region of Homs. For its part, Mari followed a similar strategy in the area of the Middle Euphrates. Its greatest victory was that over Haddu, TM.75.G.2429 (I.Z. 37. See the Abarsal treaty, ARET XIII 5 (2): Kab-lu5-ul ki ù bàd-bàd-sù in su en Ib-laki “Kablul and its fortresses belong to the king of Ebla.” 38. Archi, Iraq 61 (1999) 148–52.

A VICTORY OVER MARI AND THE FALL OF EBLA

10) obv. xxiii 16–xxiv 3: 30 babbar:kù wa 40 gín D. babbar:kù 1 gú-li-lum 32 gín D. babbar:kù subal-ak 8 gín D. kù-gi nu11-za-sù Wa-ru12-tum nígmul Ma-rí ki åÀ-duki til “30 (shekels) of silver and 40 shekels of silver for 1 bracelet; 32 shekels of silver, value of 8 shekels of gold: a sheet (for) its (covering), (for) Warutum (who) brought the news (that) Mari had defeated Haddu.” The passage in the MAT TM.75.G.2274 must refer to the same event (the total of silver reported is the same), but the syntactical order has strangely been inverted, obv. IV 13–19: 1+1+1 garments 1 níg-lá-sag 1 gú-li-lum-II babbar:kù-gi 40 30 bar6:kù níg-ba Wa-ru12-tum Ma-rí ki níg-mul åÀ-duki Ma-rí ki til “1+1+1 garments, 1 headband, 1 bracelet of 40 shekels (and) 30 shekels of silver (for) Warutum (who) brought the news (that) Haddu had defeated Mari / Haddu was defeated (by) Mari.” 5) The Defeat of Mari An uninterrupted series of victories made Ebla con˜dent of its own power. Twenty-˜ve years of sworn peace, maintained through the regular exchange of envoys and ceremonial gifts, had not eliminated its rivalry with Mari. Two routes linked northern Syria with the eastern regions and the great centers of Babylonia. The northern route was controlled by Ebla as far as the region of Nagar, while the southern one, which ran along the Euphrates down to Babylonia, required one to pass through Mari. As for Mari, the deeds of Iblul-Il demonstrate that the city considered it vital to extend its hegemony over the entire region of the Middle Euphrates. If Mari did not seek to subjugate Ebla, it was merely because it was incapable of winning a de˜nitive victory over Nagar, which had to be kept away from the valley of the Euphrates. Mari had also been under pressure for years from the east by the threat represented by Kis. No other political event is represented in the Ebla archives so richly or frequently as the campaign against Mari that took place during the last years of king Isar-damu. This event was preceded by careful diplomatic preparations while the army was being equipped. Then, ˜nally, the departure was announced by minister Ibbi-zikir, who led

13

the troops, and we learn of his journey to Mari, and then of the celebrations upon his victorious return. Then, to consolidate this success, Ebla hurried to form alliances with the ruling houses of Nagar and Kis for whom it had done the favor of weakening a fearful rival. The AAMs of the last ˜ve years of Ebla are fragmentary and contain many large gaps. Apparently this resulted from an attempt to remove some of the documents from the central archives during the ˜nal sack of the city. Twenty-two tablets or larger fragments were found on a plank of burnt wood in the Audience Hall, ˜fteen meters north of the entrance to the central archive. Seventeen of these tablets are MATs, three fragments belong to two AAMs; there are also an annual mu-DU document and a lenticular tablet. Some of these texts can be joined to fragments from the central archive and all of them come from the ˜nal years of Ibbi-zikir. Since they are of diˆerent types and must, therefore, have been kept on diˆerent shelves, whoever oversaw their removal must have known which texts to choose and from which sector of the archive. However, their fragmentary state indicates that the removal was incomplete and carried out in great haste.39 Fortunately, a record of allocations of cereals and ˘our covering seven years, ARET X 100 (TM.75.G.427), enables us to place even the most fragmentary AAMs in chronological order.40 I mu ARET X 100 obv. i 11–12: níg-kas4 Zabùki; iii 2: (níg-kas4) {u-sá-umki; iv 25426: níg-kas4 Si-da-luki; v 28–30: DI† mu níg-kas4 Si-da-luki • TM.75.G.2508 (AAM I.Z. 9): níg-kas4 Za-bùki and {u-sá-umki; TM.75.G.2429 (AAM I.Z. 10): níg-kas4 Si-da-luki. II mu ARET X 100 obv. vii 11–12: níg-kas4 Asdar ki • TM.75.G.2429 (AAM I.Z. 10): níg-kas4 dAsdar-lumki. III mu ARET X 100 obv. xii 29–30: níg-kas4 {ar-ba-tumki; xiv 12–13, 22–23: níg-kas4 Kab-lu5ulki, níg-kas4 Ib-al6ki • TM.75.G.2507 (AAM I.Z. 11): níg-kas4 {ar-ba-tumki. 39. See A. Archi, AfO 44–45 (1997/1998) 109–10. 40. TM.75.G.427 has been already published by G. Pettinato, AfO 25 (1974/77) 1–35.

14

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

IV mu ARET X 100 rev. iii 4–5: níg-kas4 Ibal6ki • TM.76.G.534+ (rather fragmentary text) (AAM I.Z. 12): obv. 1u–4u: àga-kár udu-udu Wa-zaru12ki lú Ib-al6ki. V mu ARET X 100 rev. vi 6, 15–16: (níg-kas4) Ib-al6ki; níg-kas4 Ma-rí ki • TM.75.G.10202 (fragmentary) (AAM I.Z. 13): several people traveling between Ebla and Mari; war against Ibal. VI mu ARET X 100 rev. vii 24–26: è níg-kas4 Ma-rí ki “departure (to) the Mari expedition.” viii 33–34: níg-kas4 Ma-rí ki • TM.75.G.12450 (a small fragment) (AAM I.Z. 14): obv. iii 10–13: ì-ti nígkas4 Ma-rí ki lú nídba ì-gis “(who) took part at the expedition (against) Mari which (ended with) the oil oˆering.” VII mu ARET X 100 rev. xii 3–6: kú SA.ZAxki wa Ma-rí ki “(cereals) for the meal of (men) of the Palace and Mari.” • (?) TM.75.G.12278(+)15772 or TM.75.G.12373 (three small fragments) (AAM I.Z. 15). The expedition against Mari (or its preparation) began in the thirteenth year of the minister Ibbi-zikir, immediately after the campaign against Ibal (to be located east of Qatna), and ended with a peace treaty the following year. The two events furnish the date for two tablets. The ˜rst comes from archive L. 2712, which contained documents relating to food supplies for the Palace over the last three or four years of the city, ARET X 47 rev. i 2–3: DI† mu níg-kas4 Ma-rí ki “year of the expedition (against) Mari.” The second belongs to a small number of texts (from room L.2586) that date to the ˜nal years of Ebla,41 TM.74.G.102 rev. i 2–5: DI† mu nídba ì-gis Ma-rí ki.42 We have no record of these events by Ibbi-zikir himself, who led the campaign, nor are there any letters from any other participating o¯cials. Apart from ARET X 100 and the two fragmentary AAMs mentioned above (I.Z. 13 and 14), the other documents relating to the expedition against Mari are some MATs and one tablet TM.75.G.2426 that lists consignments of metal objects (all unpublished). 41. M. G. Biga, in Eblaite Personal Names and Semitic Name-Giving, ed. A. Archi (ARES 1; Rome: Missione Archeologica Italiana in Siria, 1988), 285–87. 42. The text is published below, at the beginning of section 6.

As usual, the events that marked the history of Ebla have only left their traces in the archives insofar as they aˆected the accounts of goods distributed: garments and objects in precious metals The MAT TM.75.G.2277 (month name broken) records a quantity of cloth given on the occasion of the expedition against Mari. An initial summation relates to garments ordered by type; there then follows the overall total (the last three columns of the rev. concern other deliveries): rev. xiv.

1. su-nígin 4 mi-at 46 túg-túg 21 gu-mug-TÚG 3. è níg-kas4 5. Ma-rí ki 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+I-TÚG sa6 gùn 7. Du-bù-hu-ma-lik lú Da-du-ud 9. níg-mul Ma-rí ki 11. til

“Final total: 446 garments, 21 blankets: issue (for) the expedition (against) Mari. 1+1+1 garments (for) Tubuhu-malik, the man (i.e. the son ?) of Dadud, (who brought) the news (that) Mari was defeated.”43 The ˜rst deliveries (generally relating to the usual set of garments: a mantle, a tunic, and a kilt, occasionally a silver or gold bracelet, gú-lilum, or another piece of jewelry) are for people from well-known cities, who were involved in the diplomatic negotiations that preceded the military con˘ict. A man (Gú-na) from Haddu, an important center on the border with the territory of Mari,44 traveled to Mari. Perhaps in connection 43. Tubuhu-malik announced to Ebla also the victory on Sidaåù by Ibbi-zikir, TM.75.G.2418 (MAT) obv. i 15–ii 1 = TM.75.G.2429 obv. iii 22–iv 3 (MEE XII 35; AAM I.Z. 10): 1.3 gín kù-gi 2 gestu-lá Du-bù-hu-ma-lik lú Da-du-ud níg-mul Sida-ùki su-ba4-ti. 44. Haddu is perhaps to be identi˜ed with Tall Malhat edDer¿, situated about 70 km north of Der ez-Zor, a Kranzhügel described by H. Kühne, in Beiträge zur Altorientalistik. Festschrift K. Bittel (Mainz am Rhein: Verlag Philipp von Zabern, 1983), eds. R. M. Boehmer and H. Hauptmann, 299–308.

A VICTORY OVER MARI AND THE FALL OF EBLA

with this trip, three others from Haddu also received gifts (obv. i 1–15). Two “representatives,” maskim, of minister Ibbi-zikir went to Kis (i 17– 22). Kis, in turn, repeatedly sent its own messengers (viii 20–21, ix 7–10, and xi 19–23: 1+3+2 men; one of these, Sumalum of Kis, also received gifts after the victory over Mari, TM.75.G.2426). Various people from cities belonging to Ebla or to her allies, among them Manuwat, NIrar, Dub, Arhadu, and Kablul, also visited Ebla. A number of puri˜cation ceremonies were recorded (ì-gis-sag) for the death of people involved in these activities. Eleven are “the charioteers,” ugula surx-BAR.AN, who accompanied Ibbi-zikir during the expedition, including the faithful Enna-BE (xi 11–xii 16), known also from other texts. In a group of twenty individuals, we also ˜nd Tubuhu-malik of Dadud (xiv 4–24) who, as noted above, brought news of the ˜nal victory (rev. xiv 6–8). Twelve sets of garments and 130 kilts were delivered by a representative of Ibbi-zikir to {a(l)abitu, probably a place on the Euphrates in the area of troop operations,45 obv. xv 6–13: 10 åà-daum-TÚG-II 2 gu-dùl-TÚG 12 aktum-TÚG 2 íb+ III-TÚG sa6 gùn 1 mi-at 30 íb+I-TÚG sa6 gùn siin {a-a-bí-duki Ne-si maskim I-bí-zi-kir su-[mu“tag4”].46 The fragment ARET XII 694 refers to Ibbizikir leaving Tuttul to march towards {alabitu , ii 1–5: (garments) I-bí-zi-kir lú è si-in {a-a-bí-duki. TM.75.G.2277 then has a lengthy list of garments and cloth sent to both the sovereigns and the elders of Ebla’s long time Syrian allies who regularly received similar gifts at least once a year:47 the kings of Raåak, Burman, [. . .], [. . .], Garmu, [. . .], Ibubu, Utig, Iritum, Kakmium, NIrar; the badalum and the elders of {arran, Sanapzugum, Ursaåum (obv. xvi 1–rev. iv 9). In this way they were reminded of their promise of ˜delity to the dominant city. Some of these may

45. Astour, Eblaitica 4, identi˜es {a(l)abitu with Halÿbÿyeh, on the Western bank of the river, about 85 km downstream from Tuttul. 46. Nesi, a representative of minister Ibbi-zikir, is also mentioned in TM.75.G.2426 obv. vii 4 // TM.75.G.2335 obv. vii 17. 47. See ARET 1, 219–25.

15

also have provided military contingents and, after the victory, they received additional gifts. We then have two groups of “messengers,” kas4kas4, one group of seven and another of ten men, respectively under the leadership of Kitir and Isgibarzu, who participated in the entire military campaign.48 Alongside these there is {ubanu, a “charioteer of the king,” ugula surx-BAR.AN en, who provided them with chariots (iv 10–v 7). By reinforcing pacts with longstanding Syrian allies Ebla at this point reaped the fruits of its past diplomatic contacts and relations with the major centers surrounding Mari, that is to say Haddu (see the beginning of this document), Nagar, and Kis. The references to Haddu exemplify the kind of pressure brought to bear on a city that was important but second-rank, situated between Ebla and Mari and caught up in the lethal battle between the two cities. 49 The memorandum, TM.75.G.2561, that an informer sent to Ebla, reveals the trap laid by Mari in order to force Haddu to side with her.50 To the messenger from Mari, †uwama-wabar, who tried to convince the envoys of the king of Haddu to accept an alliance with Mari, the king of Haddu at ˜rst answers that he is bound to respect the sworn promise that ties him to Ebla. “(As concerns) myself and Ebla, (the ceremony of the) oil (for) the pact and the swearing of the pact is before the god Kura and the god Hadda!” (obv. vi 11–rev. i 11). The envoy from Mari then points out that Haddu is already a rather ambiguous ally of Ebla: “Good soldiers you have not placed at the disposition of Ebla. You have given poor soldiers (for) the expedition alongside Ebla!” (rev. ii 3–14). This passage demonstrates that allies also had to provide troops in the event of war. Mari believes that Haddu has 48. Kitir also appears also in other documents pertaining to the campaign against Mari: TM.75.G.2426 rev. viii 14; TM.75.G.2337 obv. vi 10; TM.75.G.2250 rev. viii 5–12: “6+6+6 garments (for men) of Agagalis: messengers of Isgi-barzu (kas4-kas4 Is11-gi-bar-zú); 4+4+4 (?) garments (for men) of Luåatum: messengers of Kitir.” 49. See above note 44. 50. See L. Milano in Studi e Testi II, eds. S. de Martino and F. Imparati (Eothen 10; Firenze: LoGisma editore, 1999), 133– 48 (with previous bibliography). The text has been republished as ARET XIII 19.

16

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

good cause to distrust Ebla: “Why do you consider Ebla an ally? Ebla, like a woman, is not reliable.” In the end, Haddu is won over: “The king of Haddu called his men to an assembly and said: the alliance with Ebla is not good! For us, instead, an alliance with Mari is better.” It is possible, although by no means certain, that this refers to the period immediately preceding Ebla’s victorious expedition. If that was indeed the case, Haddu had changed its mind again at the last moment and aligned itself with Ebla. The fragment ARET XII 568 may record a battle near Tuttul during the march on Mari: x+1. ì-ti níg-kas4 3u. Ma-rí ki ás-ti 5u. Du-du-luki [. . . x+ii’. x+1. àga-kár x+i.

“[x garments for PN] who took part to the expedition (against) Mari, by Tuttul [. . . (when) Mari] was vanquished.”51 TM.75.G.2277 (Appendix, Text 1), although terse like all the other MATs, gives us a clear picture of the diplomatic activity that continued throughout the war. During the march on Tuttul Ibbi-zikir met with messengers from Kis, including NIrisum, brother of the king of Kis, accompanied by one of his representatives. Gis{I brings “meat,” (uzu) from Kis (v 8–vi 4). Ennani-il arrived from Haddu (v 15–21). Then Isdubu of Kis arrived in Tuttul (vi 5–10). Later, along the road that ran along the Euphrates, they are joined also by the king of GudadaLUM and Saåaåa of Haddu (vi 25– vii 5), while Warutum of Kis brought more meat (vii 25–viii 1). A hundred and ninety kilometers from Tuttul, the army reached Terqa (Ti-rí-gaki, Ter5(BAN)gaki; el-Ashara), only ˜fty-˜ve kilometers upstream

from Mari. Here Ibbi-zikir was joined by two Eblaite princes: Igsub-damu and Ze-damu (viii 2–6). A pair of caravans from Kis, led respectively by Warutum and Islum-ahu, brought further supplies of meat (vii 25–viii 1, 13–20). Another notable from Nagar, Saåum, also joined the Eblaite army (viii 21–26). The battle between Ebla and Mari was fought near Terqa. Given the nature of these kinds of accounts, there is only this one brief mention of the fact that Ebla emerged victorious from the con˘ict and then only because the messenger Tubuhu-malik brought to Ebla “the news (that) Mari was defeated,” received a set of garments: obv. xiv 6–11: 1+1+1 garments Du-bù-hu-ma-lik lú Da-du-ud níg-mul Ma-rí ki til. Other administrative documents that record the triumph of Ibbizikir and the diplomacy brought into play by Ebla (in particular, the inter-dynastic marriages between two Ebla princesses and the sons of the kings of Nagar and Kis), leave us in no doubt about the fact that Ebla, on this occasion, won a decisive victory. Apparently minister Ibbi-zikir and his counselors were not con˜dent about advancing further and laying siege to Mari. The document registers gifts of clothing for Ebla’s main allies, undoubtedly entrusted to the messengers who bore news of the victory: the king of Nagar; the king of Kis, together with his father and his brother; the king of Haddu (viii 29–ix 13). An initial peace treaty would be discussed with NE.NE and Al6-ma of Mari (ix 14–18; NE.NE was a brother of the king of Mari, i.e. of {I-daåar, according to TM.75.G.2335 rev. vi 14–21; see below, at the end of § 7). It appears that a brother of the king of Kis, Bususum, also arrived to meet in person with the Eblaites (x 23–29). During the march homewards, a gift was also left for the king of Kakmium in DaNEnatu (xi 13–18), and one for the king of Emar in Gabaganetu (xii 19–xiii 5). 6) After the Victory

51. Another fragmentary account of garments delivered on the occasion of the expedition against Mari is ARET XII 984 obv. I 1–5:] (garments) ses-sù in ud níg-kas4 Ma-rí ki.

Certain prosopographic factors, together with repeated references to wartime deliveries of cloth to Tuttul and {a(l)abitu, permit us to ascribe a

A VICTORY OVER MARI AND THE FALL OF EBLA

further text to the Mari campaign dossier. The ˜rst of these is TM.75.G.2401, which covers a number of months. Although it does not contain any explicit reference to the campaign, the ˜nal section reads: rev. viii 14–23: su-nígin 2 mi-at 93 túgtúg 56 gu-mug-TÚG gaba-ru níg-kas4 in {a-a-bíduki ás-du-ma iti i-rí-sá a-ti-ma iti ig-za “Total: 293 garments, 56 blankets: receipt (for) the expedition, in {a(l)abitu. From month VI to month II.” More than a hundred of these garments were given as “sets of clothing (for people of) the Palace in {a(l)abitu,” mu4mu-mu4mu SA.ZAxki in {a-abí-duki (obv. i 1–8); see also rev. vii 3–10: (2 garments) 1 su-kesda hi-mu-DU SA.ZAxki wa Ib-laki in {a-a-bí-duki. {a(l)abitu is the place where, according to TM.75.G.2277 obv. xv 6–13 (mentioned above), numerous garments had been sent in expectation of the Eblaite army passing through on its way towards Mari.52 The second of these accounts, TM.75.G.2401, must date to roughly the same period. It records some contacts with Mari but, above all, diplomatic communications with Kis. A large delegation from Kis is headed by ZiNE{AR (see also ARET VIII 540 [19] 21–22: Zi-NE-{AR Kiski[!]) obv. vi 18–vii 3: (garments, jewels) Zi-NE-{AR Is-dub-ì Kiski (8+8+8 garments) maskim-maskim-sù. Another person from Kis, whose name is not given, and two of his representatives receive clothing, together with two people from Nagar returning from Kis, obv. ix 6–14: (1+1 garments) Kiski (2+2+2 garments) maskim-sù, (2+2 garments) 2 Na-gàr ki DU.DU siin Kiski; certain people travel to Kis, according to obv. x 12–19: (5+5+5+1 garments) maskimmaskim En-na-ni-il lú Sa-ù-um DU.DU Kiski. The MAT TM.75.G.2250 (Appendix Text 2) is dated to the second month (igza), that is the same month that closes the multi-month text TM.75.G.2401. It contains numerous references to Ibbi-zikir’s campaign against Mari and to the subsequent peace. We are, therefore, already in the “year of the oˆering of oil of Ebla and Mari.” This document likewise provides witness to the intense activity of messengers traveling between

52. On the location of {a(l)abitu, see above note 45.

17

Ebla, Nagar, Haddu and Kis, as well as merchants from various countries (including Mari) who deliver cloth to the areas involved in the war. The textiles registered in this document as “a receipt (for) the expedition (against) Mari” amount to a “total of 245 garments (and) 9 blankets,” rev. x 10–14. TM.75.G.2250 rev. x 1–14 1. an-sè-gú 2 túg gùn 50 lá-2 åà-da-um-TÚG-II 3. 71 aktum-TÚG 42 gu-dùl-TÚG åà-da-um-TÚG-I gada-TÚG 5. 1 níg-lá-gaba 81 sal-TÚG 7. 1 íb+II-TÚG ú-háb 93 íb+III-TÚG sa6 9. 76 íb+3–TÚG gùn su-nígin 2 mi-at 45 túg-túg 11. 9 gu-mug-TÚG gaba-ru 13. níg-kas4 Ma-rí ki These garments must have been given to some members of the expedition during the journey home, obv. iii 6–13: Ru12-zi-lum Ar-si-a-hu mas– kim I-bí-zi-kir ì-ti mi-nu níg-kas4 Ma-rí ki “(2+2+2 garments) (for) R. (and) A., representatives of (minister) Ibbi-zikir, returning from the military expedition of Mari.” In obv. iv 14–V 7, a messenger is mentioned, but his is a rather common name: Ingar, who “has brought the news” that a peace has been agreed upon with “(the ceremony of) the oˆering of oil of Mari:” (1+1+1 garments) In-gàr [. . .]ªxº [níg-mu]l [ì?]-gis Ma-rí ki nídba. Immediately after this we ˜nd the delivery of cloth for two girls of Mari (probably spoils of war) destined to serve the princess Kesdut (who will later go as a bride to Kis), obv. v 8–12: (2 garments) 2 dumu-mí Ma-rí ki ses: pa4 Kés-du-ut. It is not clear whether the delivery of garments to “the charioteers,” ugula surx-BAR.AN, of six “sons of the king,” Gadum, Sag-damu, Ibte-damu, Ílzi-damu Ìrkabrizu, Nehar-damu, obv. v 19–vi 10, is in some way related to the movement of the Eblaite troops (according to TM.75.G.2277, only the princes

18

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

the messengers of Agagalis and Luåatum (rev. vi 19–viii 12). Explicit reference is made to the campaign against Mari in rev. ix 18–22: 20 lá-3 sukésda su-ra su-ra in níg-kas4 Ma-rí ki “17 nooses . . . for the campaign (against) Mari.” The missions to Kis by people from Mari (three people according to rev. v 28–vi 7) and As(s)ur are recognized by Ebla insofar as these are also provided with garments, rev. ix 3–6: (2+2 garments) A-su-ru12ki DU.DU Kiski. This may be the Assur on the Tigris, also involved in this wide˘ung diplomatic game.54 The mission had been preceded by an exchange of messengers, rev. iii 21–iv 4: (2+2 garments) A-hu-lum Ma-ma-da-ru12 A-su-ru12ki (2+2+2 garments) I-in-zé maskim Ir-ti Ib-dur-ma-lik maskim Bar-za-[ma-]ù DU.[DU siin] A-su-ru12ki; iv 22–29: (1 cloth) I-da-ì Ma-rí ki (1+1) maskim-sù DU.DU si-in A-su-ru12ki.

Igrub-damu and Ze-damu had previously joined the army at Terqa; see above). Starting from obv. viii 4 and until rev. i 5 the deliveries would appear to all have the same destination, that is, towards the region of Mari, since Eblaite o¯cials were based at {a(l)abidu and Terqa, where the army of Ebla had arrived. There contact is made again with the representatives (lú-kar) of Mari, and envoys arrive from Kis (åAtum and Ida-i). A lengthy section deals with Nagar (Appendix Text 3), with which Ebla consolidates her relations. The receivers of gifts are, apart from the king, Nizi (second highest authority in Nagar) and the son of the king. Other towns in the region were also involved: A-bù-la-duki (rev. i 12) = Aba4-la-daki of the texts of Tell Beydar; Nu-ba-duki (rev. iii 13) = (?) Na-ba-ti-umki / Na-ba4-daki = Tell Beydar).53 A triple alliance is forged against Mari, consisting of Ebla, Nagar, and Kis. Nagar becomes directly involved in this diplomacy, sending two envoys to Kis, together with two couriers (rev. II 12–23). It would appear that the lead in this political game was taken by minister Ibbizikir when he was still in the area of Terqa. Ibbizikir, in fact, received two deliveries of beer from Nagar, rev. ii 5–11: “2+2 garments (for) Zuwadumuda of Nagar (who) has handed beer (over to) Ibbi-zikir: two receipts.” Ibbi-zikir received additional gifts from Nagar: “meat,” uzu, and “a mule,” BAR.AN, and he gives garments in exchange, rev. iv 30–v 4: (1+1 garments) Ù-zu-rí Na-gàr ki su-mu-“tag4” uzu I-bí-zikir; rev. v 17–21:] ugula [. . . .] Na-gàr ki hi-mu-DU 1 BAR.AN I-bí-zi-kir. There are then garments for allied towns which had already received gifts at the start of the campaign: pairs of anonymous people from the cities of Raåak, Burman, Emar, Utik and Iritum, the badalum of {arran, a person from each of the cities of Garmu, Kakmium, NIrar and Haddu, as well as

The peace was sealed with the oil libation ceremony. Minister Ibbi-Zikir, who had led the army to victory and conducted the peace negotiations, received a number of gifts on this occasion, TM.75.G.12450 (AAM I.Z. 14, very fragmentary) obv. iii 2u–12u: . . .] [1] gír mar-tu . . . 7 gín D. kù-gi zi-du-sù I-bí-zi-kir lú ì-ti in níg-kas4 Ma-rí ki lú nídba ì-gis “. . . 1] mar-tu dagger . . . 7 shekels of gold (for) its z. decoration: (for) Ibbi-zikir, which (is the gift for) having taken part in the expedition (against) Mari, (and for) the oil oˆering.” Towards the end of the campaign and with victory already in sight, †ema-Kura, “the overseer of the property of Ibbi-zikir,” went to the minister with a “goblet,” zi-bar of ˜ve shekels (39 gr.) “for (the ceremony of) the oil oˆering” that sealed the peace, ARET VII 109: 1 zi-bar 5 babbar:kù al6 †è-ma-dKu-ra ugula zax I-bí-zi-kir lú níg-kas4 Marí ki lú níbda ì-gis. This man is also mentioned in

53. The list of toponyms of the tablets from Tell Beydar are listed in Subartu 2, 191–92. For the identi˜cation of Na-ba-tiumki / Na-ba4-daki with Tell Beydar, see W. Sallaberger, NABU 1998/130.

54. This town is also mentioned in TM.75.G.2278 (which concerns the victory over Mari, see below) rev. iv 1 (A-su-úr ki) in the following sequence: NE-sa-u9ki, Kablul, As(s)ur, DU (rev. iii 11–iv 3). For A-su-úr ki see also ARET III 159 obv. iii.

7) The Peace with Mari and the Triumph of Minister Ibbi-zikir

A VICTORY OVER MARI AND THE FALL OF EBLA

the MAT TM.75.G.2277, dated to the end of the campaign (see above), obv. xii 5–9: 1 sal-TÚG è I-bí-zi-kir Si-ma-dKu-ra su-ba4-ti. A short period of time must have passed between victory in the ˜eld and the o¯cial peace agreement. The MAT that records deliveries of garments from the beginning of the campaign until the start of peace negotiations is (as we have seen) TM.75.G.2277. The one that registers gifts for the peace ceremony has the next inventory number, TM.75.G.2278. These two tablets were, therefore, placed either next to or very close to one another on the shelves of the archive. The month name is broken in the ˜rst of these texts, whereas the second is dated to month II-bis of the local calendar: rev. x 12–16: su-nígin 72 túg-túg 1 gu-mug-TÚG è (unwritten space) iti se-gur10-II. Next in sequence is the MAT TM.75.G.2335, dated to the III month (iti za-åà-tum) of the year of the peace, recording garments, together with objects in precious metals, destined for the principal protagonists in the victory over Mari. This tablet was drawn up a month later than TM.75.G.2278 and TM.75.G.2250 (iti ig-za, month II of the Semitic calendar; see the previous paragraph). TM.75.G.2426, parallel to TM.75.G.2335, is a lengthy text dealing exclusively with objects in precious metals. It starts with the gifts given to Ibbi-zikir on his return from his great undertaking. Next are gifts for Syrian cities long allied with Ebla that had assisted the expedition against Mari. Lastly are listed gifts for the three great states of Haddu, Nagar and Kis that had supported Ebla in its attempt to suppress once and for all Mari’s expansionist impulses. This document should be seen as an addition to the extremely fragmentary AAM TM.75.G.12450 from the same year, an anomalous administrative practice imposed by the far greater outgoing quantities of precious metals rendered necessary by the extraordinary nature of events. The last section, with a justi˜cation of the issues and possibly including the date, is lost. The text includes numerous direct references to the Mari campaign, while other deliveries or disbursals relate to internal

19

matters and have no apparent connection with this event. According to TM.75.G.2335 the textiles distributed included 452 garments (túg) and 55 blankets (gu-mug-TÚG), and the silver objects “issued,” è, according to TM.75.G.2426, weighed 399 minas (187.53 kg). The list that we can construct from these two texts provides a map of the political geography of Syria at the time. Some of the elements registered in these two documents already appeared in TM.75.G.2278. The minister received a new war chariot from the king, as was the custom in Ebla on such occasions.55 A small text from the lot of tablets found in L. 2586 (already mentioned above), TM.74.G.102, records part of the equipment of this chariot. TM.74.G.102 obv. i. 1 [. . .] 1 gaba kù-gi ” níg-anse-ag 4 kù-sal ti8musen [kù]-gi ii. 1 [. . .] 1[+1] eskirix 3 I-bí-zi-kir en rev. i. 1 ì-na-sum DI† mu 3 nídba ì-gis 5 Ma-rí ki ii. 1 2 níg-anse-ag 4 kù-sal ti8musen kù-gi Na-gàr ki 3 en ì-na-sum 5 [. . .] “[. . .] 1 breastplate of gold (with) reins (and) 4 k. (in form) of an eagle, of gold; [. . .] 2 bits: the king has given to Ibbi-zikir. Year of the oˆering of oil of Mari. 2 reins, 4 k. (in form of) eagle, of gold: the king has given to (the king? of) Nagar [. . .].” Other pieces of equipment for a chariot (possibly the same one) given to Ibbi-zikir by the king at the same time as “the oil oˆering of Mari,” are 55. See Archi and Biga, “Wars at Ebla” (in preparation).

20

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

recorded in TM.74.G.128, which belongs to the same lot as the previous tablet.

of {aba, received two light earrings for having delivered these gifts.

TM.74.G.128 obv. i. x+1. [. . .] GI†-ªxº babbar:kù nídba 3u. ì-gis Ma-rí ki ii. x+2. 1 GI†-gigir-II 2 lu-bù 4 gu-sal kù-gi 1 GI†-PA [. . .] iii. x+1. Wa-ri-KAki I-bí-zi-kir 3u. en ì-na-sum rev. (traces of signs)

TM.75.G.2426 obv. i 1–ii 17 i. 1–4 [. . .] 5. [20(?)+]ª22º ma-na babbar:kù su-bal-ag 7. [x]+1 ma-na kù-gi+IV [x] eskirix ti8musen [x] ªmaº-na 9. [x] [ní]g-anse-ag [x] kù-sal [x ma]-na [x z]i-kir-[ra]-tum [. . .] ªxº 11. [. . .] [. . . ii. 1 [. . .] [4 babbar:]kù 3. [su-bal-a]k 1 kù-gi 5. 2 géstu-lá Ba-ga-ma 7. lú {a-ba su-mu-“tag4” 9. igi-du8 en 11. wa I-bí-zi-kir 13. mi-nu níg-kas4 15. Ma-rí ki ma-lik-tum 17. in-na-sum

The MAT TM.75.G.2278 (Appendix Text 4) opens with the gift made to Ibbi-zikir at the time of the peace agreement (which included the chariot recorded in detail in the previous texts): “1+1+1+1 garments; 1 plate (of gold) of 2 minas (920 gr.); 1 belt (with) sheath (and) dagger of 1.30 minas of gold (705 gr.); 1 mar-tu dagger of gold; 4 reins (with decoration in form of) eagle of gold; 1 chariot, the 2 wheels (decorated with) gold with r.: which the king has acquired (for) Ibbi-zikir, which is (for the ceremony of) the oil oˆering of Mari” (obv. i 1–ii 2). The “carpenter,” nagar, who made the chariot received a set of garments, and another was given to Amurum, son of the king of Mari, who had represented his city during the ceremony (ii 3–6, iii 7–10). Another person bearing the same name, Amuru, son of Ibbi-zikir, also receives a garment (iii 11–14).56 TM.75.G.2426 starts with a (fragmentary) list of chariot ornaments apparently sent by the queen for “the meeting,” igi-du8, of the king (Isar-damu) with Ibbi-zikir upon his return “from the military campaign (against) Mari;” Bagama, the son

56. Amuru, the son of minister Ibbi-zikir, appears in later documents such as TM.75.G.2593, which mentions the marriage of princess Kesdut (to the king of Kis), obv. ix 6–9:] é I-bí-zikir lú níg-mu-sá bur-kak Kés-du-ut. This text has in obv. ix 6– 9: (1 cloth) A-mu-ra dumu-nita I-bí-zi-kir; in obv. xi 23u–rev. I 5: (1+1 cloth) A-mu-ru12-um dumu-nita lugal Ma-rí ki.

ARET II 15 (1), which records a garment for “the king when he moved to Mari,” i 1–6: 1 níg-lá gaba en in ud è si-in Ma-rí ki, may refer to this concluding phase of the military campaign. Here no mention is made of a campaign conducted by the king against Mari, nor do we know of any other documents recording a journey of the king to Mari. The king simply goes out to meet the army which is returning from its expedition against Mari. Prosopographic elements favor dating the text to this period; Ibbi-zikir was at Tuttul (in Du-du-luki) because wool was sent to him there for mule harnesses as well as other equipment for his chariot (section 21). The following fragmentary section of TM.75.G.2426 can be partially restored on the

A VICTORY OVER MARI AND THE FALL OF EBLA

basis of TM.75.G.2278 obv. viii 10–ix 2: (1+1+1 garments) Ì-lum ses:pa4 dBE Du-du-luki lú ì-ti nígkas4 in Ma-rí ki. This concerns “Ilum, the attendant of (Dagan) the Lord of Tuttul, who came back from the expedition (against) Mari.” Ilum had apparently been persuaded to follow the Ebla army so as to ensure the blessing of the god Dagan. TM.75.G.2426 obv. ii 18–iii 3 ii. 18. 15 babbar:kù

iii.

21

19. [1] íb-lá [1] gír-kun [x babbar:]ªkùº 21. [. . .] ªxº 1. [Ì-lum (?)] ªsesº:[pa4] (?) 3. ªdºBE Du-du-luki

The next section relates to gifts that Ibbi-zikir received on his return from Mari.

TM.75.G.2426 obv. iii 6–iv 9 iii. 6. 1 ma-na sa-pi 7 kù-gi-IV 7. 1 íb-lá 1 si-ti-tum 1 gír-kun 1 gi-ba-du 1 GI†-sal 1 ma-na babbar:kù 9. níg-sa10 1 gír mar-tu kù-gi gibil 11. ap 1 dib ma-na libir 13. wa 1 gaba 2 ma-na libir 15. wa eskirix libir 10 lá-1 ma-na 17. [. . .] iv. 1. [. . .] su-bal-ag 3. 1 ma-na 35 kù-gi-IV 1 íb-lá 1 si-ti-tum 1 gír-ku[n] 1 gi-ba-du 1 gis-sal 5. I-bí-zi-kir lú ì-ti 7. in níg-kas4 9. Ma-rí ki

TM.75.G.2335 obv. i. 16. 1 ma-na sa-pi 7 kù-gi ii. 1. 1 íb-lá 1 si-ti-tum 1 gír-kun 1 GI†-sal 1 gi-ba-du gibil

The ˜rst part of TM.75.G.2335 gives a particularly detailed list of the trappings and decorations of the chariot, and is therefore given here in its entirety: obv. i 1–ii 5: 1 túg gùn 1 åà-da-umTÚG-II 1 åà-da-um-TÚG-I 1 aktum-TÚG 1 íb+ III-TÚG sag 1 íb+III-TÚG ú-háb 10 lá-1 ma-na sú+sa kù-gi-IV eskirix 1 ma-na kù-gi 2 níg-anseag 16 kù-gi 4 zi-kir-ra-tum 8 babbar:kù 3 zú-zúba-tum 1 GI†-gigir-II ra-åà-tum kù-gi 1 gaba 2

ma-na kù-gi 1 dib 2 ma-na 1 gír mar-tu kù-gi lú en wa 1 ma-na sa-pi 7 kù-gi 1 íb-lá 1 si-ti-tum 1 gír-kun 1 GI†-sal 1 gi-ba-du gibil níg-ba I-bí-zikir lú til Ma-rí ki. A precious gift is made to Amur-damu, probably one of Ibbi-zikir’s sons who had accompanied his father during the campaign (one of Ibbi-zikir’s sons-in-law bore the same name: the husband of Bu-KUBABBAR, daughter of the minister).

i.

12.

1 dib 2 ma-na

i.

11.

1 gaba 2 ma-na kù-gi

ii.

2. 3. 5.

níg-ba I-bí-zi-kir lú til Ma-rí ki

22

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

TM.75.2426 obv. iv 10–17: 10 lá-3 ma-na sa-pi 5 babbar:kù su-bal-ag 3 ma-na 6 kù-gi-II1/2 2 níganse-ak 4 kù-sal 4 zi-kir-ra-tum A-mur-da-mu su-du8-más. The two documents then continue in an identical manner with lengthy passages that record gifts for friendly kings, for their elders and for other o¯cials (badalum and ugula) of the cities of NIrar, Raåag, Burman, Dub, Imar, Garmu, Lumnan, Ibubu, Ursaåum, Utig, Kakmium, Iritum, {arran, Sanapzugum, GudadaLUM, Nagar, Haddu, and Kis (Appendix Texts 5 & 6). No direct reference is made here to the campaign against Mari, but the overall data provided by these two texts makes it likely that these gifts were related to the role that these states had played during the war. Twelve kings each received a gold plate weighing one mina (470 gr.), v 1–26. Each of the gold plates for ˜ve badalum weighed 50 shekels (392 gr.) (v 27–vi 12), while those for four people respectively from Kis, Haddu, Nagar and Kakmium weighed 40 shekels (313 gr.) (vi 13–25). There then follow gifts consisting of three plates of 30 shekels, three of 20 shekels, six of 16 shekels and, ˜nally, six of 10 shekels each (apart from the garments, recorded only in TM.75.G.2335).57 The text recording metals, TM.75.G.2426, continues with the usual annual oˆerings to various deities. These are followed by Ibbi-zikir’s oˆering to the god Hadda of Halab of two “plates,” dib, weighing two minas (940 gr.) in gold and by the king’s oˆering of two “bracelets,” gú-li-lum, again weighing two minas of gold, for a puri˜cation ceremony upon the return from the expedition against Mari,” obv. x 6–xi 1: 10 lá-2 ma-na babbar:kù su-bal-ag 2 ma-na kù-gi 1 dib en ásdu 1 dib-sù in-na-sum I-bí-zi-kir 10 lá-2 ma-na babbar:kù su-bal-ag 2 ma-na kù-gi 1 gú-li-lum en ás-du [1 g]ú-li-lum-sù sikil dåÀ-da lú {a-labxki lú ì-ti in níg-kas4 M[a-ríki]. 57. The representatives of Nagar and Kis, as well as the king of Kakmium, who took part in the expedition against Mari, delivered gifts to Ebla in exchange, TM.75.G.10070 rev. iv 1–10: (1+1+1+1+1 clothes 40 turtle shells) mu-DU Ni-zi “ur4” Na-gàr ki (1+1+1 clothes) mu-DU A-zaxki Kiski; v 6–11:] mu-DU en Kak-mi-umki lú ì-ti níg-kas4 Ma-rí ki.

One Line Short

After a number of gifts to women of the Ebla court, the text records small amounts of silver (47 gr.) and gold (86 gr.) to decorate six objects for the gods of Mari, in order to propitiate the gods of the defeated city. Thereupon follow ˜ve shekels of silver (39 gr.) to decorate “the tablet of the oath of the king and Ibbi-zikir for the Mountain country,”58 obv. xii 11–27: 6 babbar:kù wa su+sa 7 1/2 babbar:kù su-bal-ag 11 kù-gi-II-1/2 ni-zi-mu nu11ªzaº 6 su?-ªxº-lum [dingir]-dingir-dingir Ma-rí ki 5 babbar:kù nu11-za 1 dub nam-kud en wa I-bí-zikir in Kurki. In this context, the Mountain country must be the region of Jebel Bishri: the Eblaites felt it important to draw the semi-nomadic peoples of this region into their sphere.59 The following record concerns the death of the daughter-in-law of Irik-damu, son of the former minister Ibrium. She was “the daughter of the queen of [. . .],” obv. xiii 4–14: ª50º babbar:[kù] subal-ag sú+sa kù-gi-II-1/2 2 bu-di ì-gis-sag ma-liktum [. . .] sè ús dumu-mí-sù é-gi4 I-rí-ik-da-mu. Ibbi-zikir and the king who had met him on the “road home from the expedition (against) Mari” received “a visit,” (igi-du8) from Dadubdamu, a daughter of the king who was a “priestess,” dam dingir, and who lived at {uzan.60 The parallel text TM.75.G.2335 also adds a cloak to this gift of two toggle-pins.

58. The setting of the tablet in the temple of Kura that contained the oath between Isar-damu of Ebla and Ikun-isar of Ebla weighed 50 shekels of silver (390 gr.) according to TM.75.G.2464 obv. ii 16–iii 22 (see above, section 4 § 5). 59. Kurki quali˜es the settlements of Labanan and Angai destroyed by king Anubu in the Euphrates valley, TM.75. G.2367 obv. ii 6–8, iii 4–6 (Letter of Enna-Dagan). ARET VIII 534 (44) refers to same region: níg-mul en Ì-mar ki Mar-tuki til in Kurki “the news that the king of Emar has defeated Martu in the Mountain region.” In documents concerning the administration of agricultural products and animal breeding, Kurki could also mean the region west to Ebla, see TM.75.G.1845, published in SEb 7 (1984) 71–72, where Kurki-Kurki stands in opposition to Ebla. 60. Dadub-damu received two more toggle-pins according to TM.75.G.2426 rev. xi 39–43. She is also mentioned in ARET I 1 (78), TM.75.G.2339 rev. v 2–8; for further documents, see ARES 2, 277–78.

A VICTORY OVER MARI AND THE FALL OF EBLA

TM.75.G.2426 obv. xiii 16–29 15. sú+sa babbar:kù 2 bu-di 17. Da-dub-da-mu 1 dumu-mí 19. en {u-za-anki 21. lú DU.DU igi-du8 23. en wa 25. I-bí-zi-kir lú ì-ti 27. in níg-kas4 29. Ma-rí ki

23

TM.75.G.2335 rev. ii 12–21 12. 1 zara6-TÚG 2 bu-di sú+sa babbar:kù 13. Da-dub-da-mu dumu-mí 15. en {u-za-anki 17. in ud DU.DU 19. igi-du8 en

21. Ma-rí ki

This same event is also recorded in TM.75. G.2278, which has some sections parallel to these two documents (see above) obv. vii 1–11: [a cloth Da-dub-da-mu] dumu-mí en {u-za-anki dam dingir lú DU.DU igi-du8 I-bí-zi-kir lú ì-ti in níg-kas4 Ma-rí ki. The role and position of the next individuals is not clear, though they may all have been “traders,” dam-gàr; they each received a gold plate (dib) in recognition of their having stayed or established themselves (tus-LÚxTIL) in Mari for some time, TM.75.G.2426 obv. xiv 4–xv 5: Gadum, GuNElum, Zimina-malik the son of Enna-i of Zibada and

Buwa . . . tus[-LÚxTIL] in Ma-rí ki. The parallel text TM.75.G.2335 lists in obv. x 23–xi 7: Gadum, GuNElum, Enar-Armi of Zabalum, Ri-daåar, Ilti of {ara-i, Zimina-malik of Enna-i of Zibada and Buwa dam-gàr tus-LÚxTIL in Ma-rí ki. Two sections relate to people who had participated in the campaign: 1) six people receive two garments and a gold plate of 87 gr. each; 2) seven + two + one people receive a cloth each because “their fathers died in Mari.” These three “fathers” each had a son (listed in the previous section) who accompanied them on the campaign.

TM.75.G.2426 obv. xv 19–rev. i 12 obv. xv. 18. 1 ma-na kù-gi-II-1/2 19. 6 dib 10 (gín) rev. i. 1. Zi-kir-da-mu lú NI-za-ra-nu 3. {áb-ra-ar lú Sá-mu-ù 5. Nap-ha-ì lú Ik-na-ù 7. A-su-ur-ma-lik lú EN-†ID 9. {a-ra-ì

TM.75.G.2335 obv. xii 24–rev. i 22 obv. rev.

xii. 24. 25. i. 1. 3. 5. 7.

(6+6+6 garments) Zi-kir-da-[mu] lú NI-za-ra-ni {áb-ra-[ar] lú Sá-mu-ù Nap-ha-ì lú Ik-na-ù A-su-ur-ma-lik lú EN-†ID {a-ra-ì

24

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

11. 13.

lú Rí-ì-ma-lik Du-bí-zi-[kir] l[ú En-na-ì] [. . .

9. 11. 13. 15. 17. 19. 21.

The next section registers the gifts for the son of the king of Nagar and for Nizi, who was second only to the king in that city (the plate for Nizi weighs one mina, that for the son of the king of Nagar forty shekels). The king of Nagar and Nizi had already been included in the list in the MAT TM.75.G.2250 obv. i 13–18 (quoted above). It was, therefore, at this point that negotiations began for the marriage of Tagris-damu, daughter of king Isar-damu, to the son of the king of Nagar. 61 TM.75.G.2426 rev. ii 1–15. ii. 1. 1 ma-na sa-pi babbar:kù su-bal-ag 3. sa-pi kù-gi-II-1/2 1 dib 5. dumu-nita en 7. Na-gàr ki 2 ma-na tar babbar:kù 9. su-bal-ag 1 ma-na kù-gi-II-1/2 11. 1 dib Ni-zi ur4 13. Na-gàr ki 1 ma-na 10 babbar:kù 15. 7 ábba-sù 10 (gín)

61. M. G. Biga, “The Marriage of Eblaite Princess TagrisDamu with a Son of Nagar’s King,” Subartu IV, 2 (1998) 17–22.

lú Rí-ì-ma-lik Du-bí-zi-[kir] lú En-na-ì 10 sal-TÚG 7 dumu-nita Sá-mu-ù 2 dumu-nita En-na-ì 1 dumu-nita EN-†ID a-mu-a-mu-sù in Ma-rí ki ug7

Other gifts for the king of Nagar, his son and Nizi are listed in the MAT TM.75.G.2337 (the ˜rst section registers gifts for Ibbi-zikir) obv. i 14–ii 25: 36 kù-gi nu11-za [x] GI†-gam-gam ª1º [GI†gígir] ªéº? [x] ªdaº? [. . .] [x] [zú-zú]-ba-tum sa-pi zabar A† 1 GI†-zú níg-ba en Na-gàr ki Ni-zi ur4 su-mug-“tag4”-sù lú hi-mu-DU tus: lú Ma-rí ki (1+1+1 garments) 1 dib 1 ma-na Ni-zi “ur4” wa (1+1+1 garments) dumu-nita en Na-gàr ki. This MAT dates to the same year as TM.75.G.2426. A synchronism between the two is provided by the death of Magaradu, “a woman of the king,” dam en, who had been part of the royal household since the time of king Irkab-Damu.62 A cloak and two silver toggle-pins (zara6-TÚG, bu-di) were provided for her funeral ceremony. Furthermore, in line with normal custom, clothing was also disbursed to illustrious deceased relatives, in this case to two earlier kings, Irkab-damu and Igris-{alab, and to Dusigu “the mother of the king,” ama-gal en.63 TM.75.G.2426, which records only metal objects, has in rev. v 11–16: sú+ sa babbar:kù 2 bu-di Ma-ga-ra-du dam en si-in ÉxPAP “20 shekels of silver (for) 2 toggle-pins (for) Magaradu, the woman of the king, for the funeral ceremony.” TM.75.G.2337 registers, in-

62. See A. Archi, Amurru 1 (1996) 122. 63. On the custom of oˆering garments to important deceased persons during a funeral ceremony, see Archi, ZA 92 (2002)183–84.

