ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR: A REEXAMINATION OF OBGT VI-X Peter J. Huber (Klosters)
The Old Babylonian Grammatical Texts (OBGT) in MSL IV give a detailed picture of Sumerian verbal syntax as seen througb tbe eyes of an Old Babylonian grammarian. Tbe picture does not necessarily correspond to any "true" Sumerian syntax. Tbe relevance of tbese texts lies in tbe fact tbat tbey are tbe earliest serious grammatical documents and tbey are more interesting as sucb tban as evidence for "true" Sumerian grammar. It seems tbat Black (1991) and Zoiyomi (2000) were tbe first to empbasize tbis bistoryof-science aspect. All authors prior to Black, and to some extent even Black, seem to bave taken tbe opposite view: even while appreciating tbe bistorical aspect of tbe grammatical texts, tbeir analyses were primarily interested in tbe ligbt tbe texts migbt tbrow on "true" Sumerian grammar. Here I investigate tbe model of Sumerian grammatical structure as it was created and understood by Old Babylonian scribes. Thus, this endeavor may be compared to the investigations reported by Bellugi and Brown (1971), which deal with tbe peculiar grammars underlying tbe utterances of cbildren at different stages of language acquisition, and witb tbe cbildren s implied formalization of language structure. Metbodologically, for my purpose, only tbose texts are suitable tbat contain Akkadian interpretations and are sufficiently systematically organized to reveal tbe underlying grammatical structure. Tbe texts OBGT VI-X are tbe only ones tbat satisfy tbese requirements.
Diakonoff once remarked facetiously tbat tbere are as many kinds of Sumerian as tbere are authors of Sumerian grammars. This also holds true for Old Babylonian autbors. It is fortunate tbat tbe texts OBGT VI-X form a closely knit group. I will attempt to extract from tbis material, and from this material alone, tbe model of Sumerian grammar assumed by tbe autbor(s) of tbese texts, and will consult otber texts and modern grammars for comparative purposes only. In particular, tbe only permissible structural information is tbat wbicb can be pulled from tbese texts tbemselves; tbe evidence from unilingual texts and modern grammars must be ignored. For semantic information, we migbt look furtber, if necessary. Despite tbese methodological restrictions, by paying close attention to the grid structure, I believe I bave been able to go beyond Black in several repsects and sbarpen some structural conclusions. Tbese advances concern in particular tbe structure of OBGT VII, the relationship between OBGT VI and X, tbe pronomial infix cbain, and tbe causative constructions. Sumerian, Akkadian, and modern Englisb or German syntax are sufficiently different tbat a translation of isolated verbal forms is impossible outside of a sentence context, strictly speaking. In particular, tbe Akkadian verbal case system makes fewer distinctions tban tbe Sumerian. Somewbat surprisingly, direct objects never seem to be referenced by an Akkadian pronoun in tbese paradigms; tbe Akkadian accusative refers eitber to a comitative object or to a subordinate subject. 1
JCS 59 (2007)
PETER J. HUBER To furtber complicate matters, Sumerian is an ergative language. Thus, the subject of an intransitive sentence and the direct object of a transitive sentence are treated in tbe same way. But, like most ergative languages, it is split ergative. Tbis sbows up in transitive paradigms by tbe fact tbat tbe same Sumerian pronoun in the identical position sometimes refers to tbe subject and sometimes to tbe direct object, depending on factors sucb as tbe tense. Since I do not want to enter into a discussion of tbe meaning of tbe Akkadian terms maru and hamtu, I will generally use tbe notions of present and preterite tense, wbicb at least bave an establisbed meaning in tbe Akkadian language, and tbereby permit a classification of Sumerian forms according to tbeir Akkadian counterparts. In order to empbasize the structure intended by tbe OB grammarian, I sball quote tbe texts by paragrapbs ratber tban by lines. Tbe order of presentation is more or less tbe one suggested by tbe paradigms. The construction principles that seem to underly tbe paradigms OBGT VI-X can be summarized as follows: • Tbe paradigms were designed to illustrate specific, sticky issues of Sumerian verbal syntax (as understood by OB grammarians), most of tbem also controversial in modern grammars
(Tbomsen 1984, Attinger 1993, Edzard 2003, Micbalowski 2004). • Tbe paradigm grids were constructed on tbe basis of tbe Akkadian grammar, witb its coarser case structure. But it would be a gross oversimplification to consider eitber tbe Sumerian or tbe Akkadian column as a translation of tbe otber. • Tbere are didactic inserts to bighlight Sumerian peculiarities and otber features tbat do not fit into tbe Akkadian straitjacket of tbe grids. 1. The Texts The texts are published in MSL IV (1956) as OBGT VI-X, with important additions by Black (1991). Tbeir provenance bas variously been conjectured as Larsa {MSL IV, p. 1*) or more firmly attributed to Nippur (Black 1991: 11), but it remains unknown (personal communication from M. Givil). I sball refer to tbese tablets according to their present location as The Oriental Institute (01) recension. A substantial initial part of OBGT VII is also preserved in a recension from Ur {UET 7 97, 98, 100, 101; see Black 1991: 13748). A comparison between tbe recensions sbows tbat for tbe assumed Sumerian morphology, we often have to rely on the grid structure, as tbere are too many variations of spelling. For example:
VII §35:
ga-a-mu-un-se-en-ze-en (01), gen-am-si-ze-en (Ur) common morpbology: /gen/-/m/-/n/-/si/-/enzen/ = alkanissum = "come(pl.) to bim!" Grammatical texts can bigbligbt sticky points tbat become visible only wben seen througb the rigid grid of a paradigm. For example, the grid permits separating the prefix mu- from the graphically similar ventive and the pronoun of the 1st person. Furthermore, tbe paradigms can elucidate tbe pbonetical and grapbical vagaries of the cuneiform representation of morphological consonant clusters. A surprising amount of grammatical insight can be gleaned already from a careful scrutiny of tbe grid structure.
Tables 1 to 4 present extracts from tbe paradigms, designed to bigbligbt tbis grid structure. Gaps in tbe numbering of tbe paragraphs (VII §11, 14-15; VI §21-28, 35, 56-57, 64-65) correspond to tbe didactic inserts mentioned above; tbey sball be discussed separately. In Tables 3 and 4,1 bave suppressed tbe (mostly straigbtforward) Akkadian and Englisb interpretations.
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR Table 1. Tbe first ten paragrapbs of OBGT VII (taken from tbe Ur recension, UET 7 100). Non-indicative forms (Ni): imperative, volitive precative. An analysis of tbe Akkadian structure is given on tbe rigbt. OBGT VII. Non-indicative forms: imperative. volitive, precative
Akk. structure
gen-am ga-am-gen be-em-du
al-kam lu-ul-li-kam li-il-li-kam
come! may I come! may be come!
-
gen-am-se ga-am-si-gen be-em-si-du
al-ka-as-sum lu-ul-li-ka-as-sum li-li-ka-as-sum
come to bim! may I come to bim! may be come to bim!
3D G V Ni
5 6 §3
7 8 9
gen-am-mu-se ga-mu-e-si-gen be-mu-e-si-du
al-kam a-na se-ri-ya lu-ul-li-ka-ak-kum li-li-ka-kum
come to me! may. I come to you! may be come to you!
ID G V Ni 2D 2D
§4
10 11 12
gen-am-ma ga-am-ma-gen be-em-ma-du
at-la-kam lu-ut-ta-al-kam li-it-ta-al-kam
come away! may I come away! may be come away!
-
§5
13 14
gen-am-ma-se ga-am-ma-si-gen 15 be-em-ma-si-du
at-la-ka-as-sum lu-ut-ta-al-ka-as-sum li-it-ta-al-ka-as-sum
come away to bim! may I come away to bim! may he come away to him!
3D Gt V Ni
§6
16 gen-am-ma-mu-se 17 ga-am-mu-e-si-gen 18 be-em-mu-e-si-du
at-la-kam a-na se-ri-ya lu-ut-ta-al-ka-ak-kum li-it-ta-al-ka-ak-kum
come away to me! may I come away to you! may he come away to you!
ID Gt V Ni 2D 2D
§7
19 gen-ni 20 ga-gen 21 be-{en-}du
a-lik lu-ul-lik li-il-lik
go! may I go! may he go!
-
§8
22 gen-en-si 23 ga-en-si-gen 24 be-en-si-du
a-lik-sum lu-ul-lik-sum li-lik-sum
go to bim! may I go to bim! may be go to bim!
3D G — Ni
§9
25 gen-ba 26 ga-ba-gen 27 ba-ba-du
at-la-ak lu-ut-ta-la-ak li-it-ta-la-ak
go away! may I go away! may be go away!
-
at-la-ak-sum lu-ut-ta-la-ak-sum li-it-ta-la-ak-sum
go away to bim! may I go away to him! may be go away to bim!
3D Gt — Ni
§1
1
2 3
§2
4
§10 28
gen-ba-si 29 ga-ba-si-gen 30 ba-ba-si-du
G V Ni
Gt V Ni
G — Ni
Gt — Ni
PETER J. HUBER Table 2. Tbe next twenty paragrapbs of OBGT VII, up to tbe end of tbe singular sections (taken from tbe Ur recension, UEn 100 and 101),firstlines only. Present (Ps) and preterite tense (Pt) forms. Paragrapb numbering according to OBGT VII, line numbering according to UET 7. Tbe paragrapbs are arranged in the ordering of the Ur recension. An analysis of the Akkadian structure is given on the rigbt. OBGT VII. Indicative forms: present, preterite §16 31 §17 34 §21 37 §18 39 §19 42 §20 45 §12 47 §13 50 §22 53 §23 56 §26 59 §27 62 §31 65 §28 67 §29 70 §30 73 §24 75 §25 78 §32 81 §33 84
Akk. structure
am-du am-si-du mu-e-si-du am-ma-du am-ma-si-du am-mu-e-si-du
illakam illakas.^um illakakkum ittallakam ittallakassum ittallakakkum
be comes be comes to bim be comes to you be comes away be comes away to bim be comes away to you
3D 2D 3D 2D
G G G Gt Gt Gt
V V V V V V
l-du in-si-du ba-du ba-si-du
illak illaksuni ittallak ittallaksum
be goes be goes to bim be goes away be goes away to bim
3D 3D
G G Gt Gt
-
Ps Ps
i-im-gen i-im-si-gen mu-e-si-gen im-ma-gen im-ma-si-gen im-mu-e-si-gen
illikam illikasSum illikakkum ittalkam ittalkassum ittalkakkum
he came he came to him he came to you he came away he came away to him he came away to you
3D 2D 3D 2D
G G G Gt Gt Gt
V V V V V V
Pt
in-gen, l-gen in-si-gen ba-gen ba-si-gen
illik illiksum ittalak ittalaksum
he went he went to bim be went away be went away to him
3D 3D
G - Pt G - Pt Gt - Pt Gt - Pt
2. The Paragraph Structure of the Paradigms As mentioned above, already some simple observations of tbe paradigm grids lead to quite subtle conclusions on tbe language structure of Sumerian and Akkadian, as perceived by tbe OB grammarian. Tbe paradigms are subdivided into paragrapbs, wbose structure is based on Akkadian conjugation. Most paragrapbs bave 3 lines, in tbe order: 3d-, 1st-, 2nd-person subject. Witb non-indicative forms, tbe order is reversed: imperative(2nd), volitive(lst), precative(3d). Tbis paragrapb structure divides the Sumerian forms into two conjugation types:
Ps Ps Ps Ps Ps Ps
Ps Ps Pt Pt Pt Pt Pt
Suffix conjugation (l-du, l-du-un, l-du-un: "he goes," "I go," "you go") • intransitive verbal forms • present tense forms of transitive verbs • Non-indicative forms (implicit in singular, explicit in plural) • forms rendered by Akkadian passives • most stative forms Infix conjugation (mu-un-gar, mu-gar, mu-gar: "be placed," "I placed," "you placed") • preterite tense forms of transitive verbs • some stative forms
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR Table 3. OBGT VI, non-indicative forms. An analysis of tbe Akkadian structure is given on top and on tbe rigbt. Akk.
OBGT VI. Non-indicative forms
Imperative
Volitive
Precative
G S G S G S
— — 3A 3A 3D 3D
—
G S G S G S
— — 3A 3A 3D 3D
V
G S G S
'/2D — '/2D — I/2D V '/2A '/2A
he-ri-ib-gar
G G S
I/2A
V —
bu-mu-ri-ib-gar
s
I/2A
V
gar-ra
ga-gar
be-gar
ga-bi-ib-gar
be-bi-ib-gar
§3:
gar-bi-ib gar-ra-an-da
ga-an-da-gar
be-en-da-gar
§4:
gar-ra-ni-ib
ga-ni-ib-gar
be-ni-ib-gar
§5:
gar-ra-na-ab
ga-na-ab-gar
be-na-ab-gar
§6:
gar-ra-na-ni-ib
ga-na-ni-ib-gar
be-na-ni-ib-gar
§7:
gar-ma-ab
ga-am-gar
be-em-gar
§8:
gar-ma-ni-ib
ga-mu-ni-ib-gar ga-am-da-gar
hu-mu-ni-ib-gar be-em-da-gar
ga-ma-ni-ib-gar
be-ma-ni-ib-gar
§11: gar-mu-un-na-ab §12: gar-mu-na-ni-ib
ga-mu-un-na-ab-gar ga-mu-na-ni-ib-gar
bu-mu-un-na-ab-gar hu-mu-na-ni-ib-gar
§13: §14: §15: §16:
ga-ra-ab-gar ga-ra-ni-ib-gar ga-mu-ra-ab-gar
ba-ra-ab-gar ba-ra-ni-ib-gar bu-mu-ra-ab-gar
ga-mu-ra-ni-ib-gar
bu-mu-ra-ni-ib-gar
ga-e-da-gar ga-mu-e-da-gar
he-e-da-gar he-mu-e-da-gar
ga-ri-ib-gar ga-mu-ri-ib-gar
§1: §2:
§9:
gar-ma-da-ab §10: gar-ma-ni-ib
gar-ma-ra — — —
§17: gar-mu-un §18: §19: gar-mu-ub §20: —
Tbere is a kind of duality: in transitive constructions, tbe position before or after tbe base not used for tbe transitive subject is used for tbe direct object. Tbe above pattern sbows tbat Sumerian in tbe preterite tense bebaves as an ergative language (tbe telltale symptom is tbat intransitive subjects and direct objects are treated alike). But in tbe present tense it uses tbe nominative-accusative pattern. Hence, as in most ergative languages, we bave split ergativity, tbougb, in tbe absence of a sentence context, furtber details of tbe split are difficult to ascertain; see Micbalowski (1980), and note tbat Edzard (2003: 90-91) still takes a negative position on split ergativity. I sball now discuss tbe individual paradigms and tbe specific issues tbey appear to address.
stem obj — — — — —
V V V V V
'/2D V —
3. OBGT VII: Intransitive Verb Tbis paradigm is concerned witb an intransitive verb (gen/du = alakum = "to go"). Apparently, it was designed to exercise tbe directional prefixes m-, ba-, mma-, and tbe singular and plural pronouns. In tbe order preferred by tbe OB grammarian (3d, 1st, 2nd person) tbese are: suffixed pronouns -0, -en, -en; -es, -enden, -enzen, indirect objects -n-, -mu-, -e-; -ne-, -me-, -ene-. Its grid is very systematically organized, especially in tbe Ur recension, wbicb omits tbe didactic additions of the pther recension and applies a more logical ordering principle. However, tbe preserved part is sborter and ends witb a catcb-line to §70, just after tbe beginning of tbe section covering
PETER J. HUBER Table 4. OBGT VI, indicative forms (first lines only). An analysis of the Akkadian structure is given on top and on the right. Stative (suffix conj.)
Preterite (infix conjugation) 0 (l- prefix) 0 (mu- prefix)
Akk. stem obj
§29: an-gar §30: ba-ab-gar
§36: l-gar §37-§39: (insert)
§44: mu-un-gar §45: mu-ni-in-gar
G S
—
§31: an-da-gar §32: ba-da-ab-gar
§40: in-da-gar §41: in-di-ni-ib-gar
§46: in-da-gar §47: in-di-ni-ib-gar
G S
3A 3A
§33: an-na-gar §34: an-na-ni-ib-gar
§42: in-na-an-gar §43: in-na-ni-in-gar
§48: in-na-an-gar §49: mu-na-ni-in-gar
G S
3D 3D
§72: a-ra-gar
§73: i-ra-an-gar i-ra-ni-in-gar
§74: mu-ra-an-gar mu-ra-ni-in-gar
G S
2D 2D
t-stem
Preterite (infix conjugation) t-stem + ventive
ventive
Akk. stem obj
§50: ba-an-gar §51: ba-ni-in-gar
§58: im-ma-an-gar §59: im-ma-ni-in-gar
§66: ma-an-gar §67: ma-ni-in-gar
G S
—
§52: ba-da-an-gar §53: ba-di-ni-ib-gar
§60: im-ma-da-an-gar §61: im-ma-di-ni-ib-gar
§68: ma-da-an-gar §69: ma-di-ni-ib-gar
G S
3A 3A
§54: ba-na-an-gar §55: ba-na-ni-in-gar
§62: im-ma-na-an-gar §63: im-ma-na-ni-in-gar
§70: ma-si-in-gar §71: ma-si-ni-in-gar
G S
3D 3D
§76: im-ma-ra-an-gar im-ma-ra-ni-in-gar
§75: ma-ra-an-gar ma-ra-ni-in-gar
G S
2D 2D
—
plural objects. See tbe excerpts listed in Tables 1 and 2. 3.L Grid Structure of OBGT VII Tbe paragrapbs of tbe Ur recension are ordered rigidly (01 is less systematic). Tbey vary fastest:
person of subject (inside paragrapb) person of object G,Gt ventive, non-ventive tense (non-indicative, present, preterite) number of subject (singular, plural) slowest: number of object (singular, plural)
Tbe grid of OBGT VII is complete, apart from tbe following systematic restrictions: • no self-references (lst-lst, 2nd-2nd person: semantic problem) • no lst-person singular objects, except witb imperative (Akkadian problem: ventive coincides witb lst-person dative). • motion toward 1st or 2nd person requires ventive (Sumerian and/or Akkadian restriction) I suspect tbat tbe exceptional four imperatives witb lst-person singular objects were selectively included in order to illustrate tbat in Sumerian tbe combination of ventive and lst-person could be expressed witbout circumlocution.
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR 3.2. Additions of the 01 Version (total 8 paragraphs) 01 (but not Ur) adds 6 paragrapbs witb alternate prefixes. Tbese additions cover singular and plural, but present tense only: regular §12: regular §13: added §11: added §14: added §15:
l-du in-si-du al-du an-du an-si-du
= illak = illaksum = illak = illak = illaksum
= "be goes" = "be goes to bim" (n: belongs to prefix) (n: 3d person pronoun)
Table 4 in particular (§§29-34, §72, see section 4.2.1.), and also otber paradigms sbow tbat tbe prefixes an-, a- bave stative function (or at least: stative tendency): VIII §34: §35: §36:
l-guy-e an-guy
= ikkal = ikkal = akil
= "be eats" = "be eats" = "be is eating"'
Most modern autbors treat al- as unique among prefixes. Attinger (1993: 269) takes a-, al- as complementary expressions of tbe same prefix, written al- if it is immediately followed by tbe base, otberwise a-; see also Edzard's comments (2003: 111). Tbe OB treatment does not quite correspond to tbis. Wbile al- always is immediately followed by tbe base, our paradigms sbow tbat tbis is possible also witb a-. Tbe texts alternate between writings a- and an-. Note tbat in VII §14 (and also in VI §29) -n- cannot be a pronoun, since tbe construction is intransitive, witb suffix conjugation. I do not tbink tbat tbe OB grammarian distinguisbes between
added (§71):
ba-me-du ba-e-ne-du
= ittallak niati = ittallak kunuti
two separate morpbemes /a/ and /an/, and I would be very reluctant to assume an- to be a scribal error. Tbis induces me to conclude tbat be uses an- as a grapbic rendering of a nasalized a-, notwitbstanding tbe Sumerologists' objection tbat true Sumerian does not know sucb a pboneme. I conclude tbat be treats l-, al-, a- as tbree separate prefixes, but bas difficulties expressing tbe distinctions in Akkadian. 01 moreover adds two anomalous paragrapbs (§71 present and §74 preterite), witb unique accusative pronouns in Akkadian, and unique nonventive lst/2nd-person objects in Sumerian:
= "be goes away/rom us" = "be goes away froin you"
See Jacobsen (1960) for tbe interpretation of tbe Akkadian accusative. Witb motion toward a 1st or 2nd person, tbe ventive is mandatory, but not so witb tbe 3rd person, wbere tbe grid of tbe paradigm offers: regular (§88): ba-ne-du
= ittallak sunusi
= "be goes away to tbem"