A VICTORY OVER MARI AND THE FALL OF EBLA

stead, only the garments: 1 zara6-TÚG Má-ga-radu dam en si-in ÉxPAP 2 túg gùn 2 íb+III-TÚG sa6 gùn Ìr-kab-da-mu wa Ig-rí-is-{a-labx 1 zara6TÚG Du-si-gú Má-ga-ra-du su-mu-“tag4” “1 cloak (for) Magaradu, the woman of the king, for the funeral ceremony. 2 multicoloured garments (and) 2 kilts (to) Irkab-damu (and) Igris-{alab, 1 cloak (to) Dusigu, Magaradu has consigned.” Ibbi-zikir’s entire family participated in the victory celebrations for their triumphant relative. Gifts are received by: a) the sons of minister Ibbi-zikir: Tubuhu-Hadda, Uti, Enna-damu, Amur-damu, Igna-daåar, Irigu, Ibåa-Hadda (TM75b.G.2426 rev. viii 3–ix 13); b) the sons of the former minister Ibrium (that is to say, Ibbizikir’s brothers and step-brothers): . . .] Idi-åAdabal, Siminilum (x 2–4); c) the brothers of Ibrium: Ruzimalik, Ennani-Il, Zimina-damu, Irub-damu, Amuti, Dubi-ab, Dunaåu, In-malik (x 5–xi 4); d) three elders from Daraåum, probably the ancestral home of Ibrium’s family where he possessed estates (xi 5–13); e) Azimu and Zaåase, respectively the mother and one of the daughters of Ibbizikir; f) ˜fteen of the women of Ibrium and Ibbi-zikir, as well as Bu-KUBABBAR, one of Ibbizikir’s daughters (xi 29–xii 8); g) other women of Ibbi-zikir and the women of the king’s sons (xii 12–38). The document also records harnesses for a chariot given to the king of Haddu and 2.5 minas of silver for ˜ve messengers from Kis: Is-du-bù Ga-wa-um Na-NE Is-dub-NI Su-ma-lum Kiski (xiii 8–13, xiv 17–23). This exchange of gifts was undoubtedly a prelude to the departure of the princess Kesdut, who was soon go to Kis as a bride. The rest of the MAT TM.75.G.2335 has no further passages paralleling those in TM.75.G.2426 or relating directly to the war against Mari. We note, however, that Ibbi-zikir received a set of clothing and the next record concerns two sashes and a kilt for the king of Mari as well as a blanket for NE.NE, his brother, TM.75.G.2335 rev. vi 14–21: (1+1+1 garments) I-bí-zi-kir 1 níg-lá-gaba 1 níg-lá sag 1 íb+III-TÚG ú-háb lugal Ma-rí ki 1 gu-dùl-TÚG NE.NE ses-sù. There then follow gifts for old allies like the kings of NIrar, Kakmium, Arhadu and Dulu, and other gifts also for

25

the cities against which Ebla had conducted military campaigns in the years immediately preceding that against Mari. This would have been to reward them as a means of avoiding trouble while Ebla was occupied with Mari. These cities are: Ib-al6ki, Ib-al6ki lú Edenki “Ibal of the steppe,” Lasa-nuki, Da-ùki lú åÀ-mu, Da-ùki lú igi-tùm a, Sida-(l)uki. They are mentioned in the same order in ARET IV 1 (65)–(70) and must, therefore, have been located in the same geographical area.64 Before the ˜nal summary, we ˜nd a record of “900 skeins of wool for the people now temporarily in {a(l)abitu and Tuttul,” rev. xii 1–9: 9 mi-at KIN siki mu4mu na-se11 na-se11 al6-tus {aa-bí-duki wa al6-tus Du-du-luki su-ba4-ti. As we have seen, Ibbi-zikir had used these two locations as bases for army camps and for conducting negotiations with his allies during his march towards Mari. Reference to the expedition against Mari is also found in the fragment ARET III 188(+ 184) iii 1–2: níg-kas4 Ma-rí ki. The following document, TM.74.G.101, which belongs to a lot of tablets relating to the ˜nal years of Ebla,65 records the gift of a chariot with gold decorations, complete with harness, for the king of Haddu, made in month VI; “document of the year of the attack of (i.e. against ?) the king of Mari.” TM.74.G.101 obv. i. 1.

rev.

ii.

3. 1.

i.

3. 1.

1 GI†-gigir-II kù-gi MA†.MA† gibil wa GI†.PA eskirix en åÀ-duki iti dAs-da-bíl su-ba4-ti

64. ARET IV 1 (69) has just: Da-ùki lú igi-tùm, which has been understood as “the former, earlier Daåu,” see A. Archi, NABU 2000/64. According to the new passage, it should mean: “Daåu at the dam.” P. Fronzaroli, NABU 1999/47 suggests for igi-tùm in connection with water a meaning equivalent to Akk. maharu “to dam up,” mihru “dam.” 65. See n. 41, above.

26

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

ii.

3. 1. 3. 5.

dub-gar DI† mu su-ra lugal Ma-rí ki (unwritten)

8) The Marriage of Kesdut to the Son of the King of Kis Kesdut is the only Eblaite princess who is de˜ned not only as “daughter of the king,” (dumumí en; e.g. TM.75.G.2283 rev. iii 4–5), but also as “daughter of the queen,” (dumu-mí ma-lik-tum; TM.75.G.1917 rev. iii 12–14; 2328 rev. III 11–13).66 We are unable to establish precisely when she was born. The marriage of king Isar-damu to Tabur-damu takes place in the year Ibr. 14 (the RAM is TM.75.G.1730). The MAT TM.75.G.1896 mentions “the wet-nurse of the daughter of the queen” and the queen having given birth, obv. iii 5–15: (3+1 garments) 1 dam ga-du8 dumu-mí ma-lik-tum (1 cloth) . . . in ud tu-da ma-lik-tum. This MAT has to be dated ˜fteen years after the marriage because it has a passage parallel to the RAM TM.75.G.2507 of I.Z. 11.67 Another MAT, TM.75.G.1917, has a similar passage mentioning “the wet-nurse of the daughter of the queen” and “Kesdut, daughter of the queen,” rev. iii 12–14: (1 cloth) ga-du8 dumu-mí ma-lik-tum (4 garments) Kés-du-ut dumu-mí ma-lik-tum. If these two MATs are chronologically close to one another and both passages refer to the birth of Kesdut, then she

66. The most important lady of the earlier period was also named Kesdut. She was probably the maliktum of king Igris{alab, see A. Archi, Amurru 1 (1996) 105. As a deceased ancestor, she received garments together with a few other ladies, Archi ZA 92 (2002) 183–84; add TM.75.G.1923(+) (Ibr. 18) obv. xv 1, concerning the funeral of Ibrium (the preceding lines are missing). TM.75.G.2232 probably mentions her own former wet-nurse, rev. viii 5–7: 1 túg-NI.NI A-na-åà ga-du8 Kés-du-ut. 67. The parallel passages of these two documents were cited by M. G. Biga, Amurru 1 (1996) 62–63. Wet-nurses nursed the children for about two years, and then remained at court, see M. G. Biga, “Wet-nurses at Ebla: a Prosopographic Study,” VO 12 (2000) 59–88.

One Line Short

would have been married at the age of ˜ve or six, which seems unlikely. Kesdut could have been the ˜rst daughter of the queen Tabur-damu. In certain passages where the ladies of the court are listed according to their rank, she is found in third place, after the queen and Amaga, ARET VIII 542 (23) and TM.75.G.1225 obv. ii 21–24: maliktum, Amaga, Kesdut. In TM.75.G.1876 rev. I 3–ii 1, however, Kesdut precedes Amaga. We do not know if Amaga was a daughter of the queen or of another woman of Isar-damu. In the sixth year of minister Ibbizikir (the RAM of this year is TM.75.G.10074) Amaga becomes a priestess (dam dingir) of the god åAdabal of Luban. From this moment on, Kesdut is found for some time in second place in the hierarchy of women, TM.75.G.2649 obv. iii 19–iv 1: (3 garments) ma-lik-tum Ìr-åà-ag-da-mu Kés-du-ut. Shortly before the war against Mari, the marriage was celebrated of Zaåase, daughter of Ibbizikir, and the heir to the throne Iråag-damu. This blood link with the reigning family crowned the rise of the family of the ministers Ibrium and Ibbi-zikir. The marriage takes place in the twelfth year of Ibbi-zikir’s administration, as the RAM ARET VIII 534 (11) records “a gold and silver bracelet (for) the agreement of Zaåase,” zu-lu-mu (/sullumu/) Za-åà-sè.68 Now Zaåase together with her husband occupied the third place in the hierarchy, overtaking Kesdut, TM.75.G.2270 obv. vii 6–viii 5: en . . . ma-lik-tum Za-a-sè Ìr-åà-ag-da-mu . . . Kés-du-ut Dar-kab-da-mu 2 dumu-mí en; ARET I 1 (33)–(34): . . . ma-lik-tum Ìl-åà-ag-da-mu Za-asè . . . Kés-du-ut . . . . The hierarchical order in the lists concerning the oˆerings to the gods is identical: the king – the prince Iråag-damu – the queen 68. Two documents concern the marriage of Zaåase using the Eblaite expression: níg-mu-sá 1 bur-kak Za-a-sè. ARET VII 132 registers 20+20 garments “on account of Paåaba” (al6 Paa-ba4). This lady would be the spouse of {I-daåar, the king of Mari (she is mentioned also in ARET I 44 obv. I 6–7, rev. ii 3). This gift for the marriage of Zaåase was delivered in Daraåum, the original home of Ibrium’s family. ARET VII 117 concerns, instead, a gift of 30+21 garments delivered by minister Ibbi-zikir to “the women (dam-dam) (of his family) in Daraåum on the (occasion) of the marriage (with the ceremony of) the bur-kak vase of Zaåase.”

A VICTORY OVER MARI AND THE FALL OF EBLA

– Zaåase – Kesdut, TM.75.G.1764 r. i 1–viii 13; 2075 obv. i 1–v 19; 2238 obv. i 1–ix 11; 11010(+) i 1–viii 27.69 The plan to consolidate the alliance with Kis by means of a dynastic marriage took shape during diplomatic negotiations that were conducted immediately in the aftermath of the war against Mari. This can be determined from the signs of attention paid to Kesdut on numerous occasions. Two girls from Mari, chosen from among the prisoners, are destined for the princess’ service, TM.75.G.2250 obv. v 8–12 (quoted above), and she receives small gifts in TM.75.G.2426 obv. xi 8–13, rev. iv 1–3. While minister Ibbi-zikir stayed on in the area between Terqa and Tuttul to reinforce his victory, “Guzi, the representative of Ibbi-zikir, brought the agreement (zu-lu-mu) of (/ concerning) Kesdut,” TM.75.2278 obv. ix 13–19: 1+1+1 garments) Gú-zi maskim I-bí-zi-kir su-mu“tag4” zu-lu-mu Kés-du-ut. According to the same document, two more messengers reached the court with news of this agreement, obv. xi 6–13: (1+1+ 1+1 garments) A-bù su-i maskim I-bí-zi-kir su-mu“tag4” zu-lu-mu Kés-du-ut; rev. iii 3–10: (1+1+1 garments) åÀ-da-sa lú Is11-ga-um su-mu-“tag4” zu-lu-mu Kés-du-ut ás-du I-bí-zi-kir. At this point a representative of Tubuhu-Hadda, son of Ibbi-zikir, arrived at the Eblaite court “to bear the promise (concerning) Kesdut from Ibbizikir (i.e. Ibbi-zikir sent news to Ebla that the court of Kis has agreed to the marriage),” TM.75.G.2335 rev. v 12–20: (1+1 garments) Ib-giDa-mu maskim Du-bù-hu-dåÀ-da su-mu-“tag4” ka-as Kés-du-ut ás-du I-bí-zi-kir. TM.75.G.10281 obv. vi 8–16 could record the court’s assent to the plans for the princess: (1+1 garments) Ì-lumBALA maskim A-bu su-mu-“tag4” ka-as Kés-duut ás-du-nu I-bí-zi-kir “(1+1 garments to) IlumBALA, the representative of Abu (who) has brought the promise (concerning) Kesdut from us (?) (to ?) Ibbi-zikir.” It is again this same “Abu, the representative of Ibbi-zikir, (who) brought the

69. These documents have been published by G. Pettinato, OA 18 (1979) 85–215.

27

promise (concerning) Kesdut,” TM.75.G.1680 (MEE VII 23) obv. ii 5–11: A-bù maskim I-bí-zi-kir sumu-“tag4” ka-as Kés-du-ut (the princess appears also in obv. vi 10–10: (20 garments) ti-TÚG en (3 garments) ti-TÚG Kés-du-ut ma-lik-tum su-ba4-ti). This can only mean the ˜nal decision regarding Kesdut’s marriage.70 A lenticular tablet concerning small amounts of wool, TM.75.G.2612, is dated in rev. iii 1–4: DI† mu níg-mu-sá bur-kak Kés-du-ut: “year of the marriage (and the ceremony of) the bur-kak vase of Kesdut.” Reference to this wedding is found in the MAT TM.75.G.2593 rev. x 1–5:] é I-bí-zi-kir lú níg-mu-sá bur-kak Kés-du-ut. This document must date to the period immediately following the peace with Mari, as mention is made of both Amura, son of Ibbi-zikir and of Amurum “son of the king of Mari,” who also appear in TM.75.G.2278, a MAT drawn up following the victory over Mari.71 Later on, the text mentions a gift of 1+1+1 garments and two lapis lazuli bracelets from the king and minister Ibbi-zikir that were taken (to Kis) by Ennani-Il, an o¯cial in the personal service of Kesdut, “on the occasion of the marriage of Kesdut (to) the king of Kis,” TM.75.G.2327 obv. ix 6– 21 (see Eblaitica 1 [1987] 138 [10]): . . . 2 gú-li-lum

70. In the same years (if not months) the marriage was also celebrated between the Eblaite princess Tagris-damu and the heir to the throne of Nagar, Ultum-huhu. The procedure was the same: a messenger of Nagar by the name of “Marzam has brought the promise (concerning) Tagris-damu, the daughter of the king, from the king of Nagar,” TM.75.G.1249 obv. vi 4–14: (1+1+1 garments) Mar-zamx Na-gàr ki su-mu“tag4” ka-as-ka-as Tag-rí-is-da-mu dumu-mí en ás-du en Nagàr ki, see Biga, †ubartu 4, 2 (1998) 18 with n. 9 concerning the term ka-as (Akk. purussûm). The term ka-as always appears in the following context: (garments) PN (NF) su-mu-“tag4” kaas NP (NF). It is not possible to ˜x the year of this marriage. The MAT TM.75.G.1249 quoted above, mentions a “campaign,” níg-kas4, against the town of Armi. This campaign is quoted also in TM.75.G.2276, which also contains a long list of funerary gifts for the priestess Tiste-damu, see Archi, ZA 92 (2002) 176–77. The funeral of Tiste-damu does not appear in any sections of the preserved AAMs. Therefore, her death has to be dated to one of the three last years of minister Ibbi-zikir, because the AAMs concerning these years are very fragmentary. 71. See n. 56.

28

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

. . . lú níg-ba en lú I-bí-zi-kir ì-na-sum in ud nígmu-sá bur-kak Kés-du-ud en Kiski En-na-ni-il lú †a-ù-um su-mu-“tak4.”72 Only the left edge of the reverse has been conserved of the MAT concerning the jewels and garments given to Kesdut as her dowry, TM.75.G.17684: x+ii

x+3 5

7 left edge x+1 3 5

I-b[í]-zi-kir 40 sal-TÚG 40 gùn zu5:a 10 lá-3 sal-TÚG 10 lá-3 gùn [. . . níg-mu-sá bur-kak Kés-du-ut 1 dumu-mí en (unwritten)

The lenticular tablet TM.75.G.2283 (Eblaitica 1 [1987] 122) records animals sent as dowry to Kis: “3290 bovines, 1680 sheep, 159 mules, 1 ass, 5 pigs, 19 bisons, 14 bears. Delivery (for) the marriage of Kesdut, the daughter of the king. Month V, ” mu-DU níg-mu-sá Kés-du-ut dumu-mí en.73 Of the twenty monthly oˆering-lists to the gods, relating to the very last months of Ebla’s existence, one is dated to month XII, TM.75.2398 obv. ii 22–29, and records the departure of the princess: “7 sheep (for) the goddess Ishara of the town of Zidara, (which) the queen has oˆered when Kesdut left,” in ud Kés-du-ut è. In the next eight monthly accounts, Kesdut is no longer mentioned. The registers of foodstuˆs delivered to the court preserved on the tablets from room L. 2712 con72. En-na-ni-ì/il lú Kés-du-ut: ARET VIII 524 obv. xiii 24– 25; ARET X 50 obv. III 1–2, 64 obv. iii 7–8 and 65 obv. iv 11–v 1; TM.75.G.1246 rev. i 11–12; 2270 obv. ii 6–8. En-na-ni-il “ur4” lú Kés-du-ut: TM.75.G.2593 obv. vii 13–15. En-na-ni-il / En-na-ì ses:pa4 lú Kés-du-ut: TM.75.G.10127 obv. vi 11–13; 10184 rev. v 16–18. 73. The dowry of the Hittite princess sent to Ramses II consisted also of “prisoners, horses, cattle and sheep,” who had to cover a distance perhaps longer than that between Ebla and Kis, KUB III 37 + KBo I 17, see E. Edel, Die ägyptischhethitische Korrespondenz I (Opladan, 1994), 142–43, Vs. 20–21.

tain long lists of names of royal women as well as of several daughters of the king (ARET IX 37– 45; see also 46–59). Kesdut never appears in any of them. These documents cover a number of months, but not a complete year.74 The two dates agree: Kesdut left Ebla roughly a year before the city fell. Thus she was saved from sharing the fate of her home city. The archive of food rations from L. 2712 only pertains to the last three years of Ebla. Almost 600 litres of ˘our (27 gú-bar) were given to “men,” (gurus-gurus), who left for the military campaign (against Mari), ARET IX 86 (see also no. 90). The summary ARET X 100 records 1100 litres for the year in which the army left for Mari (rev. vi 8– 17); six months later, 1320 litres were sent (vii 16–27), and after a further six months, another 1320 litres (viii 26–35). ARET IX 80 (3) and (5) records balsams given to Ibbi-zikir for the campaign against Mari. Cereals and oils were “drawn” (íl) by Ibbi-zikir for the same purpose, ARET IX 89. Diplomatic activity in this brief period of time was extremely intense. A large representation from Mari came to Ebla for consultations in the 15th year of Ibbi-zikir, ARET X 100 rev. xii 2–9: 1 mi-at 30 sig15 kú SA.ZAxki wa Ma-rí ki lú ì-ti daba-ar-da “2600 liters of einkorn barley as food (for people) of the palace and Mari, who were present for consultations (/dabartam/, from WSem. *dbr).75 We even have evidence of the presence in Ebla of the queen of Mari, ARET VIII 533 viii 13–21: Ba-ba4 dam {i-da-ar Ma-rí ki lú gi4 mi-nu Ib-laki si-in Ma-rí ki “Baba, the spouse of {I-daåar (king) of Mari who returned from Ebla to Mari.” As we have no information regarding a princess of the Ebla area having gone as a bride to Mari, this must be seen as an exceptional event. Also åAsihu, a son of {I-daåar, probably went to Ebla:

74. Documents from the same archive concerning other kinds of goods mention some people connected with Kesdut. Zi-la-BE lú Kés-du-ut: ARET X 63 rev. iv 5–6 and 64 obv. iii 9–10 (he receives a sheep in both cases). I-sar lú Kés-du-ut: ARET X 127 obv. vi 5–7 (list of ˜elds: gána-ki). For Ennani-Il, see the previous note. This does not prove, however, that their mistress was at Ebla when those documents were written. 75. For this text, see above § 5.

A VICTORY OVER MARI AND THE FALL OF EBLA

he received bread and grains totalling 1660 litres (133 gú-bar), while 800 litres (40 gú-bar) of cereals were delivered for the king himself ({I-daåar, ARET IX 102). The minister Ibbi-zikir traveled to Mari (níg-kaskal si-in Ma-rí ki), ARET IX 20 (12); to Mari and Nagar, nos. 93 and 94 (11)–(12); even to Kis and Nagar, no. 94. 9) The Destruction of Ebla By forging an alliance with Nagar and Kis in anticipation of the confrontation with Mari, Ebla for a brief spell created a political system that embraced both northern Syria and the region of the Middle Euphrates, as well as northern Babylonia, a fact that is not re˘ected in any preserved Mesopotamian sources. Three years after the victory over Mari, however, Ebla was utterly destroyed, and this put at end once and for all to the city’s ambitions as a regional power. There is no continuity in settlement on the acropolis, where the royal palace had stood, at least in the western and southern sectors, the only areas that have been excavated in depth. The palace of the subsequent phase, roughly contemporary with the period contemporary to the Third Dynasty of Ur, was built in the northern part of the lower city. Diˆering geo-political situations created different fates for the two great cities of the third millennium B.C. Mari was conquered and destroyed by Akkad. Naram-Sin made it one of the cities of his empire, recognising a local prince as sakkanakku “governor.” Its geographical isolation ensured its independence during the period of the Third Dynasty of Ur. A daughter of Apil-kÿn, sakkanakku of Mari, married the great †ulgi, son of Ur-Namma.76 From the last decades of the nineteenth century onwards, Mari was once again one of the most important states in the Near East, until it was conquered and utterly destroyed by Hammurapi of Babylon. The former grandeur of Ebla lived on in memory for only a few decades. Naram-Sin claimed 76. On Apil-kÿn, see J. Boese and W. Sallaberger, AoF 23 (1996) 24–39.

29

his victory over Ebla amongst his greatest undertakings in Syria: “Whereas, for all time since the creation of mankind, no king whosoever had destroyed Armanum and Ebla, the god Nergal, by means of (his) weapons opened the way for Naram-Sin, the mighty, and gave him Armanum and Ebla. Further, he gave to him the Amanus, the Cedar Mountain, and the Upper Sea. By means of the weapons of the god Dagan, who magni˜es his kingship, Naram-Sin, the mighty, conquered Armanum and Ebla. . . . The God Dagan gave me Armanum and Ebla and I captured Rÿs/RÿdHadda, king of Armanum.”77 Ebla is always linked here to Armanum, which must be Ar-mi(-um)ki, one of the most frequently mentioned city in the Ebla archives. Armi could not, however, compete with Ebla, to which it was always linked (though with the odd disagreement).78 Ebla, like Mari, reemerged by Ur III times: Mesopotamian sources show that it was one of the principal cities of northern Syria. In the 18th and 17th centuries it underwent extraordinary urban growth, as demonstrated by monuments that have been brought to light. But from the nineteenth century on, the dominant center in northern Syria was Aleppo. Naram-Sin claimed to have been the ˜rst king in Akkadian tradition to conquer Ebla. If he is to be believed, not even Sargon achieved as much. The emphasis which Naram-Sin placed on his own victories in the western regions ˜nds no parallel amongst his other achievements. He clearly felt that his greatest contribution to the creation of universal domination by Akkad (sar kibratim arbaåim) lay precisely in his having taken control of those lands, as far as “Amanus, the Cedar Mountain, and the Upper Sea.” Similar dominion and control had, in fact, been claimed by Sargon: “He (the god Dagan) gave to him (Sargon) the Upper Land: Mari, Iarmuti, and Ebla as far as the Cedar Forest and the Silver Mountains.”79 Since Mari was undoubtedly under

77. RIME 2, 132–34, i 1–ii 7, iii 23–31. See further, p. 136, 1–9: “Naram-Sin . . . conqueror of Armanum and Ebla;” 167, 3– 13: “Naram-Sin . . . conqueror of Armanum, Ebla, and Elam.” 78. ARES 2, 155–68. 79. RIME 2, 30–31, 13u–21u.

30

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

Sargon’s control,80 the general nature of this expression does not exclude the possibility that Ebla had also been conquered by Sargon. On the other hand, we cannot ignore how Naram-Sin stresses that he was the ˜rst to conquer Ebla in an inscription to be displayed in a public place, as were the declarations of Sargon. It is worth underlining here the profound contrast between the kinds of historical-political information provided by Mesopotamian sources and those furnished by the Ebla texts. The royal Sumero-Akkadian inscriptions narrate events with the aim of exalting the sovereign, whose successes are attributed to divine benevolence. Chronological order and even the temporal dimension itself are unimportant. In the administrative documents from Ebla, instead, events are mentioned only insofar as they aˆect the accounting and management of goods, payments, and ceremonial gifts. It is for this reason that the Ebla texts provide an analytical view of events which, if ordered chronologically, assume the rhythm and pace of a chronicle. There is incontrovertible evidence, however, to show how the destruction of Ebla documented by archaeological excavations cannot be attributed to Naram-Sin. If the Akkadian writing had been in use at Kis during the period of the Ebla archives, the Eblaite scribes could not have been ignorant of the fact, given the frequent and direct contacts between two cities.81 The Presargonic ductus of the Ebla texts may have been used only until the ˜rst decades of Sargon’s reign. Nor is the political situation in Syria, as described by Naram-Sin, compatible with Ebla’s hegemony as revealed by the archives. It was only once Ebla had been weakened that Armi could appear as Naram-Sin’s major opponent. We may also note the importance given to the capture of the king of Armanum, RisHadda or Rid-Hadda, a name not belonging to the traditional onomasticon of the Ebla archives.

80. See RIME 2, 12, 86–93: “Mari and Elam stood (in obedience) before Sargon, lord of the land.” A year name attributed to Sargon reads: “Year (in which) Mari was destroyed/ defeated (hul-a),” Gelb-Kienast, Königinschriften, 50. 81. See below, n. 88.

It would appear, therefore that we have no alternative but to ascribe the responsability and the glory for destroying Ebla to Sargon.82 This possibility is excluded, however, by an important synchronism between Ebla and Mari, which has emerged from recent excavations at Mari, related to certain events occurring in Babylonia, news of which reached as far as Ebla. During the excavations carried out in 2000 and 2001 under the direction of J. Margueron, seal impressions were found bearing the name of Isgi-Mari (LAMgiMari) “sur le sol incendié du Palais P-1 . . . près de la porte (i.e. the monumental entrance of the Palace).”83 P-1 is the second and last level of the “présargonique” Palace. Its destruction marked the end of the “Ville II” ( J. Margueron), which is dated to between 2291 and 2200, according to carbon14 analysis. In “salle 11” of P-1 were found impressions of a seal of “{I-daåar king of Mari (ªxº-da-ar / lugal / Ma-rí ki).”84 D. Beyer states that the seal impression of an o¯cial who bore the title sakkanakku (GÌR.NÍTA) “governor” comes from the same area, near the Palace gate. This is stamped “sur un scellement de porte en compagnie d’un autre sceau, royal celui-là, mais très lacunare,” the reading of which may easily be restored: I-ku-[i-sar] / lugal / [Ma-rí ki]. Moreover, “près d’une centaine de fragments de scellements de portes provenant pour l’essentiel des salles 11 et 13 de P-1 ont permis de retrouver deux sceaux au nom d’Isgi-Mari, lugal de Mari.” 82. See the detailed discussion of the diˆerent theses concerning the destruction of Ebla by M. C. Astour, Eblaitica 4 (2002) 58–76. 83. The present authors are profoundly grateful to J. Margueron and D. Beyer for having shown them the results of their excavations during a brief visit to Tell Hariri / Mari in October 2002. J. Margueron willingly made available to us his paper “Mari et la chronologie. Acquisitions récentes et problèmes,” delivered at the meeting: From Relative Chronology to Absolute Chronology, The Second Millennium B.C. in SyriaPalestine, Rome, December 2001 (in print); D. Beyer put at the authors’ disposal his paper on the seal impressions found recently at Mari, read at the 3rd ICAANE, Paris, April 2002 (in print). 84. The remains of the sign before -da- resemble the lower part of TÚG (or something similar), and not the lozenge for {I. The presence of impressions with the names of Iku(n)[isar] and Isgi-Mari does not give any other possibilities for a name ending in -da-ar.

A VICTORY OVER MARI AND THE FALL OF EBLA

These impressions, dating to the ˜nal period of use of level P-1, ˜nally provide a precise synchronism between the Mari and Ebla palaces. The Mari Palace P-1 was already in use during the last thirty to thirty-˜ve years of the Palace of Ebla, the period corresponding to the administrations of Ibrium and Ibbi-zikir, contemporaries of Iku(n)isar and {I-daåar of Mari. It is now certain that the destruction of Ebla precedes the fall of Mari. Since {I-daåar is mentioned in the Ebla texts, he reigned until the fall of Ebla and for an unknown number of years after that event. He was succeeded by Isgi-Mari, whose end came together with that of his city. The impression of a seal, on which we may easily read: I-ku-ªdUtuº / en Ma[-ríki], comes from the “chantier H, qui corresponde à un sondage stratigraphique . . . à l’est du grand palais royal.” This seal, stylistically earlier that those mentioned above, shows that the kings of Mari originally used the title of en instead of that of lugal. We may now place with certainty Iku(n)-†amas(Utu), Iku(n)-†amagan, possibly also Iku(n)-Mari, before the eight kings of Mari mentioned in the Ebla texts.85 The destruction of Mari was total. As Margueron writes: “un niveau incendié marque la ˜n de tous les bâtiments repérés à ce jour de la Ville II: temple d’Istar, quartier du Souk . . . Palais (niveau P-1), . . . temples du Massif Rouge, de Ninhursag . . . Maisons aux installations artisanales (chantier B). L’incendie paraît bien avoir été général, et, dans ces conditions, il n’est pas déraisonnable d’y voir le résultat d’une action militaire.” There is no evidence of any rebuilding, which means that there was a fairly long hiatus in the life of the city. The only exception was found at the Palace: “ici on observe une remise en état immédiate avec exhaussement de certains sols au-dessus de la couche d’incendie, réfection des enduits sans enlèvement de celui qui avait subi 85. For the inscriptions of these kings, see FAOS 7, 8–11, nos 7–11. E. A. Braun-Holzinger, Frühdynastische Beterstatuetten (Berlin: Gebr. Mann Verlag, 1977), 27, has shown on the basis of some writing elements how the Isgi(LAMgi)-Mari’s inscriptions are the most recent among the other Mari inscriptions on statues.

31

le feu et transformation de certains secteurs – espace central de l’Enceinte Sacrée et espace IV – ou création d’un nouveau bâtiment – la salle aux piliers (. . . phase P-0). . . . Or, c’est sur le sol de l’une des parties nouvellement construites – la salle aux piliers – que de la céramique agadéenne a été retrouvée.” J. Margueron’s theory is that, if the immediate restructuring of this small area is attributed to Sargon, then pottery of the Akkadian period would hardly have been already found in this room. He therefore attributes the destruction of Mari to Naram-Sin. The destruction of Ebla, which came earlier since Isgi-Mari is unknown in the Ebla texts, could instead be attributed to Sargon, according to Margueron. The reconstruction of the list of sakkanakku of Mari, dated by J.-M. Durand,86 is not binding since the text of ARMT 22, 333 does not necessarily ˜ll completely the lacuna of T. 343. If the sychronism between Puzur-Estar of Mari and the kings of Ur III: †ulgi (˜nal years) and †u-Sîn, is thus rendered invalid, the beginning of the reign of Ididis (the ˜rst of the sakkanakku) does not necessarily date back to twelve years before Naram-Sin, coinciding with the ˜rst years of Manistusu (or Rimus87 ). Thus (according to Margueron), if we bring Ididis forward a few decades, making him roughly a contemporary of Naram-Sin, there is room to collate before him Migir-Dagan, the king of Mari of the “Great Revolt against Naram-Sin.” However, there are several pieces of evidence, none of which alone would be considered de˜nitive, but which taken together have convinced the present writers to oˆer a diˆerent historical reconstruction. Kis was the only Bablyonian city of which Ebla received regular news from the messengers of Mari for most of the period covered by the archives, and with which this city maintained direct relations for roughly the last ten years of 86. J.-M. Durand, “La situation historique des †akkanakku: nouvelle approche,” MARI 4 (1985) 147–72. 87. An Ur III manuscript of the Sumerian King List places Rimus before Manistusu, see P. Steinkeller, in Literatur, Politik und Recht in Mesopotamien. Festschrift C. Wilcke (Wiesbaden: Harrassowitz Verlag, 2003), eds. W. Sallaberger, K. Volk and A. Zgoll, 272, 278–79.

32

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

its existence. That Kis is mentioned roughly a hundred times in eighty or so documents is proof that speaks for itself.88 Of the other cities, we only have news of Aksak (Ag-sa-gúki; cited ˜ve times, three in the same text) and Adab (cited twice). All three passages mentioning Aksak from TM.75.G.2359 (Ibr. 9), as well as those in TM.75.G.1838 and in ARET III 316, relate to news of the conquest of that city brought (su-ba4ti) by messengers: one from Tuttul and two from Mari.89 As far as Adab is concerned, the ˜rst passage relates to the news of its conquest carried by a messenger from Mari. This event occurs twelve years before the fall of Ebla, TM.75.G.1918 (MEE X 29; I.Z. 5) rev. iii 22–28: 1 ma-na babbar:kù nígba †u-ga-du Ma-rí ki níg-mul Adabki su-ba4-ti “1 mina of silver: gift (for) †ugadu of Mari, (who) brought the news (that) Adab was conquered.” The other reference, in TM.75.G.2429 (I.Z. 10) concerns the mother of a young singer coming from Adab via Mari.90 During the period of the Ebla archives, therefore, Kis was the political center of northern Babylonia. Ebla held its alliance with Kis so dear as to seal it, following the victory over Mari, as described above, with the marriage between princess Kesdut and the king of Kis. At this point it should be made clear that the king of Kis is given 88. The passages are collected by A. Archi, “Kis nei testi di Ebla,” SEb 4 (1981) 77–87 (= SEb 4); Id., “More on Ebla and Kish,” Eblaitica 1 (1987) 125–40 (= Eblaitica 1); Id., “Les titres de en et lugal à Ebla et des cadeaux pour le roi de Kish,” MARI 5 (1987) 37–52 (edition of TM.75.G.10091: pp. 43–50). 89. TM.75.G.2359 obv. ii 21–iii 4: (PN)] Du-du-luki níg-mul Ag-sa-gúki su-ba4-ti; vii 6–12: 2 ma-na kù-babbar níg-ba Ab-NI Ma-rí ki níg-mul Ag-sa-gúki su-ba4-ti; ix 14–x 2: 1 ma-na kùbabbar níg-ba Wa-ru12-tum Ma-rí ki níg-mul Ag-sa-gúki suba4-ti (see already ARES 2, 147). TM.75.G.1838 has in obv. iii 1–7: (1 cloth) níg-ba Ab-ba4 Ma-rí ki níg-mul Ak-sa-gúki su-ba4ti, where Ab-ba4 could be a variant of Ab-NI. This MAT has to be dated to the same year of the AAM TM.75.G.2359 (Ibr. 9) because both mention Zanehi-Mari, “a daughter of the king,” who gave birth. The passage from ARET III 316 ii: (PN)] Dudu-luki níg-mul Ag-ªsaº-ªgúº[ki] s[u-ba4-ti], has been restored by P. Steinkeller, “Observations on the Sumerian Personal Names in Ebla Sources and on the Onomasticon of Mari and Kish,” in The Tablet and the Scroll. Near Eastern Studies in Honor of W. W. Hallo, eds. M. E. Cohen, D. C. Snell and D. B. Weisberg, (Bethesda, Maryland: CDL Press, 1993), 246–45 (see p. 244 n. 42). 90. See ARES 2, 148.

the title of en in the following passages: SEb 4 nos. 34, 35 (TM.75.G.11477 and 11495), 45 = Eblaitica 1 no 20 (TM.75.G.2643); Eblaitica 1 no. 9 (TM.75.G.2277; three times). The title of lugal Kis appears in SEb 4 nos. 6, 20, 23 (TM.75.G.1391, 2236, 2556) and Eblaitica 1 nos. 28, 30 (TM.75. G.10091 and TM.76.G.704). According to TM.75.G.10091, section (1), Is-kùnnu-nu receives a larger gift than that given to (his) father: níg-ba a-bí lugal Kiski “gift (for) the father of the king of Kis,” section (2).91 From these two passages we could deduce that Iskun-nunu was betrothed to Kesdut, and that he was the king of Kis. This would appear to be contradicted, however, by ARET VIII 540, which also derives from the ˜nal years of Ebla, vii 23–viii 2: níg-ba en Kiski wa dumu-nita-sù Us !-kùn-nu-nu Kiski “gifts (for) the king of Kis and his son, Uskunnunu of Kis.”92 The usual syntactical order should, however, be: en Kiski wa U. dumu-nita-sù. The passage SEb 4 no. 45 is also syntactically strange because the document starts with the term a-mu followed by the personal pronoun “his”: (objects) a-mu-sù en Kiski; the translation should be, however: “(objects for) the father of the king of Kis,” and not: “(for) his father, the king of Kis.” Similarly, Eblaitica 1 no. 20, has: i[n] ªaº-mu lugal (Kiski) “for the father of (?) the king (of Kiski).” TM.75.G.2277 rev. ix 4–10 (year of the campaign against Mari), in any case, makes it certain that beside the king of Kis, his father was also still active (a rather unusual situation): . . . en Kiski . . . a-mu-sù wa ses-sù. “(garments for) the king of

91. For a translation as: “(for) the king of Kis, his father,” one should have: lugal Kiski a-bù-sù. The text was published by Archi, MARI 5 (1987) 47–50. For the general evaluation of this document, notice that Ni-zi in rev. i 4 (who follows the king of Nagar) was a high o¯cial of Nagar, see Archi, Subartu 4,2 (1998) 6. 92. The edition by E. Sollberger, ARET VIII, 206, as well as that by G. Pettinato, reads: Ìr-kùn-nu-nu, followed by Archi, Eblaitica 1 no. 31. Read, instead, according to collation, U† instead of ÌR (see also the photo in ARET VIII, pl. XXXI, viii 1). This strange writing could be explained by the fact that this name does not belong to the Ebla tradition. On names formed with sakanu in the Babylonian onomasticon, see Steinkeller, in The Tablet and the Scroll, 243b. On the element nunu, see now A. Cavigneaux and M. Krebernik, RlA 9 (2001) 619–20.

A VICTORY OVER MARI AND THE FALL OF EBLA Kis; (garments for) his father and his brother.”93 We know the names of two brothers of the king of Kis: NI-rí-sum and Bù-su-sum, TM.75.G.2277 rev. v 9–12, x 25–28. It was W. Sallaberger who linked the events of Babylonia recorded in the Ebla archives with certain information derived from Mesopotamian documents.94 The starting point for his reassessment was the lowering of the chronology of Ensakusana of Uruk, made necessary by the fact that the same o¯cial, Urrani, was active both under this king and under Sargon.95 Sallaberger identi˜es the conquest of Aksak (Ibr. 9) and the victories over Kis (Ibr. 10–12) mentioned in the Ebla documents with those reported by Ensakusana, Ens. 1 8–14: “he sacked Kis and captured EnbiEstar, the king of Kis. The leader of Aksak and the leader of Kis, having sacked their cities. . . .”96 The capture of Adab, which falls eleven years later (I.Z. 5) is attributed to Lugalzagesi (years 1–7), who de˜nes himself as énsi of that city. The defeat of Lugalzagesi (year 25) took place at least 18 years after his capture of Adab. At that time, Sargon had been ruling for at least 19 years. Assuming “that Sargon year +/- 25 corresponds to Lugalzagesi 25, Sargon would have ruled Southern Babylonia for another 30 years.” Therefore, for 93. For a-mu / a-bí en Kis, see already Steinkeller, The Tablet and the Scroll, 242 n. 19, who, however, did not have this last passage at his disposal. A large section of this tablet is presented above in § 5. W. Sallaberger suggests to us that the king of Kis could have been called “father” in consideration of his high rank. One should therefore translate the passages in TM.75.G.10091 and Seb 4 no. 45 respectively as: “my father, the king of Kis;” “his father, the king of Kis.” 94. The authors are very grateful to W. Sallaberger for having put at their disposition the paper he read in Marburg, April 2000 (in press in Colloquien der Deutschen OrientGesellschaft 3), which he also presented at the XVLIe RAI, Paris 2000. 95. This was noted by A. Westenholz, OSP 1, 4. F. Pomponio, RSO 68 (1995) 1–14, stressed that since Ensakusana had weakened the supremacy of Kis, sacking this city and Aksak a few years before Lugalzagesi, Kis could no longer have been so powerful in the Ebla period, contrary to what is generally admitted. 96. See Cooper, Sumerian and Akkadian Royal Inscriptions I, 105: Uk 4.1. P. R. S. Moorey, Kish Excavations 1923– 1933 (Oxford: Clarendon Press, 1978) 171, suggests that the ˜nal massive destruction of the ED IIIB city of Kis may be associated with the conquest of Kis by Ensakusana.

33

chronological reasons, the expression “the father (a-mu or a-bí ), the king of Kis” must therefore refer to Sargon. This is a tempting scenario. Mari confronted Kis directly, a city as distant as Ebla. The Akkadian sources give the impression that the Syrian region, and Mari as well, was somewhat marginal to the state of Akkad. The detailed information gleaned from Ebla’s administrative texts shows, instead, how vital relations with Kis were for Mari, and for Ebla herself, both before and after the campaign against Mari. Mari took part in the war against Kis in the years Ibr. 10–12. It does not, however, seem likely that it was Mari that conquered Aksak and conducted a glorious three-year war against Kis. These events, and even more so the conquest of Adab, have to be explained by the rivalry between cities of southern Mesopotamia. The conquest of Aksak and the repeated defeats suˆered by Kis, with the fall of some cities of this state, has to be dated before the rise of Sargon. The order of events corresponds to the information passed down to us by Ensakusana. Certain synchronisms between Ebla, Mari and Babylonia can today be evaluated according to a highly detailed temporal grid. It thus transpires that Ebla was destroyed about twenty-four years after the fall of Kis and thirteen years after the conquest of Adab by Lugalzagesi (Table 1). In this context, the destruction of Mari—to be placed not many years after that of Ebla—must have been the work of Sargon. The anomaly represented by the fact that gifts were sent to both the king of Kis and to his father may be explained if Sargon at that time had already moved his residence to Akkad, and left one of his sons, namely Iskun-nunu in Kis (the few sons of Sargon of whom we know from Mesopotamian sources are those he had in later life). For Ebla (and Mari) the political and cultural center of Babylonia continued to be Kis. They knew, however, that now, alongside the king of Kis, his father also exercised similar authority. A formula such as that in ARET VIII 540 vii 24–viii 2: en Kiski wa dumu-nita-sù Us !-kùn-nu-nu Kiski “the king of Kis and his son, Iskun-nunu of Kis,” shows that when this document was drawn up, there was a “king

34

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

of the land of Kis,” and that one of his sons, Iskun-nunu, resided at Kis. The data from Mari does not con˘ict with the fundamental points of this theory. The two statues of Isgi-Mari belong stylistically to the last phase of the Early Dynastic Period.97 It is di¯cult to imagine that such an earlier style persisted at Mari during the ˜rst years of Naram-Sin. The same applies to the two elaborate seals of this king, which D. Beyer describes as follows: “l’un présente la version pratiquement inversée de l’autre, à quelques détails près. Le schéma iconographique associe une scène de guerre avec char et fantassins, à un héros dompteur des fauves et à une célébration de la victoire royale.” The scene of battle recalls that on the stela of Eanatum of Lagas; a few stylistic elements reveal a more realistic taste. A year name, which we can now attribute with certainty to Sargon, says that Mari was destroyed or defeated: mu Ma-rí ki hul-a (hul = nêrum “to destroy, defeat).”98 One could hypothesize that Mari had, in reality, merely surrendered, thus avoiding destruction. Sargon could then have proceded towards and conquered Ebla. What makes it di¯cult to accept this reconstruction is the entire chronological framework. We clearly cannot attribute to Sargon wars against Kis such as those in the years Ibr. 10–12. Moreover, it is impossible that Sargon had established himself at Kis only a few years later, taking Adab in the year I.Z. 5. We must, in fact, leave a few years for Lugalzaggesi. It should also be added that no document from the time of Naram-Sin mentions any conquest of Mari. The inscription that narrates the expedition of this king against Armanum and Ebla takes it for granted that Mari was already in his possession. The brutal destruction of Mari was the work of someone who wanted to annihilate his enemy entirely, and not someone aiming to draw the city into his own political sphere. Such behavior ˜ts Sargon perfectly. Naram-Sin’s policy, instead, was the opposite, favouring reconstruction of the centers under his control (building works that can be 97. Braun-Holzinger, Beterstatuetten, 57–58. A. Spycket, La statuaire du Proche-Orient ancien (Leiden-Köln: E. J. Brill, 1981), 88–89, points out some archaic stylistic elements of this statuette. 98. Gelb and Kienast, Königsinschriften, 50.

One Line Long

attributed to Naram-Sin have been found at Nagar / Tell Brak and at Tell Leilan), tying local authorities to himself where possible, as is shown by some ˜nds from Mari itself,99 and now also from Urkis / Tell Mozan, where one of his daughters, Taram-Agade, married a local endan.100 J. Margueron is perfectly correct in giving weight to the argument that structures in rough brick, especially ruined buildings, if not soon reused, quickly become unserviceable. One could easily suppose that those who survived the destruction reoccupied some of the palace rooms, preserving and adapting the structures. The pottery found there, however, could be that used some years later. The rebirth of Mari, with the construction of Ville III, has to be attributed mainly to the period of Naram-Sin, when the local dynasts took the title of sakkanakku, already used in the administrative tradition of Mari,101 which now, at Mari, indicates the function of “governor,” while at Ur III it means “general.” J.-M. Durand gave Ididis an early chronological position since he reconstructed the list of the sakkanakku on the basis not only of the synchronism Apil-kÿn – UrNamma, but also on that between Puzur-Estar and the kings of Ur III. There exist, however, good reasons to exclude an identi˜cation of the PuzurEstar of Mari with the Puzur-Estar dumu-lugal and sagina documented by several Ur III documents.102 There is, therefore, no reason to place

99. Two inscribed bronze bowls found at Mari belonged to two daughters of Naram-Sin: †umsani “entu priestess of the god †amas at Sippar” and ME-Ulmas, RIME 2, 157–58, nos. 51 and 52. 100. G. Buccellati and M. Kelly-Buccellati, MDOG 132 (2000) 139–40. 101. See the impression of the seal of Idin-Dagan sagina (GÌR.NÍTA) on a bulla together with the seal of Iku(n)-isar, mentioned above. 102. See T. M. Sharlach, “Beyond Chronology: The †akkanakkus of Mari and the Kings of Ur,” in W. W. Hallo and I. J. Winter eds., Seals and Seal Impressions (Proceedings of the XLV e Rencontre Assyriologique Internationale II; Bethesda, Maryland: CDL Press, 2001), 59–70. Doubts about the identity of these two people were already expressed by P. Michalowski, in Immigration and Emigration within the Ancient Near East. Festschrift E. Lipinski (Leuven: Uitgeverij Peeters en Departement Orientalistik, 1995), eds. K. van Lerberghe and A. Schoors, 186.

A VICTORY OVER MARI AND THE FALL OF EBLA

Ididis in the period preceding Naram-Sin. He could have been the founder of the local dynasty recognized by Naram-Sin (the round ˜gure of sixty years for the reign of this ˜rst king is, however, open to question). If the list of the rebel kings of the Geneva Version of the Great Revolt has some historical foundation, Migir-Dagan of Mari can be positioned without di¯culty in the ˜rst half of Naram-Sin’s reign.103 To whom, therefore, can we attribute the destruction of Ebla? The total destruction, sack and burning of Palace G, together with the destruction of building P4, northwest of the Palace G,104 cannot have come about accidentally. This is demonstrated by the complete abandonment of the entire central area of the city following its destruction.105 If the enemy was external, therefore, the only possible candidate is Mari. Nagar, the regional state east of the Euphrates, was never strong enough to confront Ebla. Another rival, Ibal (east of modern-day Homs), and often in con˘ict with Ebla, instead, never had much luck. No center in northern Syria could hope to defeat Ebla when the city was at the height of its powers. The most likely scenario is, therefore, the following. Mari quickly recovered from its defeat, despite the alliances Ebla had forged with Nagar and Kis. Three years later, Mari managed to carry out a successful attack on Ebla and the utter destruction wrought on the city was aimed at eliminating once and for all this great rival. At that time, {I-daåar must still have been king of Mari, as the Ebla sources make no mention of any dynastic succession in that city. He continued to reign but we have no way of determining for how many more years.

103. For a discussion of the historically documented events used in the tale of the Great Revolt, see J. Goodnick Westenholz, Legends of the Kings of Akkade (Winona Lake, Indiana: Eisenbrauns, 1997), 221–61. 104. N. Marchetti and L. Nigro, “Handcraft Production in the Public Building P4 at EB IVA Ebla,” Berytus 42 (1995–96) 9–36. 105. As Astour, Eblaitica 4 (2002) 73–76, suggests. He dates the destruction of the Palace of Ebla (with insight, not having at his disposal the new data from the Mari excavations) “prior to not only the destruction of Mari by Sargon but also the destruction of Kis by Lugalzagesi,” and thinks that the Palace was burnt down by accident.

35

The two seals of Isgi(LAMgi)-Mari, which celebrate in three registers and great detail a great victory achieved by this king, employ an iconography shared, in certain elements, by the Royal Standard of Ur and the stela of Eannatum. This could refer to the victory over Ebla. As the army of Ebla was always led by the minister (˜rst Ibrium, and then Ibbi-zikir), and not by the king, it is very probably that Isgi-Mari, before succeeding to the throne, was one of the main protagonists of Mari’s victory. Sargon of Akkad, however, in˘icted on Isgi-Mari the same punishment that the latter had wrought on Ebla. If we place the fall of Ebla in the year Lugalzagesi +/-15, the defeat of Lugalzagesi and his “˜fty governors” by Sargon took place ten years later: Lugalzagesi 25. We must leave Sargon time to impose his rule over southern Mesopotamia. Mari’s defeat at the hands of Sargon could, therefore, have occurred no earlier than thirteen years after the fall of Ebla. Consequently, we can attribute ten or so years to Isgi-Mari as a working hypothesis. A lid of an Egyptian alabaster vase with the cartouche of Pepi I was found in room L.2913 of the Palace of Ebla.106 This object presumably reached Ebla during the period of minister Ibbizikir, whose seventeen years overlapped with the ˜rst period of Sargon’s reign, according to the chronological reconstruction proposed here. Ebla establishes in this way a synchronism also between Egypt and Babylonia. Appendix: Textual Documentation Text 1 TM.75.G.2277 rev. v 8–x 29 v. 8. 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+IV-TÚG sa6 gùn 9. NI-rí-sum ses 11. en 106. See G. Scandone Matthiae, “Les relations entre Ébla et l’Égypte,” in Wirtschaft und Gesellschaft von Ebla (HSAO 2; Hirschberg: Heidelberger Orientverlag, 1988), eds. H. Waetzoldt and H. Hauptmann, 67–73 and plate XII,3.

36

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

13.

15.

17. 19.

21.

23. 25.

vi.

27. 1. 3. 5.

7. 9. 11.

13.

15. 17. 19. 21.

Kiski 1 gu-dùl-TÚG 1 sal-TÚG 1 íb+IIITÚG gùn maskim-sù 2 åà-da-um-TÚG-II 2 aktum-TÚG 2 íb+III-TÚG sa6 gùn En-na-ni-il åÀ-duki su-ba4-ti 1 gu-dùl-TÚG 1 aktum-TÚG 1 íb+IV-TÚG gùn Is11-sa-nu maskim-sù 2 gu-dùl-TÚG 2 sal-TÚG 2 íb+III-TÚG gùn Is-du-bù Gi-is-{I Kiski 2 åà-da-um-TÚG-II 2 aktum-TÚG 2 íb+IV-TÚG sa6 gùn Gi-is-{I Kiski su-mu-“tag4” uzu 2 su-ba4-ti 1 åà-da-um-TÚG-I 1 aktum-TÚG 1 íb-TÚG sa6 gùn Is-du-bù Kiski in Du-du-luki su-ba4-ti 1 gu-dùl-TÚG 1 sal-TÚG 1 íb+IIITÚG gùn 1 gú-li-lum-II a-gar5gar5 kù-gi tar (gín) Ar-si-a-hu tus-L[ÚxTIL] [5] gu-mug-TÚG 5 sal-TÚG 5 [íb+x]-TÚG gùn Ar-sè I-sar {áb-ra-ar In-NE Du-bù-hu-ì su-du8 1 gu-dùl-TÚG 1 íb+IV-TÚG gùn

One Line Short

23. 25.

27. vii.

1. 3.

5. 7. 9.

11. 13.

15. 17. 19. 21. 23. 25.

27. viii.

1.

3.

Kiski lú Ib-laki su-ra 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+IV-TÚG sa6 gùn en Gú-da-da-LUMki 2 gu-dùl-TÚG 2 sal-TÚG 2 íb+III-TÚG gùn maskim-sù 1 åà-da-um-TÚG-I 1 aktum-TÚG 1 íb+III-TÚG sa6 gùn Sa-a-a åÀ-duki 1 aktum-TÚG 1 íb+IV-TÚG gùn {a-da SA.ZAxki 1 gu-dùl-TÚG 1 sal-TÚG 1 íb+IIITÚG 1 gú-li-lum-I a-gar5 kù-gi 15 (gín) Du-da-zu Ba-sa-ne-gúki tus:lú 1 gu-dùl-TÚG 1 sal-TÚG 1 íb+IIITÚG gùn Ìr-am6-da-ar maskim Is11-gi-bar-zú su-du8 1 åà-da-um-TÚG-I I-ti-dAs-dar asgab 1 íb+III-TÚG sa6 gùn Du-ha-lum su-mu-“tag4” zax en 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+IV-TÚG sa6 gùn Wa-ru12-tum Kiski su-mu-“tag4” uzu 2 åà-da-um-TÚG-I 2 íb+III-TÚG sa6 gùn Ig-su-ub-da-mu Zé-da-mu

A VICTORY OVER MARI AND THE FALL OF EBLA

ix.

5. 2 dumu-nita en 7. 3 níg-lá-gaba En-na-BE 9. lú †a-gú-bù Puzur4-ra-a-gú 11. Íl-e-i-sar [x]-ga 13. 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+IV-TÚG sa6 gùn Is-lum-a-hu 15. Kiski su-mu-“tag4” 17. uzu in 19. Ti-rí-gaki su-ba4-ti 21. 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+IV-TÚG sa6 gùn Sá-ù-um 23. Na-gàr ki su-du8 25. in Ti-rí-gaki 27. 1 TÚG-gùn Ar-miki 1 dùl-TÚG Ib-laki lú é-ti-TÚG 29. 1 aktum-TÚG 1 íb+IV-TÚG sag 1 íb+III-TÚG ú-háb 1. 1 gada-TÚG mu4mu en 3. Na-gàr ki 1 TÚG-gùn 1 åà-da-um-TÚG-II 2 aktum-TÚG 1 íb+IV-TÚG sag 1 íb+IV-TÚG ú-háb 5. en Kiski 7. 2 åà-da-um-TÚG-II 2 aktum-TÚG 2 íb+III-TÚG sa6 gùn a-mu-sù 9. wa ses-sù 11. 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+III-TÚG sag 1 gada-TÚG mu4mu en

x.

37

13. åÀ-duki 2 dùl-TÚG Ma-rí ki 2 íb+II-TÚG sa6 gùn 15. NE-NE Al6-ma 17. su-du8 Ma-rí ki 19. 2 gu-dùl-TÚG 2 sal-TÚG 2 íb+IVTÚG gùn Is11-a-ma-lik 21. I-ti-dAs-dar Gàr-ga-me-suki 23. su-du8 1 sal-TÚG 1 íb+III-TÚG gùn 25. [T ]i-la-ì [(2/3 cases)] 1. Si-ma-dKu-ra maskim 3. I-bí-zi-kir su-ba4-ti 5. 2 íb+II-TÚG sa6 gùn dumu-nita 7. en Ma-nu-wa-atki 9. wa maskim 11. Ìr-am6-ma-lik lú A-mi-du 13. 1 íb+II-TÚG sa6 gùn I-in-zé 15. lú Na-mu-lum 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+III-TÚG sa6 gùn 17. I-da-ì ses:pa4 19. dAs-dar in 21. {a-a-bí-duki su-ba4-ti 23. [x?]+2 åà-da-um-TÚG-II [x] aktum-TÚG [x] íb+II-TÚG sa6 gùn Bù-su-sum 25. ses en 27. Kiski

38

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

Text 2 TM.75.G.2250 obv. viii 4–rev. i 5 viii. 4. 1 aktum-TÚG 1 íb+IV-TÚG sa6 gùn 5. ì-gis-sag Is-ma-da-ba-an 7. in {a-a-bí-du 9. su-ba4-ti 1 åà-da-um-TÚG-II 1 aktum-TÚG 2 ib-V-TÚG [. . .] [c. 3–4 cases] 15. [. . . x] íb+4–TÚG sa6 gùn lú-kar 17. Ma-rí ki [s]u-du8 19. in Tar5-gaki 21. 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+III-TÚG sa6 I-ti-dÌ-da-KUL 23. [ ]-ªDUº? Ki-[. . .] 25. su-du8 Tar5-gaki ix. 1. 10 lá-3 sal-TÚG 10 lá-3 íb+IIITÚG gùn Ar-si 3. A-NE-za-mu Ru12-zi 5. Du-bù-hu-ì A-NE-za-mu-II 7. Ar-sa-du Puzur4-rí 9. Nu-KA su-du8 11. in Ter5-gaki 13. in Ba-ne-zuki 15. [su-ba4-]ti [2 åà]-da-um-TÚG-I 2 íb+II-TÚG gùn 1 gú-li-lum-II a-gar5-gar5 kù-gi 17. Bù-da-ì

One Line Short

19. 21. 23. 25. 27. 29. x.

1. 3. 5.

7. 9.

11. 13. 15. 17. 19. 21. 23. 25. 27. xi.

1. 3.

Ni-ti-baki Ìr-kab-ar Ga-ra-ma-anki 1 aktum-TÚG 1 íb+IV-TÚG gùn Zu-rí-zé Ga-ra-ma-anki maskim åÀ-gu-lum su-du8 in Ter5-gaki 1 gada mu4mu åÀ-da-sa lú Zu-mu-na 1 níg-lá-gaba En-na-BE ugula surx-BAR.AN I-bí-zi-kir 1 gu-dùl-TÚG 1 sal-TÚG 1 íb+IIITÚG gùn engar-kinda A-suki 1 åà-da-um-TÚG-I 1 sal-TÚG 1 íb+III-TÚG gùn åÀ-tum Kiski in Ter5-gaki su-ba4-ti 1 sal-TÚG 1 íb+3–TÚG gùn Ar-si-a-hu Gàr-ga-mi-suki su-du8 in Ma-rí ki 1 sal-TÚG 1 íb+III-TÚG gùn Ar-si-a-hu lú-kar Ma-rí ki su-du8 in Ter5-gaki 1 åà-da-um-TÚG-I 1 sal-TÚG 1 íb+3–TÚG gùn I-da-ì Kiski

A VICTORY OVER MARI AND THE FALL OF EBLA

rev. i.

su-du8 5. in Ter5-gaki 7. al6-tus Ìr-KUki 9. 10 lá 2 sal-TÚG 10 lá-II íb+3– TÚG gùn mu4mu 11. lú-kar Ma-rí ki 13. su-ti nu-mu4mu 15. 2 åà-da-um-TÚG-I 2 sal-TÚG 2 íb+IV-TÚG gùn A-gá-a-ba 17. En-na-ì lú-kar 19. Tar5-gaki 1 gu-mug-TÚG 1 sal-TÚG 1 íb+III-TÚG gùn 21. I-ti-nu nu-KA 23. 1 sal-TÚG 1. 1 íb+III-TÚG gùn maskim-sù 3. su-du8 in 5. Tar5-gaki

ii.

Text 3 TM.75.G.2250 rev. i 9–iii 20 i. 9. 1 TÚG-gùn 1 gu-zi-tum-TÚG 2 aktum-TÚG 2 íb+4-TÚG sa6 gùn Is11-ga 11. en A-bù-la-duki 13. 1 TÚG-gùn 1 aktum-TÚG 1 íb+3-TÚG sa6 gùn en 15. Na-gàr ki 1 TÚG-gùn Ar-miki 1 íb+IV-TÚG ú-háb 17. Ni-zi “ur4” Na-gàr ki

iii.

39

19. 1 åà-da-um-TÚG-I 1 aktum-TÚG 1 íb+3–TÚG sa6 gùn Ib-lul-ì 21. dumu-nita en 23. Na-gàr ki 1. 2 sal-TÚG 2 íb+3–TÚG gùn Na-gàr ki 3. UNKEN-ak 1 TÚG-gaba I-bí-zi-kir 5. 2 sal-TÚG 2 íb+3–TÚG gùn Zu-wa-du-mu-da 7. Na-gàr ki su-mu-“tag4” 9. se+TIN I-bí-zi-kir 11. 2 su-ba4-ti 2 åà-da-um-TÚG-I 2 sal-TÚG 2 íb+x-TÚG sa6 13. Na-gàr ki DU.DU 15. si-in Kiski 17. 2 sal-TÚG 2 íb+III-TÚG gùn Ìr-ni-ba 19. Gur-da-LUM lú kas4-kas4 21. DU.DU si-in 23. ªKisºki [x] åà-da-um-TÚG-I 1 sal-TÚG 1 íb+II-TÚG sa6 gùn 25. NI-ba-ì Dar-ábki 27. maskim åÀ-tès 29. su-du8 1. 1 åà-da-um-TÚG-I 1 sal-TÚG 1 íb+II-TÚG sa6 Ù-zu-rí 3. DU.DU ás-da 5. dumu-nita Ba-ga-mi-su 7. 1 sal-TÚG 1 íb+III-TÚG gùn Sá-mu

40

ALFONSO ARCHI AND MARIA GIOVANNA BIGA

9. lú-kar Ma-rí ki 11. DU.DU si-in 13. Nu-ba-duki 1 sal-TÚG 1 íb+4–TÚG sa61 íb+III-TÚG gùn 15. Ìr-ì-ba A-rí-muki 17. ªsuº-ba4-ti 1 gu-mug-TÚG 19. DU.DU I-bí-zi-kir

3. 5. 7.

9. 11.

iii.

Text 4 TM.75.G.2278 obv. i. 1. 1 TÚG gùn 1 åà-da-um-TÚG-I 1 åà-da-um-TÚG-II 1 aktum-TÚG 2 íb+II-TÚG sag 1 dib 2 ma-na 3. 1 íb-lá 1 si-ti-tum 1 gír-kun 1 ma-na tar kù-gi 1 gír mar-tu kù-gi 5. 4 eskirix ti8musen kù-gi 1 GI†-gigir-II 2 GI†-GAM.GAM kù-gi ra-åà-tum 7. lú en níg-sa10 9. I-bí-zi-kir ii. 1. lú nídba ì-gis

13. 1.

3.

5. 7. 9. 11.

iv.

13. 1.

Ma-rí ki 1 åà-da-um-TÚG-I A-mu-rúm dumu-nita lugal Ma-rí ki 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb-TÚG sa6 gùn Mi-ga-ì lú In-ma-lik nagar 1 åà-da-um-TÚG 1 aktum-TÚG 1 íb+III-TÚG sa6 gùn 1 dib sa-pi gibil A-sè-lum lú En-na-ni-il 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+III-TÚG sa6 gùn En-na-ma-lik 1 åà-da-um-TÚG-II 1 aktum-TÚG 1 íb+III-TÚG sa6 gùn Ar-bar-zú NI-ra-ar ki 1 åà-da-um-TÚG-I 1 aktum-TÚG A-mu-rúm dumu-nita lugal Ma-rí ki 1 aktum-TÚG A-mu-ru12 dumu-nita I-bí-zi-kir

Texts 5 and 6 TM.75.G.2426 obv. v i-viii 16 v. 1. [12] ma-na kù-gi-II 1/2 12 dib 1 ma-na 3. en NI-ra-ar ki

5. en Ra-åà-ak

TM.75.G.2335 obv. ii 6–viii 7 obv. ii.

6. (1+1+1 garments) 1 dib 1 ma-na 7. en NI-ra-ar ki 9. (6+6+6 garments) 1 dib 1 ma-na ábba-sù 11. (1+1+1 garments) 1 dib 1 ma-na en 13. Ra-åà-akki (2+2+2 garments) 15. ábba-sù

A VICTORY OVER MARI AND THE FALL OF EBLA

7. en Bur-ma-anki

3. 5.

9. en Du-ubki

7. 9. 11.

11. en Ì-mar ki

13. 15. 17.

13. en iii. Gàr-muki iv

15. en Lum-na-anki

17. en I-bu-bu16ki

19. en Ur-sá-umki

21. en Ù-ti-ikki v.

23. en

41

(1+1+1 garments) 1 dib 1 ma-na en Bur-ma-anki (2+2+2 garments) ábba-sù (1+1+1 garments) 1 dib 1 ma-na en Du-ubki (3+3+3 garments) ábba-sù (1+1+1 garments) 1 dib 1 ma-na en Ì-mar ki (3+3+3 garments) ábba-sù (1+1+1 garments) 1 dib 1 ma-na en

1. 19. Gàr-muki (2 garments) 1. (2+2 garments) ábba-sù 3. (1+1+1 garments) 1 dib 1 ma-na en 5. Lum-na-anki (2+2+2 garments) 7. ábba-sù (1+1+1 garments) 1 dib 1 ma-na 9. en I-bu-bu16ki 11. (2+2+2 garments) ábba-sù 13. (1+1+1 garments) 1 dib 1 ma-na en 15. Ur-sá-umki (3+3+3 garments) 1 dib 50 (gín) 17. ábba-sù (1+1+1 garments) 1 dib 1 ma-na 19. en Ù-ti-ikki 21. (2 garments) 1. (2+2 garments) ábba-sù 3. (1+1+1 garments) 1 dib 1 ma-na en

42

ALFONSO ARCHI AND MARIA GIOVANNA BIGA Kak-mi-umki (cfr. below, vi 23–24)

5. Kak-mi-umki (1+1+1 garments) 1 dib 1 ma-na 7. ábba-sù (1+1+1 garments) 1 dib 1 ma-na 9. en Ir-i-tumki 11. (2+2+2 garments) 1 dib 50 (gín) ábba-sù

25. en Ir-i-tumki

vi.

27. 10 ma-na sú+sa 5 babbar:kù [su-bal-ag] 1. 4 ma-na 10 kù-gi-II 1/2 5 dib 50 3. ba-da-lum {a-ra-anki

5. ba-da-lum Sa-nap-zu-gúmki vi. 7. ba-da-lum Ur-sá-umki 9. En-na-ma-lik NI-ra-ar ki 11. ba-da-lum Gú-da-da-LUMki

13. 6 ma-na sa-pi babbar:kù su-bal-ag 15. 2 ma-na sa-pi kù-gi-II-1/2 4 dib sa-pi 17. Du-bí-zu-i-nu Kiski 19. En-na-ni-il åÀ-duki 21. Sá-ù-mu Na-gàr ki 23. Íl-ba-zi-kir Kak-mi-umki 25. 3 ma-na sa-pi 5 babbar6:kù vii. 1. su-bal-ag 1 ma-na tar kù-gi-II-1/2

13. (1+1+1 garments) 1 dib 50 (gín) ba-da-lum 15. {a-ra-anki (2+2+2 garments) 17. ábba-sù (1+1+1 garments) 19. 1 dib 50 (gín) ba-da-lum 21. Sa-nap-zu-gúmki (2 garments) 1. (2+2 garments) ábba-sù (cfr. above, iv 16–17) (cfr. above, ii 9–10) 3. (1+1+1 garments) 1 dib 50 (gín) ba-da-lum 5. Gú-da-da-LUMki (2+2+2 garments) 7. ábba-sù

vi. 19. (3+3 garments) vii. 1. (3 garments) 3 dib sa-pi 2. Na-gàr ki 3. åÀ-duki 4. Kiski (cfr. above, v 6–7)

A VICTORY OVER MARI AND THE FALL OF EBLA

3. 3 dib tar Ne-si 5. maskim I-bí-zi-kir

7. Du-du-wa-su Ar-miki 9. Ru12-zi-ma-lik ugula Ar-ha-duki 11. 2 ma-na tar babbar:kù su-bal-ag 13. 1 m[a-na] kù-gi-II-1/2 3 dib sú+sa (gín) 15. en {U-la-za-gur ki

17. I-da-ì ses:pa4 d 19. Ba-as-dar {a-a-bí-duki

21. A-zi Kab-lu5-ulki

23. 4 ma-na 50 babbar:kù su-bal-ag 25. 1 ma-na 56 kù-gi-II-1/2

viii.

1 dib sú+sa (gín) 1. 6 dib 16 (gín) sal ábba-ábba

43

vii. 16. (1+1+1 garments) 1 dib 30 17. Ne-si maskim 19. I-bí-zi-kir níg-mul 21. lug[al] Ma-rí ki 23. su-du8107 in viii. 1. Gú-NE-sumki rev. i. 27. (1+1 garments) ii. 1. (1 cloth) 1 dib 30 (gín) Du-du-wa-su 3. Ar-miki obv. vi. 8. (1+1+1 garments) 1 dib 30 (gín) 9. ugula Ar-ha-duki

obv. vi.

14. (1+1+1 garments) 1 dib sú+sa (gín) sal 15. en RI-la-za-gur ki obv. vii. 11. (1+1+1 garments) 1 dib sú+sa (gín) I-da-ì 13. ses:pa4 d As-dar 15. {a-a-bí-duki obv. viii. 2. (1+1+1 garments) 1 dib sú+sa (gín) 3. A-zi Kab-lu5-ulki 5. DU.DU si-in 7. Kiski

obv. vii.

5. (7+7+7 garments) 1 dib sú+sa (gín) 6 dib 16 (gín)

ábba-ábba

107. This passage is di¯cult. The usual construction is: subj. – su-du8 – obj. “x has taken y,” see, e.g. above TM.75.G.2277 rev. vi 20, vii 17, viii 24. A translation such as: “(garments to) PN, the representative of PN, who brought the news that the king of Mari has been captured in GN” is not supported by other documentation.

44

ALFONSO ARCHI AND MARIA GIOVANNA BIGA 3. Kiski 2 ma-na tar babbar:kù 5. su-bal-ag 1 ma-na kù-gi-II-1/2 7. 6 dib 10 (gín) sal nar 9. Kiski lú ì-ti 11. in níg-kas4 13. Ma-rí ki in 15. SA.ZAxki su-ba4-ti

7. Kiski

6 dib 10 (gín) 9. nar Kiski

A LU E SCHOOL TABLET FROM THE SERVICE QUARTER OF THE ROYAL PALACE AP AT URKESH Giorgio Buccellati University of California, Los Angeles

The tablet A1j1 (˜g. 1) was found in July 19921 in the ˘oor accumulation labeled A1f113, in what is now labeled room B2 (˜g. 2) of the royal palace AP at Urkesh, modern Tell Mozan. It was broken in antiquity, and its three fragments were found a few centimeters apart from each other. The ˜rst and larger fragment received the label A1.69, and the other two, which were found still joined together, received the label A1.72. Even though found close to each other, the appearance of the fragments is quite diˆerent: A1.69, the right half, is blackened by being in contact with ash, while A1.72 (consisting of the two fragments that compose the left half) is not. Joined together, the fragments yield a practically complete text, with 5 lines on the obverse and one on the reverse. The text is an excerpt from the Early Dynastic LU E professions list, of which the full text has reconstructed from sources found at Abu Salabikh,2 Ebla,3 Gasur,4 and Kish.5 The fact that our school tablet was broken in antiquity, but its pieces were preserved in the

same immediate context, suggests the presence of apprentice scribes active within the storehouse. It seems likely that objects arriving in the storehouse, which were stored in sector B, were registered in the part of the building to the South (Sectors A and C, ˜g. 2). While sector A is badly eroded and little is left besides the outline of the walls, sector C is well preserved, and may give evidence of a scribal installation.6 Another complete small tablet, an inscribed docket, and more than forty fragments of tablets, have been found within the building, and also just outside it to the West. One of the reasons why this ˜nd holds special interest is that the texts from Urkesh are the northernmost strati˜ed cuneiform material in the

2. MSL 12 1.5, pp. 16–21 (the portion corresponding to our text is found on p. 17, ll. 34–39); R. D. Biggs, Inscriptions from Tell Ab¿ Salabÿkh. OIP 99 (Chicago: The Oriental Institute, 1974), nn. 54–60 (the portion corresponding to our text is found on pl. 31 ii 34–39; pl. 33 ii 34). 3. MEE 3, 27–46 (the portion corresponding to our text is found in three out of six exemplars, shown synoptically on pp. 36–37, lines 34–39). 4. HSS 10 222; see MSL 12 1.5 (pp. 16–21). The portion corresponding to our text is not preserved. 5. MAD 5 N. 35, p. 31, 133, and p1. XIII. For the attribution (by Wilcke) to the LU E list see Biggs, Inscriptions, 82. The portion corresponding to our text is not preserved. 6. See G. Buccellati and M. Kelly-Buccellati, “The Royal Palace of Urkesh. Report on the 12th Season at Tell Mozan/ Urkesh: Excavations in Area AA, June–October 1999,” MDOG 132 (2000) 143–45.

1. The text was presented in 1993 at the national meeting of the American Oriental Society in Chapel Hill. It was brie˘y described in G. Buccellati and M. Kelly-Buccellati, “Mozan, Tall,” RlA 8 (1995) 391 and in G. Buccellati and M. Kelly-Buccellati, “Urkesh, the First Hurrian Capital,” Biblical Archaeologist 60 (1997) 94, with a photograph (upside down). A large photograph of the left half portion of the tablet was published in the New York Times of November 21, 1995, p. C1. A manuscript with the edition of the text submitted for publication in June 1996 was not published.

45

JCS 55 (2003)

46

GIORGIO BUCCELLATI

Obverse

Left Edge

Reverse

Fig. 1 third millennium, and thus the discovery of this school tablet utilizing a canonical lexical list speaks to the geographical spread of southern Mesopotamian culture. It is not so much the precise location in terms of latitude that matters (Urkesh is only some sixty kilometers north of Nagar, modern Tell Brak), but rather the fact that, in my view, Urkesh belongs to a diˆerent cultural horizon, more closely linked to the north than the rest of

the Khabur plains, and is ethnically identi˜able as Hurrian.7 7. A case in support of a speci˜cally Hurrian identity for Urkesh, in contrast with Nagar, has been made in my article “Urkesh and the Question of Early Hurrian Urbanism,” in Urbanization and Land Ownership in the Ancient Near East, eds. M. Hudson and B. A. Levine, Peabody Museum Bulletin 7 (Cambridge: Peabody Museum of Archaeology and Ethnography/Harvard University, 1999), 229–50. Recent discoveries in

A LU E SCHOOL TABLET FROM THE ROYAL PALACE AP AT URKESH

47

Fig. 2 Following the discovery of impressions of the seal of Tar’am-Agade,8 a daughter of Naram-Sin,

the area immediately to the south of the royal palace AP have brought out even more dramatically the nature of Urkesh Hurrian culture, see M. Kelly-Buccellati, “A Hurrian Passage to the Netherworld,” MDOG 134 (2002) 131–48. 8. See G. Buccellati and M. Kelly-Buccellati, MDOG 132, 139–40 and “Überlegungen zur funktionallen und historischen Bestimmung des Königspalastes in Urkes. Bericht überdie 13. Kampagne in Tall Mozan/Urkes: Ausgraben im Gebiet AA, Juni–August 2000,” MDOG 133 (2001) 71–76, as well as “Tar’am-Agade, Daughter of Naram-Sin, at Urkesh,” in Of Pots and Plans: Papers on the Archaeology and History of Mesopotamia and Syria presented to David Oates in Honour of his 75th Birthday, eds. L. Al-Gailani Werr, J. Curtis, H. Martin,

our dating of king Tupkish to the Akkadian period, and speci˜cally to early Naram-Sin or possibly even slightly earlier, has been con˜rmed. The accumulation A1f113, in which our tablet A1j1 was found, is the earliest one within the palace built by Tupkish, and the nature of its emplacement makes it clear that its period of use was fully contemporary with the reign of Tupkish. The majority of the seal impressions of Tupkish himself, his wife Uqnitum and their courtiers, come from the same or equivalent accumulations throughout the service wing of the Palace. The A. McMahon, J. Oates and J. Reade (London: Nabu Publications, 2002), 11–31.

48

GIORGIO BUCCELLATI

Abu Salabikh 6 ad-kid 7 dím-kid! 8 dím-†UB9 9 dím-ban 10 dím-garig(ZUM)10 11 dím-I

Ebla 34 ad-kid 35 dím-kid 36 dím-†UB-†UB 37 dím-gisban 38 dím-gisgarig 39 dím-I-I

Abu Salabikh 6 7 8 9 10 11

Ebla 34 35 36 37 38 39

text from Abu Salabikh dates to ED III, those from Ebla to late ED III11 and from Gasur to Old Akkadian, while for the Kish text there is an outside possibility that it may date to Ur III.129 10 11 12 A transliteration of the text,13 arranged synoptically with the pertinent portions of the Abu

9. See B. Alster, RA 85 (1991) 6–8. 10. I owe the reading garíg, “comb” (i.e., later ga-rig or ga garig) to P. Steinkeller (personal communication, for which I am most grateful). 11. See recently M. C. Astour, “A Reconstruction of the History of Ebla (Part 2),” Eblaitica; Essays on the Ebla Archives and Eblaite Language, 4, (2002), 73–77. 12. Gelb, MAD 5, p. 31: “Perhaps Ur III?”; p. xvi: “(this and another tablet) give the appearance of tablets of the Ur IIII period.” But note that every other tablet published in this volume is Sargonic or Pre-Sargonic.

Urkesh 1 ad-kid 2 dím-kid 3 dím-gis†UB-†UB 4 dím-gisban 5 dím-gisgarig!? 6 dím-gisI-I!?

“worker in reeds” “maker of reed mat(s)” “maker of throwstick(s)” “maker of bow(s)” “maker of comb(s)” “maker of ?”

Urkesh 1 2 3 4 5 6

Salabikh and Ebla texts, shows interesting minor variations between the redactions (highlighted graphically in the second table below—the gray representing identity and the hatching similarity). The ˜rst two lines are identical in all three texts, and lines 4–5 are identical in Ebla and Urkesh. A further correlation between the Ebla and the Urkesh text is in the doubling of the last sign in line 3. The Urkesh text extends the use of the determinative to two more lines than the Ebla text (lines 3 and 6). This seems to suggest that the Urkesh text is closer to Ebla than to Abu Salabikh.13 13. On the reverse, there are a number of wedges, and two signs, which go in diˆerent directions, mostly opposite to that of the text. They must re˘ect an exercise in using the stylus.

OF BOWS AND ARROWS Miguel Civil The Oriental Institute, Chicago For Jacob Klein

01

For a society often involved in war—about thirty percent of the year dates of Ur III, for instance, commemorate military events—it is surprising that the huge administrative archives of the last three centuries of the third millennium B.C. contain almost nothing about army organization and supplies. It is a pleasure to offer this edition of an unusual Sargonic tablet, with a list of bows, arrows and quivers, to Jacob Klein who has worked with so much enthusiasm on the reconstruction of Sumerian texts, above all on the hymns of †ulgi, the king who loved bows and arrows so much.1 The tablet A 2736, located in the collections of the Oriental Institute, but separated from the bulk of the rest of the Adab texts, escaped the attention of the editor of the Adab Sargonic archives.2 However, its paleographic characteristics, the mention of the governor of Adab and other onomastic parallels, leave no doubt that the tablet comes from the University of Chicago excavations at Bismaya (1904–1905);3 this fact is implied by the Oriental Institute card catalog, without however giving an accession date.

05

10

15

20

25

1. My colleagues Ben Foster and Manuel Molina were kind enough to read the manuscript. Unfortunately, the prosopographic information provided by the latter could not be incorporated on time. 2. Yang Zhi, Sargonic Inscriptions from Adab (The Institute for the History of Ancient Civilizations: Changchun, 1989), henceforth SIA. 3. On the Adab excavations, see Yang Zhi, JAC 3 (1988) 1–21.

30

49

4 gisba-n[a] ª1 é-marº-[uru5] 60 kak ba-na su-[ . . . ]-li 2 gi[sba]-na [x é-mar-uru5] 22 kak LUGAL-ra NI-ªxº-[(x)] 1 gi[s]ba-na 30 kak énsi [a]dabki 3 gisba-na [1]5? kak †E†-†E† 2 gisba-na 3 é-mar-uru5 11 kak ur-zu sagina 2 gisba-na 3 é-mar-uru5 60 kak gi-nu-mu-si 2 gisba-na 1 é-mar-uru5 sum-dma-ma 1 gisba-na 1 é-[m]ar-uru5 30 kak lugal-sag9 dumu mes-zi 40 kak HI-p[ù] 60 kak da-da sabra 4 gisba-na ur-dsára sabra 1 gisba-na 60 kak d utu-sag9 sagina [x] gisba-na ur-zu lú-gisgigir [su-nígin] 10x2+1+[2] gisba-na [su-nígin x] ªéº-mar-uru5 su-nígin 60xª6º+[10x4]+3 kak ba-na más-da-ri-[a]

JCS 55 (2003)

50

MIGUEL CIVIL

This is a late Sargonic tablet with a simple inventory-style list of a number of bows ( gisbana), quivers (é-mar-uru5) and arrows (kak-ba-na, or simply kak) with the names of the donors (the governor and other dignitaries), the totals, and the subscript más-da-ri-a indicating the purpose of the listed materials. W. Sallaberger has shown that más-da-ri-a is not a religious offering but something presented to the king and/or his family.4 In the present case, the occasion (possibly a royal visit or a contribution to a military campaign) and the ultimate destination of these weapons are subject to speculation. In tabular form, the contents of the tablet are as follows: Individuals †u-[ . . . ]-li LUGAL-ra-NI-x governor †E†-†E† Ur-zu, general Gi-nu-mu-si †um-dMa-ma Lugal-sag9 HI-pu Da-da, supervisor Ur-d†ára, supervisor d Utu-sag9, general Ur-zu, charioteer Totals

Bows 4 2 1 3 2 2 2 1 0 0 4 1 [1] 23

Quivers 1 [x] 0 0 3 1 0 1 0 0 0 0 0 6+

Arrows 60 22 30 15 11 60 15 30 40 60 0 60 0 403

Although the persons in the list do not really belong to the main archive edited in SIA, a rather restricted and well-delimited archive, some names are nevertheless found in other documents from Adab, or at least in contemporary documents from elsewhere.5 Restore perhaps †u-[me]-li (Nik 2 14 ii 2, but written with -lí), or †u-[TE]-li in line 3. LUGAL-ra (line 6) is a well-known personality (B. Kienast and K. Volk, FAOS 19 [1995], 88); it is unclear if this is the name here, and the rest of 4. W. Sallaberger, Der kultische Kalender der Ur III – Zeit (Berlin: Walter de Gruyter, 1993), 160–70. See also G. Selz, ASJ 17 (1995) 251–74. 5. The references are from the texts in Yang Zhi, SIA, which are the ones preserved in the Oriental Institute; a large number of Adab texts in the Istanbul Archeological Museum remains unpublished.

the line is an o¯ce designation, or if the name is something else, *Lugal-ra-ì-ªsag9º, for instance. The scribe seems to use non-standard writings for some personal names; thus the peculiar Ginu-mu-si (line 16) would seem to be a pronunciation writing of the relatively unusual PN Ki-namu-se (e.g., A 2889:6 [Ur III Drehem]); compare Ki-na-sè in F. Pomponio, La prosopogra˜a dei testi presargonici di Fara (Rome: Università degli Studi “La Sapienza,” 1987) 136. Lugal-sag9 is said to be son of Mez-zi in line 21; this could be an alternative writing for Me-zi, an important functionary in the Adab epistolary (SIA p. 438 s.v.). †es-ses is mentioned a couple of times (SIA 640 iii 3; 660 ii 2), but is a relatively common name, at least outside of Adab. In the archive, the sabras, like the énsi, usually are not named but designated by their o¯ce, so it is not surprising that the names in lines 23–24 do not appear on the tablets. As for the “generals” they may have been in some sense outsiders. The names of the weapons require a few comments, necessarily very brief. The name of the bow is normally written in Sumerian texts6 with the logogram PAN which has the readings: pa-an pa-na

Antagal C:247. Aa 8/4:083; Sb 1:288;7 Diri Nippur 197; CBS 11319+ iv 26; in kak pa-na RTC 221 iv 1, 222 iv 11.

6. One can ˜nd occasional writings, based on a scribal bilingual wordplay, with (GI†).TIR (Akk. qistu “forest”) instead of (GI†).PAN (Akk. qastu), lexically in Diri Nippur 196, and in context in gistir hé-e-lá = q[á-as-t]am i-si A 29975 = 2N-T343:6, and especially in Gudea Cyl. B xiv 4: gispana tir mes-gin7, with probably an underlying (Akkadian!?) word play *qastu kïma qisti . . . . For the use of GI†.PANA in Ebla, see later in the text. The equation ge-e GILIM = qastu Aa 3/1:237 refers in all likelihood to “arc” as a shape (perhaps even in architecture, where the arc is otherwise referred to as gistir). For the “bow” as a geometrical ˜gure (GÁN panakki), see E. Robson, Mesopotamian Mathematics 2100–1600 B.C. (Oxford: Oxford UP, 1999), 45–47. I leave aside the question of gisillar = tilpânu = qastu, discussed by B. Groneberg, RA 81 (1987) 115–24; 82 (1988) 71–73 (where the entries lú-ma-an-du-um [before lú-pana] OB Lu D 295 and ma-an-du-um Proto-Lu 106–108, and literary references, are overlooked); for illar as a geometrical ˜gure, see Robson 48. 7. Now preserved in BM 64370, joined to source E of Sb.

OF BOWS AND ARROWS

The reading with /p/ is the only one attested. The syllabic writing ba-na, that seems to be found only in A 2736 and SET 294,8 cannot be adduced as an argument for /b/ since it must be considered a case of graphic underdifferentiation of stops, the rule in Pre-Ur III Sumerian texts. The history of Mesopotamia archery requires a study, made di¯cult by the complete absence (quite unlike the case of Egypt) of archeological remains, except naturally for the ubiquitous arrow tips and the somewhat imprecise, and infrequent, visual representations in cylinder seals and reliefs.9 Su¯ce it to say that when the bow appears in the texts it has already a long prehistory behind it.10 The oldest lists of weapons include the bow and associated implements:11 (1) AHA 283–294: 283 e-mè 284

e.zi-es

leather implements for battle (armor) a leather piece of armor

8. And in the profession name ba-na-dím (Nik 2 35 iii 1u; 44:11; B. Foster, Or 51 [1982] 325), but pan-dím in EDLu E 37. The sens of “bowyer,” at least in pre-Ur III times is reasonably assured pace Römer, Festschrift Borger 312. But see still gír-pa-na Gudea Cyl. A 18:07. ITT 2 5728: 12 ma-na urudu-anna pa-na is unclear. 9. D. Collon, “Hunting and Shooting,” AnSt 33 (1983) 51– 56. A couple of years ago I saw in the antiquities market a stone plaque with a list of captives and their geographical origins, written with signs paleographically very close to the archaic Uruk forms, on one side, and on the other, a representation, of excellent facture, of two warriors standing, one of them holds a biconvex bow on the right hand, and a bunch of arrows on the left; no traces of a quiver. 10. A good overview in E. McEwen, R. Miller, and C. Bergman, “Early Bow Design and Construction,” Scienti˜c American (June 1991) 76–82; R. Miller, E. McEwen, and C. Bergman, “Experimental Approaches to Ancient Near Eastern Archery,” World Archeology 18 (1986) 178–95. Although centered in a different geographic area, J. D. Clark, J. L. Phillips, and P. S. Staley, “Interpretations of Prehistoric Technology from Ancient Egyptian and Other Sources. Part 1: Ancient Egyptian Bows and Arrows and their Relevance for African Prehistory,” Paléorient 2 (1974) 323–88, provides a wealth of typological information. 11. For these lists and the abbreviations used, see M. Civil, “The Early History of Har-ra: the Ebla Link” in Ebla 1975– 1985, ed. L. Cagni (1987), 131–58. The list AHB is now completed by an unpublished source MS 2340, to which G. Pettinato, L’uomo cominciò a scrivere (Milano: Electa, 1997), no. 2 can be joined. ED literary references to weapons can be found in the Zami-hymns 52–58 (OIP 99, 47).

285 286 287 288 289 290 291 292 293 294 (2) AHB 21:04 21:05 21:06 21:07 21:08 21:09 (3) AHC 170 171 172 173 174 175 176

é-mah si-gud e-su4 gis pan mar-uru5 gis ti im-dug kus NU11.BURmusen gis silig gis tukul

51

? battering ram sling? bow quiver arrow (clay) sling pellet shield war axe mace (and generic term for weapons)

pan ti sita eme-SAL.U†-ga gis silig zi-es (break) pan ti sita gis silig é-ªxº-DU.DU é-SAL é-pan Commentary (mostly on AHA)

283. The word e in ED texts is a rather general term for pieces or strips of leather with different forms and uses (see, e.g., ECTJ 154, 179). The fact that in the following line (and in other passages of these lists) it can be optionally omitted, suggests that it must be considered a determinative or classi˜er. The word disappears in the later lexical tradition, surviving only in the term e6 (see simply CAD E 1a eau) “a strip of leather used as a hinge.” The entry here means “leather implements for combat,” although it may have had a more concrete meaning. 284. = AHB 21:09. Vars. zi-is-sa, and zi-is-ás (faulty); e only in one source (therefore determinative). Survived in Sargonic texts as zi-sa (STT 7:4, 12, 15; ITT 2 5723, 4430, unclear). The word must have designated a leather part of armor or helmet. It has nothing to do with the giskak-sies (= 3) with variant kak-si-2 in N. Veldhuis,

52

MIGUEL CIVIL

Elementary Education at Nippur (Groningen: Rijksuniversiteit Groningen, 1997), 164:559. 285. Written é-máh in AHC 70–71 (before igitab);12 if there is any connection with gisgud-mah = asibu Hh 7A 88 then it must be related to the item in the following line. 286. Can be written also si-gud-dil, and siGUDxDIL. See the discussion by Steinkeller, NABU 1987 no. 14. 287. Written e-su4 in Ebla sources, but apparently du[gud] in Abu Salabikh. 291. Var. im-tak4. The writing im-dug is found elsewhere, e.g., SIA no. 977:1 (600 im-dug); †ulgi B 36 (with var. du-ug); †ulgi D 185 (“sling pellets like pounding stones”). 292. One source has a half-broken sign (looks like TAB) instead of musen. The ˜rst sign is, of course, NU11 (LAK 24) and not †IR (LAK 23), and the same happens with the logogram for the town of Lagas. The term is to be read buru4 (normally “crow”) and corresponds to the later gur(u)21 (E.ÍB.ùr); note that for the bird name there is a variant reading gu-ru etymologically preferable (AfO 25 [1974–1977] 25 and JNES 32 [1973] 60), and that for the shield the initial b- is also attested, see M. Civil JNES 32 (1973) 60 [1.17].13 For the meaning, see B. Eichler JAOS 103 (1983) 95–102.14 I strongly suspect, however, given the evidence of the lexical lists that recognize many types of this device, that guru21 is a generic term for “shield” and includes, but is not restricted to, the siege shield. 293. Var. GI†GALxURUDA for GI†GALxIGI, very common in ED texts (M. Lambert, OA 13 12. See kusgur21 igi-tab ugnim-ma Lipit-Estar A 77. 13. See also the var. kuse-bu-ùr in Lugalbanda II 320 (C. Wilcke, Lugalbanda, 250 ad 320; J. Cooper, AnOr 52, 128). Note, on the other hand, that Diri Nippur 9:40 gives for the logogram the reading ku-ru, VAT 9523 iiu 3uu gives gu-ru, and there is a variant gur7 for it in the letter from Puzur-†ulgi to †ulgi A3:47 (F. Ali, Sumerian Letters, 44 A 3:47). Nik 1 281 writes e-ùr (G. Seltz, FAOS 15/1, 508, with a translation “Schild, Setztartsche”). 14. The discussion of the writing in Seltz, FAOS 15/1, 100, note 38, is not convincing. Note especially that E is a fossilized ancient determinative (see above). It is somewhat doubtful that ùr is a phonological indicator, judging from the Nippur Forerunner to Hh 11, line 90, that writes (correctly?) kus E.ÍB-ùr-ra.

[1977] 2, note 4; P. Steinkeller RA 74 (1980) 6, note 7). The quiver15 is found often in all kinds of texts and does not present di¯culties, except for its name that raises some minor problems due to its quasi-homophony with mar-uru5 “storm.” It has often an initial é-, and note the variant a-ma-ru < é-mar-uru5 in, e.g., Lahar and Asnan 101; UrNamma A 88; etc. For details, see B. Eichler, Hallo AV 90–94, and Römer, Borger Festschrift 307–10. It is uncertain whether kak-ba-na is used here strictly for “arrow tip,” or if it is used metonymically for the whole arrow (ti). Somehow it seems more likely that the offering would consist of complete arrows, the tips mounted on the shaft and with the ˘etching on, rather that arrow tips alone. In the Isin text BIN 9 415, the sense of “arrow” is inescapable: 6 ªgiº gis-gí-da, esír-é-ªa-biº 8 sil . . . , gis kak-pana ba-an-gar “6 lance shafts, their bitumen is 8 silà, were made into arrows.” The term is current in lexical lists (Hh VII A 52 [Akkadian possibly incorrectly restored in MSL 6, 87], N. Veldhuis Elementary Education at Nippur 164:557) and in the literature (†ulgi B 83, 98; UrNamma A 88; etc.), see also RTC 221 iv 1, 222 iv 11. It would seem that kak-ti is simply a lessused, or later, synonym. Another tablet probably from the same Adab archive and with similar contents, but with no personal names and only the remark níg-gál-la “available goods, supplies,” is SET 294, published among texts from the Ur III period.16

15. For the construction of leather quivers, see B. Williams, University of Chicago Oriental Institute Nubian Expedition vol. 9 (1991), 77–117. The quivers discussed therein are of a much later date, but the discussion is nevertheless extremely informative. 16. The editors found the tablet somewhat puzzling and for reasons unknown suggest, tentatively but incorrectly, that it came form Warka. Their edition requires collation, but the tablet, at the time part of a lot of ˜fty-one tablets in the Public Library of St. Paul, MN, is now in unknown hands. This collection was sold years ago in Japan judging from an auction catalogue with photos. One of the obverse of SET 294 (with an asking price of 500,000 yen) is good enough to recognize the Sargonic paleography but insuf˜cient for collation. Undoubtedly, the tablets came originally from Banks (probably sold before 1915) and SET 294 must thus be a “misplaced” piece from the University of Chicago excavations at Bismaya,

OF BOWS AND ARROWS

1 2 3 4 5 6 7 8 9

38 urudakak-si-fisá?fl 39 urudakak zabar 10 urudaza-mi-rí-tum zabar 7 urudaza-mi-rí-tum ZU 1 urudakak-si 47 gisba-na 35 é-mar-uru5 25 gisba-na UM níg-gál-la

Indirect mentions of bows, made of se-du10wood, are found in Pre-Sargonic Girsu (DP 418, 419, TSA 29, VAS 14 57). For the Sargonic period there are a couple of texts from Susa (MDP 14 no. 85–86 = Scheil RT 35 [1913] 26–35) with military supplies, and a group of Girsu texts listing presents(?) to the royal family includes some bows (B. Foster JANES 12 [1980] 29–42), plus some isolated tablets, e.g., STT 7, ITT 2 4430, 5723. In the Ebla texts there are apparent allusions to the bow, e.g., in ARET 6 x 3 (silver for the “†IR”.ZA of a GI†.PANA). 17 The most frequent use of GI†.PANA is, however, as a logogram for qïstu “gift, present.”18 In the Ur III period, the archive

directed by Banks himself. Note that (1) níg-gál-la on the reverse is correctly translated as “stock, supplies” and is not a personal name as incorrectly suggested in H. Limet, Anthroponymie, 504 s.v., (2) the -um in line 9 is perhaps uruda, and (3) I don’t know what to do with the sign transliterated ZU in line 4. 17. It would seem that NU11(!).ZA is an ornament made of lapis lazuli, judging from an Ebla vocabulary, see the author in L. Cagni, ed., Ebla 1975–1985, 143:33. If so, ARET 6 x 3 would refer not to a bow but to a “gift”; see following note. 18. Bows were sent sometimes as gifts in Mesopotamia, e.g., TRU 384 where the governor of Adamsah sends a bow. In Ebla, however, the meaning in most cases is clearly “gift”

53

of DI.KU5–mi-ar19 contains many references to weapons. The Isin leatherworks archive has some references to the construction of bows, to glue (BIN 9 121, 122, 124), and to old leather pouches recycled for bow parts (BIN 9 415). The correspondence of king Lipit-Estar (F. Ali, Sumerian Letters [U. of Pennsylvania Ph.D. Thesis, 1964: 76–79), mentions an army composed of two thousand lancers, one thousand archers, and two thousand soldiers armed with double axes. Despite the scarcity of administrative and literary sources (partly compensated by more complete lexical material) much could be said about Mesopotamian archery. A study on the subject is sorely needed.20

(GI†.PAN = qÿstu). For references, see simply the index of ARET 1 sub GI†.PAN, where it is still translated “arco.” 19. TIM 6 33–43; AUCT 1 437, 696; 2 133, 178, 384; TCL 2 5488, 5565; RA 8 184; CSTJR 372. 20. What is, e.g., the da-NAGAR (see SIA, 179–80) of an arrow (MDP 14 86), made with 5 gín of silver? Or the túgse-er-ti gis pan (UET 3 788), strips of cloth used perhaps to make the ˘etching, or, more reasonably, to bind the tip to the shaft since they weigh from 6.6 to 5.5 gín and thus could contribute to make the fore shaft of the arrow heavier? Is there a special term for the chisel arrow tips found archeologically, and seen in a Djemdet Nasr period stele (UVB 5 pl. 12–13, see especially the detail pl. 13 a)? How long did such tips survive in historical times? The distinction between pan (self-bow, which can be doubly convex!) and illar (composite bow, in all likelihood) still presents some problems. What is the difference between uhhu and siltahu? The activities of DI.KU5–misar, in charge of army supplies, and of very active bowyers, such as Ibni-Adad and Ahansibu, in the Ur III archives need study. There are also anecdotic occasions such as TRU 350 r. 1 where the archer †elha receives two sheep for the ceremonies attendant to his induction as a gala. And so on. [See now Ch. Zuttermann, “The Bow in the Ancient Near East, a Re-evaluation of Archery from the Late Second Millennium to the End of the Achaemenid Empire,” Iranica Antiqua 38 (2003) 119–65.]

54

obv.

MIGUEL CIVIL

rev.

A 2736

SEARCHING FOR AKKADIAN LYRICS: FROM OLD BABYLONIAN TO THE “LIEDERKATALOG” KAR 158 Brigitte Groneberg University of Göttingen

literature is the type of songs called irate, which are attested from the OB period through the last century of the second millennium. A large number are enumerated in KAR 158: these incipits will be translated and provided with commentary.

The earliest pieces of Akkadian literature date from the last centuries of the third millennium.1 It is at least two centuries before further evidence of Akkadian literature is found. While during this period everyday documents were written in the Old Babylonian dialect, poetic literature, in particular lyrical literature, was written in Sumerian, with some notable exceptions. But this small group of early pieces of Akkadian literature has no descendants for centuries, as there is currently little evidence of literary activity immediately after the end of the Old Babylonian period. The evidence picks up in the last centuries of the second millennium, when large numbers of Akkadian songs are attested in literary catalogs, especially in the “Liederkatalog” KAR 158. In this survey I will provide an overview of Akkadian lyrical texts from the Old Babylonian period and attempt to establish possible links between these texts and compositions listed in KAR 158. This “Liederkatalog” oˆers a broad spectrum of classi˜cations of songs; similar categorizations are already attested in Old Babylonian times, when some songs are already classi˜ed with the help of colophons such as subscripts or headings. In this article these classi˜catory principles are investigated and compared. One main link between the older and the younger bodies of

I. Old Babylonian Lyrics and Post-Old Babylonian Catalogs of the Second Millennium The earliest Akkadian lyrical texts in a strictly poetic language are two OB hymns invoking the name of Larsa kings Gungunum (TIM 9 43) and Rÿm-Sîn (YOS 11 24). The fragmentary Gungunum hymn is so badly preserved that little can be said about it besides the fact that it treats a historical theme.2 The second hymn is unique: it consists of praise of Rÿm-Sîn of Larsa as the “sun of his people” and a love-dialog between a man and a woman.3 These rare poems provide some of our ˜rst evidence for Akkadian lyric. Aside from royal inscriptions there are very few pieces of Akkadian literature before early Old Babylonian times. The main evidence for Akkadian literary language is found primarily in incantations, which are already known from the third millennium.4

2. See Hunger and Groneberg (1978). 3. For a recent translation see Foster (1993: 98–99). 4. The earliest incantation from Mari is contemporary with the Ebla archive; see Bonechi and Durand (1992: 151–59). This

1. See Groneberg (2001).

55

JCS 55 (2003)

56

BRIGITTE GRONEBERG

These incantations cannot be treated as poetic literature since they do not bear its characteristic features,5 even if they occasionally use di¯cult metaphorical language.6 The assumption that there were no poetic texts in Akkadian earlier than the kingdom of Larsa and later Old Babylonian times is con˜rmed by other sources, in particular by catalogs that list literary compositions. A comprehensive bibliography of catalogs was published by Krecher (1976– 1980). 7 Based on the material known to him, including the earliest catalogs from the Ur III period,8 Krecher concluded that the early catalogs list only Sumerian texts, with a solitary exception which might refer to an Akkadian incantation.9 Since then new information has appeared. The late Old Babylonian catalog AUAM 73.2402 published by Cohen (1978) cites at least three incipits in the Akkadian language: l. 15: [si?-me?]-e mi-il-kam, l. 16: [(x)-x u]n li-ibbi, l. 17: [(x) x]–i mu-de-e si-tu-lim. We must also keep in mind that in regard to the majority of songs, which in antiquity were

incantation is in Sumerian with a few Akkadian words, which could point to an Akkadian background. This is also likely for quite a number of the Abu Salabikh texts among which is an Akkadian hymn to †amas; see Krebernik (1992: 72–81). MAD V 8, the earliest Old Akkadian incantation (from Kis), is obviously written in the Old Akkadian dialect. 5. See Groneberg (1996: 59–84). 6. In metaphoric poetry the image refers to an object in a one-dimensional way: the metaphors are laudatory and linked to the praised subject/object. In incantations this relationship is often not expressed through words but by means of cultural and ritualistic items that are not explained. If, for example, the uzzu, “rage” of a dog is evoked, it is not clear if a real dog is meant or rather the quality of uzzu in some components of the ritual. 7. For addenda see Wilcke (1987: 85 note 1). This subject has recently been taken up again by Tinney (1996: 17–18), who follows P. Michalowski’s classi˜cations. Catalogs are arranged as follows: (1) lists related to the school curriculum, (2) texts with the same subscripts, (3) listings by content, (4) inventories of tablets by tags. 8. Compare Hallo (1963: 169). 9. The entry is i-nu a-nu den-líl, which on ˜rst sight could refer to the astrological series of the same name, attested, however, only in the ˜rst millennium. Therefore Wilcke (1993: 139) interprets the entry as the beginning of an incantation; see also Finkel (1980).