1. For the active interpretation of akil, see AHw p. 36: a-kil-a-ti, "du bist fressend."
PETER J. HUBER Note tbat witb plural objects, tbe terminative case is not marked in tbis paradigm, creating an ambiguity witb tbe direction of motion. (Anotber paradigm, N 3513+N 3592 [Black 1991: 155-58] uses -si- also in tbe plural, demonstrating tbe multiplicity also of OB grammars!) Tbe two paragrapbs §71 and §74 are additions to a very disciplined and complete paradigm. In my opinion, tbey cannot be explained away as "errors of a careless scribe" (Black 1991: 17). I believe tbey are targeted to illustrate semantic differences between deceptively similar non-ventive Sumerian forms witb tbe separative ba-: in tbe absence of tbe ventive, tbe first two of tbe forms quoted above default to motion away from, tbe tbird to motion away toward. 3.3. The Directional Prefix: m-, ha-, mma-. Tables 1 and 2 sbow clear correspondences: Akkadian: Sumerian: ventive «-»• /m/ Gt-stem ^ /ba/ ventive + Gt <-• /m/+/ba/ > /mma/ Tbe Akkadian t-infix corresponds to t-stems (not to tbe perfect, as bas been claimed by some autbors). Tbis is sbown clearly by tbe grid structure and by tbe fact tbat tbe t-infix occurs not only witb tbe ambiguous preterite, but also witb tbe present tense and witb Ni-forms, see also Black (1991: 27-28). Tbe basic meaning of tbe Akkadian t-infix is a cbange of direction (in OBGT VII it distinguisbes between "to go" and "to go away"). Tbe Akkadian ventive originally seems to express a motion towards "me," tben also toward otbers (bere it distinguisbes between "to go" and "to come"). I will denote tbe triple (/m/, /ba/, /mma/) collectively as directional prefixes, and I will use "ventive m" as a sbortband notation for a Sumerian m tbat is rendered by an Akkadian ventive. Tbis does not imply tbe existence of a Sumerian ventive, nor even tbat tbe OB scbolars responsible for tbe grammatical texts tbougbt tbere was one. But tbey certainly felt it expedient
to describe certain Sumerian features involving an m-morpbeme in terms of tbe Akkadian ventive. Table 2 sbows tbat botb m- and ba- can begin verbal forms. In tbe absence of tbe ventive m and tbe separative ba, tbe prefix i- is used, sometimes written in- (before tbe nasalized g of gen, §24, or before tbe 3d-person pronoun -n-, §13, §25), botb in tbe present and preterite tense. Before consonant clusters, a prostbetic vowel is used, a- witb Ni-forms and witb tbe present tense, i- in tbe preterite. Since, in tbe absence of consonant clusters, tbe vowel i- is used botb in tbe present and in tbe preterite tense, I believe tbat tbese prostbetic vowels are not morpbemic, but pbonemic, and serve to express a difference of pitcb or stress. But l- seems to be required witb indicative forms in tbe absence of a directional prefix. Before tbe pronoun lei of tbe 2nd person, tbe ventive consistently is written -mu-. Also Iml+lhal tben is written -mmu- (§3, §6, §20-21, §30-31). On tbe otber band, before tbe rnarti-base lei of dug4, no sucb lul appears, compare in particular IX §34: sa am-e = ikassadam = "be reacbes bere," IX §35: sa mu-e-a = ikassadaka = "be reacbes you bere." Morpbologically, we bave am-e < Iml-lel, and mu-e-a < Iml-lel-lel, witb tbe final lei being tbe maru-hase of dug4. My conclusions are, first, tbat tbe 2nd-person pronoun lei in tbese paradigms is tbougbt to be more tban a plain e, presumably *we, and, second, tbat tbe writing am seems to express a syllabic m. 3.4. Distinction Between Ventive and 1 st Person In Table 1, tbe writing gen-am-ma-mu-se = at-la-kam a-na se-ri-ya = "come away to me!" of §6 is interesting. One sbould compare tbis form to tbose of tbe preceding five paragrapbs. Morpbologically, tbe first m corresponds to tbe ventive, tbe second to an assimilated ba, and tbe tbird to a 1st person mu. It sbows tbat tbe OB grammarian bere made a pointed distinction between tbe ventive Iml and tbe 1st person /mu/. By tbe way, fully spelled-out writings witb tbree m's are rare. Tbey occur also in N 3513+N 3592 (Black 1991:
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR
155-58), but very often, sucb forms are contracted. For example, §6 of tbe duplicate UET 7 101 does not distinguisb tbe separative form in writing from tbe non-separative form gen-am-mu-se = al-kam a-na se-ri-ia = "come to me!" in §3. 4. OBGT VI+X: Transitive Verbs
OBGT VI is concerned witb gar = sakdiium = "to put." As a rule, Akkadian translations bere are given only for tbe first line of eacb paragrapb. OBGT X is concerned witb gub = izuzzum = "to stand." In Sumerian, it mostly is constructed transitively(!) as "to set up." OBGT X gives only tbe first line of eacb paragrapb. Its grid agrees witb tbe systematic part of VI, but tbe latter bas several additions. Tbis duplication of tbe systematic part of tbe grid facilitates recognition of tbe common underlying structure of tbese paradigms, wbicb is not as systematic and complete as tbat of OBGT VII (see Tables 3 and 4). Tbese paradigms exercise prefixes, tbe pronomial infix cbain, and causative constructions. Note tbat tbey group tbe paragrapbs into pairs, consisting of a non-causative G-stem and tbe corresponding S-stem causative paragrapb, tbereby empbasizing and exercizing tbe causativity transformation. 4.L Tenses and Aspects A first look at Tables 3 and 4 sbows tbat tbe paradigms of OBGT VI and X are arranged according to Akkadian tenses or aspects, in tbe order: Non-indicative, stative, preterite, and present. Tbere are present tense sections at tbe end of botb VI and X (7 and 9 one-line paragrapbs, respectively), but tbey are sbort and unsystematic. Tbey do not fit into tbe grid of Table 4 and tberefore bave been omitted. 4.2. Prefixes Tbe systematic part of OBGT VI (see Table 4) distinguisbes between tbe "stative" prefixes a- and ba-, tbe "main" prefixes l-, mu- (wbicb are indistinguisbable in tbe Akkadian rendering), and tbe
"directional" prefixes ba-, m-, and mma-, corresponding to t-stems, ventive, and tbe combination of t-stem and ventive. In addition, tbere is an exceptional prefix bi- occurring in tbe irregular insert OBGT VI §37-39, see below. So tbe OB grammarian seems to recognize at least six, and perbaps up to nine, prefixes. Tbey conveniently can be be segmented into tbree groups ("stative" a-, ba-, plus possibly al-; "main" i-, bi-, mu-; "directional" m-, ba-, mma-). Tbis segmentation is based solely on tbe (Akkadian) grid structure of OBGT VI, as exbibited by Table 4. Tbe nine prefixes seem to be mutually exclusive, so tbey all seem to bave tbe same rank. But tbeir number may bave to be reduced, since one of tbem is composite (mma-, notwitbstanding tbat tbe reduction creates conceptual problems witb tbe ranks), al- may be a mere variant of a-, and two otbers may be identical (ba-), see below. It is interesting to contrast tbe OB view of tbe prefixes witb modern views of "true" Sumerian, for example Micbalowski (2004: 44): The [conjugation] prefixes [...] constitute the most controversial part of Sumerian grammar. No two Sumerologists appear to agree fully on their form, meaning, etymology and identity; the number of ranks that they occupy is equally disputed. [...] I prefer a minimalist position according to which there are only four distinct "conjugation" prefixes: mu-, ba-, i- (or V-) and imma-. [...] I do not break these down into smaller components, as do many others. 4.2.L Stative Prefixes: a- and ba-. In tbe stative section of tbe paradigm (VI §§29-35, §72), a- corresponds to tbe ordinary Akkadian G-stem stative (witb an exceptional S-stem in §34), and ba- to tbe Akkadian S-stem stative/passive. Tbe Akkadian stative is indifferent witb regard to tense, and for transitive verbs it usually bas a passive meaning ("be is placed"), but an active meaning ("be is placing") is possible too (see GAG, 100-102). Tbe Englisb translations proposed below are tentative; lacking a context, it is difficult to be certain about tbe difference between tbe suffix conjugation in §29 and tbe infix conjugation in §35:
10
PETER J. HUBER
VI §29 an-gar an-gar-re-en an-gar-re-en
sakin
VI §30 ba-ab-gar ba-ab-gar-re-en ba-ab-gar-re-en
suskun
VI §35 ab-gar a-gar e-gar
sakin
(saknaku) (saknata) (suskunaku) (suskunata) (saknaku) (saknata)
Despite tbeir seemingly disparate functions, tbe separative ba- and tbe stative-passive ba- may be identical, also in tbe OB grammarians view, cf. tbe comments on VI §56-57, §64-65 in tbe following subsection. Gonceivably, ba- bas a basic "final" function, similar to tbe Englisb "off," indicating a move out of tbe area of immediate control. In tbe paradigms tbis function is approximated variously by an Akkadian separative or a stative-passive. 4.2.2. Irregular Inserts with ba-forms. Tbe regular grid of OBGT VI does not bave any Niforms witb tbe ba- prefix, but sucb forms are
be(0) is placed I am placed You are placed be(0) was placed by someone (b) I (en) was placed by someone (b) You (en) were placed by someone (b) someone(b) is placing 1(0) am placing You(e < a-e) are placing given in an irregular addition (§22-28). Tbe corresponding paragrapbs are omitted in OBGT X, so tbey may be later additions, perbaps intended as a kind of Ni-counterpart to tbe subsequent stative section (§29-35). In any case, tbe Akkadian side of §§22-26 offers ratber surprising N-stems, indicating a passive aspect, and tbe Sumerian forms appear to be intransitive, witb -b indicating not a direct object, but a (subordinate) subject (see Section 5). Perbaps tbere is an ambiguity as witb tbe Englisb "to bide": put away sometbing or oneself. My Englisb translations are very tentative.
VI §22 gar-ba
naskin
bave (it) put away! (?)
VI §23 gar-ba-na-ab
naskinsum
bave it put away for bim! (?)
VI §24 gar-am-ma
naskinam
bave (it) put away bere! (?)
VI §25 gar-am-ma-se- [eb] naskinassum VI §26 gar-am-ma-se
bave (it) put away toward bim bere! (?)
naskinassui OT>put away toward bim bere! (?)
§§27-28 seem to empbasize tbe separative aspect of -ba- by adding tbe ablative particle -ta-:
VI §27 gar-ba-ta VI §28 gar-/ba-na\-ta
sitkan
sitkassum
put away! put away for bim!
Two furtber additions (§§56-57, §§64-65) witb indicative forms and suffix conjugation, also omitted in OBGT X, are even more unsual, since tbey use Akkadian Nt-stems. Tbey seem to bigbligbt tbe dual separative and stative/passive aspect of -ba-:
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR
VI §56 ba-gar
ittaskan
be/it was put away
VI §57 ba-na-gar
ittaskansum
be/it was put away for bim
VI §64 im-ma-{an-}gar
ittaskanam
be/it was put away bere
VI §65 im-ma-na-gar
ittaskanassum
be/it was put away bere for bim
11
4.2.3. The Prefixes i-, mu- and bf. In tbe para- D.OBJ, tbe 3d-person direct object being invisible in suffix position); it is rendered by straigbt Akkadigms, tbe use of a conjugation prefix seems to be dian G-stems. Often, parallel forms witb l- or mandatory for indicative forms. Tbis implies tbat mu- prefix are given (VI §36 l-gar, §37 bi-in-gar, tbere must be a default prefix, to be used wben §44 mu-un-gar, all = iskun; X 26 i-gub = izziz, tbe basic, unmodified meaning of tbe verb is inX 27 bi-in-gub = usziz, probably a scribal error tended. Tbis prefix by definition tben does not for izziz, X 34 mu-un-gub = izziz; VIII §27 [l]modify tbe meaning of tbe verb. Tbe most likely guy, §32 bi-in-guy, §33 mu-un-guy, all = ikul). candidate for sucb a default prefix is l-, since it bas Sometimes only tbe bi-version is given (VIII §18 tbe widest use and can be used botb witb intrankas4 bi-in-dug4 = ilsum, IX §27 sa bi-in-dug4 = sitive and transitive constructions. iksud). Witb Ni-forms, tbe only visible conjugation preTbe second construction consists of simple fixes are tbe directional ones, and tbere is no firm causatives witb suffix conjugation (bi-ib-BASEevidence for a mandatory conjugation prefix. Tbe SUBJ). It is rendered by straigbt Akkadian S-stems, vowel appended to tbe naked base for imperative mostly in tbe preterite tense (VI §38 bi-ib-gar = forms (usually -a) migbt perbaps be interpreted usaskin; X 28 bi-ib-gub = usziz; VIII §28 bi-ibas a version of tbe prefix l-, but tbis vowel could guy = [usakil]; IX §28 sa b[i-ib-dug4l = {usak]sid), also be a mere pbonetic pbenomenon. plus a few examples in tbe present tense witb unIn tbe paradigms tbe mu-prefix occurs only witb clear construction (VI §83 bi-ib-gar-re = usaskan; transitive constructions. Sucb exclusive usage may X 69 bi-ib-gub-be = uszaz). imply tbat mu- expresses some special relationsbip Tbus, wben tbere is a definite personal pronoun between subject, action and direct object. Most (be, I, you) before tbe base, tbese bi-prefix forms occurrences are preterite tense forms using tbe in Akkadian are interpreted as ordinary transitive infix conjugation, apart from one occurrence witb constructions "be/I/you placed bim/it," witb an inpresent tense and suffix conjugation: VIII §14 kas mu-si-ib-be = i-la-sum-sum, "be runs toward bim." visible 3d-person direct object in suffix position. Witb tbe indefinite pronoun -b- (it, somebody, etc.) Tbe mu- bere is not tbe ventive; tbe latter occurs in VIII §16 kas4 am-si-ib-be = i-la-su-ma-as-sum. before tbe base, tbey are interpreted as causative constructions witb tbe minimum number of two In OBGT VI tbe ventive is distinguisbed from tbe participants, and tbe suffixed personal pronoun is prefix mu- by tbe writing ma- (see Table 4, §§66taken as tbe subject. Wbetber tbe subordinate 71, 75 against §§44-49, 74), but elsewbere mu- can subject, to wbicb -b- presumably refers, is suffering be used also for tbe ventive, witbout discernible tbe action ("be/I/you caused it to be placed") or system. Gompare in particular IX §29 sa ma-anperforming it ("be/I/you made someone eat") is dug4 = ik-su-dam, "be reacbed bere," to VIII §19 difficult to decide witbout sentence context. kas4 mu-un-dug4 = il-su-ma-ain, "be ran bere." Tbere is insufficient evidence to decide Tbe prefix bi- occurs in four different paradigms wbetber and bow tbe OB grammarians would (OBGT VI, VIII, IX, X) witb five different verbs, bave analyzed tbe prefix bi- (for "true" Sumerian, but only witb two ratber restricted constructions. it bas been proposed tbat bi- is ba- followed by a Tbe first construction consists of simple transilocative). tive forms witb infix conjugation (bi-SUBJ-BASE-
12
PETER J. HUBER
4.3. The Pronomial Infix Chain Tbe verbal forms bave a pronomial tbree-slot structure (witb strict order): (indirect object+case) + (subordinate subject) + (direct object or subject)
Tbis tbree-slot infix cbain is placed after tbe directional prefix, before tbe base. Any or all of tbe slots can be empty. I prefer tbe term "infix" to "prefix," since tbese elements apparently cannot begin a verbal form—at least, tbey never do in tbe paradigms. Tbe situation can be illustrated by tbe following forms excerpted from Tables 3 and 4.
gar-ma-ab = suknam = "place it(b) bere(m)!" gar-ma-da-ab = suknassu = "place it(b) witb bim(da) bere(m)!" gar-ma-ni-ib = suskinassu = "make bim(ni) place it(b) bere(m)!" gar-mu-na-ni-ib = suskinassum = "make bim(ni) place it(b) for bim(na) bere(m)!" ma-si-ni-in-gar = usaskinassum = "be(n) made bim(ni) place it to bim(si) bere(m)" Witb multiple Sumerian pronouns tbe Akkadian pronouns refer to tbe first element present in tbe sequence, but somewbat unexpectedly, tbe direct
object never is referenced by an Akkadian pronoun. Tbe paradigms use tbe coarser Akkadian case structure:
Akkadian accusative: Sumerian comitative or subordinate subject Akkadian dative: Sumerian dative or terminative, rarely locative On tbe Sumerian side, tbe following pronouns are used in tbe paradigms. In tbe last slot, tbe subject is: 3d-person -n-, more rarely -b-, lst-person -0-, 2nd-person -e-; -n- and -e- often are elided. Tbe direct object for tbe most part is -b-, witb a few questionable instances of -n- (e.g., in VI §17). Tbe indirect object in tbe first slot always is referenced by an Akkadian pronoun. It is: 3d-person -n-, lst-person -mu-, 2nd-person -'we- > -e-. Among tbe cases, tbe Akkadian accusative corresponds to tbe Sumerian comitative -da-, tbe Akkadian dative to tbe Sumerian dative (3d-person -na-, lst-person -ma-, 2nd-person -ra-), or to tbe terminative -si-, rarely to tbe locative -ri- (only a few 2nd-person instances seem to occur). Tbe paradigms suggest tbat tbe dative forms may bave been interpreted as <pronoun>+/ra/, namely as -n-ra- > -n-na- > -na(VI §11), -ma-ra- > -ma- (VI §13), -*we-ra- > -ra-. §2:
gar-bi-ib ga-bi'-ib-gar §4: gar-ra-ni-ib ga-ni-ib-gar §19: gar-mu-ub ga-ri-ib-gar
= suskin = lusaskin = suskissu = lusaskissu = suskinanni = lusaskikka
Tbere are isolated instances of an ablative -ta(OBGT VI 27-28, IX §11) not mirrored by an Akkadian pronoun or case. 4.3.1. Subordinate Subject: Causatives. Tbe subordinate subject requires a more extensive discussion. Eor tbe purposes of tbe present discussion I prefer to denote tbe corresponding Sumerian case by tbe name "subordinative," and to distinguisb it from tbe "locative" case, since it seems tbat tbe paradigms keep tbem syntactically distinct. In particular tbey separate tbese cases by correlating tbe Sumerian locative -ri- witb tbe Akkadian dative (VIII §20-23 and IX §42), and tbe subordinative -ri- witb tbe accusative. Gompare tbe following basic forms witbout prefix and witbout indirect object, excerpted from Table 3:
= "make someone place it!" = "let me make someone place it! = "make him place it!" = "let me make him place it!" = "make me place it!" = "may I make you place it!"
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR
13
Tbis comparison suggests tbat subordinate subjects are marked as follows: -bi-ni-mu-ri-
3rd person indefinite 3rd person definite 1st person 2nd person
no Akkadian pronoun Akkadian -su Akkadian -anni Akkadian -ka
Tbese elements presumably are composed from a pronoun and a subordinative case marker HI. Tbe correspondence between -ni-, -mu- and -ri- on one band and 3d-, 1st-, and 2nd-person subordinate subjects on tbe otber band, is establisbed witbout doubt by tbe Akkadian pronouns. But in tbe absence of Akkadian pronouns, tbe interpretation of -bi- is less secure and is based on analogy alone. In view of tbe discussion of tbe prefix bi- above, tbe grammatical particle referring to tbe subordinate subject migbt alternatively, and perbaps even preferably, be identified witb tbe second b of -bi-ib- (tbat is: gar-bi-ib = "bave something placed"), tbe otber b tben being a prefix. Eitber cboice could bave been tbe one preferred by tbe OB grammarians, but botb bave unpleasant ramifications, once one goes beyond tbe simple forms of VI §2 and §38. Witb tbe first cboice (-bi- for tbe subordinate subject), one would bave to admit tbe existence of two bomopbonous elements witb different compositions and functions, namely a prefix -bi- and an infix -bi-. Tbe second cboice (witb -b- for tbe indefinite subordinate subject) does not seem to agree witb tbe treatment of causatives in tbe indicative sections of tbe paradigm. But perbaps tbe paradigms are sligbtly inconsistent, or perbaps I am misinterpreting tbe evidence. Witb some misgivings I am leaning toward tbe first cboice, and sball assume tbat tbe indefinite subordinate subject ordinarily is referred to by a (morpbological) infix -bi-, often written -ni-. I sbould empbasize once more tbat tbis discussion is not about "correct" Sumerian, but about its OB formalization. Tbe subordinate subject is tbe only feature in tbese paradigms wbere tbe n and b tbemes are distinguisbed by a formally different treatment, and tbe treatment of tbe split is bigbly interesting and illuminating. For tbis discussion, it does not
matter by wbicb b of -bi-ib- tbe subordinate subject is referred to. According to modern grammars, "true" Sumerian does not distinguisb gender, but (depending on tbe grammarian) splits between animate and inanimate, or between person and non-person. Akkadian cannot duplicate sucb a distinction. Tbe surrogate Akkadian split used in tbe paradigms, between -su and ino pronowi>, is between definite and indefinite. Tbis suggests tbat tbe split n — b in tbe paradigms sbould be taken as: a definite person — .something else, wbicb actually comes quite close to a person — non-person split (in a sentence context, a personal pronoun almost inevitably is definite, since it refers to a person mentioned beforeband). 4.3.2. Dissimilation bi > ni after (labial)+ (vowel). Guriously, -ni- sometimes is mirrored by an Akkadian -su, sometimes not. Tbe most striking example occurs in VI §8, wbere gar-ma-ni-ib corresponds to suskinam = "make someone place it bere!," wbile in §10 tbe identical form corresponds to suskinassu = "make bim place it bere!" (see Table 3). Wby? For tbe purposes of tbe following discussion I will temporarily distinguisb between a "definite -ni-" (mirrored by -sii) and an "indefinite -ni-" (not mirrored). A first clue is found wben one inspects tbe ventive transformation tbat transforms tbe non-ventive forms of VI §§1-6 into tbe ventive forms §§7-12 by inserting -mu- or -ma-. One notes tbat tbe "indefinite -ni-" of §8 gar-ma-ni-ib originates from §2 gar-bi-ib, tbat is from an indefinite -bi-, wbile tbe "definite -ni-" of tbe identical form in §10 bas its origin in §4 gar-ra-ni-ib, tbat is from a genuine -ni-. Tbe grid permits several more comparisons of forms to tbeir "basis," (i.e., to tbe bomologous non-ventive, non-separative paragrapb):
14 Paragraphs with ni, but without -SM:
PETER J. HUBER
basis:
VI §8 (= X 8) ventive ma-, muVI §2 (= X 2) (bi) VI §45 (= X 35) mu-prefix (*) VI §51 (= X 41) separative baC) VI §59 (= X 47) ventive+separative m-maVI §67 (= X 53) ventive ma(*) VI §39 (= X 29) mi- (< bi?) IX §12 separative ba(*) IX §16 ventive+separative m-maIX §3 (bi) IX §24 separative baIX §3 (bi) IX §30 ventive maIX §28 (bi) IX §39 ventive+separative m-ma-
§28 (bi) (X 29 mi-ni-ib-gub) ('): VI §38 bi-ib-gar (X 28 bi-ib-gub), VI §39IXmi-ni-in-gar IX §28 (bi)
Paragraphs with ni and -su:
basis:
VI §10 (=X 10) ventive maVIII §5 ventive maIX §8 ventive maVIII §25 ventive ma-
VI §4 (ni) VIII §4 (ni) IX §2 (ni) VIII §25 (ni)
It tbus turns out tbat tbe "indefinite -ni-" always bas -bi- in its ancestry, w"bile tbe "definite -ni-" goes back to a genuine -ni-. Hence tbe "indefinite -ni-" seems to be notbing else tban a morpbological -bi-. We note tbat all instances of tbe "indefinite -ni-" are preceded by (labial) + (vowel). Tbe labial can bave diverse morpbological origins (prefix mu-, ventive -m-, separative -ba-). Sucb a distribution could be caused by dissimilation bi > ni after (labial) + (vowel). Tbis was proposed already by Falkenstein (1949), but rejected by Edzard (2003: 102), wbo cites, for example, mu-bi "tbis year." Edzard's argument bolds for bi, but not necessarily for bi. A searcb in ETGSL gave only a single, ratber irrelevant bit for (labial) + (vowel) + bi: i^'-'ur-za-ba-bi-tum (a type of instrument). Tbe assumption of sucb a dissimilation may irritate. Perbaps tbere is an alternative nonpbonetical, purely syntactical explanation, say tbat bi cannot be preceded by a prefix and tberefore is replaced by ni? Tbis is refuted by cases wbere tbe ventive transformation inserts -am- instead of -ma-, and wbere bi is assimilated to tbe immediately preceding ventive m. Tbus it turns IX §3:
sa dug4-ga-bi-ib = suksid = "make someone reacb!" into IX §7: sa dug4-ga-am-mi-ib = sukSidam = "make someone reacb bere!" Note tbat otber examples sbow tbat n is not assimilated to m, for example in VII §86: am-ne-du = illakam sunusi = "be comes to tbem," Tbe didactic insert: VI §37-39 (= X 27-29) is curious. It consists of tbree paragrapbs witb somewbat unexpected Sumerian renderings of iskun and usaskin (for discussion by tbe teacber?). Formal analogy witb tbe mu-prefix column migbt make one expect tbe form *in-ni-in-gar in tbis space. But if tbe proposed dissimilation model bolds, tbis Sumerian form would correspond to usaskissu, not to usaskin required by tbe grid. Tbus, my second guess was tbat tbe text sbould offer *i-bi-in-gar, witb infix conjugation. But tbe insert offers neitber of tbese forms:
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR
15
VI §37 bi-in-gar bi-/gar\ bi-gar
iskun (askun) (taskun)
bi-n-B bi-B bi-e-B
be(n) placed it(0) 1(0) placed it(0) you(e) placed it(0)
VI §38 bi-ib-gar bi-ib-gar-re-en bi-ib-gar-re-en
usaskin (usaskin) (tusaskin)
bi-b-B bi-b-B-en bi-b-B-e-en
be(0) caused it(b) to be placed I(en) caused it(b) to be placed you (en) caused it(b) to be placed
VI §39 mi-ni-in-gar mi-ni-in-gar-re-en mi-ni-in-gar-re-en
usaskin (usaskin) (tusaskin)
bi-bi-n-B bi-bi-n-B-en bi-bi-n-B-e-en
be(0) bad bim(n) put by someone(bi) (?) I (en) bad bim(n) put by someone (bi) (?) you(en) bad bim(n) put by someone(bi) (?)