One Line Long

cited by Sumerian titles like “Inninsagura” und “Ninmesara,” we do not know for certain that these referred only to Sumerian compositions. For we know that at least the song Inninsagura, which is mentioned in l. 1 in the Cohen catalog, had also been translated into the Old Babylonian language.10 In the Cohen catalog the Akkadian incipit l. 17 “. . . m¿de sit¿lim” probably refers to l. 70 of the great Marduk-hymn, which was rather popular in the ˜rst millennium but has its origins in Old Babylonian times (Lambert (1959–60: 57; see our text no. 26).11 The incipit in l. 15, probably to be read simi milkam, brings to mind a younger composition attested in several exemplars from Meskene and in one text from Ugarit.12 If this is correct, that composition, called “A Dialog Between †¿pe-ameli and his ‘Father’,” goes back to Old Babylonian times. Another Old Babylonian tag catalog, published by Shaˆer (1993), lists no fewer than six literary texts in Akkadian as well as two Sumerian compositions. The songs are identi˜ed in line 11 as ama-ér-ra-ku-tim. This term is also found as a subscript to an unpublished lament to Mama of Kis13 that ends with the colophon: 53 (lines) a-mira-ku-tum. If, as I assume, there is a relationship between the catalog and this hymn, then it should belong to seven texts of dullÿ bÿtim, “rituals of the house.” The hymn contains a lament, namely bÿtam suppuham parakka masi, “a dispersed house, a forgotten sanctuary.” It is interesting to note that this Kish lament is part of a ritual procedure, so that for once we seem to have extra-textual hints of the use of this kind of wailing in a particular cultic setting. The picture emerging from Old Babylonian catalogs provides evidence of a small number of lyrical texts in the Akkadian language that were written down already in the ˜rst part of the sec10. See Michalowski (1998), who also cites a post-OldBabylonian exemplar of this hymn! 11. Both are cited in CAD †/3 p. 144b sub sit¿lum. The lines do not have the exact parallel wording, however. 12. See Kämmerer (1998: 9 and 176). The bibliography of Old Babylonian literary texts on 15–24 is unsatisfactory. 13. The indication to im-gíd-da in Shaˆer’s catalog is a reference to the shape of the tablet, that is, a long tablet written in a single column.

SEARCHING FOR AKKADIAN LYRICS

ond millennium. The relatively small number of tablets found with Akkadian literature con˜rms this.14 Even incantations are not numerous when compared to Sumerian ones.15 We know only some thirty lyrical texts in Akkadian.16 These are hymns, laments, dialogs and narrative-lyrical texts,17 which can be listed as follows: Hymns and laments 1) BM 29624+BM 109164 (lament to Mama of Kis) 2) CT 15, 1–2 (hymn to Belet-ili)18 3) Groneberg (1997) pp. 3–54 (hymn to Istar with the description of a ritual) 4) Groneberg (1997) pp. 97–120 (lament to Istar) 5) HS 175 = HS 1880 (hymn to Mama of Kis, unpublished)19 6) JRAS Cent. Suppl. (1924) 63–86 (hymn to Papulegara) 7) MIO 12/2 (1966/67) 52–56 (VAT 17107) (lament to?) 8) OECT 11, 1 (hymn to Martu und Asratu)20 9) PBS 1/1 2 (lament to Anuna)21 10) RA 22 (1925) pp. 169–77 (hymn to Istar with dedication to Ammiditana) 11) RA 86 (1992) pp. 79–83 (hymn to Marduk)22 12) RB 59 (1952) pp.139–50 (lament to the personal god, ilÿ)23 13) UET VI/2, 404 (hymn to Nanaya) (fragmentary) 14. These are forerunners of the extensive ˜rst millennium myths and epics like Atramhasis, Etana, Anzû and Gilgames and some shorter compositions, as well as several legends about Old Akkadian kings. 15. See the bibliographical notes in Cunningham (1997: 149–56). 16. There are a few Sumero-Akkadian bilingual lyrics yet not collected. 17. Published texts are cited according to their edition in the usual manner. For further references see the bibliography at the end of this article. 18. See Römer (1967/68: 12–28). 19. A study by M. Krebernik is in progress. 20. See Gurney (1989: 15–19). 21. See Lambert (1989: 321–36). 22. Even though the text has some mimation in the form of CVC signs I doubt that this tablet goes back to the Old Babylonian times. There are other indications, which will be treated in another place, that suggest a ˜rst millennium date. 23. See Lambert (1987: 187–202).

57

14) VS 10, 213 (hymn to Istar) 15) VS 10, 215 (hymn to Nanaya with dedication to Samsuiluna)24 Dialogs:25 16) Groneberg (1999) (irati-songs) 17) MIO 12/2 (1966/67) 48–51 (VAT 17347) (love-dialog between Abi-esuh and Nanaya?) 18) PRAK I B 472 (+ PRAK II C 3?; love dialog between a man and a woman)26 19) ZA 49 (1950) 151–94 (SI 57) (debate between a man and a woman, called “The Faithful Lover”)27 20) YOS 11, 24 (dialog between “a man and a woman”)28 Lyrical and narrative compositions 21) CT 15, 3–4 (hymn to Adad)29 22) VS 10, 214 + V. Scheil, RA 15 (1918) 169–82 (to Agusaya)30 23) CT 15, 5–6 (to Sîn und Isum)31 The following lyrical texts also probably belong to this period: 24) OrNS 60 (1991) 339–43 (HS 1879) (hymn to Istar) 25) Black (1984) 25–34 (“an elegy”) No. 24 is Late Babylonian, copied according to the subscript from an original from the time of Hammurabi.32 No. 25 is an elegy, dated by Black to the Middle Babylonian period “on paleographic and linguistic grounds.”33 There are also a small number of lyrical texts that became rather popular in the literar y 24. See Von Soden (1938). 25. The classi˜cation of the second text published by Stol (1987: 383–84 and 385–86; CBS 1399), is uncertain (Edubba?). 26. See Westenholz (1987: 415–25 and note 6 to PRAK II C 3). 27. See Groneberg (2002). 28. See for the latest translation Foster (1993: 98–99). 29. See Römer (1967). 30. See Groneberg (1997: 57–93). 31. See Römer (1966: 138–47). 32. See Oelsner and Von Soden (1991). 33. See Black (1984: 25 and note 2).

58

BRIGITTE GRONEBERG

tradition of the ˜rst millennium, with Old Babylonian predecessors: 26) CT 44 21 (hymn to Marduk)34 27) Iraq 42 (1980) pl. 47 (BM 36296) (hymn to †amas)35 28) RA 32 (1935) 179–87; ZA 43 (1936) 30636 (prayer to the Gods of the Night)37 29) JCS 48(1996) 57–60 (CBS 574) (prayer to the Gods of the Night)38 Very few literary texts have survived from the three hundred years or so between the end of Old Babylonian and Middle Babylonian and Middle Assyrian times.39 We have one catalog in “Middle Babylonian script” that mentions the Old Babylonian king Ammisaduqa. It was published by Finkel (1988), and according to him it lists ˜ve Akkadian iratu-songs. I would suggest, however, that the line might be interpreted as: “one irtum of ˜ve (strophes).” Since king Ammisaduqa is mentioned, it certainly belongs to the realm of Old Babylonian literature even if it was probably written down after the Old Babylonian period.40 The next and most elaborate catalog of literary works known to us from the later second millen-

34. See Sommerfeld (1982: 129–34). 35. With many unpublished duplicates identi˜ed by W. G. Lambert. For the Standard Babylonian version see Lambert (1960: 121–38), for an interpretation and masterful translation see Reiner (1985: 68–84). 36. See Horowitz (2000: 195–98) for a copy and a new edition. 37. The texts include a ritual; for the ˜rst-millennium parallels see Mayer (1976: 427–29). 38. See also Stol (1987: 384, 387; CBS 574). 39. One of the earliest late second millennium literary texts is the “Tukulti-Ninurta-Epic,” still only available to me in the unpublished dissertation by Machinist (1978) [the dissertation by Kuk Won Chang (1981) could not be consulted]. Some literature is known from the so-called library of Tiglathpilesar. This assessment may, however, change because quite a number of literary texts from Assur still wait to be classi˜ed and studied. 40. Finkel (1988); compare also Groneberg (1999: 169–95). The western scribal centers, such as the ones in Ugarit, did produce lexical texts, but scribes did not compose any literary catalogs. Catalogs of literature are also not attested in Amarna. This, as well as literature from Bogazköy, should be the subject of another study.

nium is KAR 158.41 This text belongs to a group of tablets from Assur that has been designated as “the library of king Tiglath-pilesar I.”42 It most likely dates from the years 1114–1076 B.C.; this would mean that it would have been compiled four to ˜ve centuries after the collapse of the Old Babylonian state. The catalog, however, seems to enumerate both post-Old Babylonian and Old Babylonian literary texts. One of the cataloging criteria relates to musical characteristics so that the entries are a valuable source for musicologists. The catalog has been studied most recently by Limet, who discusses some formal criteria and oˆers an explanation and discussion of a number of its entries.43 Surprisingly, apart from about sixty Sumerian entries/subscripts of songs,44 this catalog contains about 350 Akkadian items. Since the tablet was found in a temple district, these songs may well have been used in a cultic context.45 At regular intervals in the catalog the incipits/subscripts are counted, summarized and classi˜ed, with a ˜nal summary in the last column so that we obtain an almost complete inventory of all songs listed. The separate section headings are of great interest for several reasons. First, there is the question of formal “types” that seem to govern the ordering of the hymns. Second, we must consider the possibility that the order may re˘ect their use as part of a ritual sequence, and that therefore functional criteria may have been more important than typological ones.46 Clearly the writer sorted a great number of literary works, most of them still not

41. See the remarks to this catalog and the reading of some of its entries in Finkel (1988), Black (1984: 25–29), Matsushima (1987: 164–75); for further literature see Pedersén (1986: 21a). 42. Weidner (1952–53: 197–215). 43. Limet (1996); see especially p. 151 for the arrangement of columns i–iv and v–viii (the numbering of v–viii should be reversed). 44. Col. ii 49 to iii 46+ and col. iv? 33 + a further 26 in the broken parts (iv is almost wholly broken oˆ ). 45. See Pedersen (1985: 31–42) for the reconstructed library, found in a Neo-Assyrian context in the southwest court of the Assur temple in Assur. The author of this catalog, however, seems to be unknown. 46. Ritual is not only here meant in a formal religious context but also in a private, everyday sense.

SEARCHING FOR AKKADIAN LYRICS

available to us, as only a handful can be recognized with certainty. This catalog reveals our limited knowledge of literature from late Old Babylonian times onwards, since we have to assume that there was a large quantity of Akkadian songs from Old Babylonian and Kassite times down to the middle Assyrian period that has not been preserved. Before analyzing the categorization system of KAR 158 it is useful to consider the headings, sub-headings, and subscripts that are known from Old Babylonian lyrical texts. Both can be called “colophons.” Headings Headings are attested in two Old Babylonian hymns to Istar. These are Agusaya A (VS 10, 214, Groneberg [1997: 59]), which has the remarkable heading ú-ta-ar mus, and Louvre Istar Hymn with the heading [sèr?] ta-na-ti dIstar (Groneberg 1997: 3–4). The heading of Agusaya A remains problematic. One might tentatively attempt the translation “it/I reversed the snake” and relate it to the placatory character of the story. The heading of the second hymn apparently refers to its content, but since the end of the text is broken, we do not know if that text also had a subscript or, more likely, ended with a dedication to a Babylonian king. The classi˜catory markers inhu and ser kummi precede the laudatory passage of the Adad-hymn CT 15, 3 i 1–3: [in]hu seme ikribÿ lunaåid/ al.ti seme ikribÿ lunaåid/ ser kummi ana dIM, “song of sighs, the listener of prayers I will praise. Finished: the listener of prayers I will praise. Song of the holy chamber, to Adad.” Subscripts Typological subscripts have been found in several Old Babylonian texts as well as in literary catalogs. (a) A text from the reign of king Ammiditana lists iratu-songs (Groneberg 1999: 177–81). Its subscript reads:

59

4!47 i-ra-a-tum, is-ka-ar e-es-ra-a-mi-{X} su-qú-úr: 4! iratum “of the series: where is my beloved, he is cherished.” The serial subscript corresponds with the ˜rst line of the text itself: [e?]-[es?] ra-a-mi-i su-qú-úr, “where is my beloved, he is cherished” (b) Another classi˜catory marker, which occurs twice, is amirak¿tu. One hymn to Mama/Beletili48 shows the subscript: “53 amirak¿tum,” which corresponds to the number of lines.” A similar subscript is found in the Old Babylonian catalog published by Shaˆer (1993: 209– 10): (l. 9) 2 dub (l. 10) 37 im-gíd-da ama-ér-ra-ku-tim sa dingir-mah, “2 tablets and 37 tablets of the type “amaåirrak¿tum,” adressed to Belet-ili.” (c) The classi˜cation parum can be found in two Old Babylonian songs, once in a colophon and once in a heading. The ˜rst is the parum-song to Istar (Oelsner & Von Soden 1991). It can be recognised as a parum by its colophon: naphar 20 pa-rum sa Istar mu lugal Hammurabi lugal-e, “in total twenty parum of Ishtar; year in which king Hammurabi became king.” The text has twenty lines. This particular parum-song seems to be a Neo-Babylonian copy of an Old Babylonian hymn of praise to Istar in her function as goddess of love; each line ends with the refrain: rÿsatum-ma isdum ana alim, “jubilation is the foundation of the city.”

47. It would be tempting to presume that the ˜ve irate of the Finkel catalog (Finkel 1988) are referred to in this text, especially as the sign 4 is not absolutely certain, although collation by A. Cavigneaux suggests that this is indeed the reading. 48. Still unpublished.

60

BRIGITTE GRONEBERG

The second parum is directed to the god Papulegara (Pinches 1924).49 In the opening line are the following three markers: 1 pa-ru-um a-na dPap-ul-e-gar-ra / 2 sèr ta-niit-tim a-na dPap-ul-e-gar-ra / 3 sèr / dPap-ul-egar-ra Similar classi˜cations of the mentioned songs are repeated on the reverse of the tablet at the end of the last column. (e) In the second lament, which W. G. Lambert published in MIO 12 (1966/67) 52–56 (VAT 17107), we ˜nd in line 23 a classi˜cation that is only partially identi˜able as indicating the text’s length: (22u) [x x x] si-it-tum hu-um-mu-tú? [ ]“?” (23u) 60+40 mu-bi-im im-gíd-da [ ], “100? Lines, excercise tablet of [PN].”50 (f) The Middle Babylonian “ballad” published by Black (1984: 25–34) includes an identi˜cation line in which the lines of the composition are counted and the series is called by its name: 39 [mu.s]id.bu.im? és.gàr ma-ru-um-ma ra-imni im-gíd-da PN, “39 lines; series: marumma raimni; imgida of PN.” This series, which contained 31 songs in the Akkadian manner, is mentioned in KAR 158 i 42– 45:51 [is-ka-ra]-tu 31 za-ma-rumes [is-gar ma-ru(-ma) r]a-im-ni ak-ka-di-ta] II. Patterns of Classi˜cation in KAR 158 The subscripts found before and after Old Babylonian lyrical texts can be compared to some of 49. Only part of this text has been copied by the author. 50. The SB hymn to Marduk, published by Lambert (1959/ 60: 60), lists in line 207 a classi˜catory marker just above the catchline: un-ni-[ni . . .] sa dMarduk: “praye[rs?] to Marduk.” 51. For the identi˜cation of its catchline in the catalog KAR 158 see Black (1984: 25 note 6).

the classi˜cation principles of KAR 158. But before a comparison can be made, some principles need to be elucidated. These seem to re˘ect two major ordering criteria: (1) according to musical modes; (2) according to geographical, cultural, or “ethnic” identity. These classi˜cation patterns crisscross and cannot be separated from one another. (1) Classi˜cation by Musical Category One classi˜catory principle relates to musical characteristics such as the mode or the type of instrument that was associated with particular songs. Some of these musical terms can be found in viii 45u to 51u: gaba.mes sa esirte, sa kitme etc. All these songs are in Akkadian, which contrasts with the songs in iii 9u and iii 17u that are provided with the description tigû sumerû.52 Two other musical classi˜catory markers are indicated by the expression 13 sit-ru sa eb-bu-be akkadi ki (viii 14u) and with kimin (= sit-ru) sa pi-i-te akkadi ki (viii 15u).53 Several classi˜catory markers in the catalog use the term zamaru, indicating the style of delivery or the manner of recitation. All the initial lines of songs in the ˜rst and second columns of which we also know the title of the series are marked with this term.54 In the Akkadian series maruma

52. Krispijn (1990: 3), who was the last to comment on the term tigû de˜nes it as a stringed instrument. He argues that it has a dal “rod,” which he identi˜es the neck of a lute. I am not completely convinced by this argument; at least the Akkadian tallu points to various kinds of rods and bars, so that it also could refer, for example, to a drum stick. Krispijn assumes that tigi-music is of a merry nature, which is often connected with royal hymns and which was performed by kalû-priests. 53. In rev. col. v the opening words of songs of this sitru sa ebb¿be have been preserved: they are in Akkadian. The word ebb¿be could refer to a song in the “D mode” or it could be related to the accompaniment of the ebb¿bu-˘ute in that musical code. The ebb¿bu is a long tube that, when blown upon, produced a sound. Krispijn (1990:15–16) is of the opinion that it can be of plant material, but it can also be made of metal. The sa pÿte-songs have not been preserved as opening words of songs. Since we do not know of an instrument called pÿtu, the expression might well refer to a musical key, in this case the key of the G-string. For sitru see the remarks of Limet (1996: 155 note 11). 54. On the obverse there are the zamar¿: “songs,” which in turn are subdivided in iskarate, so that several zamar¿:

SEARCHING FOR AKKADIAN LYRICS

raåim-ni, murtami and rêåÿ: rêåÿ: these zamaru are further determined as akkadÿta and as istar¿ta. The Sumerian tigû songs are also designated as zamaru. These are followed by the za-ma-ru ada-pumes 55 of which it is not clear whether they are in the Sumerian or Akkadian manner. In the expression za-ma-ru a-da-pumes the word adapu apparently refers to the accompanying musical instrument of the same name.56 It is interesting to note that singers called “adab” are already attested in Presargonic times as noted by Gelb.57 Whether or not there is any connection between that Akkadian a-da-pu and the Sumerian song type adab still needs to be determined.58 (2) Classi˜cation according to geographic, cultural, or “ethnic” identity 59 One of the main concerns of the author is the division of the works according to language or stylistic manner; this is particularly clear in the summary of the ˜nal column where entries are marked as Sumerian or Akkadian. Akkadian is often indicated by the plural masculine akkadû, but we also ˜nd the forms akkadi ki and even lú akkadû. The latter expression could be taken as going beyond an indication of the language of the

text by adding a technical (“invented by Akkadian people”) or local classi˜cation marker (“in the region where Akkadian people live”). The terms “Sumerian” and “Akkadian” iii 8u–9u (10+?) [zamar¿:] ti-gi-i su-me-ra [am-nu] iii 16u–17u 15 zamar¿ te-ge-e su-me-ra [am-nu] iii 28u–29u 23 zamar¿ te-ge-e su-me-ra am-nu (see also l. 31u) iii 37u–38u il-ta-at is-gar a-da-pa su-me-ra am-nu ii 48 su-nígin 33 za-ma-ru ak-ka-du-ú iii 31u su-nígin 23 te-gu-ú su-me-ru rev. vi 4u su-nígin 8 su-me-ru 3 ak-ka-du-ú see further in column viii: viii 14u viii 15u viii 16u viii 17u–19u viii 21u–23u viii 24u viii 25u–27u viii 28u

“songs” belong to an iskaru: “section.” A section can contain an unde˜ned number of songs, hitherto up to nine (iii 19u– 27u). When a series is mentioned that has been subsumed under one heading, the author uses the word is-gar, as in ii 46. The series a-da-pà of column iii 37u is futher identi˜ed as il-ta-at is-gàr . . . , “˜rst series . . .” and therefore there must have been other series of “adapa.” 55. See viii 11u. 56. For this type of song see Wilcke (1974: 262). According to Krispijn (1990: 3–4), adapu is a stringed instrument. I doubt his identi˜cation because of the equation urudu.a-da-pà = ada-pu = ma-zu-ú. Since mazû/manzû is a roaring drum, the “lion drum,” the instrument could well be a large drum. 57. See Gelb (1975). 58. So CAD A/1 p. 102 with both of them under one heading. D. Shehata, who is preparing a study on Old Babylonian music, will discuss these terms in detail. 59. There seeems to be no way to analyze precisely what the classi˜cations “Akkadian” and “Sumerian” mean. Sometimes this clearly refers to the language, but at other times it could refer to speci˜c Akkadian or Sumerian items such as a certain way of life, cultic setting or style of the hymns.

61

[1]3 sitru sa ebb¿be akkadi ki kimin (= sitru) sa pÿte akkadi ki 5 pa-a-ru ak-ka-du-ú 1 su-me-ru 10 ak-ka-du-ú su-nígin 11 za-mar se-e-ri 9 su-me-rumes, 1 ak-ka-du-ú mes su-nígin 10 za-mar dnin-gis-zi-da 12 za-mar lú ak-ka-du-ú 8 su-me-rumes 3 ak-ka-du-ú mes su-nígin 11 sìr-ku-gu-ú mes 5 kír-re-tu ak-ka-du-ú (!)60 (in rev. vi 11 akkadû is missing)

In viii 29u–34u the following songs (and in one case the shape of a tablet) are classi˜ed as akkadû: gangittu, n¿ru, elÿlu kurrûtu, inhu, b¿ru and dgusatu and probably also si-sa-a-tu and riip-qu but in viii 35u–39u arahhu, sutanÿdû and siqa-tu are classi˜ed as sumerû. In column viii 40u–42u is quoted: 2 su-me-rumes 3 ak-ka-du-ú su-nígin 5 qu-urdu” and in viii 43u–45u: 1 kar-su-ú akkadi ki 4 me-er-ru su-me-ru 23 gaba.mes sa e-sèr-te akkadi ki

60. As kirretu is di¯cult to analyze, the construction kirretu akkadû in l. viii 28u is unclear.

62

BRIGITTE GRONEBERG

The di¯cult terms akkadÿtu and istar¿tu appear in the ˜rst two columns of the catalog. These can be related to one or more categories.61 Akkadÿtu may refer not simply to language, but designate a song’s origin, or perhaps even its ethnic and/or ideological background. Similarly, the terms marÿtum and sabÿtum may not simply refer to stringed instruments that were invented in, and consequently imported from, the regions of Mari or Sabum. When heading a group of songs they may indicate that those are somehow connected to those countries’ cultural heritages, probably expressed through art, fabrication or tuning. The expression istar¿ta may be read in a similar manner. These could be songs related to the Dumuzi-Istar-cult, which includes the series rê åÿ: rê åÿ.62 This would explain why parts of the songs (ii 38u–39u, 42u) are directed to a masculine being (god or king). The expression n zamar¿ istar¿tu amnû could therefore be understood as: “n songs, belonging to the type used in the worship of Istar, I enumerate.”63 Of the eleven serkugû (viii 27u; vi 5u) only ˜ve have been preserved as the beginnings of songs. The translation of serkugû as “sacred song” is

61. CAD A/I p. 272 sub a) interprets n zamar¿ “minûta” akkadÿta amnû as: “I enumerated the n songs in the Akkadian ‘meter’.” The parallel of akkadÿta to istar¿ta makes this interpretation unlikely, as already noted by Black (1985: 25 note 6). 62. See already Black (1984: 25–28 note 6). 63. With this perspective in mind we can also study the subscript ama-èr-ra-ku-tim on the tag catalog that was published by Shaˆer (1993). It has already been mentioned that the same expression occurs in the unpublished lament by Mama over the destruction of Kis. The expression a-mi-ra-kutum is undoubtedly derived from the Sumerian (úru) àma-irra-(bi): “(Town) that was plundered” + the Akkadian abstractsu¯x -¿tu. The expression therefore can be understood as describing the song’s content: these songs are laments over devastated towns or cultural loci. The form of the amirak¿tu-songs seems to be refrain-songs, but this could be a coincidence in the hitherto known examples. The most elaborate lament of the amirak¿tum type that is known to us is the úru àma-ir-ra-bi-lament, which is a lament for Uruk-Kullaba connected with the cult of Istar. It is sung by the gala-priest in the Mari ritual as the ˜rst song in the morning, while the muhhû goes into an ecstatic state, see Durand and Guichard (1997: 52–58 col. i–ii).

rather vague and undetermined, as are the songs sèr-gíd-da and sèr-an-ga-la-sè, which can only be found as summaries in the last subscript column (viii 12u–13u). The concept sergida must probably be understood as a simple reference to the form of the songs, that is a “lengthy song.” The narrative poem Angimdima belongs, for example to this type of song. The category sernamgalasè contains songs of the kalû-o¯ce, which are therefore classi˜ed functionally. Likewise, sernamenakk¿ are “songs of the en-o¯ce.” Certain subscripts point to the time or location of performance, such as “11 za-mar se-e-ri (viii 19u): “11 morning songs,” while other instances probably refer to the local or ethnic character of the song: 11 za-mar a-la-li akkadiki (viii 20u): “11 songs of ‘Work (on the Fields?) of Akkade?’ ” 10 za-mar dNin-gis-zi-da (viii 23u): “10 songs ‘Of the God Ningiszida’ ” 12 za-mar lú ak-ka-du-ú (viii 24u): “12 songs ‘Of the People of Akkade’ ” Of these four, the ˜eld songs and the song on the Akkadians are in the Akkadian language, while it is not clear of the remaining two whether they are in Akkadian or Sumerian. This question of linguistic determination has hitherto not received proper attention. The Sumerian term serkugu applies both to Sumerian and Akkadian entries. Therefore, the use of this Sumerian term as the heading of a list does not automatically indicate that we have to do with a list of Sumerian songs. The reverse is also true: Akkadian headings do not necessarily indicate that we are dealing exclusively with Akkadian material, as can be shown by arahhu-songs and sutanÿdû-songs, which are in Sumerian. Such headings, whether Sumerian or Akkadian, refer to other aspects of the songs, which may be musical, thematic or situational. This proposition is supported, for example, by the fact that the Old Babylonian hymn to Adad (CT 15, 3–4) is designated as belonging to the ser kummi or “chamber song,” a rubric that must apparently be interpreted as situationally determined.

SEARCHING FOR AKKADIAN LYRICS

In addition, there are classi˜catory words that are more di¯cult to de˜ne. Though many of these terms are hapax legomena, most of them can be

elilu arahhu b¿ru gangittu g¿satu inhu iratu karsu kirretu(?) merru n¿ru paru qurdu ripqu siqatu sisâtu sitru sutanÿdû

connected with known roots, which provide an etymological basis for analyzing the function of songs:

2 e-li-lu kur-ru-ú-tu 1 a-ra-ah-hu 5 b[u-ú-ru] 2 ga-an-gi-it-tu 2 gu-sa-a-tu 2 in-humes n i-ra-a-tu (sa esirte, sa kitme, sa ebb¿be, sa pÿte, sa nÿd qabli, sa nis mehri, sa qablÿte) 1 kar-su-ú akkadi ki 5 ki-ir-re-e-tu 4 me-er-ru su-me-ru 2 nu-ú-rumes 5 pa-a-ru (akkadû) 5 qu-ur-du 2 ri-ip-qu 2 si-qa-tu 2 si-sa-a-tu 13 si-it-[ri (sa ebbube]/ sa pÿte) 1 su-ta-ni-du-ú

One song is classi˜ed by the shape of its tablet, such as “gangittu.” The term could be selfexplanatory and mean “a one-column tablet.” Two songs are described in this manner. (3) Other Classi˜cations of Hymnic Content It is noteworthy that in the Papulegarra hymn (Pinches 1924) all three songs, including the parum, are headed by the more general term ser tanitti, from which it follows that the parum belong to praise-songs. The term serum is used quite generally for any type of song. From a formal point of view the category parum provides an unclear picture.64 It probably derives from pârum “to search,” which suggests the meaning of pa64. The parum of Istar seems to have nothing in common with the parum of Papulegarra. I have also found no parallels in the contents.

63

vi 21, see viii 31u (missing, only in summary column viii 35u) vi 30; see viii 33u vi 15; see viii 29u viii 34u; see [vi 33] vi 24; viii 32u (partly in vii 6 and vii 24; complete in summary column viii 45u–51u) (missing, only in summary column viii 43u) vi 11; see kír-re-tu viii 28u (missing, only in summary column viii 44u) vi 18; see viii 30u viii 16u (missing, only in summary column viii 42u) (missing, only in summary column viii 39u) (missing, only in summary column viii 37u) (missing, only in summary column viii 38u) v 16; see viii 14u and 15u (missing, only in summary column viii 36u)

rum as a “searching-song,” possibly in reference to a deity who has turned away from an individual or sanctuary. Several other interpretations are also possible, however.65 Another general category consists of ser tanitti: “song of praise.” The “song of praise” also occurs in the catalog with the word sutanÿdû in the summary column viii 36u and thus suggests a lament. This raises the question of whether lament can be a subcategory of “songs of praise.” In this case the Louvre Istar text [ser?] ta-na-ti Istar might be such a variant of the type. A similar categorization according to content can be designated by the word inhu, “song of sighs.” Two song incipits have this heading in KAR 158 (vi 22 and 23):

65. The term parum is not carried over to the ˜rst millennium.

64

BRIGITTE GRONEBERG

Istar sarrat un.mes raåum-tu, “Istar, queen of the people, the beloved one,” and dIstar mannu balukki belitÿ, “Istar, apart from you, who is my mistress?” d

The laudatory character of the inhu is con˜rmed in a description of the inhu by Assurbanipal,66 as well as from the ˜rst line of the ser kummi song to Adad (see sub I 2) above). However, from the ritual context it is clear that the song was meant as a song of grief, so that it is likely that apart from passages of praise a lament was also uttered. This is the case in the DumuziIstar ritual KAR 42:29,67 in which the assinnupriests perform a “sighing”-(song?) before carrying away evil from the city. The inhu is also be found in the Old Babylonian Louvre Istar text, in which at a particular moment the “assinnu utters sighs and cries, till his cries become loud.”68 In the ritual KAR 154, directed to Adad, qadisatuwomen also “sigh,” and in the Seleucid ceremony for rebuilding a destroyed temple the nar sighs the sighing song.69 Such songs were used in a ritual that was intended to pacify the deity. According to Maul, this “sighing” may indicate that the performer had fallen into a “sacred rage.”70 Based on this and on the fact that the text is mentioned without its incipit, I would suggest that the “sighing song” had a speci˜c form or a ˜xed sequence that could be recognized without a speci˜c title. The last group relating to the song’s content are the n¿ru-songs from vi 16–18 and viii 30u. The opening words of these “light songs” are: dar sillÿ la tezenni, “eternal is my shadow, do not be angry (?)”71 mannu en gis.má mannu en gis.má.gur, “who is the master of the boat, who is the master of the large boat?”

66. See CAD I p. 148a: inhÿja s¿nuh¿ti Istar ismema: “Istar heard my sorrowful inhu-songs.” 67. See Farber (1977: 57, 19). 68. See Groneberg (1997: 28) col. ii 31: innah inanhu rigmu rigmasu isappi. 69. Racc 44: BE 13987, 5. 70. Maul (1992: 164–66). 71. This translation is based on zenû: “to be angry” but other words are equally possible.

These songs are presumably directed to the moon god.72 Some of these classi˜cations seem to have been inspired by the song’s mode of performance. Among them the kirretu-songs (vi 11) present an interesting puzzle. A derivation from kirretu “meat (in a special form)” does not appear to make sense, but perhaps it refers to the kirretuvessels, known only in plural form in texts from Mari and Amarna.73 We also know of the use of the kirru when drawing up contracts,74 attested as a ritual gesture in a marriage ceremony, oˆering or drinking (out of ?) the kirru.75 The relevant opening words in these songs are: hudÿ bÿt-ni s¿lilÿ, “rejoice for our mistress, jubilate!” e-ni76-ti emqÿtÿ amarat nise (unmes), “my wise punishment watches over the people” rasubtu ina ilane anaku, “the mightiest among the gods am I” s¿qa abaåima 2 mí-usmes ú-ta: “I stroll along the street (and) I ˜nd two assinn¿” It should be clear that at least some if not all these songs are connected with the practice of the Istar cult. However, these four opening lines refer to as yet unknown hymns; perhaps they may be found among the many unpublished texts from Assur. No further information has been found to assist with a determination of the meaning of this category.77 A tentative approach based on the contents of the lines as well as a third possible

72. To a similar category could also belong the qurdu: “hero-song,” which we have only found as a title in the summary of column viii. 73. See CAD K p. 409 sub 1. 74. See CAD K p. 410 a sub 2 b). 75. Laws of Esnunna § 27, see Roth (1997: 63). There is a qerÿtu-feast in later Assyrian times, presumably to be connected with the root qerû: “to invite,” known from several rituals. Compare for the feast Menzel (1981: 21–23 et passim); qerÿtu is a feast of Tasmetu and Nabû. 76. The sign NI is partly erased (collated). 77. It is remarkable that the songs are addressed either to a feminine person or are kept in the ˜rst person. The only Old Babylonin Akkadian song of self-praise known hitherto is a formal laudatory song to Istar (VS 10, 213); see Foster (1993: 74).

SEARCHING FOR AKKADIAN LYRICS

etymological derivation would be: “happy songs of praise connected with a ceremonial feast.” The g¿satu- or whirling dance-songs (vi 33 and viii 34u) must also be seen as a category determined by the function of the song. These are presumably directed to Istar in her hypostasis as

15. 16. 17. 18. 19. 20.

[li-ib-si sa-at-ti?]-sa li?-is-sa-ki-in gu-us-tu-ú i-pa-ar-sí-im sa-at-ti bi-it-ri-i [ni-si]-i gi-im-ra-as-sí-in li-me-el-lu i-sú-qí-im si-me-e ri-gi-im-si-in

65

Agusaya and were probably used in connection with ecstatic whirling dances. Unfortunately, only the ˜rst word of these songs (dgu-sa-ia[ ]) has been preserved,78 but a description of the festivities is given in Agusaya B col. iv (the speech is adressed to Istar):79

It shall be yearly, A whirl-dance shall be held in the rites of the year. Keep (fem.) an eye upon all mankind! They shall dance in the street! Listen to their (excited) screaming!

All the terms that relate to agrarian feasts, such as alalu/elÿlu, arahhu, ripku and karsu, can be taken as functionally determined classi˜cation words. The expressions alalu/alÿlu/elÿlu and arahhu are found in Old Babylonian texts as well as in texts of the ˜rst millennium, while the other two are hapax legomena.78 79 The term alalu is de˜ned in the CAD as “refrain of a work song.”80 This song is often connected with ˜eldwork, particularly with plowing.81 In the catalog, two short (kurrûtu: vi 21) elÿlu are mentioned, and it would appear that they are distinguished from the zamar alali Akkadi ki of column viii 20u by their brevity. A similar song is the arahhu, which was already connected by the CAD to arahhu: “granary,” and therefore identi˜ed as “perhaps a harvest song.”82 Also ripqu is already identi˜ed as a working song in the dictionaries.83 In connecting the incipit karsu to karasu “to bind” (also in connection with binding objects during agricultural work), karsu would mean a similar type of (work)song.

78. The substantive in the plural would exclude the possibility that the cataloging principle was songs to the goddess Gusaya and possibly points to whirling dances. 79. See Groneberg (1997: 86). 80. CAD A/1 328–29. 81. CAD A/1 328:b). 82. CAD A/2 220. 83. AHw 987 sub ripqu 2) from rapaqu, “to work, to hack a ˜eld” see also CAD R 366 sub ripqu C.

Apart from the term irtu, which will be dealt with below, the other classi˜catory words in KAR 158 are all termini hapax legomena that cannot be further identi˜ed.84 The term siqâtu and and the word b¿ru remain enigmatic. The ˜ve opening words of these songs from KAR 158 vi 25–? are cited in CAD B p. 343 under b¿ru D (hapax): VI 25ˆ. sammarÿtu tuttê tuttê, “O untamed lady, you have found, yes, you have found” etlu istu amuru[ka], “Hero, since I saw [you]” surbûta ana nisÿ (unmes) azamm[ur], “I shall sing (your) greatness to (all) men” sillulu sa nise (unmes) a[ttÿ], “You are the protection of men” ina subat bêl ilÿ izammura[ssi], “He will praise her in the dwelling of the Lord of the gods” A derivation from b¿ru: “calf ” (as a literary term in praise-songs) appears unlikely. Perhaps there is a connection to the jar called p¿ru. The lament úru-àm-ma-ir-ra-bi mentions that ingredients like uhhulu “soda water” and samnu tâbu

84. merru (viii 44u) and sisâtu (viii 38u) probably relate to natural phenomena. The possibility that merru could be related to mer¿saru: “north wind” should not be excluded (see also AHw p. 646 sub 2) and sisâtu could be a song devoted to a metereological phenomenon of the name sisÿtu, which is found in connection with Dilbat (Istar) (AHw p. 1250 sub 3b). For neither subscript do we have any opening words of songs.

66

BRIGITTE GRONEBERG

“best oil” for Inanna’s cosmetics are stored in such a jar.85 4: III. Songs of “the breast” (irate) Apart from the “songs of praise,” a term that de˜nes most of the Old Babylonian hymns, one sub-type of song, the irtu or “breast-song” is particularly popular. Since the term has recently been studied it deserves special treatment. The Finkel “catalog” mentions the opening words of what he considers to be ˜ve iratu-songs, but which are in actuality ˜ve lines of one irtusong. A good clue for the language and nature of the iratu-songs has come from an Old Babylonian tablet from Geneva (Groneberg 1999), which mentions several iratu-songs belonging to a series (iskaru).86 This type of song is also found in KAR 158, where at least 69 opening lines of songs are designated by this term. Some of these entries have already been cited by Held (1961), Black (1985), Matsushima (1987), and more recently by Limet (1996), as well as by the dictionaries.87 It had already become clear from the Old Babylonian collection of iratu-songs that these are songs about human love. It is only logical that the entries in KAR 158 center on sexual arousal: love, lovemaking at night, in the garden, in other places and at various times. The incipits cited are: Col. vii 1: [x x ul?]luskama[ “. . . swollen to you . . .” ** 2: an(coll.)!-ha ki-i mul-BU[ “they are weary like . . .” 3: ina sÿhti sa se[ri?]: 85. Volk (1989: 85). See also similar preparations in the love songs, newly translated in Sefati (1998: 132–50, especially lines 3–7). 86. See Groneberg (1999). 87. Readings taken from Black are marked with a (B), those from Held with (H), Matsushima with (M), and Limet with (L) respectively. As translations of these colorful passages in KAR 158 can be rather speculative, translations may diˆer slightly. The * indicates speech by a woman, ** indicates speech by a man and no star(s) means the speaker is not speci˜ed.

5:

6: 7: 8:

9:

10: 11: 12: 13:

14: 15:

16:

17: 18:

“in the happy laughter of the out[skirts?]”(L) annû sibûtu sa h¿du kabatti: “these are the wishes which are in the pleasure of the heart” epsu [p]ÿka atmu pi?(sign: ti)-ja summe hanimma “speak! interrupt the speech of my mouth(!): he is blooming”** su-nígin iratu sa e-sìr-te ke sÿhaku ana nahsi “how I smile at the voluptuous”(B)88** ¿m en imnÿja ishitanni: “on the day, when my right eye twitched”(B)89 ileqqe amassi ana marÿ: “he will take and I will be su¯cient for my darling”(H)** matema belu terruba inanna: “when, oh lord? you will now come in!”(H)** harpaku ana dadÿka: “I am early (quick) for your ‘darlings’ ”** itti simtat m¿siti: “with the . . . (?)90 of the night” m¿sa maru usamsakku: “I let you stay the night, small darling!”(B)91** a ammar saman erseti (KIti) “I ˜nd the earth’s oil” ¿m¿ ubla bussuratuma h¿d libbi: “on the day, when the news portend deep pleasure”(L)92* elaja maru lu namer erba: “not at all! My darling is brilliant, enter!”(L)** la mehirtÿ isnunanni: “there is no rival equal to me”(L)* uru nippurÿtÿ lahannatu dasuptu: “my girl from the city of Nippur is a sweet lahannatu-bowl”(L)*

88. Also Limet (1996: 151). 89. Also Limet (1996: 158). 90. simtat is St. constr. Pl. of simittu: “a precious object” (CAD †/3 6–7) or of simuttu/sumuttu: a plant (= beetroot?) (CAD †/3 301). 91. See Limet (1996: 157) “toute la nuit je veillerai, chéri!” 92. h¿d libbÿ could have a sexual meaning.

SEARCHING FOR AKKADIAN LYRICS 19: suppÿ hullÿ: “cause to be silent, creep along”* 20: sa?-[a]m?-sa-ás-ni tillani “our . . . (?),93 our tassel”* 21: samm¿t erini ramka belu “the cedar’s fragrance is your love, my darling lord”(H)** 22: ana bab(ká)eni tallikamma “You approach the doors” ** 23: ana musÿti annÿti ana lÿlâti annâti: “to this night, to these evenings”(B)94 24: su-nígin iratu sa kitme 25: ke nahsat ke namrat: “how voluptuous she is, how shining”(B)* 26: i-se-eå-ì!(koll.)-ma kirî ( giskiri6) lalîka: “indeed she looks for the garden of your sexuality”(L)95** 27: ¿ma libbÿ melula nigûta: “today my heart is (full of) playing and music”(M) 28: arid kirî ( giskiri6) lugal hasibu ereni “he descends into the garden, the king, who cuts the cedar”(L)** 29: atta maru raåimu dadÿni: “you are the beloved who loves our ‘darlings’ ”(H)96** 30: ebirtu nari(íd) alu melilÿ “across the river is the city of playing”(M) 31: asrunni tag¿sa97 arah sÿhti “to our place you come, month of joyful laughing”(L) 32: ulla alik maru “no, go, darling son!”(H)** 33: ke i-sa-ra-ap! sat hatti “how she burns (sexually) that terrifying one” 34: is-sur-tu ti-it-ku ur-ri lallaru rigimki 98 93. †a-am-sa-ás-ni is unclear to me. 94. So also Limet (1996: 157). 95. Limet (1996: 158) reads: i-se-e’-as-ma kirî la-li-ka: “elle languit dans le jardin de ton désir.” However, the sign NI emerges instead of AS. 96. Some similarity to the iratu-song published by Westenholz (1987: 415–42). 97. This verb is following Limet (1996: 157) taken from gâsu A: “to come” but it could also be derived from gâsu B: “you whirl around,” see CAD G s.v. 98. I do not understand the ˜rst half of this sentence.

67

“Titku-lady-bird (?),99 your cry during the day is (like) a lallaru”* 35: sandannak kirî ( giskiri6) sÿhatÿ “gardener of the garden of happylaughters!”(L)** 36–37: ¿m DU100 ina urula-ar-sa-an sÿhatu isÿqa (L) “on the day of . . . (?) in Larsa101 he/she sti˘es joyful laughter” 38: rÿsÿ dNanaya ina kirî ( giskiri6) é.babbar sa tarammi (H) “rejoice Nanaya in the garden of Ebabbar which you love”(L)* 39: mêrtu ubla libbasa suara: “the darling daughter decides for the happy playing”(L)* 40: kajanamma ke saktakuma “How I keep staying silent!” 41: asÿhkuma ilu (dingir) sehheru (tur.tur) “I laugh at you, small darling god”(L)102** 42: barmatu igi.mes-ia imdalasitta: “my speckled eyes have ˜lled with sleep” 43–44: ra l¿ na4KA 103 sÿhatuka l¿ hurasu: “your love shall be obsidian, your smiles shall be gold”(B)104** 45: ramÿ nÿru musnammeru attalî (an.mi). “my love is the nÿru-constellation, which illuminates the eclipse”(L)105 46: m¿sa ihsuskama: “during the night she remembers you”**

99. The dictionaries sub issurtu prefer issurtu titkurru (titkurru as a hapax legomenon). 100. The logogram DU can have the meaning alsaku: “to go” or izuzzu: “to be present,” but it could also be the shortened month name DU for Dumuzi. 101. Compare CAD S 186: ¿m x ina Larsan sihatu izÿqa: “when I was in the city of Larsa, dalliance came wafting through the air;” see Limet (1996: 158): “quand je vais à Larsa les ris ˘ottent dans l’air.” I also took la-ar-sa-an for a variant to normal la-ar-sa, as the reading: la-ar-sa dingir does not seem to make sense, but I cannot explain the ending -an on Larsa. 102. Limet (1996: 158) interprets: a-si-ih-ku ma-an tur-tur: “j’aurais ˘irté avec toi, petite.” 103. Compare CAD R 137a: ramka lu surru (na4ZÚ). 104. Limet (1996: 157) translates in the same manner. J. Black (1984: 25) renders the ˜rst part diˆerently as “your love shall be a jewel . . .” 105. Limet (1996: 157) translates: “mon amour est une lumière qui illumine une éclipse.”

68

BRIGITTE GRONEBERG

47: epsu pÿka “speak!”** 48: istu sallaku ina s¿n mari “after I was falling asleep in my darling’s lap”(H)** 49: ribi kâsa na4za-gìn-gìn “grow big! for you the lapis lazuli”** 50: hadis aksa sarru: “joyfully consummate, king!”(B)106** 51: migru hamru: “the favorite, the selected”** 52: ina lalîki hunbÿ: “in your voluptousness be rich”* 53: ukâl ram[ka]: “I (?) control [your] love” 54: ina sa-ar mu-r[i? . . .]107 55: ti-ka-[. . .] The rest is broken away. (1) The language of love in the irate songs The themes evoke the joy of lust and eroticism, often using well-known metaphors and metonyms. Keywords are: - laughter, joy, fun like sÿhtu: “happy, joyful laugh” (vii 3. 7. 31. 37. 41. 44) h¿d libbi (vii 15), h¿du kabatti (vii 4), riasu “rejoice” (vii 38) nigûtu “joyfullness” (vii 27) and hadis “in joy” (vii 50); - love: ramu (vii 21. 29. 38. 43. 45. 53); - intimate parts of the body: dad¿, which means genitals or other erotic organs (vii 11. 29), s¿nu “lap” (vii 48) - eroticism or sex appeal: lalû “plentifulness” (vii 26.52), nahsu “lushness?” (vii 7. 25); - brilliance: namaru (vii 16. 25. 45).

- to enter something or someone: erebu (vii 10. 16); - going or staying: akasu, alaku, la alaku (vii 22. 32. 50); - being silent: suppû, sakatu (vii 19.40) or - incite to talk: epsu pÿka (vii 5.47). Love is compared to the hardness of obsidian (na4KA vii 43) and to gold (kù.sig17 vii 44). The “syrup-sweetness” dusuptu, which is often a theme in Sumerian love songs occurs only once (vii 18).109 The time for love seems to be usually during the night: musÿtu (vii 12. 23), lÿlâtu (vii 23), m¿su (vii 13. 46); the beloved stays at night (vii 13) or sleeps in someone’s lap: salalu (vii 48). This presumably happens in the bedchamber,110 but the place of love is also often the garden: kirû ( gissar; vii 26. 28. 35. 38). The cherished and valuable cedar (of a garden?) or its resin is also mentioned (vii 21. 28). One of these iratu-songs refers to joy taking place in the city: alu (vii 30); and one tells of joyful amusement in the steppe: seru (vii 3). The terminology for the beloved is taken from terms of a¯nity (as well as hierarchical vocabulary) as in the Dumuzi-Istar songs.111 Akkadian terms are maru, “beloved son” (vii 9. 13. 16. 29. 32. 48) or belu “darling” (10. 20. 21). Once a mertu, “beloved daughter” (vii 39) is addressed. Twice a king is evoked, probably to be understood as a model of a king. The metaphorical expression “my/your (heart’s) king” (as the beloved master of a person) does not seem to be attested in Mesopotamia.112 Ony two concrete places are mentioned: the cities of Nippur (vii 18) and Larsa (vii 36). Among the gods only Nanaya is named in a context of joy and the garden (vii 34). The reference to a very

Part of the play of love is - to play: melulu/melilu (vii 27. 30);108 106. Limet (1996: 157) translates: “vas en liesse ô roi.” 107. There are several possibilities for interpreting this line: “in 3600 xy . . .” or: “in the wind of [. . .].” 108. The “playing” in erotic context is used in DI H (line 20) und in DI Z (Rev. 5u) but in a context of the (real) play of young people in DI H (16).

109. Sefati (1998) passim and see index p. 409 sub làl. 110. This is based on the fact that the love in the sacred marriage ceremonial takes place on a bed. It forms an important part of the love scene. See also Sefati (1998). 111. Sefati (1998) DI B (4), DI C; DI D DI G (ses and nin9), but see ses in the meaning natural brother in DI A!; ama and passim in DI D. 112. In the Dumuzi-Inanna-songs Dumuzi is called: mulu sà-ab-mu (DI A 49) and see D. The title dumu-lugal does appear in DI B.

SEARCHING FOR AKKADIAN LYRICS

small god in vii 41 could be understood as intimate talk to a young person or perhaps to a baby. The description of love and love-making in these love songs, though still obscure as to their addressee,113 is discrete: The texts use metaphors for sexual parts of the body. The place of love alludes to love-making. Many of the topics call forth the same pictures in our mind due to common human behavior of feeling and making love in advanced urban cultures. On a discourse level the songs reveal several characteristic features. Nearly all of them address someone. Usually this is a male person (**24) who is addressed intimately by a woman, but some love songs also address a female (at least *9). A few incipits are styled in the ˜rst person (6) and rarely a plural ˜rst person form is used. Only some are descriptive and begin in the third person. The “we” comes into play as an interjection of one of the lovers or as an address to one of the protagonists from an unspeci˜ed group.114 Direct speech expresses itself by means of emotional, abrupt and excited language in incipits, such as “how I keep staying silent!” (vii 40) or “speak!” (vii 47). In terms of the verbal forms used, only stative and present forms are to express either durative situations or actions, which are happening immediately or are expected to happen. The most normal level of communication is the nominal sentence, used to state eternal truths. (2) The cultural setting of the songs The mention of the king sets some of these songs in a more o¯cial context as does the incipit naming Nanaya. This o¯cial and exemplary lovemaking considered to take place in a formal context of marriage between Dumuzi and Inanna is the functional background to the Sumerian 113. Especially for the female person: e.g., who is the “beloved person” referred to by the male voice? 114. Westenholz (1987: 417) comments on the “we” as “found elsewhere in amatory context usually with Istar or Inanna as speaker.” In the Dumuzi-Inanna-songs, however, “we” is rarely used (e.g., Sefati [1998] DI C 39. 40). More usual is direct speech and self-reference in the ˜rst person. Her suggestion therefore has yet to be substantiated.

69

Dumuzi-Inanna-songs. The main themes describe ceremonies of wooing and clothing the bride (Inanna) up to the ceremony of preparing the bridal bed. Because the Dumuzi-Inanna songs evoke the world of the gods, they must be contextualized within the o¯cial cult, although the rites probably mirror normal human courting rituals.115 The songs address the mythical king Dumuzi and his lady, the goddess Inanna, but in the ritual setting they are applied to the cultic personnel that played these parts in the festival. The incipits from this part of KAR 158 are clearly not addressed to Dumuzi and Inanna but to an anonymous man and woman, but occasionally they seem to address a king or a god. They diˆer from the Sumerian love songs even if the ideas of love and the setting where love is described are similar. It is very di¯cult to discern the scenic setting of these songs by their opening lines alone. Whether the iratu-songs are meant to be used in a secular or in a cultic setting or if some of them are cultic and others belong to courtpoetry remains unanswered. IV. Conclusions Who could have been the author of KAR 158? One may surmise that it was either someone connected to the court or a temple o¯cial. Is it possible that he was a musical performer? This would help explain the texts’s unusual preoccupation with musical instruments and other musical notations. Indeed, one of the most interesting features of the enumeration in KAR 158 is the number of diˆerent iratu-songs cited in what may well be diˆerent musical modes. The variety of musical notations of iratu-songs is astonishing. KAR 158 mentions at least seven diˆerent terms (viii 45u– 51u).116 Is it possible that these songs were recited and accompanied by instruments that were set in particular musical keys or that some other musical principles were used? All of them appear to 115. See Alster (1993) with previous literature. 116. The musical aspect of these songs will be dealt with elsewhere.

70

BRIGITTE GRONEBERG

have been related to diˆerent “moods” or diˆerent occasions or settings. There is no apparent classi˜catory order according to the meaning or the person addressed. Songs in the musical notation irate sa kitme,117 which are cited in vii 7– 23, address either a man and a woman (vii 19). Alhough these are often love songs stressing sexuality and erotism—such as vii 7. 8. 9. 11. 13— others seem to stress diˆerent subjects. From intertextual analysis these short opening lines do not help to clarify the setting, the mode, or the musical mood. There remains the question of whether or not all the songs are cited by their opening line, or incipit. This would be the usual way that Mesopotamian scholars referred to literary texts. This appears to be the usage adopted in this catalog, judging by the example published by Black (1984) and identi˜ed by its opening line.118 If this usage was applied throughout the catalog, then none of the iratu-songs refer to any known Old Babylonian songs—at least as far as its ˜rst line is preserved—and not even one of the irate of Ammiditana (Groneberg 1999) is enumerated here. Still, it seems quite probable that Akkadian irate songs were already popular in Old Babylonian times. Could the earliest known example be the song of Rÿm-Sîn, YOS 11, 24, which is one of the earliest pieces of literature in the Akkadian language? Could it be that some of the other iratesongs mentioned in KAR 158 likewise originated in the Old Babylonian period? The grammar of the irate-songs does not provide any chronological information, as their short sentences do not oˆer any dialectal information. Only one word, mêrtum (vii 39), clearly points to a stage of language later than the Old Babylonian period. While we know that most irate-songs were formulated in post-Old Babylonian times, it is a fact that older ones existed. But this type of religious and/or secular love song would not have a future in the ˜rst millennium in the written compendia. The love songs from the ˜rst millennium feature gods and goddesses, presumably in the context of 117. Probably in the minor mode. 118. Black (1984: 25) with note 6.

the “sacred marriage.”119 These are styled in a completely diˆerent semantic and stylistic manner. They are comparable to some of the popular Sumerian Dumuzi-Istar songs, which were written down and perhaps even ˜rst formulated during the Old Babylonian period. The catalog KAR 158 thus demonstrates that many Akkadian hymns have been lost to us. It seems that Akkadian literature began to be formulated in written form on the orders of the kings of the Hammurabi dynasty. The small corpus of attested polished Old Babylonian lyrical texts consists of unique compositions that are without duplicates. Judging by their style and their themes they are to a large extent dependent on Sumerian hymns. They could very well be literary exceptions, a new way of presenting royal power in the Akkadian language to an Akkadian population. Yet some Akkadian language forerunners of a more uncomplicated kind of literature, the love songs, are already attested from the time of the Larsa kings through the late Old Babylonian period, and judging by the KAR 158 catalog, they seem to have remained quite popular up to the time of the Middle Assyrian courts. To this point the data that have been brought together have shown that among the categorizing terms and expressions the words zamaru and seru, “song,” are used in a very general sense. It would appear, however, that both these words often indicate the manner of presentation, namely the singing or recitation and that they are not used in designations that refer to the form or function of the texts. These two terms are used to describe well-known Old Babylonian hymns, such as the hymns to Agusaya, to Belet-ili,120 to Istar,121 the Standard Babylonian hymn to sarrat Nippuri,122 the hymn KAR 361//105 to †amas, but also the mythical narrative text to Erra.123 These terms of presentation are further speci˜ed by functional, situational, or thematic indicators. Apart from them we have identi˜ed many 119. 120. 121. 122. 123.

See, for an example, TIM 9, 54; compare SAA III No. 14. CT 15 1–2, see Römer (1967–68: 12). Thureau-Dangin (1925:171, 53). Lambert (1982: 202–3: 41). Cagni (1969: 126: col. v 49. 59).

SEARCHING FOR AKKADIAN LYRICS

other song-classifying words and have examined them in some detail. As far as the terms could be understood, they all relate to the function or content of the songs. No evidence has been found of classi˜catory terms that relate to the formal structure of the hymns. A classi˜catory structure in which terms and subterms are all interconnected and form subordinate groupings has not emerged. The only subdivision that can be found is that between hymns possessing refrains or repetition of strophes and those that do not have those features. It does not seem that such a lack of classi˜catory

71

structure is related to the fact that at that point Akkadian literature had not long been put to writing. Perhaps the same lack of general principles of categorization will be found in hymns of the ˜rst millennium, but this requires further research. Probably we are simply dealing with the phenomenon that lyrical works in a written form are presented and perceived as having great variety and that types and sub-types are determined with characterizations that do not re˘ect our linguistic criteria but relate to a wide spectrum of cultural aspects.

References Alster, B. 1993 Marriage and Love in the Sumerian Love Songs. Pp. 15–27 in The Tablet and the Scroll. Near Eastern Studies in Honor of William W. Hallo, eds. M. E. Cohen, D. C. Snell, D. B. Weisberg. Bethesda: CDL. Black, J. 1984 Babylonian Ballads: a New Genre. Pp. 25–34 in Studies in Literature from the Ancient Near East by Members of the American Oriental Society Dedicated to Samuel Noah Kramer. American Oriental Society 65, ed. J. Sasson. New Haven: The American Oriental Society. Bonechi, M., and Durand, J.-M. 1992 Oniromancie et magie à Mari à l’époque d’Ebla. Pp. 151–62 in Literature and Literature Language at Ebla. Quaderni di Semitistica 18, ed. P. Fronzaroli. Florence: Dipartimento di Linguistica Università di Firenze. Bottéro, J., and Kramer, S. N. 1989 Lorsque les dieux faisaient homme, Mythologie Mésopotamienne. Bibliothèque des Histoires. Paris: Gallimard. Cagni, L. 1969 L’Epopea di Erra. Studi Semitici 34. Rome: Istituto di Studi del Vicino Oriente. Universita di Roma. Chang, K. W. 1981 Dichtungen der Zeit Tukulti-Ninurtas I von Assyrien (unpublished doctoral dissertation, Seoul). Cunningham, G. 1997 Deliver Me from Evil. Mesopotamian Incantations 2500–1500 BC. Studia Pohl Series Major 17. Rome: Ponti˜cio Istituto Biblico.

Durand, J.-M., and Guichard, M. 1997 Les rituels de Mari. Pp. 19–78 in Florilegium marianum III. Recueil d’études à la mémoire de Marie-Thérèse Barrelet. Mémoires de N.A.B.U. 4, eds. D. Charpin and J.-M. Durand, Paris: SEPOA. Farber, W. 1977 Beschwörungsrituale an Istar und Dumuzi. Attÿ Istar sa harmasa Dumuzi. Wiesbaden: Harrassowitz. Finkel, I. 1980 Bilingual Chronicle Fragments. JCS 32: 65–80. 1988 A Fragmentary Catalogue of Love Songs. ASJ 10: 17–18. Foster, B. 1993 Before the Muses. An Anthology of Akkadian Literature. 2 Vols. Bethesda: CDL. Gelb, I. J. 1975 “Homo ludens” in Early Mesopotamia. Pp. 43– 74 in Haec Studia Orientalia Professori Assyriologiae et Filologiae Semiticae in Universitate Helsingiensi Armas I. Salonen, S.Q.A. Anno 1975 Sexagenario. Studia Orientalia 46, ed. I. Kärki. Amsterdam, Oxford, New York: North Holland/American Elsevier. Greengus, S. 1979 Old Babylonian Tablets from Ishchali and Vicinity. Uitgaven van het Nederlands Historisch-Archaeologisch Institut te Istanbul 44. Leiden: Nederlands Historisch-Archaeologisch Institut te Istanbul. Groneberg, B 1996 “Towards a De˜nition of Literariness as applied to Akkadian Literature. Pp. 59–84 in Mesopotamian Poetic Language: Sumerian and Akkadian. Cuneiform Monographs 6, eds.

72

BRIGITTE GRONEBERG

M. Vogelzang, H. M. L.Vanstiphout. Groningen: Styx. 1997 Lob der Istar. Gebet und Ritual an die altbabylonische Venusgöttin. Tanatti Istar. CM 8. Groningen: Styx. 1999 “Brust” (irtum)-Gesänge. Pp. 169–96 in Minuscula Mesopotamica. Festschrift für Johannes Renger. AOAT 267, eds. B. Böck, E. CancikKirschbaum, Th. Richter. Münster: UgaritVerlag. 2001 Die Liebesbeschwörung MAD V 8 und ihr literarischer Kontext. RA 95: 97–113. 2002 “The Faithful Lover” Reconsidered: Towards Establishing A New Genre. Pp. 165–83 in Sex and Gender in the Ancient Near East. Proceedings of the XLVII e Rencontre Assyriologique Internationale. Helsinki: The Neo-Assyrian Text Corpus Project. Gurney, O. R. 1989 Literary and Miscellaneous Texts in the Ashmolean Museum. OECT XI. Oxford: Clarendon Press. Hallo, W. W. 1963 On the Antiquity of Sumerian Literature. JAOS 83: 167–76. Horowitz, W. 2000 Astral Tablets in The Hermitage, Saint Petersburg. ZA 90: 194–206. Hunger, H. and Groneberg, B. 1978 Review of: Texts in the Iraq Museum, IX. JAOS 98: 521–23. Kämmerer, Th. 1998 simâ milka. Induktion und Reception der mittelbabylonischen Dichtung von Ugarit, Emar und Tell el-çAmarna. AOAT 251. Münster: Ugarit Verlag. Krebernik, M. 1998 Die Texte aus Fara und Tell Abu Salabikh. Pp. 237–427 in Mesopotamien, Späturuk-Zeit und Frühdynastische Zeit. Orbis Biblicus et Orientalis 160/1, eds. P. Attinger, M. Wä˘er. Freiburg Schweiz, Universitätsverlag; Göttingen: Vandenhoeck & Ruprecht. Krecher, J. 1976/80 Kataloge, literarische. Pp. 478–85 in Reallexikon der Assyriologie Band 5, ed. D. O. Edzard. Berlin, New York: de Gruyter. Krispijn, Th. 1990 Beiträge zur Altorientalischen Musikforschung. 1. †ulgi und Musik. Akkadika 70: 1–27. Lambert, W. G. 1959/60 Three Literary Prayers of the Babylonians. AfO 19: 55–60. 1960 Babylonian Wisdom Literature. Oxford: Clarendon.

1966/67 Divine Love Lyrics from the Reign of Abiesuh. MIO 12/2: 41–56. 1982 A Hymn to †arrat-Nippuri. Pp. 173–218 in Zikir †umim. Assyriological Studies Presented to F. R. Kraus on the Occasion of his Seventieth Birthday, eds. G. van Driel, Th. J. H. Krispijn, M. Stol, K. R. Veenhof. Leiden: Brill. 1987 A Further Attempt at the Babylonian “Man and his God.” Pp. 187–202 in Language, Literature and History: Philological and Historical Studies presented to Erica Reiner. American Oriental Series 67, ed. F. Rochberg-Halton. New Haven: The American Oriental Society. 1989 A Babylonian Prayer to Anuna. Pp. 321–36 in DUMU-E2-DUB-BA-A. Studies in Honor of Å. W. Sjöberg, Occasional Publications of the Samuel Noah Kramer Fund, 11, eds. H. Behrens, D. Loding, M. T. Roth. Philadelphia: The Samuel Noah Kramer Fund. 1990 A New Babylonian Descent to the Netherworld. Pp. 289–300 in Lingering over Words. Studies in Ancient Near Eastern Literature in Honor of William L. Moran. Harvard Semitic Studies 37, eds. T. Abush, J. Huehnergard, P. Steinkeller. Atlanta: Scholars. 1992 Prostitution. Pp. 127–57 in Aussenseiter und Randgruppen. Beiträge zu einer Sozialgeschichte des Alten Orients. ed. V. Haas. Konstanz: Universitätsverlag Konstanz. Lambert, W. G., and Millard, A. R. 1969 The Babylonian Story of the Flood. Oxford: Clarendon. Limet, H. 1996 Le texte KAR 158. Pp. 151–58 in Collectanea Orientalia. Histoire, Arts de l’éspace et industrie de la terre. Études oˆertes en hommage à A. Spycket. Civilisations du Proche Orient. Serie 1. Archéologie et environnement, eds. H. Gasche and B. Hrouda. Neuchâtel-Paris: Recherches et Publications. Machinist, P. 1978 The Epic of Tukultÿ-Ninurta I: A Study in Middle Assyrian Literature. Michigan: University of Michigan. Matsushima, E. 1987 Le rituel hiérogamique de Nabû. ASJ 9: 131–75. Mayer, W. R. 1976 Untersuchungen zur Formensprache der babylonischen Gebetsbeschwörungen. Studia Pohl Series Major 6. Rome: Ponti˜cio Istituto Biblico. Maul, S. 1992 Kurgarrû und assinnu und ihr Stand in der babylonischen Gesellschaft. Pp. 159–71 in Beiträge zu einer Sozialgeschichte des Alten

SEARCHING FOR AKKADIAN LYRICS

Orients, ed. V. Haas. Konstanz: Universitätsverlag Konstanz. Menzel, B. 1981 Assyrische Tempel. Untersuchungen zu Kultadministration und Personal. Studia Pohl Series Major 10/I–II. Rome: Ponti˜cio Istituto Biblico. Michalowski, P. 1998 Literature as a Source of Lexical Inspiration: Some Notes on a Hymn to the Goddess Inana. Pp. 65–74 in Written on Clay and Stone. Ancient Near Eastern Studies Presented to Krystyna Szarzynska on the Occasion of her 80th Birthday, ed. J. Braun, K. Lyczkowska, M. Popko, and P. Steinkeller. Warsaw: Agade. Oelsner, J., and von Soden, W. 1991 Ein spät-altbabylonisches parum-Preislied für Istar. OrNS 60: 339–43. Pedersén, O. 1986 Archives and Libraries in the City of Assur. A Survey of the Material from the German Excavations. Studia Semitica Upsaliensia. Part II. Uppsala: Almqvist & Wiksell. Pinches, Th. G. 1924 Hymns to Pap-due-garra, IRAS Centenary Supplement, 63–86. Römer, W. Ph. 1966 Studien zu altbabylonischen hymnischepischen Texten (2). Ein Lied über die Jugendjahre der Götter Sîn und Isum (CT 15, 5–6). JAOS 86: 128–47. 1967 Studien zu altbabylonischen hymnischepischen Texten. Pp. 185–99 in Heidelberg Studien zum Alten Orient. Adam Falkenstein zum 17. September 1966, ed. D. O. Edzard. Wiesbaden: Harrassowitz. 1967/68 Studien zu altbabylonischen hymnischepischen Texten (3). Ein Lied mit Bezug auf einen †ubartum-Feldzug Hammurapis. WdO 4: 12–28. Roth, M. T. 1997 Law Collections from Mesopotamia and Asia Minor. Society of Biblical Literature. Revised edition. Writings from the Ancient World 6. Atlanta: Society for Biblical Literature. Sefati, Y. 1998 Love Songs in Sumerian Literature. Critical Edition of the Dumuzi-Inanna Songs. Ramat Gan: Bar Ilan University Press. Shaˆer, A. 1993 From a Bookshelf of a Professional Wailer. Pp. 209–10 in The Tablet and the Scroll. Near Eastern Studies in Honor of William W. Hallo, eds. M. E. Cohen, D. C. Snell, D. B. Weisberg. Bethesda: CDL.

73

Sommerfeld, W. 1982 Der Aufstieg Marduks. Die Stellung Marduks in der babylonischen Religion des zweiten Jahrtausends v. Chr. AOAT 213. NeukirchenVluyn: Neukirchner Verlag. Stol, M. 1987 Two Old Babylonian Literary Texts. Pp. 185– 99 in Language, Literature and History: Philological and Historical Studies presented to Erica Reiner. American Oriental Series 67, ed. F. Rochberg-Halton. New Haven: The American Oriental Society. Tinney, S. 1996 The Nippur Lament. Royal Rhetoric and Divine Legitimation in the Reign of Isme-Dagan of Isin (1953–1935 B.C.). Occasional Publications of the Samuel Noah Kramer Fund, 16. Philadelphia. Volk, K. 1998 Die Balag-Komposition úru àm-ma-ir-ra-bi. Rekonstruktion und Bearbeitung der Tafeln 18 (19uˆ.), 19, 20 und 21 der späten, kanonischen Version. Freiburger Altorientalische Studien 18. Stuttgart: Franz Steiner. von Soden, W. 1938 Altbabylonische Dialektdichtungen. ZA 44: 26–44. Thureau-Dangin, F. 1925 Un hymne à Istar de la haute époque babylonienne. RA 22: 169–77. Westenholz, J. G. 1987 A Forgotten Love Song. Pp. 415–25 in Language, Literature and History: Philological and Historical Studies Presented to Erica Reiner. American Oriental Series 67, ed. F. RochbergHalton. New Haven: The American Oriental Society. 1997 The Legends of the Kings of Akkade. Mesopotamian Civilizations 7. Winona Lake: Eisenbrauns. Weidner, W. 1952/53 Die Bibliothek Tiglatpilesars I. AfO 16: 197– 215. Wilcke, C. 1974 Formale Gesichtspunkte in der sumerischen Literatur. Pp. 205–316 in Sumerological Studies in Honor of Thorkild Jacobsen on his Seventieth Birthday June 7, 1974. Assyriological Studies 20, ed. S. J. Liebermann. Chicago and London: The University of Chicago Press. 1985 Liebesbeschwörungen aus Isin. ZA 75: 188– 209. 1987 Die Inschriftenfunde der 7. und 8. Kampagnen (1983 und 1984). Pp. 83–120 in Isin-Isan Bahrÿyat III. Die Ergebnisse der Ausgrabungen

74

BRIGITTE GRONEBERG

1983–1984, ed. B. Hrouda. München: C.H. Beck’sche Verlagsbuchhandlung. 1993 Politik im Spiegel der Literatur. Literatur als Mittel der Politik im älteren Babylonien.

Pp. 29–75 in Anfänge politischen Denkens in der Antike. Schriften des Historischen Kollegs. Kolloquien 24, ed. Kurt Raa˘aub. München: Oldenbourg.

EVIDENCE FOR POST-GLOTTALIZED CONSONANTS IN ASSYRIAN1 N. J. C. Kouwenberg University of Leiden

1.1

seems to be a growing consensus that the latter case is most likely.3 There are many types of ejective consonants in the languages of the world. In Semitic, there is a diˆerence between Ethiopic languages such as Amharic and Tigrinya and modern South Arabian languages. The former have ejectives in which the glottal closure is simultaneous with the oral closure;4 the latter have stops and fricatives with a glottalized release following the oral articulation, for which M.-C. Simeone-Senelle uses the

It has been a controversial issue in the reconstruction of Proto-Semitic phonology whether the so-called emphatic consonants had the pharyngeal pronunciation characteristic of Arabic, i.e., with a constriction of the upper pharynx and concomitant velarization, or the ejective or glottalized pronunciation they have in Ethiopic Semitic and Modern South Arabian languages, i.e., with a concomitant glottal closure in the case of stops, or a glottal constriction in the case of fricatives.2 There

Phonology (Oxford: Blackwell, 1996), 23–24; J. Laver, Principles of Phonetics (Cambridge: Cambridge University Press, 1994), 171–72 (ejective), 326–28 (pharyngeal). 3. Adherents of the glottalized option in recent literature include A. Dolgopolsky, “Emphatic Consonants in Semitic,” Israel Oriental Studies 7 (1977) 1–13; A. Faber, Genetic Subgroupings of the Semitic Languages, (Ph.D. dissertation of The University of Texas at Austin, 1980), 114–70; A. R. Bomhard, “The Proto-Semitic Consonant System,” in Y. L. Arbeitman (ed.), FUCUS, A Semitic/Afrasian Gathering in Remembrance of Albert Ehrman, Current Issues in Linguistic Theory 58 (Amsterdam/Philadelphia: Benjamins, 1988), 115–19; I. M. Diakonoˆ, Afrasian Languages (Moscow: Nauka Publishers, 1988), 35; R. Stempel, Abriß einer historischen Grammatik der semitischen Sprachen (Frankfurt am Main: Peter Lang, 1999), 64–67. Earlier literature is listed in J. Aro, “Pronunciation of the ‘Emphatic’ Consonants in Semitic Languages,” Studia Orientalia 47 (1977) 5–18. A recent dissenting voice comes from E. Lipinski, Semitic Languages, Outline of a Comparative Grammar, Orientalia Lovanensia Analecta 80 (Leuven: Peeters, 1997), 105–6. 4. See for instance E. Ullendorˆ, The Semitic Languages of Ethiopia (London: Taylor’s (Foreign) Press, 1955), 153, and L. E. Kogan, “Tigrinya,” in The Semitic Languages, ed. R. Hetzron (London/New York: Routledge, 1997), 425.

1. Research for this paper was done as part of a project “The Akkadian verb and its Semitic Background,” which is ˜nanced by the Netherlands Organization for Scienti˜c Research (NWO). I would like to thank Prof. F. Kortlandt for comments on an earlier version, and Jan-Gerrit Dercksen and Klaas Veenhof for the same, and for providing me with additional (partly unpublished) Old Assyrian material. Abbreviations of text editions are those of The Assyrian Dictionary of the Oriental Institute of The University of Chicago (CAD), vol. 14 (R), with the following additions: KKK = K. Hecker, G. Kryszat and L. Matous, Kappadokische Keilschrifttafeln aus den Sammlungen der Karlsuniversität Prag (Prague: Filozo˜cká fakulta Univerzity Karlovy, 1998); PAOT = W. C. Gualtney Jr., The Pennsylvania Old Assyrian Texts (Cincinnati: Hebrew Union College, 1983); St. Veenhof = Veenhof Anniversary Volume, Studies presented to Klaas R. Veenhof on the Occasion of his Sixty-˜fth Birthday, Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul 89, eds. W. H. van Soldt et al. (Leiden: Nederlands Instituut voor het Nabije Oosten, 2001); TPAK = C. Michel and P. Garelli, Tablettes paléo-assyriennes de Kültepe, volume I (Kt 90/ k) (Paris: de Boccard, 2001). 2. For phonetic descriptions of these sounds, see for instance J. C. Catford, Fundamental Problems in Phonetics (Edinburgh: Edinburgh University Press, 1977), 68–71; A. Spencer,

75

JCS 55 (2003)

76

N. J. C. KOUWENBERG

term “post-glottalization.”5 Anticipating the conclusions of this paper, I will use the term postglottalized rather than ejective for the Assyrian consonants that we are concerned with here. However, there do not seem to be any clear-cut borderlines between diˆerent kinds of ejective or glottalized consonants—nor between such consonants on the one hand and clusters of consonant plus glottal stop on the other. All these types of phonemes appear to have a similar acoustic eˆect, namely that they are followed by a kind of glottal stop.6 Their IPA notation is by means of an apostrophe after the consonant symbol: [p’], [t’], [k’], [s’], etc. A dead language like Akkadian obviously has little if any direct evidence to oˆer in matters of pronunciation, yet it has played a minor role in the discussion. First, on the basis of the idea that the accumulation of glottalized consonants in a word is avoided, Knudsen has argued that dissimilatory phenomena operating in Akkadian, such as Geers’ Law7 and the change of the sequence q - q to k - q in words like qaqqadu or kaqqadu, “head,” suggest that the emphatics and q were glottalized in Akkadian.8

Second, Aro has argued that the peculiar Middle Assyrian (hence: MA) and Neo-Assyrian (hence NA) paradigm of the verb nasaåu, “to lift, to carry,” points to a glottalized pronunciation of the emphatic sibilant s.9 If the second and third radical of this verb come into contact as a result of the vowel elision rule,10 the resulting cluster /så/ is spelled as a simple or geminate <s> in MA and NA.11 This occurs in the stative (e.g., na(-as)-saku, “I carry,” for earlier nasåaku, na(-as)-su, “they carry,” for naså¿), in the perfect (e.g., it-ta-su, “they have lifted, brought,” for ittaså¿, it-ta-sa, “he brought here,” for ittasåa(m), and in the imperative (e.g., is-sa or i-sa, “bring (sg.) here!,” for isåa(m)).12 Since there is independent evidence for a sound change /s/ > /s/ in later Assyrian, this cluster was pronounced [så] from MA onwards,13 and the fact that it was written with a simple or geminate <s> suggests that /s/ was pronounced as a glottalized consonant, i.e., as [s’]. Thus Aro concludes that “Assyrian s before and after 1000 B.C. was pronounced more or less like s with a following (or concomitant) glottal stop.”14

5. M.-C. Simeone-Senelle, “The Modern South Arabian Languages,” in Hetzron, The Semitic Languages, 382. 6. See P. Ladefoged and I. Maddieson, The Sounds of the World’s Languages (Oxford: Blackwell, 1996), 369: “there is no sharp distinction between ejectives and plosives accompanied by a glottal stop.” Similarly, Catford, Fundamental Principles, 69. 7. Geers’ Law stipulates that in Proto-Semitic roots with two emphatics only one of these may remain emphatic in Akkadian, see F. W. Geers, “The Treatment of Emphatics in Akkadian,” JNES 4 (1945) 65–67. E.g., Akkadian sabatum “to seize” corresponds to Arabic dabata, and Akkadian qatnu “thin, small” to Hebrew qaton, see W. von Soden, Grundriß der akkadischen Grammatik, 3., ergänzte Au˘age, Analecta Orientalia 33 (Rome: Editrice Ponti˜cio Instituto Biblico, 1995), 65 § 51e. 8. E. E. Knudsen, “Cases of Free Variants in the Akkadian q Phoneme,” JCS 15 (1961) 84–90. His proposal was approved by Edzard, ZA 73 (1983) 134, and Bomhard, “The Proto-Semitic Consonant System,” 115–16. Knudsen’s other arguments for the glottalized nature of emphatics in Akkadian are refuted by Faber, Genetic Subgoupings, 146–47. For the tendency to avoid accumulation of glottalized consonants, see P. J. Hopper, “Glottalized and Murmured Occlusives in Indo-European,” Glossa 7 (1973) 141–66, and Faber, Genetic Subgoupings, 145 (who correctly points out, however, that these dissimilation

phenomena can provide no more than “meager, but not conclusive, evidence for ejectives in Akkadian” (147)). Moreover, F. Wedekind, “Glottalization Constraints and Ethiopic Counter Evidence,” Folia Linguistica 24 (1990) 127–37, cites numerous examples from Ethiopic languages that contradict the existence of such a tendency. 9. Aro, “Pronunciation of the “Emphatic” Consonants,” 8. See also R. Voigt, “A note on the Alleged Middle/Neo-Assyrian Sound Change så(*så) > ss<s>,” JNES 45 (1986) 53–57. 10. The vowel elision rule dictates that a short vowel in the penultimate syllable is elided if the preceding syllable is also short, see GaG 16–17 § 12, K. Hecker, Grammatik der Kültepe-Texte, Analecta Orientalia 44 (Rome: Ponti˜cium Institutum Biblicum, 1968), 68f § 44, and especially the extensive discussion in E. L. Greenstein, “The Phonology of Akkadian Syllable Structure,” Afroasiatic Linguistics 9/1 (1984) 13–42. 11. The discovery that the forms with s are part of the paradigm of nasaåu is due to Parpola (S. Parpola, “The alleged Middle/Neo-Assyrian Irregular Verb *nass- and the Assyrian Sound Change † > S,” Assur 1/1 (1974): 1–10). Parpola’s claim that na-sa stands for phonemic /nassa/ is justi˜ably criticized by Voigt, “Middle/Neo-Assyrian Sound Change,” 53–54. 12. For references, see Parpola, “The alleged Middle/NeoAssyrian Irregular Verb *nass-,” 6–10. 13. See section 9 with note 64. 14. Aro, “ ‘Emphatic’ Consonants,” 8.

EVIDENCE FOR POST-GLOTTALIZED CONSONANTS IN ASSYRIAN

The aim of this paper is to show that this peculiarity of the verb nasaåu ˜nds an exact parallel in Old Assyrian (hence: OA) in the paradigm of the verbs wasaåum, “to go/come out,” masaåum, “to be(come) able, su¯cient,” and kasaåum, “to be(come) cold,” and in the noun kusåum, “cold, winter.” This provides strong evidence for the post-glottalized nature of the emphatic sibilant /s/ in OA. I will further argue that a comparable peculiarity in the OA conjugation of the verb nadaåum, “to lay down,” does the same for the emphatic dental plosive /t/. 2. The verbs masaåum, wasaåum, kasaåum and nadaåum and the noun kusåum show a peculiar spelling pattern in those forms in which the second and third radical are adjacent. The resulting cluster of s or d plus aleph can be expressed in writing in three diˆerent ways. The ˜rst and most common way is a broken spelling, in which s or d is immediately followed by a V or VC sign.15 The second way, which is also very common, is what I will call the “alephless” spelling, in which the glottal stop (aleph) seems to be completely ignored. In some forms a third type of spelling may be used, which I will call the “glide spelling.” In a glide spelling the vowel that is subject to elision according to the vowel elision rule is preserved in the orthography. To give an example, one of the few verb forms that are found in all three spellings is the preterite usi of wasaåum when it is followed by an ending, such as the ventive ending -am: this form can be spelled us-am or usa-am with a broken spelling, ú-sa-am with an alephless spelling, and ú-sí-am or ú-sí-a-am with a glide spelling (see below for references).

15. This use of a V or VC sign in the middle of a word violates the basic syllabi˜cation rule that every syllable (at least in the middle of a word) should start with a consonant, and thereby reveals the presence of a preceding consonantal element. If we know from other sources that the word in question contains aleph, we are justi˜ed in assuming that the consonantal element to be supplied is aleph (otherwise, it could also be j or w).

77

The evidence for these spellings in the four verbs concerned is as follows:16 a) Masaåum, “to be(come) able, su¯cient,” frequently shows alephless spellings in the stative besides broken spellings, and once in the imperative plural besides more common glide spellings: Stative: 1s ma-sa(-a)-ku, Kienast ATHE 59: 18+, ma-sa-ku-um (< -ak(u)-kum) PAOT 18: 35; ma-saki-im (< -ak(u)-kim) Kienast ATHE 44: 22; 3sf ma-sa-at BIN 6, 30: 10, 1p ma-sa-a-ni BIN 6, 8: 14; 3sm subj. ma-sú-ni BIN 4, 76: 7+. The same forms also occur with broken spellings, e.g., 1s ma-as-a-ku PAOT 14: 33+, 1p ma-as-a-ni CCT 5, 3a: 28; 3pm subj. ma-as-ú-ni TCL 4, 27: 12, etc. No glide spellings of the stative (**ma-sí-a-ku, etc.) are known to me. Imperative plural: an alephless spelling is mìsa CCT 2, 41b: 16; more common is the glide spelling mì-sí-a AKT 3, 87: 14+. b) Wasaåum, “to go/come out,” has alephless spellings interchanging with broken spellings in the perfect, alephless spellings interchanging with both broken and glide spellings in the preterite, and a case of alephless spelling in the stative: Perfect: 3s vent. i-ta-sa-am CCT 3, 20: 32+, 2p subj. ta-ta-[s]a-ni BIN 4, 83: 14; 3pm vent. i-ta-súnim TPAK 1, 27: 16. Cf. the corresponding broken spellings i-ta-as-a(-ma) KKK 762: 8u+; i-ta-as-akum VS 26, 16: 7+ and i-ta-as-ú-nim ib. 6 (but ita-sa-am in line 5!). I am not aware of any glide spellings of the type **i-ta-sí(-a)-am. Preterite: 3sm vent. ú-sa-am AKT 2, 26: 26+; 3sm vent. subj. ú-sa(-a)-ni VS 26, 54: 14; KTS 1, 15: 25+; 2sm subj. tù-sú(-ni) AKT 3, 92: 3 and 6+; 2sm vent. subj. tù-sa-ni KKK 443: 28; 2sf vet. e la tú-sí RA 51, 7: 38; 1s prec. vent. lu-sa-ma AKT 3, 94: 36; 3pm vent. ú-sú-ú-nim KKK 683: 5; 3pm prec. lu-sú VS 26, 54: 28; 2p tù-sa TPAK 1, 57: 12. Most of these forms can be matched with corresponding broken spellings, such as 3sm vent. ús-a-am JCS 41, 37: 11+; 3sm vent. subj. us-a-ni JCS 14, 17 no. 11: 31+; 3pm us-ú ICK 2, 290: 5u;

16. “+” after a reference indicates that more instances of the form are attested; s = singular, p = plural, du = dual, m = masculine, f = feminine, vent. = ventive, subj. = subjunctive, prec. = precative, vet. = vetitive.

78

N. J. C. KOUWENBERG

3pm vent. ús-ú-nim KKK 762: 5u; 3pm prec. luus-ú RA 81, 14 no. 3: 12u+. Occasionally we ˜nd glide spellings of the preterite: 3sm vent. subj. ú-sí-a-ni TCL 4, 9: 4; úsí(-a)-am BIN 4, 112: 6, KKK 430: 43 and TPAK 1, 190: 6; 3sm subj. ú-sí-ú KKK 550: 4. This is also the normal spelling for the present ussiå¿, ussiåam, and some of the forms mentioned here may indeed be present rather than preterite.17 Stative: 3pm wa-sú kt k/k 102: 10.18 c) The verb kasaåum, “to be(come) cold,” has a 3pm stative (¿m¿) kà-sú, “the weather is (lit. the days are) cold” kt 91/k 530: 21,19 thus with an alephless spelling. In this case the main evidence comes from the derived abstract noun kusåum, “cold, winter.” It is once attested with a broken spelling: ku-us-um CCT 4, 29a: 10; elsewhere it has alephless spellings: ku-sú-um CCT 4, 45b: 6. 22+, gen. ku-sí-im JCS 14, 12 no. 6: 27+.20 d) Nadaåum commonly has alephless spellings in the stative, the perfect and the imperative of the G-stem and in the preterite of the N-stem, interchanging with broken spellings:21 Stative: 1s subj. na-dá-ku(-ni), 2sm. subj. na-dátí(-ni), 3sf subj. na-dá-at(-ni), 1p na-dá-ni, 3pm subj. na-du-ú KKK 574: 23+; na-du-ni CCT 4, 33b: 26+. These forms interchange with, for instance, 1s na-ad-a-ku CCT 1, 38b: 3+; 2sm na-ada-tí VS 26, 55: 28; 3sf na-ad-a-at BIN 4, 4: 12+; 3pm na-ad-ú VS 26, 56: 39+. Perfect: 3pm i-ta-du-ú kt n/k 204: 22;22 3du ita-dá. Corresponding broken spellings are 3du + 17. The easiest way to explain the unusual glide spellings of masaåum and wasaåum is to assume that they are caused by occasional interference of III/ÿ verbs, as a result of the fact that these two verbs belong to the vowel class i/i, which is the proper domain of III/ÿ verbs. In these forms aleph may have weakened into a palatal glide j (a common development for a glottal stop) with an epenthetic support vowel to dissolve the cluster: /misåa/ > /misja/, phonetically [mis ija]. 18. Unpublished, courtesy K. Hecker. 19. Unpublished, courtesy K. R. Veenhof. 20. Cf. Hecker, Grammatik, 46 § 28d. 21. Forms that contain the sign are unreferenced here, because a more complete enumeration including references will be given in section 5. 22. Quoted in J. G. Dercksen, The Old Assyrian Copper Trade in Anatolia, Uitgaven van het Nederlands HistorischArchaeologisch Instituut te Istanbul 75 (Leiden: HistorischArchaeologisch Instituut te Istanbul, 1996), 80 note 272.

accusative su¯x “me” i-ta-ad-a-ni (/ittadåani/) JCS 14, 8 no. 4: 35, and 3sm + -anni, “me,” i-ta-ada-ni (/ittadåanni/), Kienast ATHE 65: 29. Imperative plural: i-dá, but also with broken spelling id-a(-a) TCL 19, 31: 21+. Preterite N: 3sm subj. i-ni-du BIN 4, 31: 17. Normally broken spellings are used: 3sm subj. i-niid-ú(-ni) BIN 6, 28: 10; Jankowska KTK 68: 3; 3sm vet. + accusative su¯x –anni, “me,” (e) i-ni-id-a-ni KKK 433: 9, but there is also an instance of a glide spelling: 3sm subj.23 i-ni-dí-ú-ni Hecker Giessen 49: 8. No other glide spellings of the type **na-dí-a-ku, **na-dí-ú, **i-dí-a, **i-ta-dí-a-ni, etc. are known to me in the paradigm of nadaåum. 3. The peculiarity of the four verbs and the noun of section 2 consists in the fact that they are the only forms that regularly use broken spellings and alephless spellings. All other third weak verbs of OA almost exclusively use broken spellings and glide spellings, in slightly diˆerent proportions depending on the nature of the weak third radical. So there are two questions to be answered: ˜rst, why is the alephless spelling so common in these particular words and virtually unused elsewhere? Second, what is the relationship between the broken spelling and the alephless spelling, which at face value seem to be contradictory, since one of them indicates the presence of aleph (albeit indirectly), and the other one apparently ignores it? To answer these questions properly we ˜rst have to know what happens to post-consonantal aleph in general in OA. It is well known that aleph is a weak consonant that is often dropped or replaced by a glide. However, as a general rule aleph is preserved in Assyrian in the position with which we are concerned here, namely immediately following a consonant.24 This can be estab-

23. Construction obscure: possibly 3pm subj., an N-stem with active meaning like nasaåum N? 24. This is not easily inferred from the literature. The main handbooks on Akkadian and OA grammar, Von Soden’s GaG and Hecker’s Grammatik, are incomplete and contradictory

EVIDENCE FOR POST-GLOTTALIZED CONSONANTS IN ASSYRIAN

lished beyond any doubt from the regular use of broken spellings. The most telling evidence comes from the paradigm of the II/aleph and III/aleph verbs, especially if we restrict these terms to verbs whose aleph goes back to an etymological aleph (å1 in traditional Assyriological notation).25 The preterite of II/aleph verbs such as saåamum, “to buy,” saåalum, “to ask,” and maåadum, “to be(come) numerous,” shows the presence of post-consonantal aleph through its consistent use of broken spellings,26 e.g., is-a-am /isåam/ CCT 5, 22a: 23+ “he bought,” 3pm is-ú-mu /isåum¿/ ICK 1, 1: 25, ta-ás-e-li /tasåelÿ/ OIP 27, 57: 4 “you asked me,” e im-i-da-ku(-ma) /imåidakkum(ma)/ CCT 4, 6d: 21 “may it not become (too) much for you,” with plene writing im-i-id /imåid/ AAA 1, 56 no. 3: 18u. Relevant forms from other verbal stems include the imp. Gt of saåalum, “to ask,”: síit-a-al KTS 2, 34: r.10u, or sí-it-al KKK 472: 30 (both /sitåal/), and the †-stems lu-sa-am-i-id /lusamåid/ Archivum Anatolicum 3, 136: 64 “I will multiply,” and nu-sa-ás-a(-ma) /nusasåam(-ma)/ kt n/k 122: 13 “we will make (PN) buy.”27 Likewise, genuine III/aleph verbs such as kalaåu, “to detain,” malaåum “to be(come) full,” ka-

about the development of aleph in Assyrian (see note 44 below). This being said, data on post-consonantal aleph can be found for OA in Hecker, Grammatik pp. 42–44 § 27a/c; for MA in W. Mayer, Untersuchungen zur Grammatik des Mittelassyrischen (Kevelaer: Verlag Butzon & Bercker, 1971), 20–21 § 16, and for NA in J. Hämeen-Anttila, A Sketch of Neo-Assyrian Grammar, State Archives of Assyria Studies 13 (Helsinki: The Neo-Assyrian Text Corpus Project, 2000): 13. See also D. O. Edzard, “Silbenschließendes [å] im Altassyrischen?,” in St. Veenhof (see note 1) 133–35. 25. The re˘exes of the other Proto-Semitic gutturals (h, h, ç and À, i.e., å2–5) will be left out of account for the present moment. In a separate article which is in preparation (provisional title: “Aleph in Old and Middle Assyrian”), I will argue that—in contrast to what seems to be the communis opinio— the Proto-Semitic gutturals did not develop all in the same way in OA, but that only the pharyngeal plosive ç became aleph, whereas the other gutturals, which were all fricatives, became zero or a palatal glide. Therefore, broken spellings in words with original h, h and À (in so far as they are attested) are more likely to stand for j than for aleph, and cannot be adduced here as evidence for post-consonantal aleph. 26. K. Hecker, Grammatik, 152 § 91c. For the exceptions he mentions (the “weak forms”), cf. note 32. 27. Quoted in Dercksen, Copper Trade, 143 note 447.

79

taåum, “to take as a security,” tamaåum, “to swear,” and nasaåum, “to lift, to carry,” consistently show the presence of post-consonantal aleph in those forms in which the penultimate vowel is dropped as a result of the vowel elision rule. This applies, for instance, to the perfect forms ta-ak-ta-al-a /taktalåa/ KTS 2, 36: 32 “you (pl.) have detained,” and im-tal-ú /imtalå¿/ ICK 2, 147: 10u “they have become full,” to stative forms such as na-ás-a-ku /nasåaku/ BIN 6, 179: 4 “I have with me” (and similar forms passim), kà-at-at /katåat/ OIP 27, 35: r.9 “it (i.e., un¿tum, a collective for household utensils) has been taken as security,” from kataåum, and ta-am-ú /tamå¿/ TCL 20, 143: 4u “they have sworn,” from tamaåum, and to imperative forms such as ki-il5-i-su /kilåÿsu/ KT Hahn 6: 15 “hold (fem.) him!” and ta-am-a-am /tamåam/ CCT 4, 22a: 22 “swear to me!.” Post-consonantal aleph is also present in deverbal nouns of these II/aleph and III/aleph verbs, such as a pirs form nisåum, “gift,” from nasaåum in (ana) ni-is-e-em /nisåem/ KTS 1, 50c: 6, and timåum, “oath,” from tamaåum in ti-im-a-am /timåam/ KKK 735: 21u, and the mapras(t) forms namåudum, “majority, increase,” from maåadum28 and naråamtum, “love,” from raåamum, “to love,” as a feminine proper name Na-ar-am-tum KKK 467: 3.29 There seem to be very few genuine exceptions to the rule that etymological aleph in postconsonantal position is preserved in OA.30 In fact,

28. For references see CAD N/1 207b s.v. namåadu. 29. This name is also attested in a form without aleph: Nara-am-tum ICK 1, 17: 3 and cf. also Na-ra-am-diskur BIN 6, 29: 15 and Na-ra-am-[. . .] CCT 6, 2b: 7. These forms can be explained as examples of the well-known fact that Akkadian proper names show many kinds of shortenings that are not attested in other kinds of words. Numerous instances occur already in Old Akkadian, see the extensive account in M. Hilgert, Akkadisch in der Ur III Zeit, IMGULA 5 (Münster: Rhema, 2002), 65–79. This is exempli˜ed in OA by the many diˆerent forms of the name †alim-ahum, see, for instance, the index of AKT 3, 205 s.v. Another possible solution is to assume that they are in˘uenced by the common OA name Naram-Sîn (spelled Na-ra-am-ZU), which is doubtless a Babylonian name borrowed from that of the famous king NaramSîn of the Akkad dynasty. 30. In words in which aleph goes back to ç (see note 25) there is a stronger tendency for aleph to be dropped, but even

80

N. J. C. KOUWENBERG

apart from some instances in proper names alluded to in note 29, I am only aware of one pair of words in which this aleph may be dropped: the nouns ma/eråum, “son,” and meråutum, “daughter,” of which forms such as ma-ru-su TCL 4, 64: 7 “his sons,” and me-re-tí, “my daughter,” KTS 1, 1b: 3 are occasionally attested.31 Hecker mentions a number of other putative exceptions, but most of these are better explained otherwise.32 What is important is that Hecker’s conclusion “Nach Konsonanten stehen Formen mit und ohne Aleph nebeneinander” 33 misrepresents the actual state of aˆairs: indisputable forms in which post-consonantal aleph is not preserved are exceptional. 4. Consequently, among the forms in section 2 with a cluster of s or d plus aleph, those with a broken spelling are in agreement with the rule on post-consonantal aleph, whereas those with an alephless spelling seem to violate it. The problem is, therefore, how to explain these alephless spellings. If we omit nadaåum for the moment, they all have the emphatic sibilant s as their second radical. This can hardly be coincidental, so that any solution we might propose should in a non-ad

here the number of exceptions is small. Interestingly, most instances I know of are imperatives, a verbal category that is known for its predilection for short forms: 2sm vent. tí-ba(-ma) AKT 2, 30: 33 and KTS 2, 65: 3u “stand up!” (elsewhere tí-iba(-ma) /tibåam(ma)/ VS 26, 83: 47+) from tabaåum “to stand up, to rise” (Sem. TBç, if we may go by Ugaritic tbç, “to rise, to depart”), and sí-ma-(ma) TCL 19, 30: 9 “hear!” (pl.) (for normal sí-im-a-(ma) /simåa(ma)/, e.g., VS 26, 89: 3+ (pl.) from samaåum (Sem. †Mç). 31. For other instances see Hecker, Grammatik, 46 § 28d and I. J. Gelb, OIP 27, p. 21–23. For å1 in these nouns cf. their Arabic cognates imruåun, “man,” and imraåatun, “woman,” which become al-maråu and al-maråatu after a de˜nite article. 32. They are scattered over various sections of his Grammatik, especially p. 46 § 28d, p. 152 § 91c and p. 162 § 95d. Most of them can be explained as scribal errors and doubtful or incorrect interpretations of ambiguous forms. I intend to discuss them in detail in the article mentioned in note 25. 33. Hecker, Grammatik, 46 § 28d.

hoc way account for the fact that s plus aleph behaves in a diˆerent way from other consonants plus aleph. This can be attained by establishing a causal relationship between some inherent feature of s and the absence of a following aleph. Such a feature can be found in the post-glottalized pronunciation of s. If we assume that s was pronounced with a glottal closure following its basic alveolar articulation, in other words, as a post-glottalized consonant [s’] comparable to those in Modern South Arabian (see section 1), this explains both the irregular absence of post-consonantal aleph in these particular words and the use of the two seemingly contradictory spellings. For instance, the alternative spellings ku-us-um and ku-sú-um will then represent [kus’åum] and [kus’um] (or rather [kuss’um], see below), respectively. In kuus-um the original cluster [s’å] is expressed analytically, as it were, whereas in ku-sú-um it is expressed synthetically, with aleph subsumed in the post-glottalized consonant. Of course we have no idea of the actual pronunciation of OA, but we may hypothesize that the post-glottalized closure of [s’] and the subsequent glottal stop could merge into a single articulatory gesture, so that the acoustic eˆect of the cluster [så] was identical to, or at least very similar to, that of [s] by itself. It is clear from the phonetic literature that it is often hard to distinguish between consonants with a double or secondary articulation and parallel consonant clusters, such as the cluster /kw/ and the labio-velar /kw/; this also applies to glottalized or ejective consonants and the corresponding cluster of plain consonant and glottal stop.34 Since a cluster is prosodically equivalent to a long consonant, it seems plausible that /så/ interchanges with a geminate /s/, i.e., with [ss’] rather than [s’]. For OA this cannot be veri˜ed directly, because OA does not normally indicate consonant length, but it is suggested by the fact that in a form such as i-ta-sú-nim TPAK 1, 27: 16 there is no vowel assimilation (which would give **/ittu34. See, for instance, the quotation from Ladefoged and Maddieson in note 6, and also Catford, Fundamental Problems, 69–70.

EVIDENCE FOR POST-GLOTTALIZED CONSONANTS IN ASSYRIAN s¿nim/).35 This implies that the syllable with /ta/ is long; the easiest way to account for this is to assume that the consonant following a is geminated: /ittass¿nim/, phonetically [ittass’¿nim]. The ˜rst spellings with an explicit geminate appear in MA (see section 8).