Tbe parallel OBGT X bas an analogous insert but offers only tbe first line of eacb paragrapb: 27: bi-in-gub = usziz (error for izziz) 28: bi-ib-gub = usziz 29: mi-ni-ib-gub = usziz (note tbe b in X against tbe n in VI)
Here, VI §37 and X 27 appear to be bi-prefix alternatives to tbe immediately preceding noncausative l-prefix versions VI §36 and X 26, witb tbe same Akkadian rendering. Tbe forms of VI §38 and X 28 bave been discussed in section 4.2.3.; tbey may offer tbe irregular standard version of a causative construction witb tbe minimum number of participants, replacing a regularly formed *i-bi-in-gar. Tbe mi-ni- of VI §39 and X 29 is unique in tbese paradigms. I suspect tbat it bides a dissimilated prefix + infix combination bi-bi- > mi-ni-. I do not tbink tbat mi-ni- sbould be interpreted as a pbonetical variant of mu-ni- (wbicb occurs in §45 in tbe murprefix section), even tbougb §39 and §45 are synonymous from tbe Akkadian point of view. Listing a pbonetical variant of mu- in tbe i-prefix section would amount to an error (in botb VI and X), namely to an unusual morpbological misclassification. Moreover, VI §39 uses suffix conjugation, wbicb is compatible witb tbe bi-prefix, but does not seem to be used witb tbe mu-prefix. Admittedly, I cannot claim to understand tbe construction of VI §39. A construction tbat combines bi-prefix, suffix conjugation and bi-infix looks like a second order causative to me. Sucb a construc-
tion migbt reasonably be expected witb intransitive verbs sucb as "to rise," wbere tbe first order causative is "to make rise" = "to raise," and tbe second order is "to make raise," but not witb an originally transitive verb like "to put." 5. OBGT VIII+IX: Tvvo-Part Verbs Tbese paradigms cover tbe two-part verbs: kas4 ... dug4 = lasamum = "to run" (OBGT VIII) and sa... dug4 = kasadum = "to reacb" (OBGT IX). Tbey are less systematic tban tbe otbers. Like OBGT VI+X, tbey illustrate tbe formation of causatives. In fact, OBGT IX conspicuously groups tbe paragrapbs into pairs, eacb pair consisting of a non-causative and tbe corresponding causative paragrapb, but only tbe Ni-section arranges tbose pairs in a systematic fasbion. A remarkable feature of tbe paradigms OBGT VIII + IX is tbat tbey exercise tbe differences between transitive and intransitive constructions. Ordinarily, botb verbs are constructed transitively, witb -b- referencing tbe direct object kas4 or sa. Tbe forms are transformed into causative ones by inserting bi or ni:
16 VIII §1: VIII §3: IX §4: IX §5:
PETER J. HUBER kas4 dug4-ga-ab kas4 dug4-ga-bi-ib sa dug4-ga-na-ab sa dug4-ga-na-ni-ib
= lusum = sulsim = kusuzzum = suksizzum
• "run!"
: "make someone run!" • "reacb for bim!"
• "make bimj reacb for
However, witb ventive + terminative case, tbey are preferentially constructed intransitively, witbout reference to a direct object: = lusmassum VIII §7: kas4 dug4-ga-am-se VIII § 11: kas4 dug4-ga- [mu] - se = lusmam ana seriya IX §21: sa dug4-ga-mu-se = kusdam ana seriya
• "run toward bim bere!'
• "run toward me bere!" : "reacb for me bere!"
If -b is added to tbese forms, tbey are made causative, but apparently -b bere does not reference kas4 or sa, but an indefinite subordinate subject: VIII §9: kas4 dug4-ga-am-ma-si-ib VIII §12: kas4 dug4-ga-am-mu-se[-ebl IX §22: sa dug4-ga-mu-se-eb
= sulsimassum = "make someone run to bim bere!" = sulsimam ana seriya = "make someone run to me bere!" = suksidam
= "make someone reacb for me bere!"
Tbe double-m in VIII §9 and §12 seems to be merely grapbical and bas no morpbemic significance, as a comparison witb tbe non-causative forms and witb IX §22 sbows. Tbese constructions raise tbe question wbetber tbe causativity marker -b- bere is meant as a sbortened -bi-, or ratber as a kind of ergative counterpart of a transitive direct object -b-. It does not seem tbat tbis question can be decided on tbe basis of tbe paradigms. To complicate matters, tbe OB autbor of OBGT Vlll and IX seems to play deliberately witb multiple Sumerian renderings of Akkadian structures. In several of tbe paragrapbs quoted above, be switcbes tbe Sumerian construction in mid-
paragrapb. Inside VIII §11, be keeps tbe terminative case but switcbes from an intransitive hamtu-hased imperative to transitive marii-based volitives and precatives (kas4 ga-mu-e-si-ib-be = "let me run toward you bere!"). In tbe seemingly parallel IX §21, be switcbes instead from an intransitive terminative to a transitive dative construction (sa ga-mu-ra-ab-dug4 = "let me reacb for you bere!"). In VIII §12 and IX §22, tbe most likely interpretation is tbat be keeps tbe intransitive construction, but switcbes from tbe terminative to tbe locative case (kas4 ga-am-ma-r[i-ib]-dug4, sa gamu-ri-ib-dug4 = "let me make someone run/reacb to you bere!").
References AHw = von Soden, W, 1965- Akkadisches Handworterbuch. Harrassowitz, 81 Wiesbaden, Attinger, E 1993 Elements de Hnguistique surnerienne. Gottingen: Vandenhoek & Ruprecht. Bellugi, U., and Brown, R. (eds.) 1971 The Acquisition of Language. Chicago: University of Chicago Press.
Black, J. A. 1991 Sumerian Grammar in Babylonian Theory. 2nd edition. Rome: Pontificium Institutum Biblicum. Edzard, D. O. 2003 Sumerian Grammar. HdO 1.71. Leiden: Brill. Falkenstein, A. 1949 Grammatik der Sprache Gudeas von Lagas. AnOr 28-29. Rome: Pontificium Institutum Biblicum.
ON THE OLD BABYLONIAN UNDERSTANDING OF GRAMMAR
GAG = von Soden, W. 1952 Grundriss der Akkadischen Gramniatik, AnOr 33. Rome: Pontificium Institutum Biblicum. Jacobsen, Th. 1960 'Itallak niatir JNES 19: 101-16. (Reprinted in Toward the Image of Tammuz, ed. W. L. Moran [Cambridge, MA: Harvard University Press, 1970].) Michalowski, P. 1980 Sumerian as an Ergative Language. fCS 32: 86-103. 2004 Sumerian. Pp. 19-59 in The Gamhridge Encyclopedia of the World's Ancient languages.
17
ed. R. D. Woodard. Cambridge: Cambridge University Press. MSLIV = Landsberger, B.; Hallock, R; Jacobsen, Th.; and Falkenstein, A. 1956 Materialien zum Sumerischen Lexikon IV Rome: Pontificium Institutum Biblicum. Thomsen, M.-L. 1984 The Sumerian Ijinguage. Copenhagen Akademisk Forlag. Zolyomi, C. 2000 Structural Interferences from Akkadian in Old Babylonian Sumerian. Acta Sumerologica 22: 335-60.
[-ATR] HARMONY AND THE VOWEL INVENTORY OF SUMERIAN Eric }. M. Smith (University of Toronto)
It has long been accepted that Sumeriim exliibited some form of vowe! harmony, and early description.s of Siimerian vowel harmony (Poebel 1931; Kramer 1936) framed that harmony data in the context of a six-vowel inventory. While the arguments in favor of vowel harmony are now accepted, the expanded vowel inventory has largely been dismissed for the past seventy years. Instead, modern descriptions of the Sumerian vowel inventory are restricted to the four vowels that are visible through the filter of Akkadian. After decades of neglect, Keetman (2005) reopened the discussion of the relationship between Sumerian vowel harmon\- and tho languages vowel inventory. Although he is definitely on the rigbt track, the solution he suggests, which requires two forms of vowel barmony to be operating simultaneously, is overly complicated. This paper shows how the choice of appropriate phonological features can simplify Keetman's analysis, and explain the data as the result of a single vowelharmony process, The goal of the paper is to reexamine the vowel harmony described by Poebel and Kramer, and situate it in n modern phonological framework. Contrastive and typological factors indicate that tbe observed vowel harmony behavior is best explained by a seven-vowel inventory. By examining Old Babylonian lexical texts and tbe Sumerian lexicon, we hnd additional support for a sevenvowel inventory. By drawing together the data from vowel harmony, from lexical texts, and from (loss-linguistic imiversals, I conclude that Sumerian actually had a seven-vowel inventory.
1 begin in section 1 by discussing exi.sting arguments for an expanded vowel inventory tbat draw upon orthographic data. In section 2, I present the vowel harmony patterns that led Poebel and Kramer to propose their six-vowel inventory. In section 3, the core of the paper, I reframe the earlier analysis of vowel harmony in modern terms, and shows how a seven-vowel inventory is the most plausible one. Section 4 ties together the lexical and phonological evidence to provide additional support for a seven-vowel inventory. 1. Orthographic Evidence Much of our knowledge of how Sumerian was actually pronounced comes from the lexical texts of the Old Babylonian period. Since tbe Old Babylonian scribes were no longer native speakers of Sumerian, they took extra care in representing the full phonology of the words being written. These lexical lists are invaluable to modern scholars because they spell out syllabic writings for Sumerian words wbose pronunciation would otherwise be completely opaque. Although lexical lists are vital to our understanding of Sumerian phonology, most of them come with two significant limitation.s. First, most of the exemplars available to us appear to bave been exercises used as part of the .scribal training process (Civil 1979), whicb means that they are not without errors. More fundamentally, the problem is tbat tbese lists represent tbe pronunciation of Sumerian as filtered through the Akkadian phonological system. 19
JCS 59 (2(107i
20
ERIC J.M. SMITH
The limitations of the Akkadian inventory are particularly problematic for our reconstruction of the Sumeiian vowel system. Since Akkadian has a system with four vowel-qualities (/a/, /e/, /i/, and /u/), any Sumerian vowels that cannot be represented within that system are hidden to us. Scholars of Sumerian have tended to he extremely conservative when it comes to admitting new phonemes into the inventory (e.g.. Black 1990). Hence., the most recent accounts of the Sumerian vowel system, such as Michalowski (2004) and Edzard (2003) have restricted themselves to the /a/. k'L /i/, /u/ system, which can ,securely be supported by Akkadian, Some scholars, notably Edzard f2003) have argued that Sumerian, like Akkadian, had a distinction in vowel length, but this view does not seem to have garnered much support. In any event, the existence of vowel length is orthogonal to the questions of vowel quality that are the focus of this paper.
by repeating the vowel. So for instance, the in line 17 told the scribe that the sign is to be read /be/ and not /bi/, Consider lines 10 and 11 of the fragment transcribed in (1), which clearly indicate that and represent two distinct pronunciations of the grapheme. (1) A Fragment of the Lexical List Proto-Ea (Landsberger 1951,1955; Civil 1979) Line
Pronunciation
10
ku-U3
KU
11
ku-U2
KU
12
SU2-US
KU
13
tu-us
KU
15
su-uh2 si-i
KU
16 17
be2-e
18
bi-id
KU KU
19
da-ab
KU
20
du-ur
KU
/./. Evidence from Lexical Texts
21
du-ru
KU
It has long been accepted that certain signs that shared a single phonetic value in Akkadian had multiple values in Sumerian, with the best example being the signs and and conceal two different consonants in Sumerian, it is possible that some of the apparently homophonous vowel graphemes miglit also conceal two different vowels. The strongest proponent for this theory was Lieberman (1977, 1979), who proposed that some of the graphemes read as /u/ in Akkadian properly had a reading of /o/in Sumerian. The evidence for Lieberman's theory comes largely from copies of the Old Babylonian lexical list Ea. In particular, Lieberman focused on eariy versions of the lexical list FM from the city of Nippur. These early recensions consist of only two columns, with a grapheme in the right column and the corresponding pronunciation(s) in the left column. The practice in this lexical list was to make the pronunciation of a CV grapheme clearer
22
nu-U2
KU
23 24
bu-U2
KU KU
tu-kii-iil
Sign
KU
There are several similar examples in Profo-Ea, although not all of them are as clear-cut as the case of the grapheme. Throughout the lexical list there does appear to be a distinction being made between pronunciations that are indicated with the grapheme and pronunciations that are indicated with the , , or graphemes.' Lieberman claimed that is being used to indicate /u/, while the other three graphemes are being u,sed to indicate a pronunciation of /o/. These are summarized in Appendix A. Supporting this position, IJeberman claimed that in Nippur Akkadian of this period it was also the practice to make a distinction hetween
1. Keetman (personal communicrttion) notes that Proto-Ea also contains the sign , Keetman (200.3) devotes considerable discussion to the possible phonetic value of . but the question remains outside the scope nf this paper,
[-ATR] HARMONY AND THE VOWEL INVENTORY OF SUMERIAN on the one hand and or on the other hand. Poebe! (1939) had noted that when writing Akkadian, Nippur scribes of this period would use for /i+u/ and /u + u/ contractions, and would use or for /a+u/ contractions. Note that in all cases, these contractions are phonemi(all> a long /u:/, and not a diphthong; evidently though, there was an ailophonio variation depending on the underlying origin of the /u:/, and this \ariation manifested itself in written Akkadian. The same scribal tradition that used and to distinguish an Akkadian allophone of /u/ used the same graphemes to render the Sumerian phoneme lol (Lieberman 1977). Somewhat more controversial is the attempt by Bohrova and Militarev (1989) to build upon Liebermans work and extend the Simierian vowel system even further, Bobrova anfi Militarev made an analysis of spelling variations in the pronunciation guides provided in other lexical lists. So for example, in one lexical list the <MIN> grapheme is indicated with a pronunciation of <mi-in> while in another it is indicated as <ma-an>, and they argued that this a:i variation is evidence for the existence of a front /a/ phoneme. In similar fashion, they used u:i, i:e, and a;u variations to argue for the vowel system shown in (2). (2) Extended Vowel Inventory for Sumerian (Bobrova and Militarev 1989) i, 0
u 3 or i e
0 a
a
Undercutting their analysis, however, is the fact that the spelling alternations in question were largely drawn from later texts that post-date the presumed extinction of spoken Sumerian. Moreover, their methodology depends on comparing prontmciation guides from different lexical texts, which could be subject to variation for reasons other than the attempt to render distinctions in vowel quality. While such an inventory cannot be ruled out, there does not appear to he sufficient evidence to support it Lieberman is on some-
21
what more secure grounds, since the variations he observed are all within a single lexical text. 2. Vowel Harmony Data When considering data from lexical lists, it mu.st be remembered that, with the exception of a handful of early lists from Ebla, such lists date from the Old Babylonian period at the earliest. Since they were written by and for Akkadianspeaking .scribes, they might well have heen composed in a period when there were no longer any native speakers of Sumerian. Fortunately, there is other orthographic evidence that clearly dates from Sumerian texts written by Sumerianspeaking scribes. Specifically, there appear to be a fair number of orthographic patterns that can best be explained as a consequence of vowel harmony. Within two-syllable stems, there appears to be a tendency towards having the same vowel in both syllables (Poebel 1923; Michalowski 2004). A typical example is the Sumerian word for "bronze," zahar. This appears to he a Kulturwort common to a number of languages in the region. In the neighboring Akkadian, the word is siparru, while in Eiamite (southwestern Iran) the word is zuhar. This suggests that vowel harmony operated on this word after it had been adopted into Sumerian, turning the high vowel of the first syllable into a low vowel. To date, no tomptehensive study has been done on stem-internal vowel harmony in Sumerian, although a brief outline of some of the data will be provided in §4.2. Most of the attention, rather, has been devoted to studying vowel harmony in the Sumerian "verbal chain." Sumerian is an agglutinative language, and verbal stems are generally accompanied by a number of prefixes and suffixes. For many decades scholars have noted that these affixes in the verbal chaiTi tend to exhibit vowel changes depending Oil the stem to which they are attached. The general structure of the Sumerian verb complex is shown in (3). The exact forms, functions, and ordering of these affixes has been the subject of much-heated debate among Sumerologists. However, since we are interested here
22
ERIC J. M. SMITH
only in the phonological combinations that these affixes set up, all that is important here is the general rule that within each "slot," the affixes are
mutually exclusive. A number of the affixes have allomorphs, which are indicated in (3) separated by slashes.
(3) Order of Elements in Sumerian Verbal Chain (Thomsen 1984; Michalowski 2004) 1. Mood 2. Conjunction 3. Conjugation . i f — I
4. 5. 6. 7. 8. 9. 10.
Indirect object Dimensional prefixes Agreement prefixes Verbal stem -ed suffix Agreement suffixes Nominalization suffix
I he-, nu-, ha-/he2-/hu-, bara-, u-, na-, ga-, sa-} | inga-} j mu-/ma-, ba-/bi9-/be9-, i-!-/e-, al-} 1
^
i
-
^
^
i
j
'
'
{ a-, ra-, na-, me-, ne-} { da-, si-/se3-, ta-, ra-, ni-/ne-, i-} | e-, n-, b-) {-ed ) | -en, -enden, -enzen, -ene, -es } | -a |
Already in the first comprehensive study of Sumerian (Poebel 1923), a wide range of both regressive and progressive vowel harmony was identified in the verbal chain. Subsequent studies have found further instances of vowel harmony. Upon closer examination, it appears that the term "vowel harmony" is being applied rather broadly here, and that there are two somewhat different phenomena involved. In the first type of vowel harmony, the vowel of the affix assimilates completely to the trigger vowel. So, for instance, the conjugation prefix muappears with the 2nd-person singular dative prefix ra- as <ma-ra>; with the prefix 7U- we get <mi-ni> (Poehel 1923). This t\pe of harmony also operates progressively, as is the case of the verbal suffix -ed, which changes to -ud when following a verbal stem containing a lul vowel, producing a form like rather than \tum-ed-a> (Poebel 1923; Thomsen 1984). While this phenomenon is interesting, it does not provide much insight towards reconstructing the vowel inventory. 2.1. ilannouy in the Conjugation Prefixes -le- and hi-lbe- (Poehel 1931 and Kramer 1936) The dearest type of vowel harmony is illustrated in Old Sumerian texts from southern Mesopotamia, particularly from the city state of Lagas. Early studies by Poebel (1931) and
Kramer (1936) showed that the conjugation prefixes I- and hi- become e- and be% respectively, whenever the verbal stem to which they are attached contains an /a/vowel. The conjugation prefix i- is ordinarily written as .^ This in itself is rather curious because there is a very common grapheme, but this particular morpheme is consistently written as instead. It has been argued that the underlying form of the morpheme may actually be III (Thomsen 1984), but this is not widel\- accepted. Whatever the case, this l + nasal] feature does not appear to have any effect on the prefixs behavior with respect to vowel harmony. Before stems with a lul, /i/, or lei vowel, the prefix is written as . So we observe forms like , , , , and However, before stems with an /a/ vowel we observe the prefix being written <e>, as in <e-ak>, <e-ba>, <e-bal>, and <e-gal2>. There are isolated exceptions (e.g., appears once instead of the usual <e-gaz>), but these form a tiny fraction of the corpus (Poebel 1931). There is a similar patterning for the conjugation prefix bi-. Before stems with lul and HI vowels, 2. An aniMiyinoLis reviewer has pointed out that the prefix is oftfn written in the [.sin-Lai'.sa and later periods. However, this is perfectly understandable, sincv it would reflect scribal practice after the period in whith harmony was an active phenomenon in Sumerian.
[-ATR1 HARMONY AND THE VOWEL INVENTORY OE SUMERIAN the prefix is written as ; before stems with /a/ vowels it is written as and in contrast with (Kramer 1936). The cooccnrrence pattern for conjngation prefixes i- and hi- in the Early Dynastic texts
23
from southern Mesopotamia is summarized in (4). A more detailed and complete presentation of this data is included as Appendix B. The data is also discussed in somewhat more detail by Keetman (2005).
(4) Co-occurrence of Conjugation Prefixes with Verb Stems Stem vowel
With or
With <e> or
/a/ lei
lul
ag2, ak, ha, halia), har, dah^, gaz, ga/2, gar, had2, hal, Ia2, nag, sag, sar, tag, tag^ deQ, gen, me, §e
I, telti
ru, sur, u/Q, ur,
du^, dufi, duii, duh,guh,gul, hug, ku^ir].
The stems with /a/ and HI are the most clear-cut. All the /a/ stems are written with the <e> or versions of the prefixes, while all the HI stems are written with the or versions. Poebel argues that this is evidence of vowel harmony: the "open" vowel /a/ in the stem causes the appearance of the "open" /e/ in the prefix. Before stems containing a "closed" vowel HI. the underlying HI prefix is found. In most cases, stems containing the vowels lei and lul are found with the HI forms of the prefix, but there are a number of verbs that unexpectedly take the lei prefix. For instance, Poebel notes that before the verbal stems and <me-a> the /- prefix is con.sistently written as <e>, even though the stem vowel would lead us to expect a writing of . The solution he presented was to have two /e/-type vowels in the inventory. In addition to the ordinary "closed" /e/'* there is an "open" lei, which is the vowel that is found in the stems and <me-a>. ln addition to noting stems written with an leltype vowel that take the wrong prefix, there are
3. Although Poebel distinguishes these vovv-els using diacritics nornmlly associated with vowel length, it is clear from his (Hscussion that this is merely a typographic convention, and that the distinction between /f/ and /f/ is definitely not one of vowel length.
also stems written with a AiAvowel that break the pattern of prefixes. Unexpectedly we see <e-ur4> rather than ' and <e-sur> rather than * (Poebel 1931). The same can be seen with the hi- prefix, where we see rather than (Kramer 1936). The solution proposed by Poebel was that in addition to the "closed" lul, there must an "open" equivalent lol. The vowel inventory proposed by Poebel is shown in (5). With the addition of these two vowels to the four attested from Akkadian, there is a simple phonological explanation of why the conjttgation prefixes are written with the and graphemes before some verbs, and with the <e> and prefixes before other verbs, namely that the choice of "open" or "closed" vowel in the prefix is determined by the vowel used in the stem. (5) Vowel Inventory for Sumerian (Poebel 1931) i
u
"closed"
e 6
6
"open"
2.2. Behavior before Other Prefixes In addition to the conjugation prefix, other prefixes can also be involved in the vowel harmony
24
ERIC J. M, SMITH
process. For instance, the terminative-case prefix ,§/- can be written either <si> or <se3>. Poebel noted that the choice of forms is conditioned by the vowel in the verbal stem, with si- being written <se3> before an "open" vowel. If the verb happens to start with the /- conjngation ptefix, then the prefix will also harmonize with the stem vowel, so we observe forms like and , in contrast to <e-se3-gar> and <e-se3-deg>, Poebel observed the same pattern for the prefixes mi- and ni-.^ We find forms like , , and , contrasting with forms like <e-me-ed>, <e-ne-gar> and <e-ne-la2>. Once again, Poebels division of the /e/-vowels into lei and /g/ provides a tidy explanation for the otherwise inexplicable difference in morphology between /imie/and <e-me-sed> /emesed/. The actual vowel that is governing the form of the prefix need not be part of the verbal stem. If there is a prefix containing /a/ intervening between the prefix and the stem, it is the /a/ that takes effect, regardless of the vowels in the stem. Hence with the prefixes tui-, da~, ta-, and tna- we observe forms written with an initial <e> regardless of the stem vowel, such as <e-iia-de2>, <e-dasii2>, <e-ta-zi>, and <e-ma-diir> (Poebel 1931). The third-person plural dative prefix ne- also triggers the <e> form of the prefix regardless of the stem vowels. This suggests that the prefixs underlying form must be /ne/ with an "open" vowt'l. [f the tmderlying form had been Inl, we would have expected to see *, but we see <e-ne-si3> instead, 3. Phonological Analysis We are now in a position to consider which phonological features would explain the phenomena descrihed here. We shall work from the assumption that harmony effects are a consequence of
4. Care must be taken in identif\'ing these forms. The prefixes mh and ;ii- have altoniorphs me- and ne-, which can potentially be confused with the dative-case prefixes me- and ne-. Fortunately, it is almost always rlear from the context which prefix was intended.