81

In the same way as the verbs with s as their second radical provide evidence for a glottalized pronunciation of s, the alephless spellings of nadaåum provide evidence for a glottalized pronunciation of the emphatic dental plosive /t/. The crucial evidence lies in the fact that, if the alephless spelling is used and the cluster /då/ is followed by the vowel a, the resulting syllable is virtually always written with the sign . Cf. the following list of attestations:36 1) Stative: na-TA-ku BIN 4, 116: 5; TuM 1, 27b: 9; kt c/k 260: 33; 264: 2. 4. 9. 17. 25; na-TA-ku-ni RA 59, 43: 4; TCL 14, 22: 12; na-TA-tí Belleten 218, 44: 6; na-TA-tí-ni ICK 1, 1: 31; na-TA-at TCL 19, 81: 28; kt n/k 1270: 14; na-TA-at-ni CCT 4, 35b: 13; KT Blanckertz 10: 6; na-TA-ni KTS 1, 25b: 16.37 2) Imp. pl.: i-TA(-ma) Hecker Giessen 30: 19; TCL 19, 68: 13; Innaya p. 110 no. 79: 11; CTMMA 1, p. 105 no. 77: 29; RA 85, 106: r.4u; Or. 50, p. 103 no. 3: 25. 3) Perfect: i-ta-TA(-ma) (3du) kt a/k 339: 31.38

The only form of this verb in which a sequence /da/ occurs which is not from -dåa- is the in˜nitive nadaåum. Interestingly, in this form /da/ is spelled more often with than with : I know six instances of the genitive na-da (-i)-im (BIN 6, 113: 14; KKK 619: 8; TCL 4, 19: 26; TCL 14, 15: 7; JCS 14, 15 no. 10: 18; Innaya p. 413 LB 1226: r.9u), but only two of na-TA(-i)-im (Hecker Giessen 45: r.10u and AKT 3, 90: 16). Assuming that this is representative, it suggests that the alephless spelling with in the stative, imperative and perfect is not merely a convention connected with this particular verb, but has a real phonological background.39 When these forms are quoted in the dictionaries and elsewhere, is assigned the value (as I did in section 2), which seems the obvious thing to do. In OA the sign has two common uses: for /ta/ (with a voiceless dental plosive) and for /ta/ (with its emphatic counterpart), conventionally indexed ; it may also be used for /da/ (indexed ), but this is far less common; normally /da/ is rendered by .40 This means that there is nothing problematic about the use of in itself for the alephless forms of nadaåum: the question is, why these forms employ this sign with such remarkable consistency, and why they avoid with the same consistency. The answer is clear if we assign the normal value to the forms of nadaåum listed above

35. For the vowel assimilation rule, often inappropriately called vowel harmony in Akkadian grammar, see, for instance, Hecker, Grammatik, 19–21 § 10. Note that in the perfect of the G-stem (iptaras) only forms without ending are aˆected by this rule, e.g., imtuqut < *imtaqut “he has fallen,” but pl. imtaqt¿, because the presence of an ending causes the vowel elision rule to apply, which in its turn blocks vowel assimilation. This accounts for i-ta-sú-nim versus the regular 3sm i-tí-sí /ittisi/ÿ/ < *ittasiå (e.g., Hecker, Giessen 41: 6). 36. It is doubtless not exhaustive, but it is based on a su¯ciently extensive perusal of OA texts to be regarded as representative. Only one exception is known to me: AKT 2, 23: 19 i-da (imp. pl.); however, this text is only published in a (not always accurate) transcription, so that it may be a misspelling of i-dá. 37. Incorrectly quoted in Hecker, Grammatik, 167 § 97c as na-da-ni. 38. Unpublished, courtesy J. G. Dercksen.

39. If aleph is followed by u or i, there is no comparable diˆerence, because OA uses only one sign for any dental plus u or i (DU and DÍ). Instances of -dåu- were quoted above in section 2; the sequence -dåi- is only possible in the 2sf (imp. idåÿ, perfect tattadåÿ), of which I have found no instances in OA. 40. Cf. Veenhof, Aspects of Old Assyrian Trade and its Terminology, Studia et Documenta ad Iura Orientis Antiqui Pertinentia (Leiden: E. J. Brill, 1972), 226, and K. Hecker, “Altassyrisch DA-TA-an ‘sie gibt’,” in St. Veenhof (see note 1), 173– 79. Hecker has collected quite a number of instances of the use of with the value on pp. 174–77, but he insists that “ihre Zahl im Verhältnis zum Gesamttextbestand äußerst gering ist” (174). He succeeds in establishing some patterns and some possible motives, but the results of his study are in general inconclusive. One such pattern is the frequency of for /dá/ in the present iddan of tadanum “to give.” His suggestion that the spellings in the paradigm of nadaåum are caused by the occasional confusion of tadanum and nadaåum sounds rather desperate.

5.

82

N. J. C. KOUWENBERG

rather than the unusual value , and assume that emphatic t had a glottalized pronunciation. This turns the interchange of na-ad-a-ku and naTA-ku into an exact parallel of that of ma-as-a-ku and ma-sa-(a)-ku, so that it can also be interpreted as the interchange of a cluster of a plain dental plosive plus aleph and a post-glottalized dental plosive. It is no problem that there is no emphatic d in Akkadian, because in a cluster such as /då/ it seems likely that /d/ was devoiced through assimilation to å,41 so that we should actually interpret na-ad-a-ku phonetically as [natåaku].42 Thus the broken and the alephless spellings of this verb represent [natåaku] and [nat’aku] (or rather [natt’aku]) respectively, forms that are completely parallel to the “analytic” and “synthetic” spellings in the words with s as second radical. This means that we should transliterate the forms of nadaåum presented above as na-tá-ku, na-tá-at, i-tá, i-ta-tá, etc., and spellings with DU as na-tùni, etc., however strange this may seem at ˜rst sight. 6. In conclusion, the interchange of broken and alephless spellings in the forms listed in section 2 provides strong evidence for the claim that the emphatic consonants s and t had a post-glottalized pronunciation in OA. This claim establishes a direct causal link between the anomalous behavior of post-consonantal aleph in these particular words and the phonetic nature of the consonants

41. It is, however, far from certain that the diˆerence between t and d was one of voicedness rather than some other realization of a fortis vs. lenis distinction, see Lipinski, Semitic Languages, 104. Glottalized consonants are almost always voiceless, since the glottal closure prevents voicing, see Spencer, Phonology, 23; J. H. Greenberg, “Some generalizations concerning glottalic consonants, especially implosives,” International Journal of American Linguistics 36 (1970) 124–25. 42. Actually, this is conjectural: normally such assimilation processes are invisible because (C)VC signs do not distinguish between voiced, voiceless and emphatic in syllable-˜nal position, see Von Soden, Grundriß, 23 § 19a. The only cases of assimilation that may be visible in cuneiform writing are found in syllable-˜nal sibilants and nasals, see GaG 38 § 30f/g and 39–43 § 31–33.

involved, and thus explains this behavior in a non-ad hoc way. Of course it is possible to think of other solutions. We might claim, for instance, that there is an optional vowel contraction rule (as in later Old Babylonian) after /s/ and /d/ (/ittasâm/ < *ittasiåam and /nadâku/ < *nadiåaku),43 or that postconsonantal aleph may be elided exclusively after these consonants. Such solutions are conceivable from a theoretical point of view but are inferior, since they do not make it clear why this putative process of contraction or elision is restricted to this particular environment.44

43. There is no regular vowel contraction in OA, as there is in Babylonian, but it does occur sporadically, see Hecker, Grammatik, 31–33 § 19. Most instances concern contraction of identical vowels separated by a glide; Hecker mentions (la) ú-sé-lu < /useluå¿/. Note that another putative instance of vowel contraction mentioned by Hecker on p. 27 § 16f, sá-haa-at BIN 4, 34: 9, should be canceled: according to collation by K. R. Veenhof, the text actually reads em-sa-at “(the land) is hungry” (emasum). In any case, vowel contraction cannot serve to explain the systematic exception we are concerned with here. 44. A question that naturally comes to mind is what the handbooks on OA have to say about the alephless forms. As already intimated in note 24, their account of post-consonantal aleph in Assyrian is unsatisfactory. Hecker mentions some instances of alephless spellings in the appropriate sections of his Grammatik about the weak verbs, but does not comment on them. With regard to kusåum, he points to the two spellings ku-us-um and ku-sú-um and draws the following conclusion: “Wo es [i.e., post-consonantal aleph, NJCK] als 3. Radikal in pars-, pirs- und purs-Formen ausfällt, wird die vorausgehende (. . .) Silbe zum Ersatz gedehnt” (Grammatik, 46 § 28d). Thus, without stating this explicitly, he seems to interpret kusú-um as /k¿sum/ or /kussum/, as in Old Babylonian. Von Soden also discusses kusåum (Grundriß, 19 § 15b): he interprets ku-sú-um as k¿sum from *kusjum with compensatory lengthening of the preceding syllable and characterizes the alternative spelling ku-us-um as a “gelegentliche Ausnahme[n].” However, this is the exact opposite of what actually happened, because ku-us-um with aleph preserved is the form we expect, and ku-sú-um is in need of an explanation. He also comments on the perfect forms of wasaåum: ‘das anstelle des i getretene å fällt aA bisweilen aus, wobei der folgende Vokal gedehnt wird (s. z. B. (. . .) ittasam [wsÿ] “herauskommen sollte” TC 3, 25: 17 u (GaG 185 § 105j). Thus he posits a development ittasiam > ittasåam > ittasam. This is completely unacceptable: ˜rst, there is no evidence at all for any of these changes, nor for the long vowel in the ˜nal syllable of ittasam. Second, this contradicts his explanation of ku-sú-um, which implies lenthening of the preceding syllable. It seems that both Hecker and Von Soden are unduly in˘uenced by the

EVIDENCE FOR POST-GLOTTALIZED CONSONANTS IN ASSYRIAN

However, the fact that two competing spellings are used raises a di¯cult question: do these spellings represent diˆerent forms or are they merely orthographic variants of a single form? There does not seem to be a clear pattern in the distribution of the two spellings over the available OA material,45 but if we may speculate on the basis of other instances of competing spellings in OA, the most likely assumption seems to be that they do not represent diˆerent forms, but that the alephless spelling is the most accurate expression of the clusters /då/ and /så/ and that the broken spelling is morphophonemic.46 The fact that for the words in question broken spellings are no longer used in the later stages of Assyrian, when a different orthographic system is used, may provide additional support for this assumption (see further section 8). Finally, one further point should be mentioned here. Since for articulatory reasons the glottalized fricative [s’] tends to go back to a glottalized aˆricate [ts’],47 the post-glottalized pronunciation of /s/ proposed here oˆers further evidence for the claim that the sibilants s, z and s were aˆricates in Proto-Semitic and the older stages of Akkadian.48

situation in Babylonian, where post-consonantal aleph is regularly dropped with compensatory lengthening of the preceding syllable: *milåum > mÿlum “˘ood,” a pirs form of malûm, “to be(come) full,” (Semitic MLå) and where -ia- may contract to -â-. 45. In this respect it is remarkable that the writer of the letter VS 26, 16 uses two broken spellings followed by an alephless spelling of the perfect of wasaåum in three consecutive lines (quoted in section 2 s.v.). 46. See E. Reiner, Linguistic Analysis of Akkadian, Janua Linguarum, Series Practica 21 (The Hague/Paris: Mouton, 1966), 55–56. 47. See R. Voigt, “The Alleged Middle/Neo-Assyrian Sound Change,” 55–56 with further literature. For this phenomenon in Ethiopic languages, see G. Gragg, “Geçez Phonology,” in The Phonologies of Asia and Africa, Vol. 1, ed. A. S. Kaye (Winona Lake: Eisenbrauns, 1997), 174. 48. See especially A. Faber, “Akkadian Evidence for ProtoSemitic Aˆricates”, JCS 37 (1985) 101–7, the contribution of W. Sommerfeld to the 3rd edition of Von Soden’s Grundriß, 35–36 § 30, and the comments of J. Huehnergard in OrNS 66 (1997) 438–39.

83

7. Before turning to the later stages of Assyrian, it seems worthwhile to examine whether the same phenomenon is attested in other words in OA. It is signi˜cant that all four verbs involved come from genuine III/aleph roots, i.e., with etymological aleph. Masaåum and wasaåum have a reliable Semitic etymology: masaåum is doubtless cognate to Ethiopic masåa, “to come, to reach,” Aramaic mså, “to be able,” and Hebrew masa “to ˜nd.”49 Wasaåum, too, is generally derived from a root WS⁄å with aleph as third radical.50 The other two verbs do not have obvious cognates in other Semitic languages, but their primary aleph can be established on the basis of internal Akkadian evidence. Nadaåum shows the typical behavior of the III/aleph verbs in OA in that it has a strong preference for broken spellings and alephless spellings in the forms mentioned in section 2, whereas we would expect it to have a preference for glide spellings if it were a III/ÿ verb.51 Therefore, it seems justi˜ed to assume that nadaåum is a III/aleph verb as well.52 Kasaåum with its purs derivation kusåum has no known cognates in Semitic either, but is doubtless a III/aleph verb because it belongs (at least originally) to the vowel 49. For the variety of meanings of this root, the Akkadian verb kasadum, “to arrive, to reach, to acquire, to be su¯cient, to be able,” oˆers a striking parallel within a single language. See also J. Blau, “Marginalia Semitica II 4. Proto-Semitic mzå/ mtå ‘arrive’, ” Israel Oriental Studies 2 (1972) 67–72. W. Leslau, Comparative Dictionary of Geçez (Wiesbaden: Harrassowitz, 1987), 370 s.v. masåa only mentions Akkadian mussû “to make reach,” but this verb primarily means something like “to spread, to open wide,” (see AHw 1498a s.v.), and is a late form of wussûm, which rules out any connection with Semitic MTå. 50. See. AhW 1475b s.v., and W. Leslau, Comparative Dictionary, 605–606 s.v. wadåa. (Note that in his Grundriß Von Soden still considered wasaåum as a III/ÿ verb, see the quotation in note 44). 51. The only instance of a glide spelling of nadaåum known to me is i-ni-dí-ú-ni quoted in section 2 s.v. The absence of glide spellings is inexplicable if nadaåum is a III/ÿ verb. It is true that spellings with dí are very common for this verb, but only in forms that are not aˆected by the vowel elision rule, such as G pret. iddi, G present inaddi, all Gtn forms, etc. 52. Nadaåum is associated with III/ÿ verbs in other Semitic languages by Von Soden, AHw 705b s.v.: Hebrew jdÿ, Ethiopic wdj “to throw,” although the alleged cognates have w or y as ˜rst radical. Leslau, Comparative Dictionary, 605 s.v. wadaya does not mention Akkadian nadû as a cognate.

84

N. J. C. KOUWENBERG

class a/a; as far as we can tell, all third weak verbs with a/a have etymological aleph (cf. the ˜rst four III/aleph verbs mentioned as examples in section 3).53 Thus we should look for verbs with a dental or sibilant as second radical and etymological aleph as third radical. Apart from the four verbs discussed here, there are two other verbs that meet this condition, as far as I am aware: kataåum “to take as security,” and hataåum “to sin, to commit an error.” 54 However, among the few relevant

53. The dictionaries assign kasaåum to the vowel class i/i on the basis of the Standard Babylonian form i-ka-as-si Oppenheim, Glass Fragm. c 11u+. However, in Standard Babylonian forms with a occur as well: 3sm present G i-ka-sa TDP 224: 52+, Gtn ik-ta-na-as-sa TDP 220: 34. There is also a MA 3pm present i-ka-su-ú in E. C. Cancik-Kirschbaum, Die mittelassyrischen Briefe aus Tall †eh Hamad, Berichte der Ausgrabung Tall †eh Hamad/D¿r-Katlimmu, Band 4, Texte 1 (Berlin: Dietrich Reimer Verlag, 1996), 117 no. 6: 11u, which cannot be from *ikassi(å)¿, because i + ¿ does not contract in MA, see Mayer, Untersuchungen, 18 § 12 and the examples on pp. 81– 83 § 80. It points either to u/u (for which there are no other indications), or to a/a (which agrees with the Standard Babylonian a/a forms). Thus i-ka-su-ú stands for /ikassuå¿/ < *ikassaå¿ with vowel assimilation. This verb is one of the many third weak u/u and a/a verbs that occasionally show i/i forms in Standard Babylonian and later. 54. All other common third weak verbs with a sibilant or dental as second radical seem to have j, w or å2–5 as third radical. For instance, sasaåum, “to call,” and mataåum, “to decrease, to be(come) insu¯cient,” are III/ÿ. Pataåum, “to open,” (Semitic PTH) and masaåum, “to wash, to purify” (Semitic MSH) both have å3, and verbs with å3 as third radical have adopted the conjugation of the III/ÿ verbs (see note 24). Broken spellings of III/ÿ verbs conceal j rather than aleph: pá-at-a-at CCT 4, 18a: 8 is /patjat/ “it is open” and sí-is-a TC 1, 25: 6 is /sisja/ “summon!” (imp. pl.). There is also an adjective pasjum “white” (Babylonian pesûm). It has no etymology, but the fact that it is mostly written with glide spellings suggests that it is from a III/ÿ root, e.g., pá-sí-um TCL 21, 161: 1; pá-sí-ú-tim/tum Kienast, ATHE 66: 23+; a few broken spellings also occur: páas-ú-tim ICK 1, 92: 2+ (with <ú> for /ju/). A di¯cult case is radaåum “to follow, to accompany.” J. Huehnergard, “Further South Semitic Cognates to the Akkadian Lexicon,” in: Semitic Studies in Honor of Wolf Leslau, Vol I, ed. A. S. Kaye (Wiesbaden: Harrassowitz, 1991), 698–99, suggests that the rather complex meaning of redûm re˘ects the falling together of two or even three Proto-Semitic roots attested in South Semitic, one of which has aleph: Ethiopic radåa 1. “to help, support”, 2. “to pursue, hunt down”. This is semantically attractive but raises certain formal di¯culties if we compare the extant OA forms with forms from Old Akkadian and Old Babylonian. Akkadian rather suggests a root RDç. If that is correct, the imp. with

forms of these verbs and their derivations no alephless spellings are known to me.55 With regard to the nature of the phoneme q, for which a (post)glottalized pronunciation may safely be assumed,56 I have found no evidence in OA comparable to the evidence for s and t. There seem to be no verbs with a velar as second and aleph as third radical which could theoretically show alephless spellings.57 8. In so far as these verbs are attested in the later stages of Assyrian, they show the same kind of behavior as in OA. For s, MA oˆers the following relevant forms: From usaåu (corresponding to OA wasaåum) we ˜nd a preterite 3sm vent. subj. (kÿ . . .) ú-sa-an-ni AfO 17, 277: 54 “(when) he came forward,” and from masaåu a 2sm stative ma-sa-ta VAS 19, 15: 13, “you are able.” Both

ventive ri-[d]am RA 58, 126: 22 “bring here!,” which alternates with ri-id-a(-ma) /ridåam(ma)/ BIN 4, 112: 31+), can be explained along the same lines as the alephless spellings of nadaåum: /ritt’am/. 55. From hataåum, a verbal adjective hatåum is attested with a broken spelling: ana kaspim ha-at-i-im “for bad silver” TCL 4, 104: 7; from kataåum we only have the stative form quoted in section 3 s.v. 56. As for instance in the third edition of GaG 34 § 28a. The validity of this assumption follows from typological evidence alone: according to I. Maddieson, Patterns of Sounds, Cambridge Studies in Speech Science and Communication (Cambridge: Cambridge University Press, 1984), 105 and 120, almost all languages that have a dental ejective stop [t’], also have a velar ejective stop [k’]. I do not think that the NA assimilation of qt > qt can be regarded as evidence for a change of pronunciation from glottalized to pharyngealized, as J. Huehnergard, OrNS 66 (1997) 438, suggests. If we assume that q was a postglottalized velar plosive, this change represents a development from suppression of the post-glottalized release of q in qt (thus phonetically [kt]) towards a natural cluster [kt’] which is post-glottalized as a whole. 57. Laqaåum, “to take, to receive,” seems an obvious example, but it has å3 (Semitic LQH) and thus is conjugated as a III/ÿ verb, cf. note 54 (thus the imp. pl. li-iq-a KKK 433: 15 is /liqja/). Naqaåum “to sacri˜ce” is a III/ÿ verb according to the predominance of glide spellings in the deverbal pirs form niqjum “sacri˜ce”: e.g., accusative singular ni-qí(-a)-am (versus 1x ni-iq-a-am according to CAD N/2, 254 s.v. niqû b-2u), both standing for /niqjam/.

EVIDENCE FOR POST-GLOTTALIZED CONSONANTS IN ASSYRIAN

85

forms are exactly parallel to the OA forms of these verbs listed in section 2 and may be interpreted as /ussanni/ and /massata/ respectively. Note that there is no vowel contraction rule in MA, so that these forms cannot be derived from **usi(å)anni and **masi(å)ata.58 In addition, there are two instances of kusåum, one of which oˆers the earliest proof that this s was geminated: nom. ku-us-su Ebeling Wagenpferde 38: 17a, and gen. (sa) ku-se ib. 16 Ac 2. The same type of form is still used in the corresponding NA forms: from masaåu there is a stative 1s subj. ma-as-sa-ku-ni SAA 10, 294: r.28, a 1p subj. ma-as-sa-ni-ni SAA 10, 289: r.13 and a 3pm ma-su SAA 10, 278: r.5. Also the noun kusåu is often spelled with <ss> in NA, e.g., ku-us-su and ku-us-si-im-ma ABL 302: r.1–3, ku-us-si SAA 10, 236: 5.59 To my knowledge relevant forms of the verb kasaåu are not attested in MA and NA. Preterite forms of usaåu with explicit <ss> are very rare in NA: I have found a preterite 3sm subj. us-su-u-niº SAA 10, 67: 14, and a precative 3pm vent. lu-us-su-u-ni SAA 10, 259: r.12. Similar forms without explicit <ss> include 3pm ú-su SAA 15, 119: s.1 and 3sm vent. ú-sa SAA 8, 28: 4u.60

The verb nadaåum has disappeared from normal usage in MA and NA. However, a 3rd p. fem. stative form is found in MA in a legal formula in which a tablet is quali˜ed as ana hi-pi na-DA-at (once na-TA-at), i.e., “has to be broken,” lit. “is put down for breaking.”61 The use of the obsolete verb nadaåu demonstrates that it is a traditional formula; this is con˜rmed by the single instance of na-TA-at, a spelling which is directly inherited from OA, because the value of the sign is unknown in MA.62 In the other extant instances the spelling is “modernized” by means of , which is the usual sign for /ta/ in MA; so we can still interpret the forms phonetically as [natt’at].63 In one important respect, however, the situation has changed: instead of the two alternative spellings current in OA, only the alephless spelling is used in MA and NA. This suggests that the competition between the “analytic” pronunciation of /så/ and /då/ as a cluster and their “synthetic” pronunciation as a post-glottalized consonant has been decided in favor of the latter. Alternatively, if we assume that the two spellings were of a purely orthographic nature, this means that the

58. See Mayer, Untersuchungen, 18 § 12 and 82–83 § 80e. All MA statives of third weak verbs mentioned by Mayer either have a glide spelling (e.g., la-qi(-a)-at) or contain an explicit aleph (e.g., sa-al-å-at, ta-am-å-a-ta). This clearly demonstrates the special status of ma-sa-ta caused by the presence of s. 59. The MA and NA forms of this noun are identical with the Babylonian form kussu(m), but according to the proposal made here they have a quite diˆerent background. In Babylonian, post-consonantal aleph is dropped with compensatory lengthening of the preceding vowel, see GaG 19 § 15b (see note 44); therefore we expect k¿sum in Babylonian. Nevertheless, the most frequent spelling shows a geminate s: ku-ussú-(um), etc., AHw 516a s.v. D. O. Edzard plausibly explains this form on the model of its antonym ummum “heat, summer” (ZA 73 [1983] 135). 60. In NA several preterite forms of usaåu occur in which the original stem vowel i appears again, e.g., ú-si-a SAA 1, 160: r.1, SAA 13, 158: 8u, CTN 5, 120: 9 and ú-si-ú in CTN 5, 130: 1. Rather than interpreting these as instances of glide spellings, as in OA, I prefer to explain them as representatives of a new preterite conjugation built on the endingless forms, such as 3sm and 2sm forms usi and tusi. By analogy with the regular verb, e.g., iqbi “he said” å iqbi¿ “they said” and iqbia “he said to me,” new forms of the type si¿ “they went out” and usia “he (or: I) came out” were created from usi/tusi as alternatives to the inherited forms uss¿ and ussa, which were not

only irregular in structure but perhaps also too poorly diˆerentiated from the corresponding present forms ussuå¿ and ussaåa. (The original OA present ussi was replaced in MA and NA with ussa, in which a is subject to vowel assimilation, see Mayer, Untersuchungen, 88–89 § 81 sub 5.). 61. See K. Deller and C. Saporetti, “Documenti MedioAssyri redatti per annullare un precedente contratto,” OA 9 (1970) 29–59. Na-DA-at occurs on the pages 30: 11; 31: 24; 33: 15; 35: 20; 36: 12; 37: 10; 39: 9; na-TA-at on page 40: 12. An additional instance in fem. pl. is VAS 19, 38: 11 na-DA-a (in CAD N/1, 141a s.v. nahru B, erroneously referenced as VAS 19, 21: 22 (I am grateful to Klaas Veenhof for revealing the correct reference to me). 62. Apart from one (other) exception, see Mayer, Untersuchungen, 8 sub no. 102. 63. J. J. Finkelstein, JCS 7 (1953) 127b, D. J. Wiseman, Iraq 30 (1968) 178, and AhW 768b s.v. natû III 3, derive these forms from the adjective natûm, “˜t, suitable,” (178). The correct derivation is from nadaåu (see Deller and Saporetti, “Documenti Medio-Assyri”, 44, and CAD N/1, 88a s.v. nadû v. 2h); this is proved by instances from Old Babylonian Susa (quoted in CAD l.c.) and Larsa (RA 85 [1991] 17 no. 5: 8–9 tuppu ana hepê na-di (I am grateful to Klaas Veenhof for pointing out this instance to me).

86

N. J. C. KOUWENBERG

phonetic spelling as a glottalized consonant has ousted the morphophonemic spelling as a cluster. 9. However, the potential of MA and NA with regard to the issue at hand is not yet exhausted. If the emphatic consonants did indeed have a glottalized pronunciation in Assyrian, the peculiar paradigm of nasaåu in MA and NA as drawn up by Parpola (see section 1) is no longer an isolated oddity. The spellings with simple or geminate s of the forms of this verb that in OA have a /så/ cluster are quite parallel to the forms with a /så/ and a /då/ cluster discussed in this paper. Thus Aro’s observation quoted in section 1 that these spellings indicate a glottalized sibilant is fully substantiated. The diˆerence from the other cases is that they ˜rst appear in the MA period and are dependent on the occurrence of the sound change /s/ > /s/. Most evidence for this change is derived from transcriptions of Assyrian names in Hebrew and Aramaic sources and dates from NA times.64 The forms with s in the paradigm of nasaåu prove that this sound change must have started as early as the MA period, as indicated by Parpola, but that OA /s/ was not yet pronounced [s]; otherwise we should have found spellings with s already in OA.65 In this respect their appearance in MA is 64. See Parpola, “The Alleged Middle/Neo-Assyrian Irregular Verb *nass-” (see note 11), 2; F. M. Fales, Aramaic Epigraphs on Clay Tablets of the Neo-Assyrian Period, (Rome: Università Degli Studi “La Sapienza,” 1986), 61–63 and J. Huehnergard, OrNS 66 (1997) 439–40, with further literature. 65. Perhaps it was a fricative lateral, as argued by R. C. Steiner, The Case for Fricative-Laterals in Proto-Semitic, American Oriental Series 59 (New Haven: American Oriental Society, 1977) 144–48. According to C. Girbal, “Zur Phonologie des Akkadischen,” AoF 24 (1997) 172–81, <s> was pronounced /s/ already in Proto-Semitic. Although he does not mention OA explicitly, he seems to imply that this applies to OA as well. This is contradicted by the facts presented here.

an independent development, a new stage in a repetitive process in which a new form is aˆected by a change as soon as it meets the phonological requirements. Finally, nasaåu does not show any broken spellings beside the spellings with s: there is no **naas-ú beside na-(as)-su in MA or later. 66 This indicates that also in this verb the competition between the “analytic” and the “synthetic” treatment of the cluster was decided in favor of the latter. 10. The results of this paper can be summarized as follows: 1. In OA the “emphatic” phonemes s and t were post-glottalized ([s’], [t’]) as in Modern South Arabian rather than pharyngealized as in Arabic. This supports the view that in Proto-Semitic the emphatic consonants were (post-)glottalized, too. 2. The forms with alephless spellings of the OA verbs wasaåum, “to go/come out,” masaåum, “to be(come) able or su¯cient” kasaåum, “to be(come) cold,” and nadaåum, “to lay down,” cannot be construed as exceptions to the general rule that in OA aleph is preserved in post-consonantal position. 3. The use of <s> in the spelling of the cluster /så/ in the MA and NA paradigm of nasaåu, “to lift, to carry,” is parallel to the alephless spellings in the verbs just mentioned, and testi˜es to the sound change /s/ > /s/ in MA. 4. The post-glottalized pronunciation of /s/ supports the view that the sibilants s, z and s were actually aˆricates.

66. See Parpola, “The alleged Middle/Neo-Assyrian Irregular Verb *nass-,” 2.

EVIDENTIARY PROCEDURE IN THE MIDDLE ASSYRIAN LAWS Raymond Westbrook The Johns Hopkins University

Introduction1

the rapist’s punishment is imposed on the basis of a mere accusation, without su¯cient evidence. Cardascia attempts to re˜ne Driver and Miles’ approach, suggesting that the two verbs represent a gradation of evidence: burru means to gather su¯cient evidence for a presumption of guilt; kunnu means to provide irrefutable proof.6 In this way, Cardascia is able to account for ˜ve of the cases where burru is used alone (A §§3, 14, 18, 19; N §2) as having a weaker meaning, as where a man brings charges that he is unable to sustain.7 He further manages to eliminate two (A §25 and §36), by asserting that the verb has a special meaning, “to claim (property).”8 Nevertheless, there still remain two instances that Cardascia is forced to label true exceptions: A §12, the case of the rapist already mentioned, and A §7, a parallel case in which witnesses testify that a woman laid a hand on a man. Furthermore, if Cardascia’s theory of gradation is followed, the dual expression would seem super˘uous: what is the point of mentioning a prima facie case when irrefutable proof is demanded?

The phrase ubtaååer¿(s) uktaååin¿(s) appears (with slight variations) some twenty times in the Middle Assyrian Laws (MAL).2 Both verbs are well known, having the same basic meaning in the D-stem (“make ˜rm”). Nonetheless, the exact meaning of the couplet has proved elusive. In their classic commentary, Driver and Miles called it “a curious phrase” and translated “they have brought charge (and) proof against him.”3 In their interpretation, the subject is always the accusers or witnesses. The verb ubtaååer¿ refers to an accusation and uktaååin¿ to adducing su¯cient evidence.4 They admit, however, that the presence or absence of the procedure in the Laws seems arbitrary. More di¯culty is caused by the fact that in nine instances the verb burru appears alone, as in A §12, where a man is to put to death for rape if witnesses ubtaååer¿s.5 Following Driver and Miles’ interpretation, this would mean that 1. This article was ˜rst presented as a paper at the annual meeting of the American Oriental Society in San Diego on March 14, 2004, and in longer versions to the Institut für Orientalistik, Universität Wien, on March 19 and the École Pratique des Hautes Etudes, Paris, on May 25, 2004. I am grateful to all the participants for their helpful comments and criticisms. The usual caveats as to responsibility for errors apply. 2. Tablet A §§1 (/si), 9, 15, 16, 20, 21, 40, 47 (/sunu), 53; B §§4, 8, 9, 13, 14, 20; C+G §§3 (/su), 8 (/su), 10 (/su), 11 [restored] (/su); E §1; L §3. 3. G. R. Driver and J. C. Miles, The Assyrian Laws (Oxford: Clarendon Press, 1935), 339. 4. The Assyrian Laws, 341–43. 5. The other cases are: A §§3, 7, 14, 18, 19, 25, 36; N §2.

6. Les Lois Assyriennes (Paris: Cerf, 1969), 94–95 n. c. 7. Even this interpretation requires adding nuances to the literal meaning. In §18, anaku ubaååer baåura la ilaåe la ubaååer, literally “(saying) ‘I will make ˜rm, but he cannot make ˜rm, he does not make ˜rm” becomes “. . . je porterai des charges . . . , (s’il) ne peut porter des charges (et s’)il n’a pas chargé (su¯samment).” 8. “réclamer.” Cardascia’s translation of §36 actually reads (inadvertently?) “il prouvera (les faits)” (col. V 1).

87

JCS 55 (2003)

88

RAYMOND WESTBROOK

Two more recent translations adopt a diˆerent approach. Roth translates “they prove the charges against him and ˜nd him guilty”9 and Borger similarly “man es ihm beweist und ihn überführt.”10 As Driver and Miles already noted, this translation requires a shift in subject between the two verbs, between the witnesses (who prove) and the court (that convicts). They therefore rejected this solution, on the grounds that it was more likely that both verbs would have the same subject. 11 The shift in subject has the merit of accounting satisfactorily for the use of burru alone, but at the same time leaves the dual phrase with more than a whiˆ of super˘uity about it. It stumbles, however, at the hurdle of A §1, where the two verbs are used as alternatives: lu-ú ub-ta-e-ru-ú-[si] lu-ú uk-ta-i-nu-[si] (col. I 6– 7). This cannot logically apply to two successive, complementary stages in the judicial process. The context is that a god is consulted to determine the punishment of a woman who has committed sacrilegious theft. A rule that the procedure applies either when the charges have been proved or when she has been found guilty is absurd. Borger, translating correctly, walks into the trap: “man es ihr beweist oder sie überführt.” Roth, sensing the logical di¯culty, translates here: “they prove the charges against her and ˜nd her guilty.” In the same way, one interpretation in CAD takes the phrase as hendiadys: “they prove it of her by means of witnesses KAV 1 i 8 (Ass. Code §1), and passim in this phrase in the Ass. Code,” as if it were the same form as elsewhere in the code.12 The Assyrian form l¿ . . . l¿, however, can only mean “either . . . or.”13

9. M. T. Roth, Law Collections from Mesopotamia and Asia Minor, Writings from the Ancient World 6 (Atlanta: Scholars Press, 1995), 153–94 ad loc. 10. R. Borger, “Die mittelassyrichen Gesetze,” in Texte aus der Umwelt des Alten Testaments I/1: Rechtsbücher, ed. R. Borger et al. (Gütersloh: Gerd Mohn, 1982), 80–92 ad loc. 11. The Assyrian Laws, 341. 12. Vol. K 168 sub mng. 4; similarly AHw 109a bâru(m) III sub mng. D 4b: “meist vor uktaååin¿s als Hendiadyoin: sie ihn (sie) völlig überführen.” 13. Cf. A §3 (col. I 23–24): sum-ma lú lu-ú ma-ri-is lu-ú me-et. A man cannot be ill and dead at the same time.

One Line Short

Methods of Proof If the dual phrase is not to be dismissed as meaningless repetition,14 a separate role must be found for it that leaves the individual verbs with discrete functions and neither con˘icts with those functions nor renders them super˘uous. I propose the following hypothesis: the couplet ubtaååer¿(s) uktaååin¿(s) refers to the burden of proof on the accuser and more speci˜cally to the diˆerent methods of proof required to discharge that burden in court. In its use in MAL, burru means proof by rational methods, such as witnesses or documents, whereas kunnu means proof by suprarational methods such as the oath or the ordeal. It should be noted that these conclusions are unique to MAL. The combination of burru and kunnu is attested elsewhere only in reverse order, in two literary compositions in the Standard Babylonian dialect. In †urpu it occurs in a list of sins, in a context that reveals nothing of the two verbs’ function; they could simply be synonyms (II 60: ukannu ubarru usasbaru). Their context in the Assyrian Dream-Book, on the other hand, implies that the verbs are cumulative, not alternative. The omen in question reads (330 r. ii 40–41): If he walks constantly in dark waters, a hea[vy] law-suit [ ], they will sum[mon him] ana kunni u burri. Because the lawsuit is a di¯cult one, the witness will be obliged to furnish more than one mode of proof. The reverse order of the verbs in these late sources is signi˜cant, as it may re˘ect an earlier diˆerence between the Assyrian and Babylonian dialects. In Old Babylonian sources, the two verbs, appearing individually, have the same functions as proposed here, but with exactly the opposite application: burru is used of proof by supra-rational

14. This is the role assigned to it by CAD vol. B 130: “if they establish and prove that . . .”

EVIDENTIARY PROCEDURE IN THE MIDDLE ASSYRIAN LAWS

means, whereas kunnu is mostly used of proof by rational means.15 Only two forms of supra-rational evidence are positively attested in Mesopotamia, namely the evidentiary oath and the river ordeal. The ordeal is far less common than the oath and its procedure is not well understood.16 Both are imposed by the court and are decisive. The evidentiary oath is imposed on one side— on the party or his witnesses or both.17 If taken, it decides the case in favor of the oath-taker. It is, however, a burden as well as a privilege, for a person may be reluctant to expose himself to the wrath of god or king through a self-curse. Thus there are many attested instances of the imposition of the oath leading to a last-minute compromise settlement.18 On the other hand, the danger arising from the oath’s self-curse is merely potential, which might tempt a hardy or desperate soul to risk its consequences.19

15. See E. Dombradi, Die Darstellung des Rechtsaustrags in den altbabylonischen Prozessurkunden, 2 vols, FAS 20 (Stuttgart: Franz Steiner, 1996), 335 for the evidence from private legal documents. The distinction is clear-cut in Codex Hammurabi: burru: §§23, 120, 126, 240; kunnu: §§1, 2, 3, 5, 42, 106, 107, 108, 113, 116, 124, 127, 133, 141, 194, 255, 265, 282. In §2, if witchcraft is not proved, then they resort to the river ordeal. 16. Although there are references to the river ordeal from all periods (for a survey of the sources, see HANEL [= A History of Ancient Near Eastern Law, ed. R. Westbrook (Leiden: Brill, 2003)], 155, 196–97, 375–76, 495–96, 529, 575–76, 891, 925), the only accounts of the procedure come from OB Mari: Durand, “L’ordalie,” in AEM I/1 (= ARM 26), 509–39. It sometimes involved an agreement between the parties. An oath or oaths may also have been required: see F. Joannès, “La pratique du serment à l’époque néo-babylonienne,” in Jurer et maudire, ed. S. Lafont, Méditerranées 10–11 (Paris: L’Harmattan, 1997), 172–73. 17. For a detailed account based on the Old Babylonian sources, see E. Dombradi, Die Darstellung des Rechtsaustrags in den altbabylonischen Prozessurkunden I, 330–34. See also S. Lafont, “La procédure par serment au Proche-Orient ancien,” in Jurer et maudire, ed. S. Lafont, Méditerranées 10–11 (Paris: L’Harmattan, 1997), 185–93. 18. See e.g., CT 4 47a; 48 1 (Old Babylonian). 19. An Old Babylonian prayer to Nanna documents a case of successful perjury: UET 6 402, edited by D. Charpin, Le Clergé d’Ur au Siècle d’Hammurabi (Geneva: Droz, 1986), 326–29.

89

The ordeal applies to parties, not to witnesses. It appears to have been imposed on one side, but could be undertaken by substitutes.20 As with the oath, there are cases where a party refuses to undergo the ordeal. Unlike the oath, it involves immediate risks and immediate results, which would make the subject of the ordeal even more reluctant to undergo the procedure. The court has a discretion whether to impose the oath or ordeal and on which side. The most common form of oath is an exculpatory oath on the defendant. Otherwise, the sources, especially the law codes, refer to particular circumstances in which the court imposes an oath or ordeal, but never articulate the principles underlying the exercise of its discretion. The Verbs burru and kunnu Most of the cases in which the two verbs appear in MAL, whether as a couplet or burru alone, do not provide us with su¯cient context to determine what sort of evidence they refer to, whether rational or supra-rational. A few cases in which burru appears alone, however, indicate that rational evidence is meant. In A §17 a man makes the assertion that another man’s wife is notoriously promiscuous: If a man says to a man, “Your wife is always being slept with,” and there are no witnesses, they shall make a contract and go to the River. In the absence of witnesses, that is, of rational evidence, supra-rational evidence is expressly prescribed, in the form of the ordeal. In A §18 a man makes the same allegations, but then rashly declares that he will prove it (burru): If a man says to his companion, whether in private or in a dispute, “Your wife is always being slept with,” and “I will prove it,” but he cannot and does not prove it (a-na-ku ú-ba-ar ba-ú-ra

20. Durand, “L’ordalie,” 518–21.

90

RAYMOND WESTBROOK

la-a i-la-å-e la-a ú-ba-e-er), that man shall be beaten with forty strokes . . . The proof must refer back to what was stated to be missing in §17, namely the testimony of witnesses. In A §12 a man is charged with seizing a married woman as she is passing in the street and raping her. He is to suˆer the death penalty “whether he is caught upon the married woman or witnesses prove that he had intercourse with the woman” (col. II 20–22: lu-ú i-na ugu dam-lú ik-su-du-us ù lu-ú ki-i munus i-ni-ku-ú-ni se-butu ub-ta-e-ru-us). In my view, the protasis is canvassing the two standard types of rational proof: seizing in the act and testimony ex post facto.21 There would be no need to impose an oath on the witnesses, whose credibility is not in issue. This law is not about setting a minimum level of proof. The point being made by the apodosis is that the death penalty applies whether the culprit is caught in the act or not. The context is the principle commonly found in ancient legal systems that being caught in ˘agranti delicto aggravates the oˆence.22 It is a stylistic feature of MAL to give a list of alternatives in order to show that the ruling is indiˆerent to them.23 I shall return to this point below.

21. A third type, documentary, applies essentially to transactions. A confession might also be regarded as a form of proof. See S. Lafont, Femmes, Droit et Justice dans l’Antiquité orientale (Fribourg: Editions Universitaires, 1999), 165. Seizure in the act does not at ˜rst sight look like an independent type of evidence, since there will still be testimony about it at the time of trial. It becomes su¯cient proof in itself when it is such as to exclude deniability by the culprit. See below. 22. E.g., a thief caught in the act, as in CH §25, suˆers the death penalty. Cf. the case of Laban’s stolen idols in Gen. 31: 30–32, where the death penalty is to be imposed on the thief caught in possession after a hot pursuit. The distinction is expressed in early Roman law as between manifest and nonmanifest theft. See F. de Zulueta (ed.), The Institutes of Gaius Pt. I (Oxford: Clarendon Press, 1946), III §§184–85, 189–90. 23. Compare the list of diˆerent circumstances of a rape in A §55: “whether in the town or in the country or at night in the street or in a granary or at a town festival . . .” As Landsberger put it, they signify “Tatort gleichgültig”: “Jungfräulichkeit,” in Symbolae David II, ed. J. A. Ankum, (Leiden: Brill, 1968), 63 n. 1.

A §25 makes a clear distinction between rational and supra-rational modes. The text reads (col. III): 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94.

sum-ma munus i-na é a-bi-sa-ma us-bat ù mu-us-sa me-e-et ses-mes mu-ti-sa la-a ze-e-zu ù dumu-sa la-ás-su mi-im-ma du-ma-a-qé sa mu-us-sa i-na ugu-sa is-ku-nu-ú-ni la hal-qú-ú-ni ses-mes mu-ti-sa la-a ze-zu-ú-tu i-laq-qé-ú a-na re-ha-a-te dingir-mes-ni ú-se-et-tu-qú ú-ba-ar-ru i-laq-qé-ú a-na didi-id ù ma-mi-te la-a is-sa-ab-bu-ú-tu

If a woman is dwelling in her father’s house and her husband dies, her husband’s brothers have not divided, and she has no son, her undivided husband’s brothers may take any jewelry that her husband placed upon her that is not lost. As regards the remaining (property?), they shall cause the gods to move past (it) and shall prove and take (it). They shall not be seized for the River God or the oath. The brothers must engage in a procedure involving the gods, the content of which is not clear, but would seem to engage the gods as witnesses. It is not an oracular procedure, for which diˆerent terminology is used, as we shall see. Nor is it one of the two known supra-rational means of proving an assertion in court, for they are expressly excluded. The brothers must in any case still prove (burru) their claim to the items of property. The conclusion must be that burru alone is used when rational means of proof are intended. Finally, A §1 is the sole source in MAL in which both burru and kunnu appear together but not as a couplet. The text reads (col. I): 1. [su]m-ma munus [lu-ú] dam lú 2. [lu]-ú [dumu-mu[nus lú] 3. [i-na] é-dingir [t]e-ta-ra-ab

EVIDENTIARY PROCEDURE IN THE MIDDLE ASSYRIAN LAWS

4. 5. 6. 7. 8. 9. 10. 11. 12. 13.

[i-na] é-dingir [mi]-im-ma [. . . ] tal-ti-i[-]ri-iq [(lu-ú) i-na qa-ti-sa?] is-sa-bi-[it] lu-ú ub-ta-e-ru-ú-[si] lu-ú uk-ta-i-nu-[(ú)-si] ba-e-ru-ta [. . . ] [d]ingir i-[s]a-åu-[ú-lu] [k]i-i sa-a [dingir a-na e-pa-se] i-[q]a-ab-[bi-ú-n]i e-ep-pu-su-ú-si

If a woman, whether the wife of a man or the daughter of a man, enters a temple and steals something . . . , (and) it is seized [in her hand], or ubtaååer¿si or (else) uktaååin¿si, [they shall perform] a divination and shall inquire of the god. They shall treat her as [the god] orders. Notes 5. Coll. Freydank, AoF 21, 204. 6. The dubious verbal form led Postgate to suggest (tentatively) iz-za-qa[p], with an unattested intransitive meaning “come to light, turn up” (Iraq 35, 21). Otto proposes the Neo-Assyrian verb zaqapu B, “vor Gericht erscheinen (um Klage zu erheben)” but then translates “völlig überführt” (“Die Einschränkung des Privatstrafrechts . . . ,” in Biblische Welten, ed. W. Zwickel, OBO 123, Freiburg 1993, 158 n. 49). Claiming, however, is far from proving, and the form would in any case be izzaqqup. 7. A duplicate fragment (Postgate, Iraq 35, 19– 21) reads ù lu-ú ub-ta-e-ru-si [. . . 9. On the form, see Deller AfO 34, 65. In this paragraph, three diˆerent types of proof are presented as alternatives: being caught in possession of the stolen goods or perhaps in the act of stealing, and two others. In the light of the paragraphs previously discussed, it is reasonable to suppose that the remaining two would be the testimony of witnesses, and the results of a suprarational procedure, possibly an ordeal. Thus MAL presumes three standard modes of proof: seizure in ˘agranti delicto, ex post facto testimony, and referral to the divine court. As we have seen in A

91

§12, the ruling itself concerns punishment, not proof. Its point is that whatever means were used to prove the woman’s guilt, her punishment is to be determined by consulting the oracle. The Couplet ubtaååer¿(s) uktaååin¿(s) A preliminary objection to my interpretation of the dual phrase is that the oath, and more particularly the ordeal, would normally be imposed on a single person, whereas the verbs in the dual phrase are in the plural. The answer is that the third person plural of the verbs should be understood as an impersonal form. Consequently, the unspoken subject may and often will be an individual: the plaintiˆ, a witness, or the victim (as in A §16, where the victim testifying is the plaintiˆ’s wife), whoever they may be. A §40: 94–106 reveals the individual behind the impersonal phrase: If a man sees a slave woman veiled and lets her go; he does not seize her and bring her to the gate of the palace; ubtaååer¿s uktaååin¿s; he shall be struck with ˜fty strokes of the rod, his ears shall be pierced, threaded with a cord, and tied at his nape. His denouncer (batiqansu) will take his garment. He shall do the king’s corvée for a month. The denouncer is thus the hidden subject who has to testify as to the accused’s dereliction of duty and to con˜rm his testimony by submitting to the oath or ordeal. A more serious objection to my interpretation is that it has not solved the problem of redundancy. If the method of proof is to be a suprarational procedure, why require rational evidence as well? The answer lies in what I will call the condition of threshold credibility. Supra-rational procedure is the removal of the case from the human tribunal and its referral to a divine tribunal. The human court has ˜rst to decide under what circumstances it will refer a case, then which procedure is to be applied and which assertion is to be tested, the plaintiˆ’s or the defendant’s (or their witnesses’). The most obvious reason for a referral is the absence of evidence.

92

RAYMOND WESTBROOK

For example, in an Old Babylonian suit by a seller reclaiming property from a buyer, “The royal judges gave Haliyaåum (the buyer) to the oath because they had ˜xed a day and his witnesses could not be present.”24 Nonetheless, in many cases the court would only reach its decision after considering the evidence. Thus in another Old Babylonian case, where one heir sued the purchaser of land from another heir, the plaintiˆ relied on a tablet of division of the estate, but the defendant adduced witnesses to the existence of a later division. “The judges heard their statement (saptisunu) that there was a later division and the judges ordered them to speak their testimony (sibussunu) before the god. . . .”25 In a royal decision from Ugarit, the content of the witnesses’ evidence determines who is to take the oath (ll. 17–34): Takia (the debtor) said as follows: “I paid the son of Zibaia (the creditor) the 800 shekels of silver that I owed him and I have witnesses: Attanu and Aiu.” If his witnesses say, “(The debt) is settled: Takia paid the 800 shekels of silver to the son of Zibaia,” then let Takia swear together with his witnesses and let the son of Zibaia relinquish his claim. And if Takia’s witnesses say it is not so, let the son of Zibaia swear together with his (Takia’s) witnesses and let Takia pay him his silver.26 The court might therefore decide that for an assertion to be tested by supra-rational procedure, 24. OECT 13 91: 12–15: di-ku5–lugal ha-li-ia-ú-um as-sum u 4– mu is-ku-ú-ma si-bi-su ú-la i-ba-as-su-ú-ma a-na ni-is dingir [i]d-di-nu-su. The example is Old Babylonian. Unfortunately, there are no litigation documents from the Middle Assyrian period to illustrate the oath or ordeal in practice, but there is a rich fund of sources from Mesopotamia in other periods (especially the Old Babylonian period), and the basic principles of supra-rational procedure appear to remain constant throughout the cuneiform record. 25. BE 6/2 49: 25–29. The plaintiˆ thereupon agreed to a compromise settlement rather than allow the witnesses to swear. 26. RS 20.22 = Ugaritica V no. 27: 5–34. For a legal analysis of this and the preceding text, see S. Loewenstamm, “The Cumulative Oath of Witnesses and Parties in Mesopotamian Law,” in Comparative Studies in Biblical and Ancient Oriental Literatures, AOAT 204 (Kevelaer: Butzon & Bercker, 1980), 341–45.

the person making the assertion must ˜rst have attained a certain threshold of credibility, whether through testimony, documents or material evidence. The condition is more appropriate to the oath than to the ordeal, because the oath was applicable to a wider range of persons and was open to abuse by a party with a weak case or by a dubious witness. The immediate peril of the ordeal would have acted as more of a deterrent against the temptation of perjury. Nonetheless, it was not an absolute deterrent, and threshold credibility could be regarded as necessary for the ordeal, depending on the circumstances of the case.27 In many instances where the dual phrase is used in MAL, it is reasonable to infer that the court heard evidence before deciding whether to impose the oath or the ordeal and on whom. Thus the landowner who claims that a neighbor has encroached upon his land would need to bring some prima facie evidence of title before the court would impose the oath (or ordeal) on him rather than on his neighbor who also claims the land (B §§8, 9, 13, 14, 20).28 Again, the victims of fraudulent practices in C+G §§8, 10 and 11 would need ˜rst to show that the property of which they were defrauded was theirs. C+G 3 imposes a punishment on a creditor who sells abroad a son or daughter held as pledge, unless the debt equals their full market value.29 The procedure ubtaååer¿s uktaååin¿s is invoked against him. Aside from the question of sale abroad, the debtor, who is the father of the pledged child, needs to show that the child was given as a pledge and not sold as a slave, and furthermore that the child was not pledged for its full value. The issue is one of the

27. Lafont argues that A §17 shows that rational and irrational modes of proof were in theory mutually exclusive, because the ordeal is prescribed in the absence of witnesses (Femmes, Droit et Justice, 259). As I shall hope to show below, the paragraph represents a much narrower principle. 28. Where boundaries are ˜xed by tradition of usage they are hard to prove; there will be no deeds of sale, and boundary stones are notoriously movable. 29. The object of the latter clause, an Assyrian man or woman, refers to an Assyrian citizen, not an inferior social class. See S. Lafont, “Middle Assyrian Period,” in HANEL, 530–31.