contrasts within the vowel inventory. Hence, it is necessary to divide up the vowel inventory using a contrastive featute hierarchy Following Dresher (2003), we will build such a hierai'chy using the Successive Division Algorithm, as described in §3.3. The primary contrast in the vowel systetn seems to he between the vowels that Poebel referred to as "open" (/a/, /e/, and 16/) and the vowels that he referred to as "closed" (/i/, lei, and lul). We could adopt Poebels terminology and refer to the conttastive feature as being |±open| or |±lowj. The choice of [+low| as the active feature in vowel harmony seems to have been the starting point for the analysis proposed by Keetman (2005). However, Keetman recognizes that the feature |±lowl is not sufficient to account for all the vowel harmon\' patterning, particularly hannony within roots. Keetman propo.ses that a second type of harmony, involving the feature |±roundl, is simultaneously active in Sumerian. Under his analysis, the first type of harmony involves changes to the first formant, while the .second rounding harmony involves the second formant. However, as we will show, the type of vowelhannony behavior found in Sumerian strongly resembles examples from other languages, where an apparent height hannony effect can best be described as an instance of tongue-root hannony. Such hannony does not necessarily involve the first and second formants, but nonetheless can account for all the vowel-harmony behavior, both within the verbal prefix chain and within roots. If we are describing Poebels closed lei and open lei in terms of tongue-root, we would say that lei is I+ATR] (advanced tongue root) while lei is [-ATR]. To use International Phonetic Alphabet (IPA) notation, Poebel s lol would correspond to Izl while his lei would correspond to IPA lei. Since the vowels in the stems and <me-a> both triggered the e- prefix, the vowels must be the |-ATR| variant, so those stems would actually be /gen/ and AiiEa/. Poebel gave no phonetic details of what lol might be, but following our assumption that lei is I-ATR], Poebels // should correspond to a [-ATRj vowel such as hi or possibly lul. Under this inter-
I-ATRI HARMONY AND THE VOWEL INVENTORY OF SUMERIAN pretation, the verbs written , <sur>, and would actually be hrl, Iscni, and /ro/, respectively. Using this |ATR| feature, the vowel hannony system arranges itself very neatly. The |-ATR1 vowels (/a/, /e/, and h/) coincide with the |-ATRl prefix forms <e> |E1 and [bEJ, The [+ATR1 vowels {HI, lei, and lul) coincide with the |+ATR] prefix forms [i| and [bi]. Restating Poebel's inventory from (5) using this terminology, we get something like (6). (6) Distribution of [ATR| Feature in Sumerian i
u 8
0
[+ATRI
,
-ATR|
a Since we lack any access to the articulatory and acotistic details of Sumerian phonetics, we can only use the term 1ATR| in the loosest of ways. That is, we have no way of knowing whether the Sumerian |ATR[ feature was actually articulated using a tongue root or by adjusting the pharyngeal cavity in some other way. Casali (2003) notes that from an acoustic standpoint, [+ATR1 and |-ATR1 vowels cannot always be distinguished hy their height (i.e., their first and second formant values). Descriptions of African languages oltcn characterize | + ATR| vowels as "hollow," "deep," or "hreathy," while |-ATR] vowels are "tight." "choked," "muffled," "bright," or "creaky." Whatever the articulatory or acoustic details, 1ATR| is a convenient name for this feature, and is consistent with terminology used in describing other languages. A natural question would be whether it is HI or lei, which is the underlying form of the conjugation prefix. That is, do we have an |-ATR! feature that is spreading from stem vowels /a/, /E/, and hi and converting HI into {E\? Or is the active feature |+ATR| spreading from stems with lei, HI., and lul to convert an underlying IEI into ji|? The answer is given to us by the form of the third-person plural dative prefix ne-, which co-occurs with the <e> form of the conjugation prefix, so its tinder-
25
lying form must be luzl. The vowel remains unchanged even in front of a stem with a [+ATR] vowel, as seen in <e-ne-si3> [enEsi[ and <e-ne-gi4> [EnEgi]. If the [+ATR[ feature were spreading and converting /ne/ to /ne/, then we would expect to .see ' [inesi] rather than <e-ne-si3>. Since this indicates that the /E/ vowel is not subject to vowel harmony, then the conclusion is that the mutable underlying form must be HI, and the spreading feature is [-ATR[. In terms of targets and triggers, the only target we have identified is HI. This phoneme is the target both for the prefixes /- and hi-, but also for the prefixes i-, mi-, and ni-. In all cases, HI becomes [E] under the influence of a following [-ATR] vowel. As to whether the other two [+ATR[ vowels, lul and lei, are subject to vowel harmony, it is difficult to be certain. These vowels do appear in prefixes like mu- and he-, but these prefixes do not appear to be affected by the presence of a [-ATRj stem vowel. In any case, even if these prefixes did alternate, it might be difiScult to tell; /mu/ would become [m:i[ but would still be written with the <mu> grapheme, while /he/ would become [hE], which would also still be written as or Like ne-, the prefixes containing /a/ ina-, da-, ta-, and ma-) all retain their identity even when they precede stems containing an [+ATR] vowel. We have no prefixes containing a reconstructed hi vowel, so it is impossible to say what might happen to it. However, in the absence of any evidence to the contrary, it seems safest to assume that hi is also immune to the effects of vowel harmony. Again, this is consistent \A'ith the analysis that the active spreading feature is [-ATR[ and not [+ATR]. Many languages display j-ATR] harmony, including Khalka (Mongolian), and various Tungusic languages (van der Hulst and van de Weljer 1995). Chukchee (Chukotko-Kamc hatkan), a language of eastern Siberia, also displays [-ATR] harmony, which is noteworthy since Krecher (1987) noted a range of structural traits shared by Chukchee and Sumerian. Of pariicular relevance to our study
26
ERIC J. M. SMITH
of vowel inventories are Italian (Calabrese 1995) and various Niger-Congo and Nilo-Saharan languages (Casali 2003) which both display [-ATR] harmony similar to that of Sumerian.
3./. A Typology of Vowel Inventories (Calahre.se 1995) The vowel inventory Poebel came up with consists of three vowels with the [-ATR[ feature (/a/. Izl, and hi), and three vowels with the [+ATR[ feature {HI, lei, and luf). Calabrese (1995) proposes a typology of vowel inventories that is based on a
set of "marking statements," which are activated or deactivated in a given language. The deactivation of marking statements controls the complexity of phonemes that a given language will toletate, and there are some universal restrictions on which sets of marking statements can be deactivated. The tree of marking statements is shown in (7). If a language deactivates a marking statement on a given branch, it must also deactivate the marking statements higher on this same branch. Under his analysis, a language that has not deactivated any marking statements at all would have the simplest possible inventory: /a/, HI, lul.
(7) Marking Statements for Vowel Systems (Calabrese 1995)
,-back] -back, + round[ / | [+back,-tound] / [
, -l(>w[
,-low[
, +n)und[
The inventory we have reconstnicted for Sumerian most closely resembles an inventory that has deactivated the marking statements [-low; -high] and [-high, +ATR] on the leftmost branch of the tree. This is the sort of inventory found in Standard Italian, and is shown in (8). It should be noted that the inventory in (8) includes an lol, which Poebel did not reconstruct for Sumerian. This could represent a naturally occurring gap in the invetitory, or it could just as easily represent yet another vowel that cannot be distinguished from lul due to the limitations of the writing system. In fact, given the evidence in §3.2 and §41, it seems likely that Sumerian did have an lol vowel distinct from both hi and lul.
(8) Inventory with [-low,-high] and [-high, +ATR[ Deactivated (Calabrese 1995) i
e
high
-
)
low
-
back
-
-
roitnd
-
-
ATR
+
+
E -
-
-
a -
0 -
_
-
+ -
-
-I-
-
_
-
o
-
u +
_
_
-I-
+
-H
+
+
-(-
+
+
[-ATR] HARMONY AND THE VOWEL INVENTORY OF SUMERIAN 3.2. \ATR\ Harmony and Vowel Inventories (Casali 2003) Fortunately for our analysis of the interaction between vowel harmony and vowel inventories, Casali (2003) has done an extensive cross-linguistic survey of [ATR| harmony in languages of the Niger-Congo and Nilo-Saharan families. According to Casali (2003), there is an ongoing dehate in the field between those who feel that only [+ATR| can be the active feature in vowel harmony (what Casali calls "Universal [+ATR[ Dominance" theories) and those who feel that either [+ATR] or [-ATR] can be the active feature (the "Variable [ATR] Dominance" theories). Casa!i is very strongly in the Variable [ATR] Dominance camp, and the main goal of his paper is to show that not only is [-ATRj dominance a possibility, it is in fact very common in languages with certain types of vowel inventories. As it turns out, the presence of [-ATR] spreading is correlated with certain types of vowel inventories. Casali divides inventories into three classes, as shown in (9). By our analysis, Sumerian should fall into the 4Ht(M) class. In Casalis ciassification, "4Ht" means that a vowel inventory has four different heights, while (Mj indicates that the [ATR] feature is contrastive only for mid vowels. (9) Vowel Inventories with Contrastive [ATR] (Casali 2003) 4Ht(M)
5Ht i
tl
I
u
e
(a)
0
u
i e
(3)
e
8
a
4Ht(H) i
u
I
u (a)
0
27
inventory proposed by Poebel (1931), lacking the lol vowel, is not particularly common. A search of the UCLA Phonetic Segment Inventory Database (Maddieson 1984) suggests that Poebels inventory is actually quite rare. Of the 317 languages listed in UPSID, there are 37 that have a contrast in the mid vowels and thus fall into Casali's 4Ht(M) category. The tendency towards .symmetry is strong, because of those 37 languages; only six have asymmetrical vowel inventories. Furthermore, of those six, four are missing one of the mid-front vowels, and two'' are missing one of the mid-back vowels. Of the two UPSID languages which are missing one of their mid-back vowels, in both cases the missing vowel appears to be hi rather than lol.^ So while it is certainly possible that Sumerian has the six-vowel inventory proposed by Poebel, a full seven-vowel inventory seems more likely. Casali's survey of over one hundred NigerCongo and Nilo-Saharan languages indicates that languages with a 4Ht(M) inventory tend to have [-ATR] as the active feature in vowel harmony. In this regard, Sumerian agrees with Casali's observation. Ca.sali classifies the environments in which [ATR] harmony can manifest itself into six different categories, as shown in the columns of (10). Languages with a 4Ht(M) inventory tend to display the weaker forms of assimilation, but not the stronger forms. Once again, Sumerian behaves like a very typical 4Ht(M) language. In particular, the type of assimilation displayed by the Sumerian conjugation prefixes falls under Casali's category of "weak assimilatory [-ATRj dominance."
E
a
tl
For the purposes of his discussion, Casali ignores front/back asymmetries where there is a gap in the inventory of either front or back vowels. He does note however that inventories with a missing front vowel are more common than inventories with a missing back vowel. This suggests that the
.5. Tn fart, it may turn out that there is only a singlp language, Washkuk (Sepik-Ramu), which is missing a mid-hack vowel and fits Poebels pattern. The other UPSID language thai is missing a mid-back vowel is .\ngas (Chadici, hut Burkctt (1973) indicates that Angas actually has a symmetrical 4Ht(M] inventory. 6. Note however that the classification ma\ not be entirely accurate. When a lilng^la^|;e has only a single mid-back vowel, there seems lo be a tendency U) call it "o" mthev than specifying it more precisely (Maddieson 1984).
ERIC J. M. SMITH
28
Allophonic [-ATR] dominance refers to cases where an underlying [+ATR[ vowel such as /i/or lul has a [-ATR] allophone, such as [i[ or [o]. If such allophones did exist in Sumerian, they do not appear to be distinguished in the writing system. However, I argue in §3.3 that the allophones |i] and [o| are not present in Sumerian.
Coalescent [-ATR) dominance would occur when adjacent [+ATR[ and [-ATR[ vowels fuse to form a sitigle vowel. If [-ATR[ is dominant, we might expect to find the fused vowel emerging as [-ATR[. However, there is no evidence that this sort of coalescence occurs in Sumerian.
(10) Sumerian Harmonv in Various En\itonments Strong assimilatorv dominance -ATR| spread across word boundaries
[-ATR[ sptead in compounds
[-ATR] spread frotn dominant affixes
no
no
no
Allophonic [-ATR1 dominance
Coalescent [-ATR[ dominance
Weak assimilatorv [-ATR[ dominance
no
no
no
Casali claims that the strotig assimilatory dominance is unlikely to be found in 4Ht(Mj inventories. This may have to do with strong assimilatory dominance only being necessary because the perceptual distinctions between the [-t-ATR] vowels (/i/ and luf) and their {-ATR[ connterparis {III and hf) are pariiculariy subtle. In a 4Ht(M) system, the perceptual distinctions between the [+ATR] vowels (/e/and lol) and their [-ATR[ equivalents (/e/ and hf) are easier to make, so strong assimilation is not necessary. As expected from Casalis analysis, there is no evidence that Sumf rian displays any sort of strong assimilatory dominance. There is no indication that [-ATR| spreads across word boundaries or between elements of a compound. Although there are affixes that have a [-ATR] vowel, such as the genitive ca.se suffix -ak and the nominalizing suffix -a, these do not appear to influence the quality of other vowels in the stem to w hich they are affixed. On the whole, Sumerian behaves like a very typical 4Ht(M) language. In line with Casalis analysis, the language displays [-ATR[ dominance rather than [+ATR] dominance. And also in accordance with Casali, Sumerian displays only weak assimilatory [-ATRj dominance.
long history, dating back at least to Jakobson and Halle (1956). The phonemic inventory is oiganized on the basis of contrastive featutes, which are specified in a hierarchical fashion. One particular contrast is deemed to he the most ftmdamenta! one within the system and that ieatttre detennines the initial binary division with the system. Subsequent contrasts within the hierarchy may only be valid for a particular value of a preceding feature. So for instance, if [back] is above [round[ in the contrastive hierarchy, [back] is said to have scope over [round[ and we would expect to find that [rotind[ is onl) contrastive for a particular value of [backj. This concept is developed further in the Successive Division Algorithm (Dresher 2003; Dresher and Zhang 2004), as summarized in (11).
3.3. Application of Successive Division Algorithm (Dresher2003)
The selected feature is contrastive for a members of these sets.
The notion that conttasts within a phonological system are organized hierarchically has a
c. Repeat step (b) in each set with the next feature in the hierarchy, dividing each remain-
(11) Successive Division Algorithm (Dresher and Zhang 2004) a. In the initial state, all sounds are assumed to be variants of a single phoneme. h. If the set is found to have more than one phoneme, a binarv distinction is made on the basis of one of the universal set of distinctive features; tbis cut divides the inventory into a marked set and an unmarked set.
[-ATRI HARMONY AND THE VOWEL INVENTORY OE SUMERIAN ing set until all distinctive .sounds have been differentiated. d. If a feature has not heen designated as contrastive for a phoneme, then it is redundant for that phoneme. The task then is to refer back to the inventory in (8), and determine which features have scope over which other features. The harmony process involves /a/ as a trigger vowel and HI as a target vowel, which indicates that [ATR[ must be contrastive for both [low[ and [high[ vowels. This suggests that, in Sumerian, the IATR[ contrast is more fundamental than either of the height contrasts, and the application of the Successive Division Algorithm shotild start with a division on the basis of [ATR]. If we were to place either [high[ or [low[ for the initial division, then the
29
algorithm would result in some vowels not being contrastive for [ATR] when they should be. Since [-ATR[ is the active featuie in spreading, we will take that as the marked value. Rather than saying that the vowels /i/, /e/, and lul are [+ATR], we will instead describe them as simply lacking the [-ATR] feature. The second iteration of the algorithm involves splitting the [-ATR[ vowels according to the [low| feature. Similarly, the vowels that lack [-ATR[ are split according to the [high[ feature. After making divisions on the basis of |-ATR], |low], and [high], the next thing to determine is whether to divide on the basis of [back[ or [round[. As it turns out, either choice produces the same results. We will arbitrarily choose the relevant feature as [round], but we could have just as easily called it [back] or [labial].
(12) Application of Successive Division Algorithm
not -ATR
-ATR
not low
low /a/
round hi
high
not round kl
not high
round
not round
round
not round
Al/
HI
lol
lei
(13) Eeature confignrations produced by (12) hi
/a/
1-ATR[
[-ATR]
Izl
hound]
|-ATH[
Under this model, the vowel harmony behavior observed in Sumerian can be seen to be caused by the spreading of the [-ATR[ feature. Adding a [-ATR] node to lei produces /eA bttt the writing system makes it impossible to distinguish the two vowels. Sitnilarly, spreading [-ATRj to an lol vowel produces hi, but once again such a change is obscured by the writing system. The one t hange that is clearly visihle through the orthography is what happens when a [-ATR]
lul
[high[
[round]
HI
lol
ihigh[
[round[
lei
node is added to an l\l \owel. Tbere is no vowel iti the inventory that has both [bigh] and [-ATR] nodes. Referring back to Calabreses tree of marking statements shown in (7), recall that we argued that Sumerian only deactivated the first two statements on the left branch (i.e., [-low, -high] and [-high, +ATR]). The marking statement [+hip;h, -ATR[ is still active in Sumerian, so a node containing [high[ and [-ATH1 will need to be repaired. The simplest repair process is simply to
30
ERIC J.M, SMITH
delink the [high] node, leaving onl\' a bare [-ATR| node (i.e., an /E/). The same sort of process must occur when a -ATR[ feature spreads onto a lul vowel, producing a configuration that has [high], [round[, and [-ATR] nodes. Again, the [high[ and [-ATR| features violate Calabreses marking statements, and a repair must be made, removing the [high[ node to produce hi.
As it turns out, this is exactly the sort of repair process that Calabrese (1995) describes as taking place in Italian dialects such as norihem Salentino and southern Umbro. In those dialects the [-ATR] feature spreads to an HI vowel, creating an illegal configuration, which is repaired by delinking the [high| node, producing an Izl. The various assimilation processes are summarized in (14).
(14) Spreading of [-ATR] in Sumerian lei
/a/
1-ATR[
HI
Id
/a/
[low]
/a/
[-ATR]
Id
/a/
T ^^^ [high[
lul
[high[
[-ATR]
[low
[htgh]
[-ATR[
[low]
hi
13.1
[round]
]-ATR[
[l{)w[
[high]
[round]
[-ATR]
[low[
(15) Allophonic [-ATR] harmony (not found in Sumerian) HI
[high]
hi
]-ATR[
HI
[low]
ihigh[
/a/
[high[
iround]
[-ATR[
/a/
[-ATR]
lul
[low
[high[
[round[
[low
/a/
[-ATR]
[low[
[-ATR] HARMONY AND THE VOWEL INVENTORY OE SUMERIAN Note that if these repair processes did not take place, then we would expect to observe the /i+a/ and /u+a/ assimilation to behave as shown in (15). This would represent the allophonic [-ATR[ dominance described by Casaii (2003). However, if this were the case, the conjugation prefix for a verb like ak "to make" would be pronounced [iak[, and we would no longer have a justification for the spelling <e-ak>. It is only by postulating a repair process that deletes the [high[ feature that we can account for the <e> and forms of the prefixes. Akkadian scribes would not have had the vowels kl, lol, or hi in their inventories. The simplest way for them to adapt Sumerian vowels would be to map all back round vowels onto lul and all midfront vowels onto lei. This is summarized in (16). (16) Adaptations of Sumerian vowels into Akkadian Sumerian Akkadian i
e e a
i e a
0
u
u 4. Further Evidence for the Vowel Inventory The two arguments in favor of expanding the vowel inventory have so far followed separate paths. On the one hand, Lieberman (1977, 1979) and Bobrova and Militarev (1989) have argued that orthographic data suggests the existence of vowels beyond the Akkadian /a/, lei, HI, lul. In a parallel but separate approach, Poebel (1931) and Kramer (1936) have argued that vowel harmony also suggests the existence of an expanded vowel inventory. Lieberman (1979) gives credit to Poebel for broaching the idea of an lol vowel in Sumerian,
31
but he fails to connect his findings to Poebel's. Moreover, he suggests that Poebel later gave up on arguing for the existence of an lei vowel." While the IEI vowel is not important to Liebermans orthographic studies, it is crucial for accounting for vowel harmony. Similarly, the hi vowel suggested hy the vowel harmony data is not seen by Lieberman as being distinct from the lol vowel. 4.1. Correlating Orthography and Harmony Unfortunately for comparing the orthographic data from Proto-Ea with the harmony data from the conjugation prefixes, only a handful of the entries in Proto-Ea can unequivocally be identified as verbs. Of those verbs, only a few of those are forms that may contain the elusive lol vowel. And of those verbs with a po.ssible lol vowel, only five are attested in the Early Dynastic texts from southern Me.sopotamia with the relevant conjugation prefixes. These are summarized in (17). Referring back to (4), we note that there are four verbs with apparent lul vowels, which must be hi since they appear with the <e> and prefixes: ru, sur, urg, and ur4. Of these, ru and sur are found in Proto-Ea with orthographies that plausibly represent an hi vowel. Line 736' of Proto-Ea, containing the sign, is found on only one tablet, and the pronunciation column is too damaged to read. Line 872, containing the sign, gives only as the pronunciation, which provides no indication of whether an hi vowel might or might not be present, Nonetheles.s, it is reassuring that the Proto-Ea data at least does not contradict the existence of those hi vowels that are suggested by the vowel harmonv behavior.
7. Apparently Poehels change in position was dup to the recognition that the <deb2> sign also had a value of . The existence of forms like <e-deb2> had been the initial motivation for Poebels arguments in favor of an /6/ (or Izf) vowrl. However, there are several otber verbs, such as gen, me, and dp^, whic'h also t-o-ot'cur with the e- form of the prefix, and hence provide evidence for Izl.
ERIC J.M. SMITH
32
(17) Correlating Proto-Ea /o/vowels with vowel harmony data Verb
Gloss '3
Ur,
to ride
SU^
to be empty, to drown
U7-U4
to sow, to ctiltivate to lay down
ru sur
Prefix forms Reconstructed phonology bio e be2
4
lol
29 /
/so/
to produce a fluitl
However, the vowel-harmony data do point towards the existence of both lol and hi. If, as Lieberman had suggested, there is only a single /o/-t\'pe vowel, then it is hard to explain why u^ and ur»4 are written with the prefix, while sur and ru are written with the <e> and prefixes. The existence of a distinct hi and lol vowels, one with and one without the spreadable [-ATR| feature, provides the necessary explanation for this diflerence. It would have been useful if the lexical lists contained similar orthographic evidence supporting the IeI vs. IEI distinction in Sumerian. However, such a distinction is not to be found in Proto-Ea, which is the earliest such li.st. It has to be remembered that the Sumeriati wiiting system was not rigorous in distinguishing even between IeI and HI; it is hardly surpri.sing that the much subtler distinction between IeI and kl is nol represented at all. Evidence from the Lexicon While evidence for the expanded vowel inventory has drawn on vowel harmony in the prefix chain, it has also been noted that vowel harmony seems to he present elsewhere in the language. In particular, Michalowski (2004J notes a tendency for both vowels of a two-syllable word to harmonize. While a full study of the harmony within the Sumerian lexicon is beyond the scope of this paper, it is worth ptesenting some of the tnore relevant observations.
lovul
3 1
H-DI
IS3VI
The electronic version of the Pennsylvanian Sumerian Dictionary (Sjoberg, Leichty, and Tinney 2004) lists 4336 lexical entries. These can be broken down as shown in (18). It should he cautioned that this is based on a ctirsory s\irvey of the contents of the PSD. A more exhaustive study should be able to reduce the residue of unexplained violations of vovi^el harmony further.'* There is an interesting class of compound verbs that refer to tnaking noise of various sorts. These imiformly violate regressive vowel harmony, typically consisting of a syllable with an HI or lul vowel followed by the same syllable with an /a/ vowel. Examples include zikzak.. .za, pudpad... za, wuwa ... za, and in a more extended template, duhuldahal... za (Black 2003). Evidently there is some onomatopoeic factor here, but it is telling that this is expressed as a violation of the normal constraints on vowel harmony. In accordance with Casali (2003), Sutnerian shotild not display strong assimilatory |-ATR] dominance in compounds. That is to say, there is no reason to consider a form such as nigsaga "goodnes.s" (from nig |derivational motpheme] + .sag, "to please ) to be a violation of the rules of vowel harmony within Sumerian. 8. An anonymous revievt'er has su^gesteH that many apparent CVCVC and CVC^IV words in Sumerian miiv actually represent an attf'mpt to render fonsonant chisters f/ccve/and hvvf respectively). He also suggests that a lar^e number of CVCVC vt'ords in Sumerian are SemiHc loanword.s. Either of these factors may help to reduce further the residue of words that appear to be violating vowel harmony.
[-ATR] HARMONY AND THE VOWEL INVENTORY OE SUMERIAN
33
(18) Breakdown of Harmony within the Pennsylvania Sumerian Dictionary 4336 1190 single-syllable words total 3146 155 with no establed reading entries multi2275 unambiguously obey vowel-harmonv rules syllable 36 appear to deliberately violate hannony for onomatopoeic reasons words 871 contain an /e-a/, /i-a/, 335 can be analyzed as compounds or /u-a/ sequence that 97 appear to have a final /a/ that originated as a suffix (e.g., genitive case -ak) may violate vowel at least 19 appear to be loanwords from Akkadian or Hurrian hannony 209 have an orthography where it is possible that an apparent lul is rules actually an hi or an apparent IeI is actually an kl 175 lack an obvious explanation
Similarly, Casali would predict that Sumerian suffixes should not be dominant. That is, a |-ATR] suffix should fail to spread |-ATRi to the stem. This is the case with the genitive case suffix -ak, as in Iu inimak "witness" {from Iu "person" + inim "word" + genitive -ak). In many cases, it may be productive to reanalvze the reconstructed prontmciation to see whether an apparent lul or IeI might actually conceal an hi or an kl. That is to say that guza "chair" might actually be pronoimced /gDza/, and erigar "fanner" might actually be /engar/. There does remain a residue of multi-syllable words that appear to disobey the rules of vowel harmony. However, this fraction is relatively small, only 176 out of the 4336 total words, or 4.1 percent of the total. These words, such as nitah, "male," and sipad, "shepherd," are difficult to analyze as anything other than an unequivocal hannony violation. The existence of such words, which clearly violate the phonological rules of Sumerian, has been used as evidence for a proposed pre-Sumerian substrate language, although the existence of such a substrate has strongly been challenged (Rubio 1999).
5. Conclusions In this manner, Sumeiian behaves in a fashion identical to niodern-da\- languages that display tongue-root hannony. The patterns of weak assimilatory [-ATR1 dominance and the 4Ht(M) inventory fit perfectly with the model described by Casali. While it is impossible to state unequivocally that the contrastive feature in Sumerian was phonetically realized as |-ATR}, from a phonological standpoint, it acts just as if it really were [-ATR]. By correlating the data from lexical lists and the vowel haimony behavior, we see that Poebels six-vowel inventory is inadequate for explaining the observed data. An additional lol vowel is required to explain the apparent discrepancies. Such an lol vowel produces a symmetrical seven-vowel inventory, which accords nicely with Calabre.ses system of marking statements. In addition, this seven-vowel inventory appears far more common cross-linguistically than the asymmetrical inventory hypothesized by Poebel. Although the feature model proposed here is necessarily speculative, it explains the observed data much better than the generally accepted fourvowel system that is visible to us through the filter of Akkadian.
34
ERIC J. M. SMITH References
Bauer, J.
Keetman, J. 2005 Die altsumerisrhe V{)kalharmonie und die 1972 Altsumerifiche Wirtachaftstexte OILS Lagasck Vokale des sumerischen. JC.S57: 1-16 Rome: Biblical Institute Press. Kramer, S. N. Black, J. A. 1936 The Sumerian Prefix Forms he- and hi- in the 1990 The Alleged 'Extra" Phonemes of Sumerian. Time of the Earlier Princes of Laga.i Chicago; RA 84; 107-18. University of Chicago Press. 2003 Sumerian Noises: Ndeophones in Context. Pp. 35-.'52 in Literatur.. Politik. uiirf Recht in Krecher, J. Mesojxitamien: Festschrift fur Clans Wilcke, 1987 Morphemeless Syntax in Sumerian as Seen eri. W. Saliaherger. Wieshaden: Hanassowitz. on the Background of Wnrd-Composition in Bobrova, L V, and Militarev, A. Yu. Chukchee. AS; 9; 67-88. Landsberger, B. 1989 Towards the Recon.struction of Sumerian Phonology. Pp. 96-105 in Liny^viNtichc.skaija 1951 Die Serie llr-e-a = naqu, Materialen zum Rekoustrukcii/a i Dreiiieyshaya Istorhja VosSumerischen Lexikon II. Rome: Pontifi})edia of the World's Ancient languages, Pontificium Institutum Biblicum. ed. Roger D. Woodard. Camhridge: Cambridge Dresher, B. E. University Press. 2003 Contrast and Asymmetries in Phonology. Pp. 239-.57 in Asymmetry in Grammar, Vol. 2: Poebel, A. Morphology. Phonology. Ac(iuisitUni, ed. A. di 1923 Grti7idziige der sumerischen Grammatik, Sciullo. Amsterdam: John Benjamins. Rostocker orientalistische Studien, 1. Rostock: Selbstveriag des Verfassers. Dresher, B. E., und Zhang, X. 2004 Phonological Contrasts and Phonetics in 1931 The Sumerian Prefix Forms e- and i- iii the Manchu Vowel Systems. Pp. 59-71 in ProTime of the Earlier Princes of iMgas. Chicago; ceedings of the R^rketey Linguistic S(x:iety 29, The Oriental Institute of the University of ed. P M. Nowak, C. Ycx^uelet, and D. Mortensen. Chicago. Berkeley: Berkeley Linguistics Society. 1939 Studies in Akkadian Grammar. Chicago; The Oriental Institute of the University of Chicago. Edzard, D 0. Rubin, G. 2003 Sumerian Grammar. HdO 1.71. Leiden: Brill. Frayne, D. 1999 On the Alleged Pre-Sumerian Substratum. forth;_CS 51: 1-16. coming Pre-Sar^^onic Period {2700-2350 BC), RIME 1. Sjoberg, S. W.; Leichty, E.; and Tinney, S. 2004 Electronic Pennsylvania Sumerian Dictionary Toronto: University of Toronto Pre.ss, van der Hulst, H., and van de Weijer, J. . 1995 Vowel Harmony. Pp. 495-534 in The Handhook Thomsen, M, L. of Phonological Theory, ed. J. A Coldsmith. 1984 The Sumenari Language: AJI Introduction to Cambridge, MA: BlackweU. Its History and Grammatical Structure, Jakobson, R, and Halle, M. Mesopotamia 10. Copenhagen: Akademisk Forlag. 1956 Fundamentals of iMnguage. The Hague; Mouton.
[-ATR] HARMONY AND THE VOWEL INVENTORY OF SUMERIAN
35
Appendix A: Possible lol orthographies in Proto-Ea {Landsberger 1951,1955; Civil 1979)
Line Pronunciation 10 38 63 66 67 69 76 77 79 94 112 113 SU-U2, su-u, SU-U4 121 SU-U2, SU-U4 128 145 151 164 SU2-U2, su-u, SU-U3 172 mu-u, mu-u4 190 230 U4 310 346 U3-gu-ur 362 368 374 408 U2, U4 416a 476 SU2-U2, SU2-U 497 du-U2, du-u 512 SU-U2, SU-U4, SU4-U2, su-u 522 539 540 541 575 U4 594
Sign
Gloss
KU
ku "to place"'' umah "blow" udu "sheep" mu4 "to get dressed" tug2 "textile, garment" umus "sagacity" ukur "poor" usar "neighbor" la'u "arrears" mu5 "good" u "hole"
LAGABxU+A
LU TUG2 TUG2 TUG2 LAL2.DU LAL2.SAR LAL2.KAK
NI U U
SU4 "red"
SU2
SUS2/SU2 "to cover"
HUSI
tnai
U5 "to ride, mount" uri "a vessel" xx
UD
U4/ud "day"
LAGARxSE
HI AH
sur/su7 "threshingfloor;to produce a fluid mu "name, vear"'' uruda/urudu "copper" U2 "plant(s)" ugu "skull" ugur "sword" lud/ulud "a cup, bowl" dub3 "knee" or dug3/dui() "good" uman "insect{s), bug(s)"'
IGLDIB
U3 "sleep"
EN
urun/uruie "to be strong, exalted" sud/su3 "to be empty, to drown" du "to go (sing.)" sub2 "to go (pi.)", sug2 "to stand (pi.)" uru4 "to sow, to cultivate" urum/uruji urujs"flood" uru2 "city" ug^ "light uras "earth"
MU URUDU Uo UKA U.GUR DUG
SUD
DU DU DU
APIN
URU URUxA
URUxUD PIRIGxUD IB
a In those later tablets that provide an Akkadian gloss, the pronunciation coliimn contains , so it is likely that Ihe meaning "tn place" is properly associated with line 11 and nol wilh line 10 . h. Line 171 has a proiiimeiiiHon of <mu-ii2> fbi the <MU> sign, suggesting thai the sign had Iwo different pronunciations (/mu/ and /m.i/ by (iiir reronstriietion). However, there are no surviving tablets with an Akkadian gloss for this sign, so it is not clear whether mu "year" and mu ""name" were pronounced the .same, or differently.
36
ERIC J. M. SMITH
597 619
ru-U3, ni-U2
RU
ru "to lay d o w n
du-U4, tu-U4
TUK
d u i 2 "to take"
669
tu-U2, tu-U4
tui5 " w i n d "
671
ku-U2
IM MI
766
U4-un
EZENxKAS
un3 "to arise, to b e h i g h "
772
U4-di-ni-im,
EZENxSIGy
u d n i m , a place n a m e
774
U3-ud-nim
lEZENxSlGyCP)]
u d n i m , a place n a m e
830
U3-bur
DAG.KISIM5XGA
ubur"breasr
ku|() (in kuiQ-ku|Q, "black
Appenrlix B: Co-occurrence Data for Conjugation Prefixes The following table represents a s>'nthesis of data from Poebel (1931), Kramer (1936), Bauer (1967), the Royal Inscriptions of Mesopotamia (Frayne, forthcoming), and the Cuneiform Digital Library Initiative (Englund and Damerow 2000-), Verbs that tend to co-orour with atid are listed first, followed by verbs that tetid to cooccur with <e> and . Data is drawn from all available Early Dynastic texts from southern Mesopotamia, chiefly from the city-state of Lagas.
<e>
Stem
1
bil " b u r n "
5
bu^(r) "to t e a r out"
1
13
de2 "pour"
11
11
120
1
duf^ "open, loosen"
3
d u g 4 / d u j | "speak, talk"
1
dub "heap up" durun "sit (pi.)" gi4 "return" gid2 "be long, measure out" gU7 "eat" gub "stand" gul "destroy" hug "hire, rent" il2 "lift, carry" kes2 "bind" ku4(r) "enter" mu7 "make noise" se3(g/k) "place" se|2 "dwell" si "be full, fill" sig7 "be pleasant" su "replace"
2
3
5 4
31
1
30 5
1 5
12
1
188 1
17 6 2
4
1
3 2
du3 "build"
[-ATR] HARMONY AND THE VOWEL INVENTORY OF SUMERIAN
SU3 "drown" su8(g) "stand"
29 2 5
24
24 1 1 3 35 19 4 1
SU2 "cover, overwhelm" te/ti "approach" til "finish, cea.se, perish" tu5 "bathe, wash" tUi7
5"
13
3
1 1
1
1
2
1
2
1 2 5
I
63
U
6
8*^ 2 1 1 1
1
37
14 4 57 2 4 14 17 14 1 3 9 5 3 3
1
tuku "have" tus "sit, dwell" U5 "ride, mount" uru4 "plough" US2 "follow. }oin, reach" zig3 "rise, stand tip" ha-lam "ruin, destroy"'' sed/sid "count, recite" ag2 "measure" ak "make, do" ba "give" bal(a) "cross, transfer" bar "open, split" dab.-j "seize, catch" deQ "hring" gaz "slattghter, kill" gaU "be (somewhere)" gar "place" gen "go, come" had2 "dry" hal "deal out, distribute" Ia2 "carrv. hang, weigh" me "be" nag "drink" a ... ru "dedicate" sagg "to be good"
a. Al! live of Ihese exceptions eome from a single tablet, RTC 76. b.On the strength of the form , Poebel (1931) reads the sign as .The value of this sign is attested in ?rot<)-Ea, so such a reading is not unreasonable. In Poebel's aeeount. the variation in conjugation prefixes then becomes a ease of one seribe pronouneing the verb as leknlaml and another scribe pronoLUicing it as |iku!ani|. This seems like a iol to infer on tbe strengib of a single variation in (brm. c. Seven of tbese exceptions come from a single inscription, tbe Stele of Tbe Vultures of Eannatum. Other tban this verb, tlip Stele i.s (jiiite consistent in representing vowel harmony in conjugation prefixes. To account for tliis, Poehel (19311 argued for a reading of for tbe sigii, but such a value is not attested until mucb later Prolo-Eu lists only Ihe pronunciation
tor the sign.
38
ERIC J. M. SMITH
1
1
19 1 9 22 17 9
sar "write sur "produce a fluid" ses2/ ses4 "anoint" 36 1
tag "touch" du]^3 "leave, divorce" "i"3 "drag" ur4 "pluck, collect, harvest"
AN EARLY SECOND-MILLENNIUM CUNEIFORM ARCHIVE FROM CHOCHA GAVANEH, WESTERN IRAN Kamyar Abdi (Dartmouth College) and Gary Beckman (Universittj of Michigan)
Archeological Context (by Kamyar Abdi)
Iran, were carried out in response to ongoing destructive activities at the site by local residents, especially recent leveling and constrtiction work which had alreadx' removed several meters of the upper parts of the high mound, exposing some architectural remains (Kordevani 1971). Chogha Gavaneh best exemplifies the deplorable fate of archaeological sites situated in urban areas. A photograph taken in 1936 by Erich Schmidt (fig. 2) during his aerial reconnaissance of western Iran (Schmidt 1940) indicates that Chogha Gavaneh may have originally occupied an area as large as forty hectares, but in recent decades the lower town has been covered bv new
The epigraphic material presented here was discovered in 1970 during a series of salvage excavations in an architectural complex at the site of Chogha Gavaneh in the middle of the town of Shahabad-e Gharb (formerly known as Harunabad, renamed Islamabad-e Gharb after the 1979 Iranian Revoltttion), about 60 km to the west of Kermanshah in western Iran (fig. 1). (For a more detailed description of tbe geographical situation, see Abdi 1999b: 34-36.) These excavations, supervised by Mahmoud Kordevani on behalf of the then Archaeological Service of
Abdi would like to thank the many people wbo were instrunitntiil in allowing the tablets to leave Iran to come to tbe United States for conservation and .study: First iind foremost I am grateful to Mr. Jalil Golsban, Deputy for Researcb at tbe Iranian Cultural Heritage and Tourism Organization (ICHTO), wbo appreciated tbe significance of the tablets and facilitated llieir loan to Abdi. Since foMowing the 1968 Resolution, no artifacts have been allowed to leave Iran, Se\yed Mohammad Bebeshti, former Direetor of tbe ICHTO, had to go to some lengths with tbe Ministry of Culture and Jraiiian Customs Bureaus to secure a sjiecial permit. Mrs. MohammaH-Reza Kiiregar and Ahmad Chaichi, Director and Deputy for Re.searcb of the Iran National Museum, facilitated the release of the tablets. Ms. Souri Ayazi, head of the Department of Historical and Luristan Antiquities, Miss Zahra Akbari, Curator of tbe latter Department, and Mr. Sbabrokb Razmjou, Curator of tbe Epigrapbir Material in tbe Museum, were all of great help, The tablets left Iran on August 22, 2001 and returned on
December 30. 2002. In the United States, we are grateful to Ms. Ulla Kasten of the Yale Babylonian Collection for conservation and preservation work on the tablets, and to Professor Henry Wrigbt of the University of Michigan, who accompanied the tablets on tbeir journeys. Abdi also wishes to express bis personal gratitude to Mahmoud Kordevani, the original excavator of the tahtet.s, who despite his advanced age and poor health agreed to make several visits to the storage facilities of the Iran National Museum in 1999 and 2000 to belp with the recovery and identification of the finds from his 1970 excavations, including the tablets, and for his generosity in allowing Abdi access to wbatever remained of bis excavation documents, especially maps and plans, and sharing with him much interesting and important anecdotal information. Unfortunately, Kordevani pa.ssed away in 2003 before Abdi could pa\- him anotber visit with more specific questions and a copy of this report.
39
JCS 59 (2007)
40
KAMYAR ABDI AND GARY BECKMAN
Fig. 2. Aerial photograph of the town of Harunabad in 1936. Note the site of Chogha Gavaneh and its high mound. Photo liy E. F. Schmidt, courtesy of the Oriental Institute of the University of Chicago.
I
I < 1.000
I
I 1.000-2.000 > 2.000
\aneh the first visible feature as one approaches the town of Islamabad from kilometers away. Therefore, it comes as no surprise that Chogba Gavaneh has long been a subject of curiosity.
Fig. 1. Map of western Iran and Mesopotamia showing the location of Chogha Gavaneh.
Research al Chogha Gavaneh
constrtiction, while the hip;h motind has contintted to be quarried and shaved off from the sides and top to clear room for building activitie.s. Consequently, Ghogha Gavaneh is now only about 4 hectares in area, perhaps as little as ten percent (if its original size (fig. 3). The most prominent part of the site today is the "high mound" (fig. 4), where one can find the longest preserved sequence of occttpational deposits (as early as the Early Neolithic Period to the Middle Bionz.e Age. according to test excavations at Operation W263 and STl; see Ahdi forthcoming), The "lower town," where one tnight have expected to find deposits of later periods, is now completely covered by the town of Islamabad. Despite extensive damage, Chogha Gavaneh is still the largest site of prehistoric and early historic times on the Islamabad Plain and one of the largest archaeological sites in the West-Central Zagros Mountains. The high mounds massive volume and towering height of over 25 meters above the plain level (fig. 4) makes Chogha Ga-
Several early travelers (cf. Jones 1857) passed throtigh the Islamabad Plain (then called the Harunabad Plain) in the nineteenth and early twentieth centuries, but none seem to have remarked on the archaeological remains in the area Th(^ first traveler/archaeologist to give a hrief description of the region was Aurel Stein, who visited the town of Harunabad in 1936 during his general survey of western Iran (Stein 1940; 420). T^ater the same year Erich Sc hniidt f!ew over the Plain during his aerial reconnaissance in western Iran (Schmidt 1940). In 1959-1960, as part of the Iranian Prehistory Project of the Oriental Institute of the University of Chicago, Robert Braidwood and his team briefly surveyed the Plain (Braidwood 1961), but they do not seem to have explored Chogha Gavaneh in any detail, A few years later, Clare Gofl' (1966) of the British Institute of Persian Sttrdies visited the Plain during her general survey of the Central Zagros Mountains. The first series of archaeological excavations at Chogha Gavaneh were carried out in 1967 by Ali-Akbar Karegar Sarfaraz, Mohammad-Rahim
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
41
Fig. 4. The high mound iu 1997.
Fig. 3. Topographic plan of the high incjuud as of 1998. Sarraf, and Ismai'l Yaghma'i (from the then Archaeological Service of Iran). This team opened a step trench at the northeast side of the high mound to study its stratigraphy, finding levels from the Chalcolithic Period to the Ir"on Age as well as disturbed remains of later periods. Following the 1967fieldwork, although the site was registered in the list of national monuments., infringements upon Chogha Gavaneh by the locals intensified. In a matter of just three \ears, as a result of the growth of the town, the slopes of the high mound were covered with houses. Early iu 1970, the town municipality cleared the conical-shaped top of the mound to establish a tea house. Notified of the.se activities, the Archaeological Service of Iran dispatched Mahmoud Kordevani to prevent further damage to the site arrd conduct excavations in cjuarried areas. For three months in the summer of 1970, Kordevani and his team cleared an area of about 0.8 hectares on the top of the high mound, exposing a major architectural complex (fig. 5). The finds from this architectural complexincluding the collection of tablets discus.sed here— and the evidence for conflagration in some places led Kordevani (1971: 46-50) to date the complex to the late Iron Age II of the Central Zagros {ca. 800 B.c:.) and to interpret it as one of many settlements the Neo-Assyrian kings claimed to have sacked and burned in the Zagros.
Despite its tremendous potential, work at Chogha Gavaneh did not continue beyond the first season. Kordevani was dispatched to work at Persepolis with Akbar Tajvidi, and Chogha Gavaneh was left to the mercy of the locals, who continued with their destructive activities. For a short season in 1980, Mahmoud Mousavi, Ismail Yaghma'i, and Ali Valinouri from the then Iranian Center for Archaeological Research returned to Chogha Gavaneh to make another attempt at preserving the site from further destruction. Kamyar Abdi began a regional archaeological project in the Islamabad plain in 1997. Abdis field research was primarily focused on the early prehistory of the area (Abdi 1999a, 1999b, 2000, 2001a, 2001b, 2002, 2003; Abdi, Biglari. and Heydari 2002; Abdi et al. 2002; Biglari and Abdi 1999; Heydari 2001; Mashkour and Abdi 2002), but a re-study of old excavations at Chogha Gavaneh was also on the agenda, especialh' the recovery of the tablets and other finds from the 1970 excavations in the storage facilities of the Iran National Museum in Tehran, where they were stored after the excavations, and limited excavations that would shed more light on the archaeological context of the tablets. Abdi is pleased to inform the reader that he has been able to locate and document most of the finds from the 1970 excavations. They are currently under study and will be published in another paper in the not too distant future.
42
KAMYAR ABDI AND GARY BECKMAN
Fig. 5. Plan of the architectural complex excavated in 1970. Redrawn after the original field map, courtesv of Mahmoud Kordevani.
The Archilectiiral Complex The architectural complex (fig. 5) in which the tablets were discovered demonstrates the characteristics of a pre-planned and well constructed compound. In terms of general layout, the complex resembles a range of public and private buildings excavated at Mesopotamian sites such as Ur (cf. Woolley 1976: pi. 128) and Tell ed-Der (Baqir and Mustafa 1945:fig.4; Gasche 1989: pi. 2), dating to late third and early second millennia B.c:. Despite relatively extensive excavations (abotit 0.8 h) it seems that only a portion of the compound has been exposed. The excavated parts seem to be the inner quarters of the compound, while the more peripheral rooms and walls that may once have surrounded it have been quarried or
eroded away. Functional interpretation of excavated rooms is difficult in absence of a detailed description of finds from individual loci, but general observation indicates that the compound consists of a number of irregularly-shaped rooms in the northeast corner, which may have been for domestic activities, and a series of residential spaces to the east, to the north of a feature which may have been an entrance to the compound at the north side of a courtyard. To the west one can see a number of larger rooms whose function is unknown, but which may have been the administrative part of the compound, where a large reception hall and behind that a repository for tablets (Room B15) were located.