EVIDENTIARY PROCEDURE IN THE MIDDLE ASSYRIAN LAWS

state of accounts between the creditor and debtor, for which some rational evidence would be required, before allowing the debtor to swear an oath (or submit to an ordeal). Supra-rational procedure, therefore, does not necessarily eliminate the need for rational evidence as a preliminary step. It follows that the dual requirement in the couplet ubtaååer¿(s) uktaååin¿(s) is not redundant. It is a gradation, although not in the sense that Cardascia meant. The law could simply have prescribed an oath or ordeal, deeming that the circumstances warrant it. It does so in A §17 where, as we have already seen, the ordeal is expressly provided where there are no witnesses. The law thus orders (or empowers) the court to apply a supra-rational procedure instead of rational evidence. The dual phrase is not used. I would argue a contrario that the dual phrase was used only in cases where the law chose not to eliminate the preliminary step, but to require (or perhaps to allow) the court to proceed to the ordeal (or the oath) once that threshold requirement had been satis˜ed. By far the most common application of the oath or ordeal was exculpatory, to allow a defendant to disprove an accusation. It is not surprising, therefore, that most of the paragraphs of MAL that prescribe the oath or ordeal in explicit terms refer to an exculpatory procedure. In six paragraphs, the law imposes an oath using the verb tamû. In two of these (B §10; C+G §1), the tablet is too broken to establish the context, but in A §§22 and 56, an exculpatory oath is imposed upon the accused. In A §5, the oath is imposed upon the accuser, but is analogous to an exculpatory oath, in that the alleged victim of theft must prove that he did not incite the theft. The ˜nal instance, A §47:16–17, also concerns an unconventional oath that is exculpatory in purpose: a hearsay witness exonerates himself (zaku) by proving what an eyewitness told him. Likewise, of the three cases of ordeal in MAL, at least two (A §§22 and 24) are undergone by the accused and are clearly exculpatory. The dual phrase, by contrast, refers exclusively to proof in support of an accusation. Herein, I suggest, lies both the rationale for its existence—

93

to indicate an application of supra-rational procedures that was less common—and the rationale for its duality. It is precisely because the oath or ordeal was being imposed upon an accuser, not a defendant, that threshold evidence was regarded as necessary.30 It is signi˜cant that A §17, which as we have seen imposes the ordeal not by the dual phrase and explicitly excludes threshold evidence, also appears to impose the procedure on the accuser.31 The Dual Phrase in Action The rationale for prescribing the procedures indicated by the dual phrase may be illustrated by contrasting two fairly straightforward cases from MAL, one that uses the dual phrase and one that uses burru alone. A §20 reads: If a man has intercourse with his companion and ubtaååer¿s uktaååin¿s, they shall have intercourse with him and turn him into a eunuch. A man claims that he was sodomized by an associate. Two factors aˆect the evidentiary procedure to be applied. First, there would be no outside witnesses to such an intimate sexual act. Second, the evidence of the accuser himself is a statement against interest. As Daube has pointed out, the oˆence is not homosexuality at large but imposing upon a free man the despicable passive role in intercourse, treating the victim sexually like a woman. Hence the punishment of the oˆender—to be treated like a woman and turned

30. A similar distinction exists in CH: tamû is used where the oath is exculpatory (§§206–7, 227), ina mahar ilim burru for accusations or claims of lost property. 31. Driver and Miles argue that the ordeal was undergone by the wife, in order to clear herself of suspicion, which would make it exculpatory (The Assyrian Laws, 68–69). In order to maintain that interpretation, they are obliged to regard the statement about the wife not as slander but as a friendly word to the husband. As Cardascia points out, the maker of the statement must be the same as in the following paragraph, where his allegation is de˜nitely an unfriendly act (Les lois assyriennes, 127–29).

94

RAYMOND WESTBROOK

into a woman.32 For the victim to admit publicly that he had played the passive role, whether willingly or unwillingly, would be deeply shaming. Accordingly, the personal, uncorroborated evidence of the victim will cross a threshold of credibility—not enough to prove his accusation outright, but enough to persuade the court to refer it to a divine tribunal, for example by imposing the oath upon him.33 A §36 considers the case of a man who has gone abroad, or perhaps just to the provinces (a-na a-sà i-it-ta-la-ak), apparently on business, and fails to leave his wife with provisions or to send her provisions from abroad.34 After ˜ve years she is free to remarry, and if the husband subsequently returns he has no right to reclaim his wife. There is one exception, however (col. IV 103–V 3): If he is delayed beyond ˜ve years but was being held against his will or a . . . seized him and he ˘ed or he was seized as a criminal and delayed, on his return he shall prove it (ú-baa-ar) and give a woman like his wife, and he may take his wife. The absentee claims that he is entitled to take his wife back because his absence beyond the statutory limit was due to force majeure, in particular judicial measures taken against him. Again, two factors in˘uence the evidentiary procedure prescribed. First, other evidence would be available, even though the events in question took place elsewhere, since they would have been witnessed by fellow merchants or be demonstrable from court records.35 Second, he has a strong incentive 32. “The Old Testament Prohibitions of Homosexuality,” ZSS 103 (1986) 447–48. 33. In this and the following example, I will assume for the sake of argument that the more common supra-rational procedure, the oath, was intended. 34. Lit.: “to the ˜eld.” See Cardascia’s discussion of the context: Les lois assyriennes, 187–89. 35. Sources from the Old Assyrian Period show that the Assyrian courts had great experience in dealing with the problems of witnesses or parties absent abroad and the time needed to gather evidence. See Veenhof, “The Old Assyrian Period,” in HANEL, 443–45.

One Line Short

to lie in his own self-interest. Accordingly, the claimant’s personal testimony su¯ces neither to prove his claim outright nor to cross the threshold necessary for referral to a divine tribunal. He must prove his case by rational means alone; he is not to be granted the bene˜t of an oath. Two more complex cases, both revolving around the de˜nition of seizure in the act, illustrate further details of the dual procedure in practice. I distinguished above between three standard types of evidence: seizure in ˘agranti delicto, testimony, and oath or ordeal. Seizure in ˘agranti does not at ˜rst sight look like an independent type of evidence, since there will still be testimony at the time of trial. It becomes su¯cient proof in itself when it is such as to exclude deniability by the culprit. Lack of deniability arises from the circumstances of the seizure, which may cover a broad spectrum of time, place, or actions: in a public place, in a private place where there can be no innocent explanation for the culprit’s presence, in possession of incriminating material in the immediate aftermath of the crime. Thus the rapist caught in the act in public by a crowd of bystanders has no deniability (MAL A §12), nor does the burglar caught trespassing, especially at night (LE §§12–13), nor the thief in whose possession stolen goods are found after a hue and cry and formal search (Gen. 31:22–35). Seizure in the act may fail to exclude deniability, on the other hand, where it is done by an interested party alone, or where an innocent explanation is available for the presence of the accused, or where too much time has elapsed between the crime and the seizure. The court will then need to revert to other modes of proof in order to establish the link between seizure of the accused and the alleged crime. The law on witchcraft in A §47 rules that seizure of incriminating materials does not exclude deniability. In laying down further evidentiary requirements, it reveals the key elements inferred by the dual phrase: duplication of testimony and oath by an individual witness in support of an accusation. The opening lines read (col. VII 1–6):

EVIDENTIARY PROCEDURE IN THE MIDDLE ASSYRIAN LAWS

If a man or a woman practices witchcraft and they (the items) are seized in their hands;36 ubtaååer¿sunu uktaååin¿sunu; the makers of witchcraft shall be killed. Witchcraft was regarded as a danger to public safety. In the war against witchcraft, the palace needed to be informed and could take extreme steps to identify and apprehend sorcerers. At the same time, false accusation was all too easy, since the link between cause and eˆect was intangible. Seizure of “incriminating” items was not enough; the accuser had to testify as to the practice of witchcraft and also take an oath or submit to an ordeal. The ˜nal section of the law con˜rms that he had in fact to take an oath. The eyewitness has informed another who in turn informs the palace but the eyewitness then denies his statement. Evidently fear of the sorcerer deters the eyewitness from testifying. The palace, however, has a remedy (ll. 20–23): The king shall interrogate him as he is able and shall see his back (lugal ki-i i-la-å-ú-ni ilta-na-å-al-su ù ku-tal-lu-su e-em-mar) and the exorcist shall cause the man to speak when he is puri˜ed . . . The meaning of this enigmatic phrase is in my opinion that the king will interrogate the reluctant witness under torture, so as to counteract fear of the sorcerer with an equal terror.37 If the purpose were merely to gain information, there would be no problem, but it emerges from the continuation of the law that as a result of his “interrogation” the eye-witness has sworn an oath by the king and the prince (ll. 24–31):

36. i-na qa-ti-su-nu is-sa-ab-tu. The reference is to magical preparations or products: see Cardascia, Les lois assyriennes, 230–31. 37. The use of torture in judicial interrogation is attested in Neo-Babylonian times, when a “ladder of interrogation” was used to extract confessions: M. Jursa, “Akkad, das Eulmas und Gubaru,” WZKM 86 (1996) 199, 210. The phrase “see his back” could refer to a reversal of his denial. I am grateful to Professor Jursa for this suggestion.

95

He (the exorcist) shall say: “He will not release you (pl.)38 from the oath that you swore to the king and his son. You swear in accordance with the tablet that you are swearing to the king and his son.” The exorcist is on hand to reassure the witness that the sorcerer will not be able to release him from the curses attendant upon his oath, which the sorcerer might perhaps justify on the grounds that the oath was not given voluntarily. From our point of view, the signi˜cance of the law is that it reveals that not only must seizure of incriminating evidence be corroborated by testimony, but that the testimony itself must be con˜rmed by oath. Con˜rmation by testimony and oath of what at ˜rst sight seems to be covered by seizure in ˘agranti delicto may also be behind a use of the dual phrase in A §15 that has puzzled commentators. In this case, it is a cogent parallel that supplies the context. The paragraph reads (col. II: 41–57): 41. sum-ma lú is-tu dam-ti-su lú is-sa-bat 42. ub-ta-e-ru-ú-us 43. uk-ta-i-nu-ú-us 44. ki-la-al-le-su-nu-ma 45. i-du-uk-ku-su-nu 46. a-ra-an-su la-ás-su 47. . . . 41–46: If a man seizes a man with his wife, ubtaååer¿s uktaååin¿s, it is both of them that shall be killed; he has no liability. 47–57: If he seizes them and brings them to the king or the judges, ubtaååer¿s uktaååin¿s; if the woman’s husband will kill his wife, then he shall kill the man. If he will cut oˆ his wife’s nose, then he shall turn the man into a eunuch and his whole face shall be mutilated; if he will let his wife oˆ, he shall let oˆ the man.

38. Possibly a reference to the oath-taker’s family, who would be included in the curse.

96

RAYMOND WESTBROOK

A husband catches his wife “with” another man, that is, strictly speaking not in ˘agranti delicto but in circumstances that could be construed as indicating an adulterous relationship. The distinction may seem overly subtle, but is con˜rmed by the brutally graphic language of A §12, where the rapist is caught “upon” (ina muhhi) the woman.39 The second part of the law not unreasonably requires the husband to prove the case against the man before a local or royal court (ubtaååer¿s uktaååin¿s) and also to refute any suspicion of entrapment by punishing his wife and her lover in equal measure (ll. 47–57). The ˜rst part of the law creates a di¯culty in that seems to allow the husband to exercise summary justice (ll. 41–46), but with the same requirements, including ubtaååer¿s uktaååin¿s. Driver and Miles postulated an informal trial with the neighbors as witnesses, citing Greek parallels.40 In an earlier study on the law of adultery, I conjectured that the proving might refer to a procedure ex post facto of the husband killing the lovers on the spot, in a trial in which he was the defendant.41 The result would be that any oath sworn by the husband would be exculpatory. The conclusions of the present study, that ubtaååer¿s uktaååin¿s refers only to an accusatory procedure, render my earlier conjecture, which was di¯cult to reconcile with the syntax, untenable.42 At the same time, they reveal a closer parallel in the Gortyn Code, one of the Greek sources mentioned by

39. Driver and Miles noted the diˆerence in terminology, but dismissed it as unimportant (The Assyrian Laws, 48); Cardascia took the same view but was troubled by the fact that proof by witnesses should also have been required in a case of seizure in ˘agranti delicto (Les lois assyriennes, 120 note a). Native sources were conscious of the distinction, however. The report of a trial from Old Babylonian Nippur relates that the husband actually tied his wife and her lover caught in the act to the bed and brought them, bed and all, to the court. See S. Greengus, “A Textbook Case of Adultery in Ancient Mesopotamia,” HUCA 40–41 (1969–1970) 33–44. Cf. also Lafont, Femmes, Droit et Justice, 68, arguing in favor of the distinction. 40. The Assyrian Laws, 45–46, 50. 41. “Adultery in Ancient Near Eastern Law,” Revue Biblique 97 (1990) 552–53. 42. See also the criticisms of Lafont, Femmes, Droit et Justice, 69–70.

Driver and Miles, which gives added credibility to the scenario that the latter postulated.43 The Gortyn Code discusses the case of a man seized while committing adultery (moikion) with a free woman in the house of her father, brother, or husband (col. II: 20–28).44 The text continues (28–45): Let him (the captor) before three witnesses declare to the relatives of the one seized that he is to be ransomed within ˜ve days. . . . If he is not ransomed, it is for the captors to do (with him) as they wish. If he (the captive) claims to be the victim of fraud, the captor is to swear . . . with ˜ve others, each cursing himself solemnly, that he was taken committing adultery, and was not the victim of fraud. It emerges from this account that the adulterer is held by the accuser, who makes a formal declaration before witnesses but does not bring him before a court. If the accused claims, presumably before those witnesses, that it was a fraudulent trick, then the captor must swear a declaratory oath. Apparently this procedure also takes place in the captor’s home, where the accused is being held, not before a court. It thus conforms with Driver and Miles’ reconstruction of the circumstances in the ˜rst part of MAL A §15. From our point of view, what is important is that this extrajudicial procedure is twofold: a declaration before witnesses of the circumstances of the seizure and a supra-rational stage.45 43. Ed. R. Willetts, The Law Code of Gortyn (Berlin: De Gruyter, 1967). 44. The Greek term moicheia covers a broader scope than adultery in the ancient Near Eastern sense, being illicit intercourse with a woman under the authority of any close relative. 45. The Gortyn law requires oaths not only by the captors but also by four others, whom I take to be members of the captor’s family who would be suspect of conspiring with him to entrap an enemy (and who probably stood to pro˜t with him from the ransom). The older view that they were oathhelpers whose sole task was to establish the principal witness’s credibility has been disproved. See M. Gagarin, “The Function of Witnesses at Gortyn,” Symposion 1985, 29–54, esp. 51–52. A further illuminating feature of the Gortyn law is its shift between single (captor) and plural (oath-takers) subject,

EVIDENTIARY PROCEDURE IN THE MIDDLE ASSYRIAN LAWS

There is, however, an important diˆerence between the Greek and the Assyrian law. The Greek oath is in fact two-fold: an accusatory positive oath that the oath-taker seized the accused in adultery and an exculpatory negative oath that it was not a fraudulent scheme to entrap the accused. Adultery was an oˆence against the husband; if he agreed to another sleeping with his wife, he would have no claim against him.46 The second oath is not present in the Assyrian version because a diˆerent safeguard is employed against entrapment, namely the requirement (emphasized by the enclitic –ma in line 44) that both the wife and the lover be killed. The result is the same: the husband will be immune from future suit by the relatives of the man he killed, a suit in which he would be the defendant. Nonetheless, in the present extra-judicial procedure the oath is accusatory, not exculpatory.47

and impersonal verbs. A similar background would account for the switch between singular and plural verb forms in MAL A §15 that has disturbed commentators: see Driver and Miles, The Assyrian Laws, 48–50, and Lafont, who argues on the basis of the 3rd person plural verb forms in lines 42–45 (ubtaååer¿s uktaååin¿s . . . idduk¿sunu) that the ˜rst part of A §15 is a formal procedure before the court just like the second part (Femmes, Droit et Justice, 71–72, 90). As I have argued above, these forms can not only be impersonal but can hide an individual subject. 46. See Westbrook, “Adultery,” 564–68. 47. Note that the husband’s immunity is described as aransu lassu (cf. A §59), whereas in A §47: 17 the exculpatory oath makes the oath-taker zaku.

97

The Assyrian procedure is a hybrid born of the special circumstances of the case. The husband is allowed to execute summary justice on the lovers without recourse to a court. As a safeguard against abuse, he must ˜rst demonstrate their guilt in two stages. Firstly, by rational means, namely presentation of the lovers to witnesses in the compromising circumstances in which they have been found. This is insu¯cient to prove the case de˜nitively (should the alleged paramour deny wrongdoing) but is su¯cient to cross the threshold of credibility, thus opening the way to proof by supra-rational means, presumably an oath. In that respect there is a trial, before a divine tribunal. The procedure may therefore justi˜ably be called ubtaååer¿s uktaååin¿s, in spite of its informal setting. Conclusions MAL express the requirement of a suprarational evidentiary procedure in two ways. They either set out the procedure explicitly, whether by oath or ordeal, or they employ the enigmatic phrase ubtaååer¿(s) uktaååin¿(s). The latter is used in order to substantiate an accusation, when both rational and supra-rational means of proof are required. Accordingly, I would propose as a (non-literal) translation of the phrase ubtaååer¿(s) uktaååin¿(s) wherever it occurs in MAL the following: “(if) the burden of proof against him(/her/ them) has been discharged by all means, human and divine.”

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK IN THE LIGHT OF THE TEXT NBC 48971 Stefan Zawadzki Adam Mickiewicz University, Poznan

1

Animal husbandry in Neo-Babylonian times has been the subject of many important studies, including the pioneering treatise by M. San Nicolò2 and G. van Driel’s studies published in the last decade.3 A great many among the numerous published texts concern the delivery of animals to

the temple and the receipts given to individual herdsmen. Texts accounting for the development of a speci˜c herd during a longer period, which could facilitate the study of Neo-Babylonian bookkeeping practices, were unknown until 1999, when R. Sack published the remarkable tablet NBC 4897.4 In 1993, the text was discussed by G. van Driel,5 and again by the same scholar with K. Nemet-Nejat in 1994.6 The text is an account of the herd of sheep and goats which the managers of the Eanna temple entrusted to Nabû-ahhe-sullim, son of Nabû-sumiskun, and covers the period from his ˜rst account in year 37 of Nebuchadnezzar to the spring of year 1 of Neriglissar. Sheep and goats are dealt with separately, each species broken down by sex and age groups. Thus, the ˜rst set of columns in the account refers to rams, ewes, male lambs and female lambs, and concludes with the total number (the “horizontal total” as van Driel calls it); the second set of columns accounts for full-grown hegoats, she-goats, male kids and female kids, and is followed by their total number (the second

1. I would like to express my profound gratitude for the discussion on the issues of the breeding of sheep and goats to A. Gut, Head of the Department of Sheep and Goat Breeding of the University of Agriculture in Poznan, Michael Jursa, John MacGinnis, and to my collaborators at the Department of the History of the Ancient Near East: Jaroslaw Maniaczyk, Witold Tyborowski, Joanna Paszkowiak, as well as to Rafal Kolinski of the Department of Archeology of the Adam Mickiewicz University. I would like to give particularly warm thanks to Paul-Alain Beaulieu, who kindly collated certain passages at Yale. The results of his collation are acknowledged where applicable. I also owe thanks to John MacGinnis, who kindly improved my English. 2. M. San Nicolò, “Materialien zur Viehwirtschaft in den neubabylonischen Tempeln,” I Or 17 (1948) 273–93; II, Or 18 (1949) 288–306; III, Or 20 (1951) 129–50; IV, Or 23 (1954) 351– 82; and V, Or 25 (1956) 24–38. 3. G. van Driel, “Neo-Babylonian Sheep and Goats,” BSA 7 (1993) 219–58; “Cattle in the Neo-Babylonian Period,” BSA 8 (1995) 215–40. See also E. Gehlken, Uruk. Spätbabylonische Wirtschaftstexte aus dem Eanna-Archiv. Ausgrabungen in Uruk-Warka, Endberichte 11 (Mainz: Philipp von Zabern, 1990), 19–55 and A. C. V. M. Bongenaar, The Neo-Babylonian Ebabbar Temple at Sippar: Its Administration and Its Prosopography. (Leiden: Nederlands Historisch-Archaeologisch Instituut te Istanbul, 1997), 415–22. An important study concerning animal husbandry in Sippar in early Neo-Babylonian times has been published by R. Da Riva, Der Ebabbar-Tempel von Sippar in frühneubabylonischer Zeit, AOAT 291 (Münster: Ugarit-Verlag, 2002) 173–309.

4. R. H. Sack, “Some Notes on Bookkeeping in Eanna,” in Sudies in Honor of Tom B. Jones, eds. M. A. Powell and Ronald H. Sack. AOAT 203 (Neukirchen: Neukirchener Verlag, 1979), 111–18. 5. BSA 7 (1993) 233–35. 6. G. Van Driel, K. R. Nemet-Nejat, “Bookkeeping Practices for an Institutional Herd at Eanna,” JCS 46 (1994) 47–58. For convenience the transliteration and translation of the text is appended to the end of the article.

99

JCS 55 (2003)

100

STEFAN ZAWADZKI

“horizontal total”). The ˜nal item of the account is the grand total (the sum of the two “horizontal totals”). Separate entries in the columns refer to the dead animals (called “hides”) and the payments of the wages (idÿ ) to the herdsmen and for the shearing. Van Driel has solved many problems, yet he has failed to explain several signi˜cant points, or has proposed interpretations that require reevaluation. In particular, we will scrutinize van Driel’s calculations of the numbers of stock in the herd and the average yield of wool per sheep, and demonstrate that they are partly based on a mistaken reading of the text, which aˆected the quality of his calculations. According to the extreme right column, two accounts were delivered each year in Nbk 37–38 and Nbk 41–42, one each year in AmM 1 and AmM 2, and none in Nbk 39 and 40. The acceptance of two yearly accounts would entail, in turn, that virtually each ewe gave birth twice a year. In fact, in the so-called uncontrolled mating, when rams are allowed to stay with ewes all the time (as was the case here), some ewes may give birth more often than once a year, but it is certainly impossible for them to have two lambs repeatedly each year.7 The dating of the individual accounts is contradicted by line 39, according to which the document covers the period from Nbk 37 to Ner 1, which must refer to the entire time rather than selected years only. Although the fact, which misled the accountant, cannot be recognized,8 the dates of the annual accounts must unavoidably be emended in the way that van Driel did it.9 The 7. Still, a ewe may well have bred three times in two years. Although lambing usually took place in the months Adaru– Aiaru, we also know of births in the summer, see J. N. Postgate (with a contribution of S. Payne), “Some Old Babylonian Shepherds and Their Flocks,” JSS 20 (1975) 14. 8. Van Driel’s explanation, JCS 46, 54, that “if the ‘accountant’ had a complete ˜le, he would ˜nd the same data in tablets dealing with the consecutive years: once at the end of one text and again at the beginning of the succeeding year” is not satisfactory because it does not clarify why some years were completely omitted. 9. See, BSA 7, App. IV on p. 258: text year suggested 37 36* (asterisk denoting corrected year) 37 37

phrase in line 39 that the account covers the period “from 37th year of Nebuchadnezzar . . . until 1st year of Neriglissar,” although the herd was entrusted to Nabû-ahhe-sullim in year 36, suggests that the accountant meant the period from the ˜rst to the last account. Van Driel’s discussion of the accountant’s method of reckoning is correct. The starting point of each subsequent account is the number of stock in the herd speci˜ed in the account for the previous year,10 from which the scribe subtracted (from the corresponding items, i.e., rams from rams, ewes from ewes etc., and from the horizontal totals) the dead animals (called KU† = masku, “hides”), the animals given as wages (idÿ) and for shearing (referred to as “x animals ina gizzi” in “Grand total”). This indicates that shearing provided the opportunity to count the stock and that the herdsmen were paid for the shearing after its completion. Most important, male and female lambs (and kids) mentioned in the previous account were shifted to the rams or ewes (and he- and she-goats) in the next account. In light of this observation it is certain that when they appeared for the ˜rst time in the text they were about a year old. During the second year they reached maturity and male lambs (and kids)11 were shifted to the ram (and full grown he-goats) category, while young ewes (and young she-goats) gave birth to a new generation of male and female lambs (and kids) which,

38 38 38 39* 41 40* 41 41 42 42 [42] (my reconstruction based on the observation of two accounts each year)

[43]* (van Driel’s reconstruction is not valid, but emendation is right)

No further emendations. 10. So it was already in the Old-Babylonian period, see F. R. Kraus, Staatliche Viehhaltung im altbabylonischen Lande Larsa (Amsterdam: Nederlandse Akademie Van Wetenschappen, 1966), 25. 11. Incidentally, modern goats are able to breed at a much younger age than ewes, becoming fertile at 5–7 months only: thus, a one-year-old goat may have her ˜rst kid(s). The discussed text suggests that the age of maturity of the Mesopotamian goats was the same as that of the ewes.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

again, appeared in the text only after a year. For this reason, for example, 52 male and female lambs mentioned in l. 5 cannot be the oˆspring of the 101 ewes mentioned in the same line, but were born of ewes that were mentioned in the previous report, i.e., of 90 ewes. The lack of data on the numbers of animals born in the herd in the spring litters may lead to the conclusion that the temple was surprisingly indiˆerent to this issue. In fact, some information contained in the text allows us to establish how many of the new-born animals were to be assigned to the temple herd. According to both YOS 6, 155 (a contract between the Eanna temple and its naqidu) and the Arsam contracts,12 the herdsman was obliged to supply to the owner 66%-6 (or 66@-3) new-born animals per 100 ewes each year. The owner’s pro˜t had to be speci˜ed in such detail because in practice it was impossible to supervise the herdsman during a period of several months in a year, when the herd was grazing in distant pastures; another advantage of this stipulation was that it precluded possible contention or fraud. The terms of the contract being what they were, the owner did not need to record the numbers of new-born stock, and was interested only in the quota to be supplied, which was speci˜ed as a percentage of the number of ewes. The texts of the contracts also suggest that if the number of the new-born animals was lower than expected, the holder was nevertheless obliged to provide the agreed number of new stock. Accordingly, we infer that a yearly account in the “horizontal total” (col. V) does not cover all the stock in a herd, since the number does not include new-born animals. The quota to be supplied to the owner was calculated based on the number of ewes at the moment when the herd was entrusted to the herdsman (in this document, based on 90 ewes). Furthermore, the contract provided for losses for which the holder could not be blamed, and which could not exceed 10% of the total number of the stock. In order for such losses to be recognized, the holder had to document 12. Concerning these documents, see van Driel, BSA 7, 222–23.

101

them by producing the hides and tendons of the dead animals. The numbers of male and female lambs speci˜ed in the accounts refer to the animals assigned to the owner and remaining from the previous year’s quota. It was very important to record all the losses, which decreased those numbers, since this allowed a calculation of the updated size of the owner’s herd. Another reason to register the losses of young female lambs (and the associated expenses) was that the number of the females left in the herd and taken care of by the holder was included in the number of ewes, which in turn provided the basis for determining the number of the young to be supplied to the owner by the herdsman during the following year (66%-6 per year and per 100 ewes). Thus, we infer that not only the losses (“hides”) and “wages” (“idÿ ”), but also the “payment for shearing” was deducted from the quota to be supplied to the owner. This leads to a very important conclusion, which could not be deduced from either YOS 6, 155 or the Arsam contracts: That the quota of 66%-6 young animals per 100 ewes was a gross amount, from which losses and expenses were subsequently deducted. We will ˜nd further proof of this conclusion if we compare the herdsman’s share13 with the losses of male and female lambs and the associated expenses. His quota of new-born animals hardly covers losses and expenses. The herdsman, who acted as an entrepreneur, obviously would not accept a contract whose terms did not oˆer to him a realistic opportunity for pro˜t, even if losses—as was the case of the discussed document—were within the ordinary limits. When negotiating the terms of the contract, the herdsman must have also considered the possibility that the losses might exceed the acceptable number, in which case he still wanted to realize a pro˜t, albeit a small one. Doubts arise when we recall that the entry on the payment for shearing (idÿ ) appears in the last column (“grand total”) rather than in a separate 13. The herdsman’s pro˜t is the diˆerence between the estimated lambing percentage and the quota to be delivered to the owner (66%-6 sheep per 100 ewes).

102

STEFAN ZAWADZKI

row, as losses and wages were recorded. Although the layout of the text diˆers from those known from the contracts,14 a more likely explanation is that idÿ was included in the “grand total” column because it refers almost exclusively to one category of stock, i.e., male lambs (BAR.GAL) culled regularly at the beginning of a year. The Quality of the Accounts In order to construe the text properly, it is essential to establish whether the account calculations are correct. Below we will discuss only accounts that contain accountant’s errors or whose interpretations were based on erroneous transliterations or bad copies of the texts. According to van Driel, the data in lines 2–4 refer to year 36, but the accountant made many mistakes: (a) the total number of sheep in line 2 col. V, or the horizontal total of the ˜rst four items, should be 124, while the text speci˜es it as 133;15 (b) l. 5 speci˜es 101 ewes, while according to van Driel’s calculations there should be only 93: 81 (90 reduced by 6 hides and 3 ewes given as wages) plus 12 ewes (15 younger ewes16 reduced by 2 hides and 1 given as wages, l. 4).17 Although it is not my purpose to uphold the reputation of the ancient accountant, let me explain how he came up with the ˜gure of 101 ewes in l. 5 and the horizontal total of 133 in l. 2. “101 ewes” might be the result of the subtraction of 8 hides (6 hides of ewes and 2 hides of female lambs, l. 3) from the 90 ewes and 15 female lambs in l. 2 (a total of 105), and then of adding to the result (97 ewes) 4 sheep given as wages (3 ewes and 1 female lamb, l. 3). This assumes that the accountant’s error resulted from adding the ˜gure in l. 4 to the subtotal

14. “wages,” or “idÿ,” were not reported in other contracts from that period. 15. Considering the ˜gure 124 to be the scribe’s error, van Driel emends (BSA 7, 258) “6 hides” to “9” and “15 parrat ” to “25” which gives the horizontal total of 134 sheep; yet, he does not justify his corrections. 16. Younger ewes, rams etc., were animals, which in the report in question had been shifted to the proper group of fullgrown animals from a group of young animals. 17. JCS 46, 56.

of 97 sheep instead of subtracting it. If one accepts this explanation, it follows that both the horizontal total of sheep in l. 2 (133 instead of 124) and the grand total (137 instead of 127) are wrong. This is how van Driel construed the data in l. 2. Still, the ˜gures in ll. 2–5 may be correct if one assumes that the accountant’s procedure was slightly different than in the subsequent accounts: 101 might also be the result of the subtraction of just the wages in l. 4 (3 ewes and 1 female lamb) from the total of 90 ewes and 15 female lambs (105 total sheep).18 The horizontal total of sheep in l. 2 would be the result of the addition of the item 1– 4 in this line to the “hides” in l. 3. Further proof of the correctness of the horizontal total of 133 in l. 3 is the fact that the sum of this ˜gure and the horizontal total of goats is the grand total of 137 animals. Thus, unlike in the subsequent accounts, the data in l. 2 refers not to the size of the herd in the year 36, but only to those animals in the herd of Nbk 36 that had survived until the time of the drawing up of the account early in Nbk 37. Accordingly, to reconstruct the size of the herd at the moment when Nabû-ahhe-sullim took it over, one has to add the ˜gures from col. 1–4 in l. 2 (total 124 animals, not 133) and from col. 1–4 in l. 3 (9 animals),19 coming up with the number 133, as in the horizontal total of l. 2. If we assume this interpretation of the ˜gures in l. 2, then in order to establish the number of sheep at the beginning of the year 37, it is enough to subtract the data in l. 4 from the total of the ewes and female lambs. Regardless of the problems created by ll. 2–4, line 5 is obviously a statement of the size of the herd at the beginning of the year 37, after the birthing and the shearing. At that juncture, the herd consisted of 171 sheep and 5 goats, or a total of 176 animals. The 2 lambs (BAR.GAL) given as

18. However, this explanation cannot apply to the 18 rams in l. 5, since the sum of 7 rams and the 12 male lambs (shifted to rams) in l. 2, is 19. Apparently, the accountant reduced the number of rams by one “hide” (l. 3). The subsequent calculations base on the ˜gure of 18 rams. 19. Interestingly, the horizontal total of hides in l. 3 is 9 rather than 12. The error must have resulted from adding the hides and wages from ll. 3 and 4 and leaving out the hide of one ram.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

wages for shearing must further reduce this number. Thus, in the spring of 37, the temple actually owned 169 rather than 171 sheep. The next accounts are essentially correct, if only one reduces the ˜gures in the respective columns by the numbers of animals given as wages for shearing. I will now discuss only those accounts where the accountant’s calculations are mistaken, and attempt to explain the probable reasons for these errors. Note that the ˜rst line in the account is the result of the last settling and point of departure for the next one, seen in the last line of the account; see for example the second account below, where l. 5 comprises the result of former account (spring 37) and l. 8 the new settling made in the spring 38. The second account: spring 38 (lines 5–8) and third account: spring 39 (lines 8– 11), are correct. In the fourth account: spring 40 (lines 11–14), the number of 54 rams seems incorrect: 36 old rams (40 reduced by 3 “hides” and 1 ram given as wages) + 21 (23 younger rams reduced by 2 lambs for shearing) = 57 instead of 54 as in the text. Since Beaulieu’s collation con˜rms the accuracy of Nemet-Nejat’s copy, we can only emend either 54 to 57! or 2 BAR.GAL to 5! BAR.GAL. The argument justifying the latter improvement of the text is that otherwise the wages for shearing would be reduced dramatically. The ˜fth and sixth accounts (lines 14–17 and lines 17–20) are correct. The seventh account (lines 23–26) is correct if we assume that the number of 3 ewes in the grand total in l. 23 was really recorded by mistake. As van Driel has already noted in his transliteration, the scribe tried to erase it, although 3 UDU is still legible. The eighth account (lines 26–31), including the augmentation of the ˘ock by the irbu sa Adari AmM 0), could be considered correct; however, because of certain complications, we discuss here in detail the method of counting used by the accountant: Rams: 119 (131 old rams augmented by 5 old rams from the irbu sa Adari AmM 0 but reduced by 13 hides and 4 given as wages) + 51 (48 former male lambs reduced by 5 given for shearing and

103

augmented by 8 former male lambs from the irbu sa Adari AmM 0) = 170. Ewes: 298 (275 old ewes augmented by 68 old ewes from the irbu sa Adari AmM 0, but reduced by 34 hides and 11 given as wages) + 92 (90 younger ewes augmented by 20 younger ewes from the irbu sa Adari AmM 0, but reduced by 11 hides and 7 younger ewes given as wages) = 390. The horizontal total of 55 hides in l. 29 is 58, as noted already by van Driel.20 Evidence of the accountant’s error is that the grand total of the sheep plus 3 goat’s hides would be 61. The interpretation of line 27, concerning 5 rams, 68 ewes, 88 male lambs and 200 female lambs (total 101 animals) “from the income of the month Adaru AmM 0” added to the ˘ock in the ˜rst year of Amel-Marduk, presents certain di¯culties. It should be noted that at the time when the sheep were entrusted to Nabû-ahhesullim, most of them were considered fully-grown, except for 28, which were listed as lambs. The high number of 68 ewes suggests that at some point between Adaru and the time in AmM 1 when Nabû-ahhe-sullim took custody of them, at least some of them had been shifted to the category of “full-grown.”21 Still, we are unable to establish the precise ages and numbers of animals in each category as of Adaru AmM 0. It is utterly inconceivable that these animals should be the oˆspring of the ewes of Nabû-ahhe-sullim’s herd, since the number of newly-born lambs was higher than that of the potential ewes in the herd; besides, the lambing season only begins in Adaru. Thus, it is more likely that the managers of the temple decided, for reasons unknown to us, to increase the herd by animals from other sources.22 The ninth account (Simanu Ner 1) is correct: Rams: 148 (170 reduced by 16 as hides and 6 given as wages) + 61 younger rams (66 reduced by 5 lambs taken in the spring of AmM 2 for shearing) = 209. 20. Beaulieu’s collation has con˜rmed the accuracy of Nemet-Nejat’s copy in this case as well. 21. Likewise, 2 goats were classi˜ed as she-goats, and 1 as a female kid, which is further evidence in favor of the ˜rst half of AmM 1. 22. Possibly from the king’s irbu.

104

STEFAN ZAWADZKI

Ewes: 340 (390 reduced by 39 hides and 11 ewes given as wages) + 113 (13323 reduced by 13 as hides and 7 female lambs given as wages) = 453. The ˜nal reckoning (lines 34–36) is correct, but several errors appearing both in the copy and in van Driel’s transliteration must be recti˜ed. Line 34 speci˜es the size of the herd before the ˜nal reckoning. It must be borne in mind that the lambs given in Uruk and those given for shearing in the spring of Ner 1 (cf. grand total) must be subtracted from the ˜gure of 888 sheep. In fact, the accountant extracted the number of rams instead of the number of male lambs. Unfortunately, we cannot establish the reasons for the additional remuneration amounting to 8* lambs paid in Uruk. Line 35 speci˜es in detail the part of the herd, which was actually brought to the inspection in Uruk, and l. 36, the unsettled portion; Rams: 209 is reduced by 8*24 lambs, 3 lambs given for shearing and 5 which were inspected, which resultes in 193 uninspected rams. Ewes: out of 453, only 13825 were inspected, i.e., 315 ewes remained uninspected. Male lambs: out of 80, only 14 were inspected, i.e., 66 male lambs remained uninspected.

23. Van Driel reads mistakenly 193 female lambs while the copy gives clearly 133. The horizontal total of 759 is correct. Thus his calculations in JCS 46, 57 from point (3) to the end of the article are wrong. (Incidentally, the name of the editor of The Animal Life Encyclopedia is Grzimek, not Grizmek). 24. The result of substraction in ll. 33–35 would be correct if the number of lambs given to the shepherd were emended to 8 and not 9 lambs, or if 9 is correct we should emend 3 lambs to 2! lambs. Another argument to support this improvement is that 922 sheep and goats (grand total in line 34) reduced by 11 lambs (8!+3 or 9+2!) and 208 inspected sheep and goats (grand total in l. 35) equals exactly 703 sheep and goats (grand total in l. 36), (not 714 as suggested by van Driel, JCS 46, 55). Beaulieu’s collation speaks for the reading 8*+3. Below I cite in extenso his observations: “The tablet apparently has 8 UDU BAR.GAL instead of 9 as copied by Nemet-Nejat, but there is a crack running through the number which was not copied, and therefore we cannot be certain if it is 8 or 9, although 8 seems slighty more likely in my opinion. Further down the line the tablet has a clear 3 BAR.GAL, as copied.” 25. This number 138 (not 198 as in van Driel copy and translation) is in Nemet-Nejat’s copy.

Two Lines Short

Female lambs: out of 146, only 4126 were inspected, i.e., it remains 105 uninspected female lambs. The Herd and Its Condition To assess the condition of the ˘ock one must analyze such factors as the absolute and relative growth of the ˘ock, the lambing percentage, and the time of the reproduction of the original animals. The Absolute27 and Relative Growth of the Flock “Absolute growth” is the actual number of sheep left in the ˘ock, which in all the cases equals the number of animals given in the horizontal total (col. V) minus the animals given “for shearing.” “Absolute growth A” and “Relative growth A” are the absolute and relative growth of the number of sheep since the previous year, while “Absolute growth B” and “Relative growth B” are the absolute and relative growth of the number of sheep since the date of the ˜rst account (Table 1). As we can see, the average yearly growth of the herd (excluding the addition of new animals in AmM 1) was about 18%. Counting from the ˜rst account from spring 36, when it was 133 sheep in the ˘ock,28 its number was duplicated during about four years, i.e., before the spring Nbk 40 and again after next four year, before the account of spring AmM 1 has been made. The

26. In this place Sack’s copy is better than Nemet-Nejat (in her copy 51). The correctness of Sack’s copy may be proven by a simple arithmetical operation: 5 rams, 138 (!) ewes, 14 male lambs and 41 female lambs add up to 198, i.e., exactly the same ˜gure as in the horizontal total. Delete the numeral 268 written in bold type in Van Driel’s translation. Beaulieu comments, “The tablet has a clear 41, indeed, but the scribe has written 51 and then erased one of the Winkelhaken to make 41. This is re˘ected on Dr. Nemet-Nejat’s copy, which has dots encircling the ˜fth Winkelhaken.” 27. The absolute growth means the actual number of sheep left in the ˘ock, i.e., without losses and all expenses. 28. However, without new-born animals.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

105

Table 1 Year

Number of sheep

Absolute growth A

Relative growth A

Absolute growth B

Relative growth B

Nbk 36 Nbk 3729 Nbk 38 Nbk 39 Nbk 40 Nbk 41 Nbk 42 Nbk 43 AmM 1 AmM 1a AmM 2 Ner 1

133 16930 20331 24132 28633 33934 39635 46236 53937 640 75439 87740

36 34 38 45 53 57 66 77 178 114 123

ca. 27.06 20.01 15.76 15.73 15.63 16.8 16.66 16.66 ca. 33.038 17.8 16.3

36 70 108 153 206 263 332 399 577 691 814

ca. 27.06 52.6 81.2 115.0 154.9 197.7 249.6 300.0 433.8 519.5 612.0

29 30 31 32 33 34 35 36 37 38 39 40

rapid growth in AmM 2 was caused by the augmentation of more than a hundred sheep as irbÿ of AmM 0, but during the whole period the ˘ock grows quite quickly and its condition seems to have been very good. The Lambing Percentages, the Sex of the New-Born Animals Assigned to the Owner’s Herd and the Holder’s Pro˜t The data contained in the texts refer only to the part of the herd that was assigned to the owner, and therefore it may appear risky to try to establish the lambing percentage based on the known number of the young animals left in the herd a year after they had been born, although we can calculate the size of the yearly quota of new-born

29. Nbk 37, Nbk 38 etc. means exactly “the growth of ˘ock till spring Nbk 37, Nbk 38 etc.” 30. I.e., 171 animals reduced by 2 animals given for shearing. 31. I.e., 207 animals reduced by 4 animals given for shearing. 32. I.e., 246 animals reduced by 5! lambs (see supra). 33. I.e., 292 animals reduced by 6 lambs given for shearing. 34. I.e., 341 animals reduced by 1 lamb and 1 ewe given for shearing. 35. I.e., 403 animals reduced by 7 lambs given for shearing. 36. I.e., 469 animals reduced by 7 lambs given for shearing. 37. I.e., 544 animals reduced by 15 lambs given for shearing. 38. Including the irbu of AmM 0. 39. I.e., 759 animals reduced by 5 lambs given for shearing. 40. I.e., 888 animals reduced by 18* lambs “given in Uruk” and 3 lambs given for shearing.

animals for the owner based on the number of ewes. Accordingly, the following calculations are estimates only, even if one may limit the possibility of error by making certain assumptions. First of all, one must decide whether the temple expressed any preferences concerning the sex of the animals in the quota of 66%-6 per 100 ewes, or whether the natural ratio of sexes was preserved. As possible evidence of the former possibility, we may quote above all the fact that one-year-old male lambs were used as sacri˜cial animals, and most of them were collected during the ˜rst year of their lives. In fact, after a year the number of male lambs usually amounted to slightly less than half of the number of female lambs. 41 During the second year, the managers of the temple agreed to assign several male lambs as the payment for shearing, but not as the payment of the “wages” (“idÿ ”); furthermore, they did accept any “losses” in this group.42 Thus most of the male

41. Likewise in other texts from Uruk and Sippar; see BSA 7, 253–55. 42. It is altogether incredible that there should be no losses among the male lambs during the second year of their lives. Apparently, the occurring losses were either replaced by the herdsman, or, equally possibly, the owner agreed to a reduction of the number of stock of another category (rams, ewes or female lambs). It is striking that the losses in the group of ewes usually amount to 10%. Perhaps when a male lamb died, it was the accepted practice to deduct an old ewe from the stock.

106

STEFAN ZAWADZKI

Table 2

Number of ewes

66 young animals per 100 ewes for the owner

Number of the owner’s female lambs a year later

Owner’s maximal possible number of male lambs

Owner’s young animals a year later

Taken away from the owner’s young animals during the ˜rst year (mostly male lambs)

90 101 119 138 158 182 209 237 343 390

60.1 67.5 79.5 92.2 105.5 121.5 139.6 158.3 229.1 260.5

36 40 45 53 60 65 80 90 133 146

24.1 27.5 34.5 39.2 45.5 56.5 59.6 68.3 96.1 114.5

52 58 68 80 91 105 121 138 199 226

8.1 9.5 11.5 12.2 14.5 16.5 18.6 20.3 30.1 34.5

lambs that lived to be one year old must have been transferred to the ram herd; we also infer that male lambs were used as sacri˜cial animals (or possibly slaughtered for food, albeit rarely) only during the ˜rst year of their lives.43 On the other hand, one may question whether the temple actually demanded more male than female lambs, and whether the markedly lower number of the males in comparison with the females did result from so many male lambs being collected for sacri˜ces.44 The temple might have indeed needed considerably fewer male lambs from its herds than one may expect. Since the Eanna temple kept many herds, sacri˜cial male lambs might have been collected from all of them simultaneously, if only in order to prevent an undesirable future increase or decrease of the number of the rams. The sexes of the new-born animals assigned to the temple’s quota may be established with much reliability in the following way: Knowing the number of ewes, we begin by calculating the total quota to be supplied to the temple, and since the losses among the female lambs were negligible during the ˜rst year, this allows us to ˜nd out the maximum number of male lambs, which in fact 43. Apparently mature rams were slaughtered for food only when they were old or sick. 44. JCS 46, 55 with n. 7.

may be somehow overestimated because of the reasons outlined above. The diˆerence between the original number of male lambs and their number reported after a year is the number of the male lambs collected from the herd during the ˜rst year of their lives, most probably as sacri˜cial animals (Table 2). It turns out that among the new-born animals delivered to the owner each year, there were more female than male lambs.45 This must have resulted from the essential task assigned to the holder, which was to raise the size of the herd quickly. The only means of carrying out this task was to leave a fairly high number of female lambs in it and to start taking them away only when it was decided that no further male lambs should be taken away. If, as we have assumed, the losses among the female lambs were negligible during the ˜rst year, and usually amounted to 15% by the second year (7.5% per year), than 80–85% of the original number of female lambs were transferred to “ewes.” Another surprising conclusion is that virtually all new-born female lambs were probably 45. The lambing percentages were most probably a little higher, because the losses among new-born animals must also have occurred during the ˜rst year of their lives, even if they are not recorded in the text. The conclusion is in opposition to Van Driel’s statement that “The owner was especially interested in young male lambs” (BSA 7, 234).

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

107

Table 3

Number of ewes

Estimated total number of the newborn animals

90 101 119 138 158 182 209 237 343 390

72 80 90 106 120 130 160 180 266 292

Lambing percentages 80% 79.2 % 75.6% 76.8 % 75.95 % 71.42% 76.5 % 75.95 % 77.55 % 74.9% ca. 76.4%

Number of new-born animals for the owner

Number of new-born animals for the holder

Percentages of young animals for the holder

60.1 67.5 79.5 92.2 105.5 121.5 139.6 158.3 229.1 260.5

11.9 12.5 11.5 13.8 14.4 8.5 20.4 21.7 36.9 31.5

16.5 15.6 12.7 13.0 12.0 6.5 12.7 12.0 13.9 10.8 ca. 12.6%

assigned to the temple herd. We draw this conclusion from the assumption that statistically the numbers of new-born male and female lambs during a long period was almost equal. We have also pointed out that according to the accounts, the number of female lambs at the beginning of the second year of their lives was the same as after the lambing season or only insigni˜cantly lower. If losses occurred within this group, or if only a part of the new-born female lambs were assigned to the owner, then by doubling this number and comparing it with the number of the ewes, we would arrive at a lambing percentage lower than the actual ˜gure. In fact, however, the average lambing percentage determined in this way amounts to 76.4% and is close to that from Old-Babylonian Larsa (75–80%).46 It seems that in the herdsman’s share the female lambs were extremely rare and in a very small amount. The diˆerence between the total number of the new-born animals and the quota supplied to the owners is the holder’s pro˜t. If our calculation is correct, then Nabû-ahhe-sullim’s net pro˜t was fairly low. We may infer that holders decided to conclude such contracts only because the entire cost of the tending of the stock was normally paid 46. F. R. Kraus, Staatliche Viehhaltung, 126–27, 139–40 (cited already by van Driel, JCS 46, 54 n. 4).

Year Nbk 36 Nbk 37 Nbk 38 Nbk 39 Nbk 40 Nbk 41 Nbk 42 Nbk 43 AmM 1 AmM 2

by the owner; since the holder tended the owner’s herd (or herds) jointly with his own, this procedure might have reduced the cost of keeping the former’s herd (Table 3). The Reproduction in the Herd, or the Rate of the Replacement of the Original Animals We assume that the animals removed from the herd, whether due to death or as herdsmen shearling wages, were primarily the oldest ones. Although we realize that this is inaccurate, since among the ewes in particular the dead animals certainly included the younger ones, it is interesting to consider the period after which the entire original population of the ewes was replaced. In the spring (Nbk 37) the herd consisted of 90 ewes. The following numbers of animals were removed during the subsequent years: 1. 87 (90 minus only? 3 given as idÿ ) Nbk 37 2. 74 (87 minus 10 dead and 3 given spring Nbk 38 3. 59 (74 minus 14 dead and 4 given spring Nbk 39 4. 40 (59 minus 16 dead and 5 given spring Nbk 40

till spring as idÿ ) till as idÿ ) till as idÿ ) till

108

STEFAN ZAWADZKI

Table 4

Year

Rams

Absolute growth

Nbk 37 Nbk 38 Nbk 39 Nbk 40 Nbk 41 Nbk 42 Nbk 43 AmM 1 AmM 2 Nrg 1

18 30 40 54 68 89 111 131 170 209

12 10 14 14 21 22 20 39 39

Growth % 66.6 33.3 35.0 26.0 30.9 24.7 18.0 29.65 23.0

Ewes

Absolute growth

Growth %

Total rams and ewes

Total growth

101 119 138 158 182 209 237 275 390 453

18 19 20 24 27 28 38 115 63

17.8 16.0 14.5 15.2 14.8 13.4 16.0 41.8 16.2

30 29 34 38 48 40 58 154 102

25.3 19.5 19.1 17.9 19.2 13.4 16.6 27.5 15.4

5. 23 (40 minus 18 dead and 5 given as idÿ) till spring Nbk 41 By the next, sixth year (Nbk 42), the last of the original rams must have been removed from the herd. This is to be expected, as the reproductive potential of sheep diminishes sharply after their seventh year. The original sires were replaced after the same period of time. It is particularly enlightening to compare the rates of the growth of the numbers of fully-grown rams and ewes (Table 4).47 Although in terms of absolute numbers the number of the rams increased at a slower rate than that of the ewes, their percentage growth was quicker than that of the females (except for AmM 2, when the herd increased suddenly as new animals were added to it); thus, the ratio of ewes to rams decreased. The fact that there were more rams in the herd than mere reproduction would warrant is indirect evidence of the essential purpose of the breeding: It seems that the ewes were kept only at a number that ensured steady and quick increase of the size of the herd. The fast growth of the number of the rams proves that the purpose of the multiplication of the herd was not the supplying of milk and dairy products

47. In view of the di¯culties with the interpretation of the data in ll. 2–4, our considerations will base on the ˜gures for the year 37 (l. 5).