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
Fig. 6. Room B15 prior to excavation. Re-excavating Room B15 According to Koidevanis report, the tablets were discovered in Room B15. In this room was also found a clay figurine of an animal (4 cm in length), a nude female plaque ( 7 x 5 em)., and a bronze arrowhead (about 6 cm long). In the excavation report there is also reference to a bronze blade inscribed with the words "Palace ... day" (Kordevani 1971: 43), but it is not clear whether the blade came from this room or elsewhere in the complex. In any event, Abdis attempt to locate this blade met with no success; it may have undergone conservation over the years and been transferred to one of many departments in the Iran National Museum or have been sent on long-term loan to a regional mu.seum. The cylinder seal discussed iu this paper (figs. 33-34) was also di.scovered in room B15. In order to study the archaeological context of the archive., Abdi re-excavated Room B15 on Jtily 21-25, 1998.' The primary goal of reexcavation was to study the internal organization of the room and di.scover any clues as to the context of the tablets. Finding additional artifacts, including tablets, was not a possibility, since Kordevanis team had already cleared the room down to the floor level and removed all thefinds.Since Kordevani was hoping to return to the site for a second 1, Abdi is grateful to Abbas Motarjeni (tbe lepresentative of the Imntan CuilunU Heritage Organization), Jebreil Nokandeh, Ali Farahani, and Mamid-Reza Valipour for tbeir as.si.stanre with the excavations.
43
l-'ig. 7. Krnuniug tlu^ underbrush prior lo excavations. Note in the background the hummock erected during the Iran-Iraq War as support for an anti-aircraft battery.
season of excavations (which never took place), he did not refill his excavations, but over time the complex had been partially reburied due to natural erosion. Further, there have been several changes in the topography of the area on top of the mound, inehrding a hrrmmock erected during the Iran-Iraq War of 1980-1988 to install an antiaircraft battery. This hummock has since eroded away and become part of the mound's topography (see fig. 7), making discernment of the layout of the compound and the location of individual rcroms more difficult. The only clue as to where to look for Room B15 was its known genera! location towards the north .side of the ccmipound and patterns on the surface that might be interpreted as remnants of the erosion of the ancient mudbrick walls some thirty years after they had been excavated. After some probing, we settled on an area as the possible location of Room 15 (fig. 6). After clearing the area of underhrush (fig. 7), we sprayed water to mark the mudbrick walls from the fill (fig. 8), an archaeological trick that proved successful, as color differentiation immediately showed us where the mudbrick walls were. Hence, we began excavation from the south side of the room (fig. 9). The first features to be exposed were a doorway with a pivot to the right-hand side as one entered the room and a block of stone to the left (fig. 10), presirmahly for a wooden door to sit on the pivot and be closed with the block.
44
KAMYAR ABDI AND GARY BECKMAN
Fig. 10. Doorway to Room B15 with a stone pivot to the right and a door-stop to the left.
Fig. 8. Spraying water reveals mudbrick walls.
Fig. II. Southeast corner of Room B15 showing the position of the doorway vis-^-vis stone slab, presumably used as a bench.
Fig. 9. Excavation begins at the southern end of Room B15.
Fig. 12. Traces of burning visihle iu tiie area around the stone bench on the eastern wall of Room B15.
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
Fig. 13. The eastern side of Room B15 with the location of the doorway, stone bench, and mudbrick platform.
Fig. 14. Northern side of Room B15 with the lot ation of the nuidhrick platform (NE corner) vis-a-vis ashy and silty feature (NW corner).
Fig. 15. Ashy and siltv feature in the NW corner of Room B15 prior to excavations.
As excavations progressed, we fotind another block of stone about 2 meters north of the doorway along the eastern wall of the room, presumably a bench (fig. 11). The area aiound this hlock and the wall behind it showed clear evidence of fire in the form of discoloration (fig. 12). Farther up, in the northeast corner of the room we found a mudbrit k benc h with some traces of ash (fig. 13). As excavations progressed along tbe northern wall of the room, we encountered a peculiar ashy/silty
45
deposit opposite the mrrdbrick bench on the northwest corner of the room (fig. 14). Upon further excavation and removal of the top la\er of deposit, we encountered a large ashy deposit (fig. 15). Once this deposit was removed, a peculiar feature emerged; It consisted of a fragmentary pottery vessel placed in a pit with .sloping sides leading to a ridge marked with mudbricks. To the west was found a supporting buttress behind this feature and the room wall, with two baked bricks intact (fig. 16). The function of this feature still eludes us, but the abundance of ash in and around it would make some sort of fire installation a feasible explanation. However, the only way we can explain the amount of ash still preserved at this area is to assume that Kordevanis excavators backfilled this area with its own fill after they were done with the excavations. Unfortunately we never had the opportunity to ask Kordevani about this. Excavations continued along the western wall, where we ohserved some patterns in the brickwork that could be interpreted as a coved cornice (fig. 17). Once the entire interior- surface of the walls was exposed, we C'ontiuued to (>robe the floor; Only 20-25 cm below the surface a r"ough floor made from hardened plaster was recovered (fig. 18). As expected, no finds were made in the (ourse of excavation. Room B15 is roughly rectangular in shape and about 12 x 15 m in size (fig. 19), with mudbrick walls preserved in some places to a height of 70 cm. The walls of the room ave regular and made from standard-size mudhricks of roirghly 35 X 35 X 10 cm, often plastered and in some cases painted. Small-scale clearing of the plaster suggests that mudbricks were geuerally laid in stretcher-stretcher form on the exterior surfaces (fig. 20). Scraping ofF the rrpper snrface of a wall section indicates that beyond the exter ior courses of full bricks, the interior of the wall is filled in a multitude of ways with full, half, or fragmentary hricks(fig.21). According to the original excavation report, the tablets were discovered sitting on the floor along the southern wall of Room 15 (Kordevani 1971: 45), but we had uo way to verify this, nor have we any information on their arrangemerrt. Our reconstruction of the room (fig. 22), its internal
46
KAMYAR ABDI AND GARY BECKMAN
Fig. 16. The feature in the NW corner after excavation.
Fig. 18. General view of Room B15 seen from SE to NW.
Fig. 17. SW corner of Room B15 with traces of coved cornice.
Fig. 19. Schematic ground plan of Room B15 and cross-section of the NW corner feature.
organization with benches and fire installations, and its position vis-a-vis the other rooms in the compound, suggest that this was the seribes oflEice as well as the tablet archive for the administrative apparatus functioning out of this complex. We now tnrn to the tablets to find out more about this administration and its function. The Tablets (by Gary Beckman) Dedicated to the memorij ofHarald Ingholt 11896-198.5), excavator of Tell Shemsliara
In accordance with the Iron Age II date assigned to the excavated material bv Mahmoud Korde-
1
I
I
Fig. 20. Schematic profile of a wall in Room B15 showing stretcher-stretcher brickwor k.
vani (see above). Dr. Abdi had told me to expect tablets fr-om Neo-Assyrian times, but paleographic analysis dates this archive to the Old Babylonian period, more precisely to the early eighteenth century B.c;. (middle chronology). The script is similar to that of the letters of Hammurapi to Samas-
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
Room B15
N
t Fig. 21. Overview plan of the eastern wall of Room B15 showing irregular brickwork on the interior section of the wall held together by regrrlar brickwor'k on the exterior
Fig. 22. Three-dimensional reconstr'uction of Room B15.
hazer, for example, or to that used in economic records from the reign of Rim-Srn I of Larsa. Mimation is for the most part still present. In his
47
work on Old Bahylonian material from the Rania plain, Jesper Eidem noted; "The finds from Shemshara indirectly prove that contemporary written sources should be available at other sites in the western Zagros" (Eidem and L8ess0e 2001: 58). The Chogha Gavaneh tahlets, discovered before Eidem wr'ote but unknown to philologists until now. confirm tbe perspicacity of this prediction. All of these doctrments are written in Akkadian, with a frequency of Sumerograms comparable to that of contemporary Mesopotamian usage. The syllabary is that of the Diyala region: note the use of WI for walivilwn (ChG 20 rev. vi 13' and passim) and QA (ChG 1;22), The archive consists of fifty-six tablets, many very worn and practically all incomplete, another twenty-eight fr-agments deemed worth copying, and one cylinder seal. Jirdging fVom the number of half tablets and fragments that could be joined neither to another tablet nor to one another, the corpus of texts must have originally been significantly larger than what came down to me. Assuming that the building in which the tablets were found has been correctly dated to the first millennium, the Old Babylonian texts must have found their way thither as fill, and may therefore already have been in poor and fragmentary condition when deposited. The primary concerns of the archive are agricultural and pastoral. Letters deal with barley rations (ChG 3) and with draft animals (ChG 2), and we find an account of work performed by slaves and asses (ChG 16) and another of sheep (ChG 12), Among the few professional designations appearing in these records are SIPA, "shepherd" (ChG 20 v 4', 12') and NA.GADA, "chief herdsman" (ChG 20 i 17'; 23 i 7'). Other texts are concerned with textiles (ChG 3; 4; 17?) and with soldiers bearing the unusual designation A dozen small tablets (CbG 6-10; 11?) record the receipt of a commodity, apparently seed gr'ain. There are a large number of lists—ration li.sts and simple lists of names, probably duty rosters. It is interesting that many of the personal names recorded here are those of women. These are the
48
KAMYAR ABDI AND GARY BECKMAN
records of a redistributive economy, in which connection note also the occtjrrente of the term SUKU, "food allowance" (ChG 33: 9', 12'; 44: 2', 3', 4'). The onomasticon i.s overwhelmingly Akkadian, with a small admixture of Amorite names (13 of 180 complete names, or 7,2 percent—see Index 1), There is no clear sign of Elamites or Hurrians (with the possible exception of Zuzzu), or for that matter of Gutiansor LuUn. The single mention of li^i^m^Ijapirf, (ChG F3:8') is uncertain. Besides the seal legend, which mentions ''ISKUR—certainly to be read Adad here, we encounter no divinities in these texts. In theophoric names Sin is by far the hest-represented god, followed at some distance by Istar, Amurrum, Samas, and Adad. Also attested are Ea, Gula, Ishara, Lahma, Mama, Namar, Tispak, and Tutu, See Index III. Around thirty place names appear in the records—most only once—but the great majority of these cannot yet be identified with toponyms attested elsewhere. See Index II. The towns of Nikkum (RIA 9: 569-70), Me-Turan(?) (RIA 8: 150), Haburatum (Wafler 2001: 82), Agade, and D?r lead us down the eastern Diyala drainage through the Hamrin basin to the Tigris and into Mesopotamia, (For the historical geography and archaeological sites of this region .see Postgate 1979; Muhamed 1992: 10-25; and Saporetti 2002: 144-65). There is little evidence in the tablets to allow us to determine the ancient name of Chogha Gavaneh itself, but it is just possible that it was called Palum (see commentary to GhG 5 below). Our archive comes from further up the Great Khorasan Road than any other published group of cuneiform records; see map in Levine (1974: 101) for a particularly clear indication of its location (as Shahabad). We are literally in unknown territory. Consequently, it is not certain just what the region in which the mound is located was called in ancient times. However, if Nikkum is to be located at or near Haninqln (so Rollig, RlA 9 [1998|: 92; cf. also Frayne 1992: 64), then Chogha Gavaneh may well have been part of the land of
Namar/Namri, In this connection, note the occurrence of the deified geographic term in the personal name Su-Namar (ChG 19:17) and cf. map 11 in Parpola and Porter (2001). Most of these records are undated, and the few that do bear dates include only tbe month and day. The three attested month names; Kinunu(m) (ChG 8:5; 9:5), Tamhirum (ChG 7:5; 10:5), and Saharatum (ChG 11:6), indicate affinities, as might be expected, with the calendar of the Diyala region (see Cohen 1993: 251-54). We can conchide that the settlement only partially excavated at Chogha Gavaneh was inhabited by Mesopotamians linked to the towns of the lower Diyala, and most likely to the kingdom of Esnunna in particular. It seemingly sustained itself through the raising of sheep and the cultivation of grain, as well as perhaps by the production of textiles in workshops staffed primarily by women. Catalog No. 1 No, 2 No, 3 No, 4 No. 5 No. 6 No. 7 No. 8 No, 9 No. 10 No. 11 No. 12 No, 13 No, 14 No. 15 No, 16
No, 17 No. 18
letter letter letter letter receipt of seed(?) receipt of seed receipt of seed receipt of seed receipt of seed receipt of seed receipt of seed(?) account(?) of sbeep account of slaves disbursement of seed grain and barlev rations receipt(?) of flour(?) ledger account of days worked by slaves and asses(?) disbursement of linen cloths(?) list of soldiers
41x68x22 35x42x21 37x23x19 40x29x20 21x26x17 23x25x16 25x24x16 27x26x16 29x28x20 28x22x16 19x26x17 29x21x18 34x57x18 27x31x14 25x12x18 27x40x19
60x62x30 56x45x21
A CUNEIFORM ARCHIVE EROM CHOGHA GAVANEH, WESTERN IRAN
No. 19 list of numbers of slave 41x83x29 girls and personal names, bv towns No. 20 list of persons 80x73x38 No. 21 list of persons 30x25x12 37x31x22 No. 22 ration list—badev No. 23 list of persons 86x121x52 No. 24 list of persons 33x58x22 No. 25 list of persons 22x32x19 No. 26 list of persons 20x35x22 ration list 31x26x20 No. 27 No. 28 ration list 32x37x17 No. 29 ration li.st 35x35x15 No. 30 ration list 48x60x17 No. 31 ration list—barley 31x43x17 No. 32 ration list 36x48x20 No. 33 disbursement of barlev 32x30x17 tinn fldiir cl J l\-l IIIL^LII
No. 34 No. 35 No. 36 No. 37 No. 38 No. 39 No. 40 No. 41 No. 42 No. 43 No. 44 No. 45 No. 46 No. 47 No. 48 No. 49 No. 50 No. 51 No. 52 No. 53 No. 54 No. 55 No. 56
ration list ration list ration list ration list ration list ration list ration list ration list ration list ration li.st bulla—disbursement of barley ration list account(?) of barley ration list ration li.st unc'ertain record of slaves bulla—disbursement of ft)odstutf ration li.st—barleyCr") ration list ration list ration list ration list
40x60x21 38x44x20 40x38x16 42x23x17 36x50x21 44x72x20 44x87x23 49x58x22 32x36x17 40x29x25 38x26 26x40x20 20x26x20 22x33x08 24x17x14 28x29x22 31x24x21 41x37 33x29x21 36x34x17 46x36x22 46x38x21 38x34x18
49
Fragments Fl F2 F3 F4 F5 F6 F7 F8 F9 FIG FU F12 F13 F14 F15 F16 F17 F18 F19 F20 F21
F22 F23 F24 F25
F26 F27 F28
list of persons uncertain uncertain list of persons list of persons; indirect join to No. 20 list of persons list of persons list of persons list of persons list of persons(?) disbursement(?) ration list uncertain uncertain list of persons list of persons list of persons uncertain list of persons list of persons list of persons(?) ration list list of persons list of persons uncertain list of persons uncertain ration list
30x36x15 39x36x21 40x57x09 30x29x12 36x23x10 34x23x12 • 24x30x14 29x22x11 18x22x11 17x29x13 40x16x19 21x15x18 26x16x12 25x29x20 22x24x16 23x13x13 23x17x15 20x13x15 27x15x23 13x21x13 18x30x22 18x21x09 19x22x10 40x32x18 33x48x19 33x37x22 31x27x24 30x19x22
50
KAMYAR ABDI AND GARY BECKMAN 10. u iS'ie-en s\a T^'Ha-bu-h-a 11. id'-hia^-as-su-nu-si-im-ma LoE 12. iY"?j-.s'»-|7i(/l li'il-li-ku 13. ANHU-/[w|"i rev. 14. at-{t\aa-b\i t]i-di 15. X X X [ 0 |-//-aH-7j/
Fig. 23. ChG 1 obv.
UE
16. 17. 18. 19. 20. 21. 22. 23. 24. 25.
ki-a-am
.sJa fa-o[/"''-^|a'-^a-/cu H as-sum *''^/c[a-ii]a-.sa-ar-' sa In-ba-x [ o' ] x x | iq'hi-a-kn-u\m /cw-uH|/-|'ri-;;ia ^'^^ka-na-^sa^-ar-Wa}-\am\
la is-ti'iii-is ^da^-[am-qa-am\ i
Say to Ammu[...], thus says Isar-|.. .|: This young child(?) of mine |.. .| Kuhburum and(?) |...] May they [...] the inspector(?). And now |. ..| and send here! Give them one of (the town) Nakitta(?) and one of (the town) Haburatu(?), and let them ^o w ith them. You, my father, know ... to me thus that you will depait. And concerning the pole pin(?) that Inba ... 1.,.) spoke to you about; "I..) and withold the pole pin(?)" Let him encircle the cart with good |.. .| oil. You shall not pour [out] the [good] oil all at once.
Fig. 24. ChG 1 rev.
ChGl (figs. 23-24) obv.
1. 'a'-MH \A\ni-mu-x [ 2 ni-b\i~nia\ 3. um-ma l-Har^-^x [ 4. an-ni-a da-aq-qf sa-x \ 5. ""Ku-ub-bu-ritni hY^ x-la6. mu-wa-m-am-ma Hi^-\ 7. u a-nu-um-ma ta-\ [ 8. u Hu-bi-Ham'^-ma X \ 9. is'-te-en sa ^'^^^'N\a^-k\r-it
The word order in this letter is often unusual; see especially lines 14 and 24f. obv. 4: For daqqum as "small (child)," see CAD D, 107. obv. 6: CAD M/II, 196, translates mumassii as "slave inspectorC?)," but since it is simply the participle of mussu, "to distinguish" {CAD M/II, 235-36), perhaps sucb a restricted rendering is not necessary. rev. 17, 21: CAD K, 143, lists attestations from Mad and Ischali for '^^^kana.ssarum as an agricultural implement, but the lexical section also indicates its use as part of a wbeel, wbich may fit better with ^-'^MARGiaiDA] in rev. 23. rev. 19: It is unclear wbere the quotation introduced here ends. rev. 20: For the rare usage of kalu in the D-stem, seeCA/iK, 102f.
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN ChG 2 obv.
rev.
1. [a-n|a Su-mu-li-si 2. qi-bi-hna^ 3. um-ma '^EN.ZV-i-Uji-sa-am^-m\a] 4. a-nu-um-ma a-n[a''] i-x-ti 5. [i]e-ti^ANSE\m.Asaln-ba'X 6. [ 0 o 1 X at-ta-ar'-da-am 7. [ o 0 \-^da'^-am' li-fa-at-\[ (probably only one line lost) 8'. x{-)a-bi{-)\ 9'. [u\m'-hna^ GAL X \
51
12'. [ o 0 1 X a-hi ma-an-\
13'. (traces) |.. .| two garments |.. .| all of them. Perhaps they will indeed convey those washed two or three times into the possession of Kinialluh. When you (f.) wrote to me; "1...) in (the town) Palum |..., or] in (the town) Almati |.. .| mv brother (...)"'
10'. [ o l x - a / x [ Say lto| Sumu-lisi, thus says Sin-iqt§am: Now I have sent 1...] with the asses of Inba... Let him/ them I.. .|
ChG 3 obv.^
r. 2'. 3'. 4'. 5'.
[ 1x1 |SlEVBA-i.HI.Arsaix| X SO" tup-pi an-tum x4 [ ol
LoE
6'. 6-.SW i-nu-ma-tum
rev.
7'. U-^ik-lu 8'. r§Ei.BA.HI.A GIR 9'. X I I X 'sa IN'.NU
I am unable to make connected sense of tbis fragmentary letter tbat seems to be concemed with disbursement of barley rations and with a large quantity of linen cloths, i-nu-ma-tum in line 6' is puzzling. ChG 4 obv.
1'. 2'. 3'. A'. 5'. LoE 6'. 7'. 8'. rev. 9'. 10'.
[ [ o o 0 ] 2 fiihba~tu [ o o ] x-i.5 ka-lu-u-m-nu \S[a 2-su m 3-su me-e-si 'mi-id-de a-na" ^Ki-ni-al-lu-uh a-na le-et bti-H-im Iti-ii i-re-ed-du-ii i-nu-ma ta-as-pu-ri-im xx-f/ia^Ti-na'^'™Pa-/i*^' [ ]-nim
Fig. 25. ChG 5 rev.
ChG 5 (fig. 25) 1. '2,1.01 NU[MUN''1 2. N|IG.S|U''EN.'ZUi-x 3. {na\m-\h]ar-Hi^ 4. "'Be-el-su-nu LoE 5. ' ™ l ^ ' P ' ' ' 6. e-zu-ub rev. 7. 7ZlZm 8. maaliri'tim' 9. [SlUNIGIN^ 15 ma-an-^di^ S1E| 10. KA/SAG NIG.GA X X
obv.
2,1.0 of seed, via Sin-..., received by Belsunu (of the town) Palum, excluding the earlier... for the mandiisoldiers (to he provisioned withO barley. Total: fifteen »iiindu-soldiers (to be provisioned withO barley ...
Gf.ChGia LoE 5: Does tbe mention of the town Palum in connection with tbe recipient perhaps indicate tbat this was the ancient name of Chogha Gavaneb? rev. 7: For mandu as a type of soldier, previously attested only in lexical lists, see CAD , 209.
52
KAMYAR ABDI AND GARY REGKMAN
ChG 6 obv.
1. 2. 3. 4.
[ NUM|UN NIG.'SU' SO-\i\m-gur-ra-a?i'ni SUTI.A "'Be-el-^su^-nu Sealing: Be-el-su-nu DUMU Da-aq-tum ARAD 1-ni-ib-sar-ti
UE
0,0.5'' .seed, at the disposal of Sin-imguranni, received bv Sri-AmiiiTum. Second dav of Tamhiniin. ChG 8 obv,
\.. .\ seed, at the disposal of Stn-iniguranni, received hy Belsiiiiii. No date.
The legend of the cylinder seal has been reconstructed from tbe partial impressions found on GhG 6-10; of the pictorial portion of tbe sealing only a single standing figure may be made out, but no details are recognizable. Since tbe personage of whom a seal owner is said to be the servant is— if not a deity—normally a human ruler, Inib-sarri must have been an important figure. The only prominent individual bearing this name known to me is the daughter whom Zimri-Lim of Mari gave in marriage to Ibal-Adad of Aslakka (Ziegler 1999; 62). But since this polity was located in northern Syria, it is unlikely that we are dealing with the same person here. Remarkably, Belsunu identifies himself as both the son of one woman and the servant of another. To Daqtum, cf. Daqatum (Ziegler 1999: 268).
6. UD2.KAM Sealing as on ChG 6
rev. UE
1. 2. 3. 4. .'5. 6.
0,0.5 NUMUN NiaSV SO-im-gur-an-ni ^nam-har^-ti |"'S]i/-''''iMAR.TU "'|K|i-n»-n» UD.2a'KAMi Sealing as on ChG 6
0,0.5 seed, at the disposal of Sin-imguranni, received bv Sn-Amurriim. Twent\'-third dav of Kinunii. ChG 9 obv.
rev.
1. 1 1 NUMUN 2. INIG.I^SUi |'*E|N.ZU-im-gwr 3. hiam'^-har-^fi^ 4. ^'"^Bt-ei-sii-nu 5. 6. 'UD126.KAM Sealing as on ChG 6
]...] seed, at the disposal of Sin-imguranni, received by Belsnnu. Twenty-sixth day of Kinunu. ChG 10 obv.
1. 0,1.0 NUMUN
2. NiG.su '^E^ZV-Hm-gur^-au-ni 3. \na\m-\har-ti\ Fig. 26. ChG 7 obv. ChG 7 (Hg. 26) obv.
rev.
1. 2. 3. 4. 5.
0,0.5"^ NUMUN mCSU QO-im-gui-an-ui nam-ljar-ti "'Sn-MAR.TU
rev.
4. [ " I 5. '[''"'J'Tam-/i/'-ru;jiUD.10,KAM Sealing as on ChG 6
0.1.0 seed, at the disposal of Sin-imguranni, received bv 1.. .1. Tenth dav of Tamhirum.
ChG 11 ohv.
1. I 2. 1271 X
A GUNEIEORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
LoE rev.
3. 4. 5'. 6'.
NfG.§U Sa-x nam-har-t\i] j '^'Sa-/ia-rla-l
ChG 15 obv.
ChG 12
LeE
1. 3,1.2
i
2. rev.
obv.
I. me 20 UDU.HI.fA 2. m^UDU.Hl.A[ 3. |m|plO^U[DU.HI. l.MES 4'.
3. \?iam-ljar-ti\ 4', [ 5'. ''^^
rev. 5': From the typical layout of receipts, one might expect a month name here, but the traces do not correspond to any attested month listed in RIA 5, 297-303, sub Kalender.
ChG 13 1. 2. 3. 4. 5. 6. 7. 8. LoE 9. rev. 10. 11.
nbv.
20 20 20 10 '12'i 16 49^1
'2'.|KA1M '3l[KA]M •4l|KA]M |5|.fKAMi fO'.KAM •7'.KAM '8'.KAM [116[9.KAM] SU.NIGIN 1 me 43 AIMD.x erasure
GhG16 obv.
11+
^2+1
5 3 3 3
4 2
rev.
I. \N Hmpa -mMU.l.KAM(?)l 2. U2^^ 2.K|AMi
4. 5. 6. LoE 7. 8. 9. rev. 10.
3.fKAMi 4.KAM 5.KAM '^6.KAMi 7.KAM •8'.KAM f9lKAM
23 31 28 14 10 [N] [SUNIGIN] x-^5 SE.NUMUN
11. 1 1 r 1
SE.BA I erasure
obv. 1: Restored after CbG 1:1.
[ 1
1
11
11
X
7
[ 1
X
11
11 11
[
[ 1
]
7'. X 1 8'. AN[SE 9'. X I
10'. 2 rue [ 11'. 2 me [ 12'. 2 mex [
ChG 14
3. fl6-i
2 2
2 4
11
Note that the sum of the individual entries, 139, does not match tbe total given in line 10.
obV.
53
13'. 7mc'x [ 14'. ARAD.GEM[E LeE
15'.
liq mt^l
,S'N3ARADMES1'X
16'. 4me86ANSE.[HI.A
54
KAMYAR ABDI AND GARY BECKMAN
9. 10. LoE II. 12. 13. rev. 14. 15. 16. 17. 18. 19. 20. 21.
I x-BV-fum \x-sa-xi-)\ -k]i-mi-xi-) ]x-tu{-)\ ]xxx I lxxx[ 1 x-'iui-x f
-s/f\a'-ma'-l\a(-) |x-x-/a'-a/jG|ADA| ]{-)x-ra'tu7n 1
GADA 22.
\ X ^ki
nu''ma-in-iz''
Eig. 27. GhG 17 obv. Tbis ratber tbick tablet is of unusual sbape; mucb of tbe surface is very worn. It seems to record tbe disbursement of linen clotbs to various individuals. Because of tbe uncertainty of interpretation, wbat appear to be fragmentary personal names bave not been included in tbe index. ChG 18 obv.
1. 7 ma-an-di^ A'mu'ur^-\ri] 3. 3 ERIN pu-hu'-um 4. fia A-ga-de 5. ^8^ Sa-li I-di
rev.
Fig. 28. CbG 17 rev.
7. 18 UGULA^ IM^ ma-an-di 8. SE
Seven Amorite mandu-soldievs from Der. Tbree .substitiite soldiers from Agade. Eight (soldiers) of Silli(ya),
ChG 17 (figs. 27-28) obv.
1. [
]
2. i 5. 1
] x-mu-^m'-^-NAMGADA' ] PU.SA-ya GADA 1 x-NI-BI GADA |-'/fl'^-/??-j/flGADA
6. [
1 'a-i-»a NIM^'^GADA
3. [ 4. [
^A^-pif-Ku'-bi'\-GADA'
7. 1 -s]a'-na-mu GADA 8. [ -s\a-pa-ar-s\a{-)
<.son of) Tdi, o
iiunidu-soldit
Cf. CbG 5. obv. 5: Cf.
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN 8. 9. 10. 11. 12.
55
19 14ARAD-''MAR.TU 12ARAD-''MA1R.T1U 17AHAD-''MAiR.T|U 17ARAD-''MAR.TU
LoE 13. 10^ GALU-me-'nSKVH' rev. 14. URU Pa-hu-nal^^'"] 15. 12 Me-er-ku 16. 15 Ma-sum '~^^^'Si-gi-y\a'\ 17. i-t-' GAL Su-^Na-mur 18. sa 19. ag'' Su -[oj 20. 21. 2 x X X
22. 2 me 1 su 17 Fig. 29. CbG 19 obv. obv. 5: For Imerum as a personal name, albeit not tbeopboric, see CAD I, 112. rev. 17-19: Tbe beginnings of tbese lines were lost in preservation but are still visible in a pbotograpb taken earlier and bave subsequently been added to tbe copy.
Fig. 30. CbG 19 rev. ChG 19 (figs. 29-30)
Fig. 31. CbG 20 obv., before join witb additional fragment on upper left.
obv. 2. \] 3. 17 Sil-lt-ya DUMU I-di 4. 19 '^VTV-se-mi sa URU.GIBIL 5. 15 ANSE-Iss-tdr DUMU Sa-ru-ru 7. 10 Is-gu-tim-lr-ra
ChC 20 (fig. 31) obv. i r . 2'. 3'. 4'.
[1 ] XraX [ |1 x-n |1 1
>K|I
KAMYAR ABDI AND GARY BECKMAN
56 5'. 1 [S[(7-/fi/o" 6'. 1 \Z\ii-ta-nim-zi
7'. [1] ^h^~ba'^~an-na-an-nu-um
9'. [1 [ X X [ 0 [ ^'"^Sa-x"
12'. [1[ h-ra-da-an 13'. 14'. 15'. 16'.
[1[ ^Ha-af-mi-ya [1[ x-7iu-us-ba-su [1[ f/.5'-m£-''lSKUR ^41 ''"^Zo-a/-//"''
17'. 18'. obv.ii 1'. 2'. 3'. 4'. 5'. 6'. 7'. 8'. 9'. 10'. 11'. 12'. 13'. 14'. 15'. rev. iv 1'. 2'. 3'. 4'. 5'. 6'. 7'. 8'. 9'. 10'. 11', 12'. 13'. 14'.
[1 [xNA.GADA (traces) [1[ ' h 1 Ba-AK-x [ 1 DINGIR-;;«-/n1 0 -x-.sV-BU-U[D Ha-at^-mi-y[a I ' [o-[s»-W(-)[ 1 G\i\-hnin-l\um] Ar^-ra-b\u-ttm\ A^o-\b/ma-s[i\ 1 Jj:i^-mi-\ 1 ([ 1 1 1 1 1 X ). x [ 1 K. \ 1 •i [ 1 U\s-1 Nu-\ 1 DINGIR1 1 DINGIH1 Ba-^za''^-[ 1 XTMU.MU[NUS 1 1<X 1 1
15'. 16'. 17'. 18'. rev. V r . 2'. 3'. 4'. 5'. 6'. 7'.
1 Sa-ma -\
1 ''G(/-/Ia' P A-bu-u\a-qar] [1[ X [ [1 -/[« [1[ ^Su^-ma-^-^a-bir^-um 11' Ib-bi-^ta'^-ti 1 Ri-i8-Ha^-HS]?A 1 Ha-am-mu-ra-pi 1 lk-ki-ya 1 ln-bi-'^\}'V\]
ft' 0.
9'. 10', 11', 12', 13'. 14'. 15'. 16', 17'. 18'. rev. vi 1'. 2'. 3'. 4'. 5'. 6', 7'. 8'. 9'. 10'. 11', 12'.
13'. 14'. 1', 2', 3'. Unplaced A r. 2'. LeE
1 DUMU.GAL?S/Tfl^-a/,/«-; 1 Ap-U-ya 1 I-li-ma-a-hu 1 DINGIR-/;/»ia--ra-fl?SIPA [4'-'"['-'rLa'"i-^a-u^a-a7]'^' [1 o-]x-we-de'e-via DUMU (traces) [ [x [ \-unt [ -mjii'-ur [1 j-m
[1
l-x-^«
[1 •.v/?[o-an-m [1 |-x-,
4'. (traces) 5'. (traces)
A GUNEIFORM ARGHIVE EROM GHOGHA GAVANEH, VVESTERN IRAN
B rt. co[. r, (traces) 2'. [1| 3'. l i 4'. 1) 5'. 1 ' 6'. [1] r . [1 8'. [1 left col. Traces of two lines obv. i 6': My reading of the town name here, which postulates metathesis of signs, is obviously verv uncertain. ChG 21 ohv.
1. \f\l-ta-ni 2. Da-mi-H 3. Vm-m 4. \V\m' 5. \A'h\ ChG 22
ohv.
rev.
1'. 2'. 3'. 4'. 5'. 6'.
(traces) "^9^^ ba di ha sa lum x di nu 1 M/-»i-to-s 1 Si-ma-tum 1 x-x~^kV-i[m'] 1[ [
7'.
1 '/-/H'-.S1»-M(/[
8'. 9'. 10'. 11'. 12'. LeE 13'.
1 ll-ta-n[i\ 1 Wa-qar-t\um\ 1 Be-ei-ta-^ni^ 1 Be-le-su-nu [1[ ^AHi-ni-HUSE.BA
6', [1[ x-x-x-mi 7'. [1 A-[/^((/-.s-n-flNA.GADA 9', [1[ x~^na-\-na-mi'X 10'. fN L'^'^'^^mu-sa-di-nw'i 11'. [^'^l^'/Cf-fZ-Za-an 12'. [1 Z\u'-uz-zu 13'. [1[ x-ffca-/ai-af 14'. 1 ^Ba'^-bi-tum ha-za-an 15'. 1 DINGIR-«-«-L-a5 16'. 4 ^-^''^"[^Yrfo-w'fl'-rfa-fl' 17'. 1 Na-aq-qu-um 18'. 1 Pa-ni-ya 19'. 1 La-^ka-ak'-kf^ 20'. IBc-'/ai-nu-fumi 2r. 1 Ab-nu-nu 22'. 1 In-bu-um 23'. 'U A-/?u-[ra-f/flr 24'. [llB»-f/a-/w[m[ 25'. [8^['^'^^^rH«-sa-rfi-n« 26'. ['-'^[^'Mi-ZI^' 27'. [ -t\u-x-[ col. ii illegible; note only ii 11': 7 The personal names here are summarized by category: ^^'^^musaddinu, "'(tax) collectors" (i 10', 25'), and ^^'^^^da-wa-da-ar-sa, "?" (i 16'). For the former, see CAD MAI, 252. The vestigial col. ii seems to have contained names of women. ChG 24 obv.
1. 1 Wu-tir-DINGIR 2. 3. II 4. 1 X X X 5. 1 X X 6. 1 X X X
ChG 23 col. i 1'. 2'. 3'. A'. 5'.
57
(traces) /r-i[s'-'5a^i-og x-x-^ar^-su x-us'''-sa-ru-ut LoE
7. 8. 9. 10. 11. 12. 13. 14.
1 [ i *Ka('-b J x-x-'su'-nw 1 \-x-Hu-nu'"^ 1 iA'-^[um| lx[ [ 1 ^A-li^-ni-\su'a/ya] 1 Ixi--in\a-sa/sa-tum]
58
KAMYAR ABDI AND GARY BECKMAN
15. [1| A-li-tum rev. uninscribed This piece is extremely worn and the persona[ names could he restored only through comparison with other lists, such as that in ChG 31. ChG 25 2'. [ 3'. [ 4'. [
Fig. 32. ChG 28 obv.
/VH-w[?-DINGIRiVw] x-na-bi-ta-U ] x-sa-i-lt
5'. I 6'. I
7'. (traces) other side uninscribed, as far as preserved
ChG 28 (fig. 32) obv.
ChG 26 obv.'^
r. [ 2'. 3'. 4'. 5'. 6'.
rev.'
r. [ 9'. 10'. 11'. 12'. 13'.
DUMU.MUNlUS--J.v,s-fdr ] Re-m-twn ]-e-si-mu-UD 'slt\a'-a.m-i-su-ur
[ [ [ [ [
] Ba-ni-tum ] GEME.30 ChG 27
r . (traces)
rev.
3'. 4'. 5'. 6'. 7'. 8'. 9'.
fO,O.l"'UTU-si-ma-x(-)[ 0,0.3 ^A^-x-b/ma--x \ 0,0.2 X X [ (traces) 5xda-[ 22 x [ (traces)
OJd.S A-ha-Tu-Tu X [ 0,0.2 n-ta-ni ii ARAD-[ o [-x 0,0.3 Wa-qar-be-(i 0,0.3 Ba-mat-'^A.A
rev. uninscribed ChG 29 obv.'''
] DUMUMUNUS A-ta-na^-a\h\ Sa-bi'twn ] Sti^^-pu-ur-tum ]X h
1. 2. 3. 4.
1'. '0,0.2' [ 2'. 0,0.2 [ 3'. 0,0.2 ^Wa^-qlar-tum]
5'. 0,0.2 x-hi-ti-t[um\ 6'. 0,0.2 A-li-ni-^m-ya^ T. 0,0.2 A'b'-twn LoE^ 8'. 0,0.2 DINGIR-foa-»i rev.- 9'. [0,0.2'r-kv-lsu-nu 10'. |0,0.2'[ ^A^liyni-su^-y{a] 11'. [0,0.2'[x-x-[Z[I' 12'. fO,0.2-' [ j-fa/T 13'. 0,0.2 x [ 14'. 0,0.2 [ 15'. '0,0.21 [ LeE 16'. [0,0.2-Ba-a\b-tum 0,0.2 Wa-*qar^-tum ChG 30 obv.
V. 2'. 3'. 4'. 5'.
[0,0.1 0,0,1 [
[-tum ]-tum
0,0.1 X [
[
0,0.1 \Wa -qary l)e^-(i 0,0.1 [A-l.I'y ni-su^-ya
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN 6'. 7'. 8'. 9'. 10'. rev. 11'. 12'. 13'. 14'. 15'. 16'. 17'. 18'. 19'. 20'. 21'. 22'. 23'. 24'.
0,0.1 Ha-m\a'-\tum 0,0.1 JM-ma-H[a\-tum 0,0.1 Ma-'Ca'-Aum 0,0.1 A-li-da-du-^ya^ [0,0.1[ x-ka-ha-[ (traces) [Oj)\ A-\li-a-[hi\ ^OS^.V La-ma-sa-t[um\ 0,0,1 Sa'ti-y\a\ 0,0.1 S[a[-al-mu-tum 0,0,1 0,0.1 0,0.1 ^ 0,0.1 Na-^ra^-am-tum 0,0.1 Mu-na-wHr 0,0.1 GEME-^'MAR.TU 0,0.1 GEME'.DINGIR [0,0.1 [ Ra-x (traces) ChG 31
obv.
r. 2'. 3'. 4'. 5 .
6'. 7'. 8'. 9'. 10'. rev. 12'. 13'. 14'. 15'. 16'.
[0,0.1 [ Wa-qar-t[um\ 0,0.1 "'TJTU^'-na-au-ir 0,0.1 Ta-a-bu-ni [0,0.1 A[-li-^a-hi^ 0 01 \Jy\j.l
l/\*'-i'Ml"l^ r\
rv|t I
fiivn'
0,0.1 nP-ta-ni 0,0.1 La-ma-sa-tum 0,0.1 /5-to-o 10,0.1] I-rsu-hu-nu [0,0.1 K[a-na-na-ya (traces) 0,0.1 A-U-da-du-ya 0,0.1 A-ha-Tu-tu 0,4.3 •ma^-on-di SE X ^mu' X ,3(r x nu
ChG 32 r. 2'. 3'. 4'.
5'. 0,0.2 Kal-bi-tum 6'. 0.0.2 l-la-,w-nu T. Ofi.2 Ka-fii-na 8'. 0,0.2 Sa-al-^mu^rev. uninscribed ChG 33 obv.
1'. 2'. 3'. 4'.
5'. 6'. 7'. EoE 8'. 9'. 10'.
(traces) 0,0.1 X [
0,0.2 Qi-^sa'-at^-x-x u^-um ^Ha'-nam-tum a- ^'f'^'^Gfail-li-kir] lSE.'GURo'-na'E' [ o o [ x fO,O.l'i ZI.DA Ta-7ia-[
[f>\um^-ma a-na KAL-x [ o o \-i-ti-ma '0,0.2-a-''-»[a[ SUKU Su-na-ak-ku X X a-na GU4!.A da-ri ANSE 'da-?/' •0,0,2' a-7ia ri-ik-si 0,0.2 SUKU E .sa lTl.l.KAM •0,0.1'"^ mu-ut-qu a-na Lu-sa-bu-um •20^1 SUTI Ls-j;if'-''lSKUR 10 a-na NINDA.HAR.RA
rev. 11'. 12'. 13'. 14'. 15'. 16'. 17'. (traces)
Ivitll
rev. 15: For "nian^/w-soldiers {to be provisioned barley," cf. ChG 5,18, and 34.
obv.
59
0,0.3-'Z/i-?[a-?ii[ 0,0.3 Ka-na-[na-y]a 0,0.3 Sa-^/-)/a Q,O.Q &i-7na-
This very vv'orn tablet records the disbursement of barley for various purposes, including provisions (SUKU-9', 12'), the fattening(?) of animals for offerings {darrum—\O'), and tbe production of sweet bread {mutqum—\Q>') and fine flour {samldu77\—\b'). Obv. 4' seems to indicate tbe occasion of an issue: "The day when Hanamtum [went(?)[ to (the town) Ga[...[" ChG 34 obv.
1. \0,0.2\Si-i-um-m\i[ 2. '0,0.2'' A-hi-gu-ul-lu-[ub[ 3. [0,0.2[ ^Be-cP-(u[m\ 4. '0,0.2•' 5. 0,0.2 [H\a-su-nu 6. 0,0.2'ri-/[a1-'.su 7. 0,0.2 'A'-Zja-a^'-' 8. 0,0.2 9. 0,0.2
60
rev.
KAMYAR ABDI AND GARY BECKMAN
10. 11. 13'. 14'. 15',
0,0.2 [0,0.2] (traces) '0,0.2^ ARAD-?a 0,0.2 iVu-ur-3|0[
ChG 37 obv.
16'. 2TUG.'HI.A' [ 17'. 9 ?na-an-rf[/§E[ rev. 1T: For restoration, cf. ChG 31:15'
1. 0,0.1 Ma-a-.sVf/MUNUS.'SIKlO 2. 0,0.1 re-u,^a-x-AN-x[ 3. 0,0.1 Bu-ne-tum Nn-ba-[tu7n[ 4. 0,0.1 Sii-kal-la-t[um'\ 5. [0,0.11 X I rev.rt. 6'. ^V-la-H[i'-7m] 7'. Ka-ra-n[a-twn[ 8', ll-ta-7ii rev.left 9'. 'SiNi-ya' 10'. (traces)
ChG 35 obv.
1'. 2'. 3'. 4'. 5'. 6'. 7'. 8'. 9'. 10'. LoE 11'. 12'. rev. 13'. 14'.
[0,0.21 llt[a-ni[ [0,0.2] 'DUMU.MUNUS-[7%'Mr[ Omn-[ta-ni[ [0,0.2/^[-'ma'-.sa-'fumi ^0,0.2' A-li-7u-su-ya '0,0.2' DUMU.MUNUS-Zsg-Mr' 0,0.2 Ha-si-nd0,0.2 Sa-bi-tum O.{)l^Tu^-ra-a '0,0.21 Sa-i|r''[-fr«i '0,0.2TKa'-'na-f(/mi '0,0.2' U7n-mi-wa-aq-7a-at 0,0.2 Ma-sa-a 0,0.1 Ba'-bi-ru
ChG 36 obv,
rev.
1'. 2'. 3'. 4'. 5'. 6'. 7'. 8', 9'. 10'. 11'. 12'. 13'. 14'.
[0,0.2] 0,0.2 Na-ra-am-tum 0,0.2 Wa-^qar^-tum 0,0.2 ^AUi-ni-^su^-y 0,0.2 x-x-s[i-(-) [0,0.2] 0,0.2 A-h'-da-du-ya (blank) (blank) "'Uq-tum ""^Kallxi {""[xxx-rum (blank) (blank)
"'Si-^bi^-ta-tum '"Da-di
'"Ma-as-tum
""Kat-bi-tum
rev. left: These few signs, seemingly erased, were missed in copying from the original, but are visible in tbe photograph and have been added later.
ChG 38 obv.
1. 0,0.1 Ma-.v|(/m[ 2. 0,4.0'EN.ZU-re-me-[m] 3. 0,1.4 ARADDINGIR 4. 0,0.1'Sr/-''MAR.|TU] 5. 0,0.1 A-/ja'r(/'-[fu] 6. 0,0.1 JM-7na-s[a-tum[ 7. 0,0.1 Mit'k[i8. 0,0.1 '5a''-[ 9. 0,0.1 X [ 10. 0,0.1 X [ 11. '0,0.1' [ rev., as far as preserved, uninscribcd
ChG 39 obv.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.
'0,0.3-' XXX KU.BABBAR [Pa'-[7ii-ija 0,0.2 X X KU.BABBAR 'A'-fc 0,0.2 '^Da-mi'-iq^'-tum '0,0,2' DUMU.MUNUS-7.ss-Mr [0,0.2[ Be-le-et-su-nu '0,0.2' n-ta-ni 0,0.2 DUMU'MUNUS^-Zsg-Mr '0,0.2'La-ma-.sa-^|«/n[ 0,0.2 x-ma-[ 0,0.2 Ma'-^kiH 0,0.2 A-/jrt-T|(W(/[ 0,0.2 Ka-7a-7}[a-tum\ [0,0.2[ Sa-bi-t[um[
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
rev.
14. 15. 16. 17. 18.
|0,0.21 Ba-ab-t[um] 0.0.2'Sa''-/a-'ma^i-f[(/m[ 10,0.2 K[a--na-
19.
36'. 37'. 38'. 39'. 40'. 41'.
/ UE
obv.
1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
0,0.2 'ZiZ' Te-ya-x[ 0,0.2 x'x-a'bu{-)\ '0,0.2" [L\a-ma-sa-twn 0,0.2 S[ 0,0.2 0,0.2 A-[/ta-aU-a/j-/ti 0,0.2La-]/a•-[^w»^ 0,0.2 /-/[a-a[.s-,s'»-nw 0,0.2 A-f/ia1-a-/ji 0,0.2 Wa-qar-tum
11. 5 x x 12. 13. 14. 15. 16, 17. 18. 19. 20. 21. LoE 22. rev. 23'. 24'. 25'. 26'. 27'. 28'. 29'. 30'. 31'.
na -hi- ra-sa
0,0.2 A-/i'-'a'-/u GAL 0,0.2 A-li't\im 0,0.2 Na-7-a-am-tu7n 0,0.2 Si-7na-tum 0,0.2 Wa-X X X X X 0,0.2 [ [ x-tum '0,0.2' Wa-qar-be-U [0,0.2] Mi-in-ta-sa-DlNGlR [0,0.2[ Me-f-KUR.NAM |0,0.2| A-/(-HI-/iu-x (traces) '0,0.2' [ 0,0.2 'Da'i-M»-[ 0,0.2 /.s-/a-a 0,0.2 U7n-7m-ta-ba-at 0,0.2 ^Ba'^-luf-m-tum 0,0.2 Ma-a[t-ta-tu[7n0,0.2 GEME-[''[MAR.'TU' 0,0.2 Ka'-r[a[-na-tum 0,0.2
32'. 33'. 0,0.2 H((-'^a'-/ 34'. 0,0.2 A-/;'-/(/-.s(/-;i-fl 35'. [0,0.2[ A-x-x-''EN.ZU x x x
61
[0,0,2] Be-[/]»-,sw-»» 0,0.2 Be-[el-[tum 0,0.2 X [ IX X X 0,0.2 Be-e[l-ta'-n\i' 0,0.2 Si-ma-a[f|-30 0,0.2 Dan-7m-a[t\-x-x
42'. 43'. 0,0.3 Nu-ur-rs-ha-r[a[ 44'. 0,0.3 Sa-mu-sa-x [ 45'. 1 UGULA '-
This text and ChG 47 both show small "check marks" next to each entry. The professional designations in lines 11 {^^'^^'hia-lji-ra-sa), 32' {^^'^^Hami-i/a'-" ""j, and 42' ("''^^^^BI-ZA-[) are unparalled elsewhere and therefore obscure. The words cannot be interpreted as personal names because of the appearance of the plural marker in line 32'. Note also that the MUNUS-sign only occasionally marks a woman's name in this archive (used with Hannamtum, Kinialluh, A-lia-at-ma-[, and Satiya— see Index I). ChG 41 obv.'^
1'. '0,0.2 A-^-'-1 2'. 0,0.2 l!-t[a-7ii] 3'. '0,0.2 Me^-' [ o ] x [
4'. 5'. 6'. 7'. 8'. 9'. rev.-' 10'. 11', 12'. 13'. 14'. 15'. 16'. 17'. 18'.