One Line Short

but of wool and meat, of which rams yield more than ewes. Wool Production In his discussion of this issue, van Driel assumes that according to the available Neo-Babylonian texts, 1!-2 minas of wool were expected per sheep (ram and ewes).48 Van Driel correctly identi˜es the ˜rst nine items of data as accounts of deliveries of wool from Nbk 37 to AmM 2, but commenting the entry of 14 talents and 5 minas in l. 35 (which according to his calculations is the yield of 563!-3 sheep) concerning the inspection of 208 sheep, he remarks that “we can reasonably assume that more than the sheep inspected were ultimately present.”49 By adding this amount to 58 talents and 36!-2 minas (l. 36), van Driel comes up with the total of 72 talents 41!-2 minas, or the yield of 2900 sheep, which is utterly incommensurable with the size of the herd of 888 sheep in Ner 1 (l. 33). Using sophisticated methods, van Driel eventually concludes that 947 animals were shorn in Ner 1, yielding 27 talents and 25!-2 minas of wool, or 1.74 minas per sheep. In fact, as the list below demonstrates, the grand total of wool for the entire period of the account was 58 talents and 36!-2 minas. If this has been 48. BSA 7, 222–23. Obviously, the production over 1!-2 minas per sheep belonged to the herdsman. 49. JCS 46, 55.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

dif˜cult to establish, it is because of the errors in both the copy and the transliteration.

Nbk 37 Nbk 38 Nbk 39 Nbk 40 Nbk 41 Nbk 42 Nbk [43] AmM 1 AmM 2 Ner 1 total

1 talent 20.5 minas (l. 5)50 1 talent 55!.[5!] minas (l. 8) 2 talents 24.5 minas (l. 11) 2 talents 45.5 minas (l. 14) 3* talents *18.5 minas (l. 17) 5 talents 39.5 minas (l. 20) 7 talents 1.5 minas (l. 23) 8 talents 43.5 minas (l. 26) 11 talents 22.5 minas (l. 30) 14 talents 5.0 minas (l. 35) 54 talents 276.5 minas (l. 36) (I.e., 58 talents 36.5 minas.)

l. 8: van Driel reads 50 [MA SÍK.H]I.A, but my reading is justi˜ed by the following arguments: (a) according to the copy of Nemet-Nejat, the text in this place is written over an erasure; (b) other passages mentioning speci˜c quantities of wool and the ˜nal total of wool (58 talents 36!-2 minas) are preserved very well; the proposed reconstruction of the badly preserved entry in l. 8 (which, to make things worse, seems to have been written over an erasure and is ambiguous) is the only means of clarifying this issue. According to Beaulieu’s collation, “the reading ª58º [MA.NA SÍG.H]I.A is possible, but not certain, as the numerals are very cramped on the tablet, and this is re˘ected on the copy. These numerals are followed by a possible erasure, and then by a break which extends to the edge.” l. 17: van Driel reads 4 (talents) 8 minas, but as the signs are very small, our suggested reading of 3* (talents) 18* minas is also possible (NemetNejat’s copy is not clear to me). This reading is supported by the fact that the wool production rose by ca. 30 minas between the two successive reports, while according to the reading known so far (4 talents 8!-2 minas), the increase would amount to 1 talent 23!-2 minas, which is too much 50. Lines are given according to the copy of Nemet-Nejat.

109

with respect to the growth of the number of animals in the herd. An additional argument in favor of the suggested reading is that the total of individual deliveries is the same as the grand total in l. 35. Beaulieu comments: “the reading 3 (GUN) 18 MA SÍG.fiHIfl.A is indeed very likely. The horizontal of 4 (GUN) is close in shape to a Winkelhaken, and should therefore be the numeral 10 going with 8 to form 18.” The Yearly Wool Production per Sheep We deduce from the grand total in l. 2 that Nabû-ahhe-sullim received the herd early in year 36, either immediately after the shearing, or after the animals had been shorn in Uruk and the wool had been immediately delivered to the temple— therefore this fact is not recorded in the account. The average production of wool per sheep depends on the age the sheep are shorn for the ˜rst time. We must assume that male and juvenile lambs were shifted to the category of “fullygrown” when they reached reproductive maturity, approximately at 1!-2 years of age.51 One may reasonably suppose that they were shorn for the ˜rst time about a year after they appeared in the text, i.e., when they were about two, but in this case the average yield of wool per sheep would be very small. Therefore, one must also consider the possibility that a sheep was shorn for the ˜rst time only a year after it had been transferred from the group of “the young” to “the mature” or that the temple demanded wool only a year after the young sheep were transferred to the mature group. We shall deal with both possibilities, since we cannot identify the correct one: If an animal was ˜rst shorn upon being transferred from the “young” to the “mature group,” then in order to establish the average production of wool per sheep we must divide the amount of wool in the grand total by the total of rams and ewes in a given line (Table A). If, however, the stock were ˜rst shorn only a 51. They were transferred to “adults” after a year, which means that they were ˜rst listed in the reports at the age of at least six months. See the diˆerent opinion of J. N. Postgate and S. Payne, JSS 20 (1975) 14–15 and 19–20.

110

STEFAN ZAWADZKI

Table A 1 talent 20.5 minas (4830 s): 1 talent 55!.[5] minas (6930 s): 2 talents 24.5 minas (8670 s): 2 talents 45.5 minas (9930 s): 3* talents *18.5 minas (11910 s): 5 talents 39.5 minas (20370 s: 7 talents 1.5 minas (25290 s): 8 talents 43.5 minas (31410 s): 11 talents 22.5 minas (40750 s): 14 talents 5 minas (50700 s): = 58 talents 36.5 minas (=210990 s):

119 sheep52 = 40.6 s each 149 sheep = 46.5 s each 178 sheep = 48.7 s each 212 sheep = 46.8 each 250 sheep = 47.64 s each 298 sheep = 68.35 s each 348 sheep = 72.67 s each 406 sheep = 77.36 s each 560 sheep = 72.76 s each 662 sheep = 76.7 s each

Nbk 37 Nbk 38 Nbk 39 Nbk 40 Nbk 41 Nbk 42 Nbk 43 AmM 1 AmM 2 Ner 1

3182 sheep = 66.3 s per sheep.

Table B 1 talent 20.5 minas (4830 s): 1 talent 55!.[5!] minas (6930 s): 2 talents 24.5 minas (8670 s): 2 talents 45.5 minas (9930 s): 3* talents 1*8.5 minas (11910 s): 5 talents 39.5 minas (20370 s): 7 talents 1.5 minas (25290 s): 8 talents 43.5 minas (31410 s): 11 talents 22.5 minas (40750 s): 14 talents 5 minas (50700 s): = 58 talents 36.5 minas (=210990 s):

year after the transfer, then we must divide the amount of wool in the grand total by the total of rams and ewes from the previous account, reduced by the hides, wages for the herdsmen and wages for shearing (Table B).52 53 The data in Table A yield the average production of wool per sheep amounting to 64.2 shekels; although the productivity increased with time, in Nrg 1, or after several years, the production was still appoximately 20% lower than the expected 1!-2 minas.

52. I.e., rams and ewes. 53. 92 sheep, if we subtract only 3 ewes given as wages and two given for shearing (on the problem with interpretation of this line, see supra). If we also subtract the hides, the result would be 4830 shekels of wool divided by 87 = ca. 60 shekels per sheep.

94 sheep = 51.4 s53 each 104 sheep = 66.6 s each 130 sheep = 66.7 s each 156 sheep = 63.6 s each 184 sheep = 64.7 s each 218 sheep = 93.4 s each 260 sheep = 97.3 s each 304 sheep = 103.3 s each 412 sheep = 98.9 s each 488 sheep = 103.9 s each

Nbk 37 Nbk 38 Nbk 39 Nbk 40 Nbk 41 Nbk 42 Nbk 43 AmM 1 AmM 2 Nrg 1

2350 sheep = 1 mina 38 s per sheep.

An important argument in favor of the second method of counting is that by dividing the total amount of wool (210,990 shekels) by the 90 shekels due to the owner per each ram or ewe, one arrives at the result of 2344.3 sheep, which is very close to the ˜gure of 2350 sheep resulting from adding, according to the second method, the number of the sheep shorn during the whole period.54 As we can observe in Table B the average production was almost exactly 1!-2 minas, increasing rapidly from a much lower value at the beginning of the period to more than the expected 1!-2 minas at its end. The relatively low initial yield may be explained to a degree by the young age of

54. The small diˆerence probably results from the mistakes made by the scribe; see above, “The Quality of the Accounts.”

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

the shorn animals and the small number of rams, which give more wool; the latter’s ratio to sheep increased rapidly over the following years. The average yield of wool decreased during the shearing of AmM 2: the addition to the herd of 73 young sheep from the “income of AmM 0” (probably shorn for the ˜rst time) lowered the average result, which rose again in the following year. Obviously, the percentage of young animals decreased with time, and the average yield of wool increased consequently. The growing number of rams in the herd had a particularly notable eˆect on this increase. The turning point may be observed after the shearing of Nbk 41: the average production of wool per head rose substantially when the number of the rams approached half of the number of the ewes. It was probably these two factors—the increasing ratio of rams among the shorn animals and the considerable fall of the percentage of the less productive young sheep— that caused such dramatic increase in the yield of wool after Nbk 41. The quota of 1!-2 minas demonstrates that the herd was fully mature, i.e., it consisted of animals several years old, just like the herd leased out by Arsam. Still, the Table shows that while in YOS 6, 155 and in the Arsam contracts the quota for the owner was ˜rmly ˜xed (1!-2 minas per sheep), under the terms of the discussed text the owner received a ˜xed share of the total wool produced per sheep, which for the whole period gave the same average amount of 1!-2 minas of wool per sheep per owner. The diˆerence is most probably caused by the fact that the herd tended by Nabû-ahhe-sullim was young, and its productivity was quite low at the beginning. Goats The herd entrusted to Nabû-ahhe-sullim at the beginning of year 37 included four goats: 1 fullgrown he-goat, 2 full-grown she-goats, and 1 female kid. As in the case of the sheep, we will discuss only the accounts where the calculations are incorrect. The fourth account (lines 11–14) is incorrect: there is only 1 full-grown he-goat, while we expect

111

2. Because the total number of goats is correct (9), the accountant apparently deducted a male kid as wages instead of female kid. Once made, the error was repeated throughout the account. The sixth account (lines 17–20): we expect 3 full-grown he-goats (1 old and 2 younger) instead of 2 in l. 20, and 8 she-goats (8 old and 2 younger, reduced by one hide and one female kid as wages) instead of 9. The extra animal should be probably moved to full-grown he-goats as wages or a hide. The ˜nal reckoning (ll. 34–36): out of 34 goats, only 10 were inspected (8 she-goats plus one male kid and one female kid). The accountant made two mistakes in the number of “remains” (l. 36): there should be 8 full-grown he-goats (not 7) but 9 she-goats (not 10). Only the grand total of “remaining goats” (24) is correct. An examination of the consecutive horizontal totals reveals that between the autumn of year 37 and the spring of year 42, the herd increased by exactly two heads a year. During the next years the rate of growth was slightly faster, amounting to 3 or 4 head per annum. It is conspicuous that throughout the period covered by the text, the number of male and female kids was twice lower than that of their she-goats, while goats almost always give litters of two kids. The fact that so few young goats were left in the herd must have resulted from a deliberate policy of keeping the population of this species low, although the ratio of goats to sheep grows steadily in the subsequent accounts: 1. 133 sheep : 4 goats, i.e., 1 goat per 33.25 sheep. 2. 171 sheep : 5 goats, i.e., 1 goat per 34.2 sheep. 3. 207 sheep : 7 goats, i.e., 1 goat per 29.5 sheep. 4. 246 sheep : 9 goats, i.e., 1 goat per 27.3 sheep. 5. 292 sheep : 11 goats, i.e., 1 goat per 26.5 sheep. 6. 341 sheep : 13 goats, i.e., 1 goat per 26.2 sheep. 7. 403 sheep : 15 goats, i.e., 1 goat per 26.8 sheep. 8. 469 sheep : 18 goats, i.e., 1 goat per 26.05 sheep.

112

STEFAN ZAWADZKI

9. 544 sheep : 22 goats, i.e., 1 goat per 24.57 sheep. 10. 640 sheep : 25 goats, i.e., 1 goat per 25.6 sheep. 11. 759 sheep : 30 goats, i.e., 1 goat per 25.3 sheep. 12. 888 sheep : 34 goats, i.e., 1 goat per 26.1 sheep. The numbers of the male and female kids kept in the herd were usually the same; if their total was an odd number, the extra animal was female. This ensured better reproduction and might also indicate a larger demand for male kids, both for religious sacri˜ces and for meat. Goat Hair The accounts dated between Nbk 37 and AmM 1 do not mention deliveries of goat hair. The text mentions supplies of this product only three times: AmM 1 AmM 2 Ner 1 Total:

8 minas 50 shekels (l. 26) 20 minas 50 shekels (l. 31) ª11º minas [10 shekels] (l. 35) 40 minas 50 shekels (l. 37)55

l. 35: van Driel reads here 1[0 MA SÍ]G.ÙZ but numeral 11 is clear in Nemet-Nejat and Sack’s copies. It seems to me that Nemet-Nejat’s copy can be read 11 ªMAº [10] ªGÍN SÍGº.ÙZ. Such a reconstruction of the original text is justi˜ed by the parallelism with the account of sheep wool, where the last amount was the grand total for the whole period covered. If the last amount of goat hair, i.e., 40 minas 50 shekels, were the production of the ˜rst year of Neriglissar, as suggested by van Driel, this would be evidence of the rapid growth of both the number of the goats and the productivity per goat, which is completely unacceptable. According to Beaulieu the passage is “in fact an erasure and should be disregarded. It has 11 at 55. Thence, the amount of 40% - 6 minas of goat hair is the total of all deliveries, rather than the last delivery, as van Driel construes it.

One Line Short

the beginning, and ÙZ at the end, neither of which were completely erased by the scribe.” Accepting this, it would be a question why the text was written and why it was erased. Because sometimes it is di¯cult to recognize if the text was erased or simply damaged, I think that the above arguments speak rather for the second possibility. Yearly Production of Goat Hair According to YOS 6, 155 and the Arsam contracts,56 the holder of the ˘ock was to receive 50 shekels of goat hair per each full-grown goat. In light of this quota, the total amount of goat hair mentioned in our text (40 minas 50 shekels) must be construed as the produce of exactly 49 animals: 8 minas 50 s (530 s) 20 minas 50 s (1250 s) ª11 minas 10º s (670 s) Total:

: 50 shekels = 10.6 goats AmM 1 : 50 shekels = 25 goats

AmM 2

: 50 shekels = 13.4 goats Ner 1 49 goats

However, the accounts also specify the actual number of goats in the ˘ock, and if we divide the amount of goat hair by the number of goats, the resulting productivity per goat is very diˆerent from that known from the contracts. Additionally, the major diˆerences between individual entries have to be noted. Here we encounter the same problem as when we tried to estimate the productivity of sheep: it is not certain whether the amount of goat hair should be divided by the number of adult goats of both sexes entered in the same line, or by the number of adult goats from the previous account, reduced by dead animals, wages and the animals given in payment for shearing.57 56. Only she-goats are mentioned in the Arsam contracts, but the ˜gures most probably also account for full-grown hegoats as van Driel, BSA 7, 223 also supposed. 57. Since the second method seems to be correct in estimating the productivity of wool, here we should expect the same in respect to goat hair.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

Table A 1 8 minas 50 s : 16 goats = 33.1 s each (530 s) 20 minas 50 s : 22 goats = 56.8 s each (1250 s) ª11 minas 10º s : 25 goats = 26.8 s each (670 s) Table B 1 8 minas 50 s : 12 goats = 44 s each (530 s) 20 minas 50 s : 16 goats = 78.1 s each (1250 s) ª11 minas 10º s : 19 goats = 35.3 s each (670 s)

AmM 1 AmM 2

113

We admit to being unable to establish which of the two possible methods of determining the amount of goat hair to be delivered to the owner was applied in this text. This ambiguity may be explained by scribal or accounting errors.

Ner 1 Wages AmM 1 AmM 2 Ner 1

Whatever the method of calculation, the yield of hair per goat is very diˆerent from the quota known from YOS 6, 155 and from the Arsam contracts. The average production, based only on the ˜rst and third items of data,58 calculated by means of method A, is ca. 30 shekels of hair per goat, while the second method gives the value of ca. 40 shekels of hair per goat. The rather large amount of hair in AmM 2 could not have been come from 16 or 22 goats. It is possible that this ˜gure comprises goat hair that was to be supplied in Nbk 37–43 but was not listed under those years. In this case, if we use the ˜rst method (Table A), then we must add the 42 fullgrown goats from the accounts of 37–4259 to 22 goats (a total of 64 goats), which would make the average yield ca. 20 shekels per goat, and if we use the second method (Table B), we must consider a total of 16 + 3460 = 50 goats, their productivity per head being 25 shekels. In both cases the calculated productivity is so low that the reconstruction seems implausible.

58. We exclude the second because the quota per goat is evidently too high. 59. I.e., 4 (l. 5) + 5 (l. 8) + 6 (l. 11) + 7 (8 reduced by one full-grown he-goat given for shearing) (l. 14) + 9 ( l. 17) + 11 (l. 20) = 42. 60. I.e., 3 (l. 2) + 4 (l. 5) + 5 (l. 8) + 6 (l. 11) + 7 (8 reduced by one full-grown he-goat given for shearing, (l. 14)+ 9 (l. 17) = 34.

Since no other published Neo-Babylonian document mentions the wages (idÿ ) paid upon the yearly settlement of accounts, such stipulations in the contract between Nabû-ahhe-sullim and the Eanna temple might have been exceptional. Most of the idÿ consisted of sheep; only in exceptional cases goats were given.61 The summary of the categories of sheep given as idÿ throughout the period between Nbk 37 and Ner 1 gives interesting conclusions. Rams 23

Ewes 58

Male lambs 0

Female lambs 36

In terms of absolute numbers, the idÿ consisted in the ˜rst place of ewes, and then of female lambs; rams were the least numerous group. The total absence of male lambs, as well as kids, in the idÿ proves that the temple was not willing to part with its young male animals. If, however, we consider the ratio of the animals given as idÿ to the total number of a certain group of animal at a given time, then the data provides evidence of a diˆerent nature (Table 5). The group from which most animals were deducted as the idÿ were female lambs (between 5 and 6.6%, 5.65% on the average), and then rams (an average of 3.2%), and ewes (2.86%). The low expenditure of ewes, in terms of percentages, must have been due to a preference for keeping the animals, which provided the greatest advantages as the suppliers of oˆspring as well as wool. Likewise, the relatively low number of rams in the idÿ may be explained by the large losses due

61. Throughout the period, a total of only two she-goats and two female kids were given, and not a single full-grown he-goat or male-kid. The temple certainly took surplus animals of the latter categories for sacri˜ces.

114

STEFAN ZAWADZKI

Table 5 Rams up to Nbk 37 up to Nbk 38 up to Nbk 39 up to Nbk 40 up to Nbk 41 up to Nbk 42 up to Nbk 43 up to AmM 1 up to AmM 2 up to Ner 1 Total

0 from 18 (0 %) 1 from 30 (3.3.%) 1 from 40 (2.5 %) 2 from 54 (3.7%) 2 from 68 (ca. 2.95%) 3 from 89 (ca 3.4 %) 4 from 111 (ca 3.6 %) 4 from 131 (ca 3.05%) 6 from 170 (3.5 %) 23 from 711 (3.2%)

Ewes

Female lambs

3 from 90 (3.3%) 3 from 101 (ca. 3.0%) 3 from 119 (2.52 %) 4 from 138(ca.2.9%) 5 from 158 (ca. 3.2%) 5 from 182 (ca. 2.75%) 6 from 209 (ca. 2.9%) 7 from 237 (ca. 2.95%) 11 from 343 (ca.3.2%) 11 from 390 (2.8%) 58 from 2025 (2.86%)

1 from 15 (6.6%) 2 from 36 (5.5%) 2 from 40 (5.0%) 3 from 45 (6.6%) 3 from 53 (ca. 5.6%) 3 from 60 (ca. 5.0%) 4 from 65 (ca. 6.15%) 4 from 80 (5.0%) 7 from 110 (ca. 6.36%) 7 from 133 (ca. 5.3%) 36 from 637 (5.65%)

to natural causes (“hides”), which, combined with the idÿ, caused a prompt replacement of the original animals of this category. Since the temple culled the majority of male lambs and some of them were additionally given for the wages for shearing, then the idÿ necessarily consisted mainly of female lambs, whose number did not decline for of any other reason than natural deaths. This accounts for the high proportion of female lambs in the idÿ. As the size of the herd increased, the amount of the wages also went up, although the criterion for these increases cannot be easily identi˜ed. The proportional rate of wages was one head of livestock for the custody of 30–33 animals. When only ewes and female lambs were given as idÿ, then there were always more ewes than female lambs (typically one or two more); when rams were given as well, then ewes accounted for a half or nearly a half of the idÿ. The payment of idÿ from the temple’s share to the shepherds (?) might have reduced the holder’s cost, as the wages were spent on the tending of the whole herds, including both the temple’s and the holder’s stock.

under discussion.62 The breakdown of the losses into groups is so intriguing that one needs to ask if these ˜gures described actual losses, or were merely calculated percentages; obviously, we cannot answer this question with any certainty. The amounts of the losses in all the groups (rams, ewes and female lambs) seem to result from the following formulae: 1 animal per 10, 2 animals per 15–25 3 animals per 26–35 4 animals per 36–45, etc. On rare occasions the number of lost animals was higher by one. The lack of losses reported in the group of male lambs throughout the account period indicates that the temple managers did not allow deductions in this category; the herdsmen must have replaced such animals that were inevitably lost over such a long stretch of time with their own livestock.63 Shearing

Losses (“Hides”) Van Driel notes that the Neo-Babylonian contracts allowed for a 10% annual death rate in a herd, under condition that the skin of the dead animals had to be produced. He also observed that losses never reached this amount in the text

The additional payment for shearing was paid in male lambs, or in animals of the most prized category (a total of 41 animals), and occasionally 62. BSA 7, 235; JCS 46, 54. 63. See M. Van de Mieroop, BSA 7 (1993) 168.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

in ewes (1), and young he-goats (1). The fact that additional wages were paid for shearing, and that the most valuable animals were given for this purpose, demonstrates that the temple managers considered e¯cient shearing a priority. The amount of the wages steadily increased, but it was not proportional to the number of the sheep shorn in a given year.64 It is interesting to note that male lambs used for shearing wages are always described as BAR.GAL rather than as kal¿mu, a term referring to male animals retained in the temple ˘ock. The Akkadian reading of BAR.GAL and its meaning are not certain. W. von Soden (AHw 834) and R. Borger (ABZ No. 74) construe it as the logogram of parru, “male lambs.” The reading par-gal would pose the question of whether it was a pseudologogram for parru or whether it should be pronounced pargallu. 65 Van Driel’s view is that “kal¿mu do not have the same meaning, a kal¿mu is younger than a parru.”66 This cannot be true, at least in the text under discussion, because both kal¿mus as well as BAR.GAL were born in the same lambing season. The only suitable de˜nition of BAR.GAL in this context is that the term describes animals removed from the kal¿mu, which cannot constitute the temple’s puhalu.67 Conclusions Why did the temple managers decide to set apart one herd and entrust it to Nabû-ahhesullim? Since the delivery of milk and dairy products are not mentioned among Nabû-ahhesullim’s obligations—and such an obligation does 64. Thus, e.g., we calculate that 218 and 260 sheep were shorn in the sixth and seventh year respectively, but on both occasions the wages were the same (7 male lambs); during the following two years, the number of shorn sheep rose substantially, but the wages fell, in both cases amounting to 5 male lambs. 65. See M. Jursa, “Neu- und spätbabylonische Texte aus den Sammlungen der Birmingham Museums and Art Gallery,” Iraq 59 (1997) 101–2. 66. BSA 7, 231. Note that van Driel reads BAR.GAL both as parru as well as pargallu. 67. See the similar opinion of van Driel, BSA 7, 231: “Male lambs retained are called kal¿mu, but those removed are BAR.GAL.ME.”

115

appear elsewhere, for example, in the Arsam contracts—it is certain that these goods were not covered in the settling of accounts with the temple, although the herdsmen might have consumed small quantities of them.68 The slow rate of the growth of the number of goats, and the fact that herdsmen received a goat as wages on one occasion only resulted from the policy of limiting the number of this species. Because the numbers of goats could easily be doubled every year, it seems that kids aged around one year, both male and female, were taken by the temple for sacri˜ces and consumption. The goats may have also provided the small quantities of milk and dairy products required for the herdsmen’s meals. Thus, the main objective of the herdsman was to breed sheep, although the text seems to provide evidence that the two essential products of sheep, meat and wool,69 were of equal importance. The text may be considered a classic example of the way in which these two purposes were pursued in the speci˜c conditions of the temple economy. Male lambs taken during the ˜rst year were obviously used as sacri˜ces, as well as food for temple personnel. During the following years, the surviving rams and ewes were kept in the herd, the former because of their higher yield of wool (their reproductive capacity being considered only a secondary feature), and the latter, because they were at the peak of their fertility. The number of animals of both sexes was reduced by natural deaths and by culling individuals whose reproductive capacity was declining. By giving the older animals to the herdsmen as wages, the temple managed to dispose of such animals at a proper time, avoiding their negative impact on the development of the herd: the older individuals, at 68. Our observation accords with the conclusion of I. J. Gelb presented in his article on animal husbandry in the third millennium B.C., see I. J. Gelb, “Growth of Herd of Cattle in Ten Years,” JCS 21 (1967) 69 (“dairy products of cattle, goats, and sporadically sheep played a relative limited role in ancient economy”). Accidentally (or not?) both texts TCL II 5499 discussed by I. J. Gelb and our text cover a period of ten years. 69. For very interesting comments of the patterns of sheep breeding, see S. Payne, “Kill-oˆ Pattern in Sheep and Goats: The Mandibles from Asvan Kale,” AnSt 23 (1973) 281–303, esp. 282–84 and ˜gs. 1–3.

1 pu-hal 2

7

116

NBC 4897* ªU8º.ME ka-lum par-rat PAP BAB-BAR-ti MÁ†.GAL ÙZ MÁ†.TUR SAL.Á† PAP GE6-tú 90 12 15 PAP 1 me 33 1 2 [1 PAP 4 6 3 1 me 1

6 2 7 8 30

10 3 1 me 19

9 3 10 1 11 40

12 3 1 me 38

12 3 13 1 14 54

14 4 1 me 58

15 5 16 2 17 68

16 5 1 me 82

18 7 19 2 20 89

18 5 2 me 9

21 8 22 3 23 1 me 11

21 6 2 me 37

24 10 25 4 26 1 me 31

23 7 2 me 75

16

2 1 36

PAP 12! PAP 4 PAP 1 me 71

18

3 2 40

PAP 15 PAP 5 PAP 2 me 7

23

4 2 45

PAP [1]9 PAP 6 PAP 2 me 46

27

4 3 53

PAP 21 PAP 8 PAP 2 me 92

31

5 3 60

PAP 26 PAP 10 PAP 3 me 41

40

6 3 65

PAP 31 PAP 10 PAP 4 me 3

41

6 4 80

PAP 35 PAP 13 PAP 4 me 69

48

8 4 90

PAP 41 PAP 15 PAP 5 me 44

1

1

1

2

1

2

3

[ [ [1

PAP 5

4

[1

[ [ [1

PAP 7

5

[ [ [1

[ [1 [2

PAP 1 PAP 9

[1 [ [2

PAP 1

6

[ [ [1

[1]

PAP 1

8

[2

[ [ [2

1

[ [ [2

[ [ [2

PAP 1 PAP 1 PAP 15

[1 [ [2

[

PAP 2

[3]

PAP 18

[1 [ [3

[ [ [3

PAP 2

9 1

3

10

1 4

PAP 11

12

PAP 13

PAP 22

STEFAN ZAWADZKI

3 1 4 5 18

PAP-ma se-e-nu NÍG.GA dGA†AN sá UNUG.KI u dNa-na-a PAP-ma 1 me 37 mdAG-†E†.ME-GI A mdAGMU-GARun MU.37.KÁM KU†.ME i-di PAP-ma 1 me 76 MU.37.KÁM 2 BAR.GAL ina gi-iz-zi IGIir 1 (GÚ) 20!-2 MA SÍG.HI.A KU†.ME i-di PAP-ma 2 me 14 MU.38.KÁM 4 BAR.GAL ina gi-iz- zi 1 (GÚ) 55! !-2! [MA SÍG.HI].A KU†.ME PAP-ma 7 i-di PAP-ma 2 me 55 MU.38.KÁM 2 BAR.GAL ina KI.MIN 2 (GÚ) 24!-2 MA SÍG.HI.A PAP-ma 22 KU†.ME i-di PAP-ma 3 me 3 MU.41.KÁM 6 uduBAR.GAL 1 MÁ†.GAL ina KI.MIN 2 (GÚ) 45!-2 MA SÍG.HI.A PAP-ma 27 KU†.ME i-di PAP-ma 3 me 54 MU.41.KÁM 1 uduBAR.GAL 1 U8 ina KI.MIN 3* (GÚ) 18* !-2 MA ªSÍG.HI.Aº PAP-ma 32 KU†.ME PAP-[m]a 11 i-di PAP-[ma] 4 me 18 MU.42.KÁM 7 BAR.GAL ina KI.MIN 5!-2 (GÚ) 9!-2 MA SÍG.HI.A PAP-[ma] 37 KU†.ME ªi-diº [PAP-ma] ª4º me 8ª7º [MU.42.KÁM] {3 UDU} (erasure) 7 uduBAR.GAL ina KI.MIN 7 GÚ.UN 1!-2 MA SÍG.HI.A [PAP-m]a 43 KU†.ME i-di PAP-ma 5 me 66 ina MU.1.KÁM LÚ-d†Ú 5 BAR.GAL ina KI.MIN 8 (GÚ) 43!-2 MA SÍG.HI.A 8%-6 MA SÍG.ÙZ

NBC 4897* 27

5

68

20

PAP 1 me 1

66

1 me 10 11 7 1 me 33

PAP 6 me 40 PAP 55 PAP 22 PAP 7 me 59

8

3 me 43 34 11 3 me 90

32 16 33 6 34 2 me 9

39 11 4 me 53

80

13 PAP 68 7 PAP 24 1 me 46 PAP 8 me 88

35 ina †À 5

1 me 38

14

41

PAP 1 me 98

36 re-hi 1 me 93 3 me 15

66

1 me 5

PAP 6 me 79

56

4 1 6

14 [3 1 [ [ 16 [4

2 1 8

7

[

[1 [4 [1 [ [4

17

[1 [ [4

[ [1 [5

8

[1

[1

10

[3

[4

PAP 1 me 4 ina ir-bi sá ITI.†E MU.SAG.NAM.LUGAL.LA LÚ-dAMAR.UTU PAP 2[5] PAP-ma 6 me 65 MU.1.KÁM PAP 3 PAP-ma 61 KU†.ME i-di PAP 30 PAP-ma 7 me 89 MU.2.KÁM 5 BAR.GAL ina gi-iz-zi {KI.[MIN]} 11 [GÚ].UN 22!-2 MA SÍG.HI.A 20%-6 MA SÍG.ÙZ PAP 3 PAP-ma 71 KU†.ME PAP 2 PAP-ma 26 i-di PAP 34 PAP-ma 9 me 22 MU.1.KÁM dU+GUR-LUGAL-URÙ LUGAL TIN.TIR.KI 8* uduBAR.GAL ina UNUG.KI! ma-hi-ir 3 BAR.GAL ina gi-iz-zi PAP 10 PAP-ma 2 me 8 MU.1.KÁM am-ru 14 GÚ.UN 5 MA SÍG.HI.A 1ª1º* [MA 10 GÍN SÍ]G.ÙZ PAP 24 PAP-ma 7 me 3 se-en 58 (GÚ) 36!-2 MA SÍG.HI.A PAP 3

40%-6 MA SÍG.ÙZ TA MU.37.KÁM dAG-NÍG.D[U-URÙ LUGAL TIN.TI]R.KI a-di MU.1.KÁM dU+GUR-LUGAL-URÙ LUGAL TIN.TIR.KI sá la re-ha-n[u sá se-e-nu la am-ru] (upper edge) NÍG.KA9 ép-su-tu sá mdKUR.GAL-LUGAL-URÙ lúqí-i-pi sá É-an-na u mNUMUN-ya lúsà-tam É-an-na it-ti lúna- GADA.ME i-pu-su ITI.SIG4 UD.28.KÁM [MU].1.KÁM dU+GUR-LUGAL-PAP LUGAL TIN.TIR.KI

* If my transliteration diˆers from that by G. Van Driel (but not from the copy by Nemet-Nejat), the signs are in bold type. Transliteration appears without emendation except for the reconstruction of the only partly preserved numerals in “grand total” in l. 8. The places where the text has to be emended are marked in translation by exclamation mark. Note that the ˜rst column gives the number of line and is added for the comfort of the readers.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

28 PAP 1 me 31 29 13 30 4 31 1 me 70

2

117

2

7

ewes male female Total: lambs lambs white animals 90 12 15 total: 133

3 1 4 5 18

6 3 101

fullshe-goats male female grown kids kids he-goats 1 2 1

total: 12 total: 4 total: 171 1

3

6 2 7 8 30

10 3 119 18

3 2 40

total: 15 total: 5 total: 207 1

4

9 3 10 1 11 40

12 3 138 23

4 2 45

total: [1]9 total: 6 total: 246 1

5

12 3 13 1 14 54

14 4 158

27

4 3 53

total: 21 total: 8 total: 292 2

15 5 16 2 17 68

16 5 182 31

5 3 60

total: 26 total: 10 total: 341 1

18 7 19 2 20 89

18 5 209

total: 31 total: 10 total: 403 2

1

40

6 3 65

9

2

21 8 22 3 23 111

21 6 237

total: 35 total: 13 total: 469 3

1

1

41

6 4 80

10

2

24 10 25 4 26 131

23 7 275

total: 41 total: 15 total: 544 4

1

1

48

8 4 90

12

3

68

8

20

total: 101

27

5

6

1

total: 5

1

1

total: 7

1

1 2

total: 1 total: 9

1

total: 1

2

total: 11

1 [1]

8

2

2

total: 1 2

total: 13

2

total: 1 PAP 1 total: 15 total: 2

[3]

total: 18 total: 2

3

total: 22

1

total: 3

Grand total: sheep and goats, the property of the Lady-of-Uruk and Nanaya Grand total: 137, Nabû-ahhe-sullim, son of Nabû-sum-iskun the 36th! year Hides Wages Grand total: 176, 37th year; 2 lambs were received for shearing. 1 (talent) 20!-2 minas of wool (was the result of shearing). Hides Wages Grand total: 214, the 38th year; 4 lambs (were received) for shearing. 1 (talent) 55!!-2! minas of wool (was the result of shearing). Hides Grand total: 7, wages Grand total: 255, the 39!th year; 2 lambs, likewise. 2 (talents) 24!-2 minas of wool (was the result of shearing). Grand total: 22 hides. Wages Grand total: 303, the 40!th year; 6 lambs, 1 full-grown he-goat, likewise, 2 (talents) 45!-2 minas of wool (was the result of shearing). Grand total: 27 hides Wages Grand total: 354, the 41st year; 1 lamb, 1 ewe, likewise, 3* (talents) *18!-2 minas of wool (was the result of shearing). Grand total: 32 hides Grand total: 11, wages Grand total: 418, the 42!nd year; 7 lambs, likewise, 5!-2 (talents) 9!-2 minas of wool (was the result of shearing). Grand total: 37 hides Wages Grand total: 487, the 43!rd year; 7 lambs, likewise, 7 (talents) 1!-2 minas of wool (was the result of shearing). [Gr]and total: 43 hides Wages Grand total: 566, the 1st year of Amel-Marduk; 5 lambs, likewise, 8 (talents) 43!-2 minas of wool and 8%-6 minas of goat hair (was the result of shearing). Total: 104 as income from month of Adaru, the accession year of Amel-Marduk.

STEFAN ZAWADZKI

16

2 1 36

Total: black animals total: 4

118

1 Rams

28 total: 131 343 29 13 34 30 4 11 31 170 390

56

3

4 1

16

4

4

total: 68 total: 24 1 total: 888 8

2

1

17

4

1 5

8

1

1

10

3

4

32 16 33 6 34 209

39 11 453

80

13 7 146

35 among them: 5

138

14

41

36 remains: 315 193

66

105

total: 198

total: 679 7

total: 2[5] Grand total: 665, the 1st year of (AmM). total: 3 Grand total: 61 hides Wages total: 30 Grand total: 789, the 2nd year; 5 lambs for the shearing, {likewise}; 11 (talents) 22!-2 minas of wool; 20%-6 minas of goat hair (was the result shearing). total: 3 Grand total: 71 hides total: 2 Grand total: 26, wages total: 34 Grand total: 922, the 1st year of Neriglissar, king of Babylon; 8* lambs were received in Uruk; 3 lambs (were given) for shearing. total: 10 Grand total: 208 (animals) were seen (=accounted for) in the 1st year; 14 talents 5 minas of wool; 11* minas [10 shekels of go]at hair (was the result of shearing). total: 24 Grand total: 703 animals (were not seen). 58 (talents) 36!-2 minas of wool.

40%-6 minas of goat hair (was the result of shearing) from 37th year of Nebuchadnezzar, king of Babylon until the 1st year of Neriglissar, king of Babylon. As for the remaning [sheep and goat the account has not been made].(?) Balanced account which Enlil-sar-usur, the resident of Eanna and Zeriya, the administrator of Eanna, settled with the herdsmen. The month of Simanu, the 28th day, the 1st year of Nergal-sar-usur, king of Babylon.

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

total: 640 4 total: 55 1 total: 22 total: 759 6

14 1

66

110 11 7 133

119

120

STEFAN ZAWADZKI

ages of approximately eight years, whose value (including their value as food) was deteriorating, were apparently not taken by the temple at all. This was probably one of the essential diˆerences between the management of a temple herd and of smaller private ones, where the eating of fullygrown individuals, whose value is declining, was conceivably a much more general practice. Appendix 1. Inspection of the ˘ocks of the Ebabbar temple at Sippar Among the texts known from Sippar BM 75503 is especially important, because its relatively good state of preservation makes it possible to make many interesting comparisons. The text, written on an elongated, pillow-shaped tablet, contains the results of the inspection of seven ˘ocks belonging to the Ebabbar temple, which had been given to “the shepherds of †amas.” The time of the inspection and the animal terminology involved (puhalu, alittu, parru and parratu), is well known from other texts from Sippar. The animals are summed up by the word senu, which reveals that they included not only sheep but also goats. The text does not mention losses, and goats are not treated separately from sheep. This circumstance reveals that the goal of the document was to establish the number of animals and their sex on the day of inspection, and to note the quantity of wool delivered to the temple, with additional information concerning its quantity delivered for cultic purposes.70 From this point of view the text diˆers from many other texts in which changes in ˘ocks which had taken place in the period from the previous inspection are included. Despite this character of the text, important observations concerning the purpose of animal husbandry are still possible. There are important diˆerences between the individual herds. The largest was the ˘ock of †amas-zer-ibni (473 animals), the smallest were

70. According to my knowledge it is the ˜rst text, which includes such data.

those of †amas-udammiq (181 animals) and Taqÿs (184+x animals). The ˘ock of Ulmasaya (236 animals) is equal to half the ˘ock of †amas-zeribni; while in the ˘ock of Nabû-eres there are only 27 animals more than half of the ˘ock of †amas-zer-ibni. The ˘ocks of Ussaya (357 animals) and of Nabû-zer-ukÿn (ca. 323 + x), although smaller than the ˘ock of †amas-zer-ibni, were recognized as viable and their augmentation was not treated as desirable or crucial. Such a conclusion is based on the fact that only one male lamb was left in each herd, while if augmentation of the ˘ock was planned we would expect an increase of rams in the ˘ock. A glance at the percentage of females left in the herds, in comparison to ewes, suggests that the ˘ocks of †amas-zer-ibni, Ussaya, Nabû-zer-ukÿn and Ulmasaya had reached their maximum desired capacity, while the ˘ocks of Nabû-eres, †amas-udammiq, and particularly of Taqÿs, should have been on the increase. †amas-zer-ibni: female lambs comprise ca. 6.1% of the ewes in the herd. Nabû-eres: 14% Ussaya: ca. 5.8% Nabû-zer-ukÿn: ca. 3.7% (probably even lower, because in the herd were certainly more than 300 ewes). Taqÿs: ca. 19% Ulmasaya: ca. 7% †amas-udammiq: ca. 13.3%. These observations suggest that the condition of the ˘ocks was quite good. The ewes were reasonably young, in good breeding condition, and their replacement with younger animals would not be necessary during the next two to three years. Only in the ˘ock of Taqÿs (and to the same degree also in the ˘ocks of Nabû-eres and †amasudammiq) was the percentage of female lambs considerably higher in comparison to the ewes, but this might have been caused by the desire to enlarge the ˘ock. The considerably smaller number of rams in each ˘ock should be noted. The comparison of the number of rams to the ewes gives the following results:

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK †amas-zer-ibni: 24 ewes to one ram; Nabû-eres: 28 ewes to one ram; Ussaya: 36.5 ewes to one ram; Nabû-zer-ukÿn: ca. 27 ewes to one ram;71 Taqÿs: only one ram to 140+x ewes! Ulmasaya: 43 ewes to one ram; †amas-udammiq: 25 ewes to one ram. It seems that an average of 30 ewes to one ram secured full impregnation of all ewes in uncontrolled putting,72 and was very close to the rational need.73 From this point of view the situation in the ˘ock of Taqÿs was dramatic. We see that in the column “rams” the numeral 8 and below it 1 udupuhal “one ram” has been added. The explanation is that the numeral 8 concerns he-goats.74 Such a small total as 140+x ewes is not the result of a conscious decision by the shepherd, but must have resulted from attacks of wild animals, disease and/or rustling. To re-establish the correct proportions eight male lambs were left in the ˘ock. In the future, when male lambs would have become rams and female lambs shifted to the ewes group, the relation of rams to ewes would be about 20+x ewes for one ram or slightly more, allowing for the possible loss of male lambs during the year. It should be also noted that in the ˘ock of †amas-zer-ibni with the highest number of rams no single male lamb was left, which suggests that this quantity was adequate. Five male lambs in the ˘ock of Nabû-udammiq could be explained by a possible decision to enlarge the ˘ock, while three in the ˘ock of Nabû-eres by the replacement of a few rams or enlargement of the ˘ock. In all seven 71. Divided here are 300 ewes by 11 rams, although it is probable that both numbers (i.e., those of ewes and rams) were higher. 72. At the beginning, when only a small number of ewes have symptoms of rut, the rams are very active and ˜ght each other, but later, when the number of ewes ready for insemination increases, the activity of the rams slows down. For this reason in such ˘ocks there were more rams than in the ˘ocks with controlled putting (in modern times). 73. The situation diˆers substantially from the texts, which were the subject of van Driel’s studies, where the number of rams was much higher than the rational need. 74. If the ram were included in this number, the phrase ina libbi, “including” would have to be added.

121

˘ocks only nineteen male lambs were left, which indicates that the main purpose was to secure the oˆering for the everyday cult in the temple. The considerably smaller number of female lambs left in the herds (149 in comparison to 1783 + ewes, i.e., only ca. 8.5%) indicates—as already mentioned above—that most of the ˘ock had reached its maximal amount, and that female lambs had been taken by the temple administration, most probably for consumption. Because it seems unlikely that the temple needed all the young animals at same time, it is conceivable that they were sent to the bÿt urê, “fattening house.” This argument, and the information concerning the destination of wool for sattukku, i.e., for the production of cultic garments, suggests that the ˘ocks belonged to the ginÿ/sattukku category, and these shepherds (not known to me from other texts) were under the power of the reåi ginê/sattukkê. One interesting observation concerns wool. The division of the amount of wool by the number of rams and ewes suggests very low productivity of wool per animal: †amas-zer-ibni:

23400 shekels of wool: 447 sheep = 52.3 shekels per sheep. Nabû-eres: 9900 shekels of wool: 229 sheep = ca. 43.2 shekels per sheep. Ussaya: 12060 shekels of wool: 337 sheep = ca. 35.8 shekels per sheep. Nabû-zer-ukÿn: 13500 shekels of wool: 311+x sheep = less than 43.4 shekels per sheep (because the number of sheep was higher than 311). Taqÿs: 9300 shekels of wool: ca. 141+ sheep = ca. 66.0 shekels per sheep. Ulmasaya: 9000 shekels of wool: 220 sheep = ca. 41.0 shekels per sheep. †amas-udammiq: 9000 shekels of wool: 156 sheep = ca. 57.7 shekels per sheep.

122

STEFAN ZAWADZKI

However, such a method of calculating productivity could contain an important mistake. It is probable that some of the animals included among the full-grown ones were recognized as being too young to be shorn during the previous year. This would mean that the amount of wool should be divided by a smaller number of animals, but what this number might be impossible to determine. The amount of wool for sattukku oˆerings from ˜ve ˜rst ˘ocks is similar, and runs from 1 talent to 1 talent 15 minas. In the ˘ock of Ulmasaya the

data concerning sattukku wool is lacking, and in the ˘ock of †amas-udammiq the amount of sattukku wool is relatively small (only 20 minas from 2 talents 35 minas of wool). However, the lack of proportion between the total amount of wool and the sattukku wool suggests that the quantity of wool delivered for manufacturing garments depended most probably on the speci˜c orders of the weavers involved in manufacturing garments for gods and goddesses.

BM 75503 (AH 83–1–18, 844) 10.7 x 4.2 cm Transliteration 1. si-e-nu a-mir-tu4 sá lúSIPA.ME† sá ªUTU sá ina garimGi-l[u-s]ú 2. am-rat ITU.SIG4 UD.11.KÁM MU.1.KÁM mKam-bu-zi-ja LUGAL E.KI LU[GAL KUR].KUR 3. pu-hal a-lit-tú par-ri par-rat SÍG.HI.A PAP m (sic) md 4. 18 4 me 29 26 6 GÚ.[UN] ª30º ma-na UTU-NUMUN5. ina lìb-bi 1 GÚ.[UN x+]3 ªmaº-na sat-tuk ªDÙº A m†u-la-a md 6. 8 2 me 21 ª3º 31 2 GÚ.UN ª45?º [m]a-na AG-KÁM 7. ina lìb-bi 1 GÚ.UN sat-tuk m 8. ª9º 3 me 28 ª1 19 3 GÚ.UN 21 ma-na ina lìb-ªbiº Us-sá-a-a 9. 1 GÚ.UN 15 ma-na sat-ªtukº A mdAG-BE-TINit md 10. x+1 3 m[e x] ª1 11 3 GÚ.UN 45 ma-na AG-NUMUN11. ina lìb-bi 1 GÚ.UN sat-tuk DU m 12. ª8º 1 me 40+x ª8 27 2 GÚ.UN ª35?º ma-na Ta-qis udu 13. 1 pu-hal ina lìb-bi 1 GÚ.UN 5 ma-na sat-tuk m 14. 5 2 me 15 ª1 15 2!-2 GÚ.UN Ul-mas-a-a md 15. 6 1 me 50 5 20 2!-2 GÚ.UN 5 ma-na UTU-SIG5iq 16. ina lìb-bi 20 ma-na sat-tuk A mRi-mut Translation 1. Sheep and goats of the shepherds of †amas, which were inspected in the tamirtu of Gil[usu]. 2. Month of Simanu, eleventh day, ˜rst year of Cambyses king of Babylon, king of Lands. 3. rams ewes male female Wool Total lambs lambs 4. 18 429 26 6 tal[ents] ª30º minas, including 1 talent x+ ª3º †amas-zer-ibni, 5. [mi]nas for the oˆering son of †¿lâ 6. 8 221 ª3º 31 2 talents ª45?º [mi]nas, including 1 talent for the Nabû-eres 7. oˆering 8. ª9º 328 ª1 19 3 talents 21 minas, including 1 talent 15 minas Ussaya, son of 9. for the oˆering Nabû-mÿta-uballit 10. x+1 300+x ª1 11 3 talents 45 minas, including 1 talent for the Nabû-zer-ukÿn 11. oˆering 12. ª8º 140+x ª8 27 2 talents ª35?º minas, including 1 talent 5 minas Taqûs 13. 1 ram for the oˆering 14. 5 215 ª1 15 2!-2 talents Ulmasaya 15. 6 150 5 20 2!-2 talents 5 minas, including 20 minas for the †amas-udammiq,

BOOKKEEPING PRACTICES AT THE EANNA TEMPLE IN URUK

123

Translation 10. x+1 300+x ª1 11 3 talents 45 minas, including 1 talent for the Nabû-zer-ukÿn 11. oˆering 1. Sheep and goats of the shepherds of †amas, which were inspected in the tamirtu of Gil[usu]. 12. ª8º 140+x ª8 eleventh 27 2 talents ª35?of º minas, including 1 talent 5 minas 2. Month of Simanu, day, ˜rst year Cambyses king of Babylon, king of Taqÿs Lands. 13. 1 ram for the oˆering 14. 5 215 ª1 15 2!-2 talents Ulmasaya 15. 6 150 5 20 2!-2 talents 5 minas, including 20 minas for the †amas-udammiq, 16. oˆering son of Rÿm¿t Comments l. 3: Instead of PAP we would have expected here rather ina IGI “at disposal” or IGI “before” or similar, but not “total”. l. 5: x + ª3º could be also [1]ª6º, or [1]ª9º minas. l. 6: the reading ª42?º minas seems less probable, but is not excluded (only the lower part of two verticals are preserved). l. 10: Because one vertical stroke is preserved, the only possible reconstruction is 3, 11, 12 or 13. Three rams in a ˘ock of over 300 hundred ewes is obviously too small, so only the numeral 12 or 13 rams could be taken into account.