[0,0.2 -i\7n'-ri-ta 10,0.2 A']-li-tum 10,0.2 ''EN.ZU'-EjN^-a/j-//n( 10,0.2 [-an-^ki'^-rum 10,0.2 [0,0.2 [-''EN.ZU [0,0.2[ ^Be'^-[er-t\ [0,0.2[ SJ-b[i-i[r-sa '0,0.2'Sa-'/ci'-n/ '0,0.2' 0,0.2 0,0.2 Wa-qar'-be-U 0,0.2 Ba-hu'[f>]a-tum f0,0.2' Lf/-'?»a'-[,s[ [0,0.2] 'A'-/f/-/um'
62
KAMYAR ABDI AND GARY BECKMAN
ChG 42 obv.
1. 2. 3. 4. 5. 6. 7.
[0,0,3[ NIN- ba' ma as sa' [0,0.3[ Ka-na-na-a 0,0.3 lt-ta-7ii 0,0,3 fLa'-<ma->.sa'-f[um[ 0,0.3 Kal-^ba'-'^-t[um[ [0,0.3] x-fa--M/"'(-)[ (traces)
rev. 9'. 10'.
[-x-a
rev.-'
8'. 9'. 10'. 11'. 12'.
ChG 46 1'. 2'. 3'. 4'. 5'. other side
'0,0.2'^' [ f0,0.2' [ 0,0.2 ARAD-M[AR.TU 0,0.1 x[ 2x1
SU- X [ 53 pa SE X [ 29 pa SE X [ 50 X [ 17pa[SE lost
ChG 43 ob •v;
rev.
1'. 2'. 3'. 4'. 5'. 6'. 7'. 8'. 9'. 10'.
(traces) 0,0.2 Nu'ur-'^UTX] 0,0.2 Sa-an-[q]um [0,0.2] Nu-ur-Hs-ha-ra (blank) [0,0.2] ^A-iw}-at-\a[h-h [0,0.2 ]-'tum' [0,0.2 [-tum [0,0.2 Si-i-nm--[mi [0,0.2 [ X
ChG 44 (bulla)
1'. 2'. 3'. 4'. 5'. 6'. 7'. 8'.
,
Tbe appearance of tbe abbreviation pa for parisum (see CHD P, 186) is unexpected in an Old Babvlonian text. ChG 47 1'. 2'. 3'. 4'. 5'. 6'. 7'. other side
M
0,1.2 fSUKU' ANSE'.HI.'Ai [ 0,0.2'SUKU ANSElHI.Axx 0,0.1 SUKU A N S K ^ L A Ba-fo 0,3.2 a-7ia NICHURRA x [ 1,2.4 X [ 1,1.4 X [ 0,2.4 [
'0,0.21 L[a-^0,0.2' Nu-[ ro,O.2'A-[ 0,0.2 X [ 0,0.2 [ 0,0.2 [ 0,0.2 [ lost ChG 48
obv.
rev.
9'. X X [
1. 2. 3. 4. 5'.
0,0.1 Ta-a-bu-7i^[i\ 0,0.1''LAMA'-x-[ 0,0.1 Su-qd-tum 0,0.1 GEME-R'n-'/;/ '0,0.2^' [
7'. 0,0.1'' 7-fc/-''UTU 8'. O,O.\x^O-re-m[e-ni] ChG 45
obv.'
1'. 2'. 3'. 4'. 5'. 6'. 7'.
ChG 49 0,02. A-ha-a[t'ah-hi[ 0,1.2 SH 0,0.2 n-t[a-7u[ 0,0.2 Mi-q[ar-tu7n\ 0,0.2 \-la-\Hi-7iaku-nn\ 0,01 Be-[
obv."
rev.-
1. 2. 3. 4. 5'. 6'.
[ DUB1 SE.BA-)»i I [ x--ti-i77i [ ] X b/77ia a ga [ -s/t\a-za-l[i[ ]X [ I ]ar "
um
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
7', 1
lHi'-'\lT]V'-di-m
8', 1
T. 8'. 9'. 10'. 11',
I AN a-na .sa-b/ma-ri-i\h] ChG50
obv.
A-k\r-tum Da-mi-i\q \-He'guv-turn A-h\a-\r\u-tu
1. DUBSE,BAEARADG|EME^ ChG 54
2. 2'^A-ha-at-Ku-\bi]
3. [11 4a^-ti
obv.
rev. uninscribed ChG 5 1 (bulla) 1'. [ 2'. 3'. 4'. 5'. 6'. T. 8'. 9'. 10'.
1. 0,0.3 Ma-as-\t\um LUKUR^.DINGIR''' 2, 0,0.3 A- li-t[uiii W]a-qar-tum 3. [ 1 X [ K\a-na-na-hja^ 4. [ Ix rev. uninscribed
X
|xARAD-''M[AR.TUj \ Ma-sum
j Z-foF'UTU ] QO-im-gttr-an-ni a-n\a'mTVKDV A-!i-ni-.su-\a/ya] a-n|aNIG,HUR.RAa-;i|a ix'6^'xKUR[ |xx[
ChG 52 obv.
1. [ 2. [ 3. [ 4. [ 5. [ UE 6'. [ rev. uninscribed
ChG 55 obv.
1', . 2'. 0,0,2 Si-ilj-t[i-rii-(id-at\ 3', 0,0.2 [A-/fn]fsw-[«-a-] 4', 0,0.2 l!-ta-\ni] 5'. Ofi.2 ^AUi-t\um\ 6'. 0,0.2 ^Wa^-(j\ar-tum\ LoE r. 0,0.2 [ rev, 8', 0,0.2 Kal-bi-tum rest uninscribed Surface flaking badly,
[ X Mu-iia-di-na-t\um]
ChG 56
1-a obv; X X [
obv. h. We migbt also understand this line as presenting a title, musaddinatum, the feminine equivalent of mmaddinu in ChG 23 i 10' and 25',
V. [0,0.2| x-x-'ni-[ 2'. [0,0.2| '"^'^^'^'SUI 3'. [0,0,21 Ua-^at^-si-na . \)SJ2 b>i-'itj'-h-ru-qa-at 5', f0,0.2' (traces) 6', 0,0.2 Ka[-UH^-t\um\ T. 0,0.2 A-/i 8', 0.0.2 A-/i, 9'. '0,0.2 A-; rev, illegible traces
ChG 53
ChGFl
obv. 2'. X X X X X
rev.
63
\-rum'
3'. [ ] xxx-'/ci-i-m4'. 1/2 SILA M/-/«-^a-,va-DINGIR 5'. (traces) 6'. |xx[
col. i
1'. 2', 3'. 4'. 5'.
-/J1'/--X-|
I x-d[u-I K-ga-mil A-p]il'ku-bi iVi/-(H r-TISPAK
64
KAMYAR ABDI AND GARY BECKMAN
6'. [ 1 x-Z\-a-bu T. [ 1 x-sa-am-mi-e
ChG F6
r. [1 Nu'-u\r
8'. [ \y 9'. [ /b--]n/-''MAR.TU 10'. (traces) col. ii traces of a few lines beginning witb DIS
2'. 1 iia-/7U-(v^''IS[KUR| 4'. liV[a-/?]/-''UT[U] 5'. (traces) ChG F7
ChG F2 1', 2'. 3'. 4'. 5'. 6'.
[ [ [ [ [
\-lu-ub/p-r\i{-) sa nu'' aV mu [.E,A [ -Ha\-am-mi-e ARAD x x x kax\
r. 2'. 3'. 4', 5'. 6'. r.
ChG F3
ChGF8
col.
2'. 3'.
X SI X o X X
4'. 5'. 6'.
]-''''lSKUR X X -Ua^'-rum X Hu-zi-rum X AN,DAH-/a-n(m
7'. O.
L U .IVIICJ (_«V
9'. [
col. i—(traces) col.ii r . '20^'X [ 2'. 1 A-lut-w[a--(iai-] 3', 25 TERIN'.MES-"' [ 4'. 1 A-hu-wa-q\ar] 5', 1 Ku-iH 6', l x [
-Ul-'Ul
ChGF9
]-/w
10'. [ -/[/ARADE.GAL 11'. [ l-x-SI 12'. l-rf/^i col. ii 13 lines in whicb only DIS is preserved
ChGF4 (traces) 2'. 3'. 4'. 5'. 6',
! I I ! (traces)
1 A-it'[i-(7'1 l'ra-h\a-lAf-/«-[ 1 A-hu-ii\m1 M?"-/?i-.vu-[DINGIR] [1 N]u-n-\}r] [1] X [
1'. 2'. 3', 4', 5'. 6',
[l|'A^-'[ 1 Ba-ni-\tum] l'^EN.ZV-e-\r 1 AN BT [ 1 ^A-nu-Ut 1 X -at-x \
r. 'P [Ni»-'" » ChG FIO
| x - i-is-sa-a\s.\ X X -
ChGF5 Now recognized as unplaced fragment of CbC. 20.
Ieftcol.1'. [ 2', [ 3'. 1-''''ISKUR 4'. \-tum 5'. 6'. \-x-ar T. Ix rt, col, 5 lines in which only DIS is preserved
A CUNEIFORM ARCHIVE EROM CHOGHA GAVANEH, WESTERN IRAN ChCFll obv,'
rev/
1'. I 3iO,
6',
sa
I a~x I
(traces) ChG FIB
obv,'' r. (traces) 2'. [ Z-(/|a'''-ma-ra-as rev.'
3' I
1 roc;!'l
4'. \
]-x-um
ChGFI2 ChG F19
r . (trace.s) 2', 0,0.3 &-W-.V[»-MM]
rev.-
3', 0,0.3 f''EN.ZU'i-i[m'-^ 4'. 0,0.3''A;>/[i-(/o) ChGF13
UE^
1'. 2' 3'. 4'. 5'. 6'.
[1[ (traces) ^1 Ru^-^x I 1 UGULA ARAD.[MES/GF.ME 1 l-la-[f>i-na/su-nu\ 9 fA-ix [ lOM
1'. [ IxBU(/afo/ZUx 2'. I \x ZI sa X [ 3'. (traces)
ChG F20
ChG F U I AN PA TE
1'- [ 2'. [
rev. ChG F15
1'. 2'. 3'. 4'. 5', 6'. r.
[ [ [ [ [ [ j
r. 2'. 3', 4', 5'.
-a\d' X \ ERIN' [X \-tum \ ^
\ x-af-kam'' 1 GEME-MAR.TU ChGF21
x-am-ba'-tum
1'. [ 3'. [ 4'. [ 5'.
(traces)
Ixx'a'/;
2;. [
\-x-du-pa'-a\yi-am{-)\ (traces)
ChGFl 6 obv.
1. 2. 3. 4'.
Mu-na-w\i-ir-tum D\a-mi-iq-tum
ChGFl 7 obv,' 1'. LoE 2'. S'. 4'.
|xx[ 1 lM-ma-s{a-fum\ I n-ta-n{i] (shallow traces, added later?)
ChG F22 1'. 2'. 3', 4'. 5'.
0,0,1 XX [ 0,0.1 A-ha-\Tu-tu'\ 0,0.1 30-|im-gu;-a?i-H/1 0,0.1 n-\ta-ni]
ChG F23
r, 2'.
65
66
KAMYAR ABDI AND GARY BECKMAN
ChG F24
r . [ 1 x-um-[ 2'. rp \M\i-nu-u\m3'. n^ Na-bi-\-\li-su\A'. ' I ' x [ 5'. •8
ChG F27
obv,^ r. 2',
col. i^ 1'. 2', 1 x-su 3'. x-x-ya 4'. 1 [ col. ii' 5', '6^ [ 6', sa a'-x 1
]x
3'. 4'.
rev.^ ChG F25
X
X'la-tum
5'.
]x
6'. 7'. 8'. 9'.
10'.
-la 11)
-lam X s/la-qa-KAM x
U[
U ChG F28
r.
1 Be-la-nu- um\
2', 0,0,3'i A'pil-Ku'bi ', X
ChG F26 obv,
1'. 1 X [ 2'. l x [
3'. l ' [ rev.
4'. 1 DUMU.MUNUS-[J%-ta'r] 5'. K'^KlSi.ZV-e-ri-\ba-atn\
A CUNEIEORM ARCHIVE EROM CHOGHA GAVANEH, WESTERN IRAN
67
r cm 0 •
Fig. 33. Cylinder seal and modern impression.
Fig. 34. Drawing of cylinder seal impression by Ryan Burkhalter, Cylinder Seal Se-mi-tum DUMUMUNUS Nu-ri-ri GEME ,50 ^ISKUR
Semitum daughter of Nuriri servant-girl of Adad
I have been able to study this seal only in photograph. No particulars as to its dimensions or material, or concerning its particular findspot within Room B15, are available. One might even question whether it belongs with the archive presented bere. Note, however, that the proper name Semitum is attested on one of the tablets, and tbat Nuriri may also be present—see Index I. Tbe iconography of the seal is very simple: A female figure does obeisance to a god who bas
placed one foot on a stool. No elements besides these two personages and the sea! legend are present, Cf. CoUon (1987: No. 166). The legend informs us that Semitum was a devotee of Adad, and the lightning-bolt held by the primary figure accordingly identifies him as the Storm-god. Since the head of the worshipper is too worn to reveal whether sbe is graced with horns, it is not clear whether this figure represents an intercessory deity, as so often on seals of tbis period, or the seal owner herself. An interesting feature of tbe text is the use of the SA-sign witb the syllabic value /Sa/, known only sporadically from Mesopotamian sources, but common in those from Elam—.see von Soden and R6llig,AuOr42,No,224.
68
KAMYAR ABDI AND GARY BECKMAN Indices I. Personal Names (* = Amorite; * = uncertain etbnicitv)
*A--ba'-an-na-an-an-nu-um: 20 i 7' ^Abnunu Ab'tiu-nu: 23:21' Abu-waqar A-bu-iva-qar: 20 iv 17'; 23:23'; 48:6'(?) Aham-arsi Ahat-ahhi "A'ha-at-ah-hi: 34:7(?); 40:6; 43:6'; 45:2'; 56:8' Ahat-Kubi A-ha-af-K\u-bi\: 50:2 Aha-Tutu A-/ia-7«-^u: 21:5; 28:1; 31:14'; 38:5; 39:11; 53:10'; F22:2'(?) Ahi-guUub A-hi-gii-ul-lu-\ub]: 34:2 Ahi-sagis \-hi-sa-gi-is: 20B:5'; FlO:r(?) Ahu-waqar A-hu-wa-qar: 20 vi 13'; F8:2', 4' Ai-ahi A-(/a-/i':36:12'(?) A-ij\a''-]a-hi: 40:9(?) Akitum A-ki-fum: 24:11(?); 31:5'(?); 39:2; 41:18'(?); Ali-ahi A-/fa-/ii: 30:12'; 31:4; 40:12 Ali-dacluva A-li-cla-dii-ya: 30:9'; 31:13'; 36:7' Ali-nisuva A-li-m-su-ya: 22:12'; 29:6', 10'; 30:5'; 33:2'; 35:5'; 36:4' A-U-ni-su-il-a: 40:34'; 56:9' A-li-ni-su-a: 23:7' A-Jf-ni-HU-l. 51:6' A-li-ni- : 24:13 \A-li-n\i-su-\\ 55:3' Alitum A-li-fum: 24:15; 40:13; 41:5'(?); 54:2; 55:5'; 56:7' A-li-tum: 29:7' Amat-Amurrum GEME-'^MAR.TU: 30:21'; 40:29' GEME-MARTU: F20:7'
Amat-ilim GEME'DINGIR: 30:22' Amat-Kubi GFME'Ku-bi: 48:4 A mat-Sin GEME.30: 26:13' Apil-Kfibi A-pit-ku-bi.H: l(?);20vi 14'(?); Fi:4'; F28:2' Apliya Anabum Ar-ra-b[u-urn\: 20 ii 8' Awat-Kobi KA-Ku-bi: 19:8 Awil-ili A-wi-
9f>•^9'• F9-9'
B^lanum Be-la-tm-um: 2S:20';F2H:l' Belessunu Be-le-ef-su-nu: 39:5 Be-(e-su-nu: 22:11' BelSunu Be-el-su-mv. 5:4; 6:4; 6-10, sealing; 9:4; F12:2' Beltani Sf'-W-m-Ni: 22:10'; 34:8; 40:39'(?) Beitum Be-el-fum: 34:3; 40:37'; 41:10'{?) Belussunu Bi'-lu-su-nu: 40-M' Budalum Bn-da-lum: 23:24' Buiietum Bu-ne-tum: 37:3
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN Dadi Da-di: 36:8' Damiqtum Da-mi-iq-tum: 21:2; 39;3(?); 53:8'(?); F16:2 Daqtum Dd'tuj-fum: 6-10, sealing Gallabtum('r') "^'^'^^SU.I: 56:2' Gim ilium G[/]-m/7-/[»m]:2()ii7' Habil-kinum Ha-bil-ki-nu'um: 20B:3' '•''Hamat{t)um
Ha-m\a-\ tum. S0:6' (cf. Gelb 1980, No. 2115) 'Hammurapi Ha-am-mu-ra-pi: 20 v 5' [see Streck 2000: 92-93) Hannamtum ^Ha^-nam'tum: 33:4' •Haradi Ha-r\a-\di: 20 vi 6' {see CAD U, 88a) Hat miya Ha-at-mi-ya: 20 i 13', ii l'(?), 5'(?) Hu-za-la-tum: 30:16'; 40:33' (see Streck 2000: 311) Huzarattim Hifza-ra-tum: 28:5 Huziium Hti-zi-rmn: F3:6' Ibbi-Sama.s ' : 33:16' : 48:7'; 51:4' Ibbitati Jh-bi-fa-ti: 20 v 3' Ibbitum lb-hi-t\um]: 20 vi 12' Ibni-Amurrum Ibni-ili Ib-ni-hh'\ 25:3' Iddin-Lahma I-dm-La'ali-m\ah
vi 11'
1-dv. 18:5; 19:3 (Gelb 1980, No. 2432)
69
Ikkiya Ik-kf-ya: 20 v 6' Ilassina I-la-at'Si-na: 56:3' I-la-si-na: 35:7'; Ilassunu hla-as-su-nu: 40:8 Ha-su-nu: 22:7'; 32:6'; 34:5,6(?) hla-HU-nu: 29:9'(?) Ui-hatu Ui-lm-tw. 17:19 Ili-ma-ahu I-li'-ma-a-hu: 20 v U' Iltani n-ta-ni: 21:1; 22:8'; 28:2; 29:4'(?); 31:6'; 32:1'; 35:1', 3'; 37:8'; 39:6,17; 41:2'; 42:3; 45:4'; 55:4'; F16:4';F17:3';F22:4';F23:2' Ilum-ay-abbas DINGIR-fl-a-Zw-as: 23:15'; Fl:8'(?) Ilum-bSni niNGIR-/w-Hi: 20 ii 3'; 29:8' Imer-lstar ANSE-lSs-tdr: 19:5 (DUMU Zaruru) 'Inbi-Samas /fr-/?r-''UTU: 20 v 7' (see Streck 2000: §3.53) Inbum hi-bu-um: 23:22' Inib-sarri l-ni-ib-sar-ri: 6-10, sealing Irra-dan Ir-ra-da-an: 20 i 12' Ispim-Irra Is-gii-um-lr-ra: 19:7 Isme-Adad /s-me-^ISKUR: 19:13; 20 i 15'; 33:8', 14' •ista /,S-to-a: 31:8; 34:9{!); 40:25' Istar-imdl Isi^-tdr-im-di: 26:10' Tsuhunu i-.V(Wi)/-/((/:31:9' Itti-Samas-dini [lHi-'^UT\V'^-di-ni: 49:7' Kalbatum Kal-ba-tum, 36:11'; 40:5{?); 42:5(!P)
70
KAMYAR ABDI AND GARY BECKMAN
Kalbitum Kal-hHum: 24:8(?); 32:5'; 34:10; 36:14'; 55:8'; 56:6' Kalmastum Kal-mas-tum: 40:31' *Kananaya Ka-na-ua-ya: 30:17'; 31:10; 32:2'; 54:3 (see Streck 2000: 352-53) 'Karanatum Ka-ra-na-^wm, 35:11(!); 37:7'; 39:12; 40:30' (Gelb 1980, No. 3994) *Kasina Ka-si-na: 32:7' ^^KHii-al-al-lu-uh: 4:5' Kin-ilum K!-m-DINGIR:20vi 10' Kubburtum Ku-bifur-tum: 26:9' Kubburum Ku-ub-biirunv. 1:5 "Lakakki La-^ka-ak'-kih 2S:\9' 'Lalatum lM-\la-]tumA0:7 (Gelb 1980, No. 4282) Lamassatum JM-ma-m-fum: 30:7', 13'; 42:4(!) La-ma-sa-tunv. 31:7'; 35:4'; 36:1'; 38:6; 39:8; 40:3; 41:17'; F17:2' La-ma'[: 24:14 Lusabum Lu-sa-bu-um\ 33:13' Man-l5-ili Ma-an-ia-i-li: 20 i 2' Marat-Btar DUMUMUNUS-Zsg-^ar: 20 iv ll'(?); 26:1'(?), ll';35:2',6';39:4, 7;F26:4'(?) 'MasSya Ma-Sfl-a: 35:13' (see Streck 2000: 352-53)
Mastum Ma-as-tum: 36:13'; 54:1 (LUKUR^) Masum Ma-mm: 19:16; 38:1; 51:3'
"Mattatum Ma-at-ta-tum: 36:10'; 40:28'(?) Ma-ta-tum. S0:8';S\:U' (Gelb 1980, No. 314) ^Me-e-KUR-NAM: 40:2 "Merku Me-er-ku: 19:15 Minsu-ilu MHn-s«-[DINGIR}:F7:5' Minta-sa-ilim MHn-ta-m-D\NGlR: 22:3'; 40:19; 53:4' Munawwir Mu-na-wi-ir. 30:20' Munawwirtum Mu-ua-ivi-ir-tunv. 41:14'; F16:l Musadinatum(?) Mu-sa-di-na-t\uin]: 52:1 Nabi-ilisu Na4)i-)-\lt~m]:F24:3' Nabi-Samas Naqqum Na-aq-qu-uni: 23:17' Naramtum Na-ra-am-tum: 30:19'; 36:2', 6'; 40:14 Nubatum Nu-ba-tum: 34:4; 37:3 Nur-ili N»-«;-DINGIR:24:l;25:2'(?) Nuriri Nu-ri-ri: F7:6'(?); cylinder seal Nur-Lshara Nu-ur-U-ha-ra: 40:43'; 43:4' Nur-Sin /Vi(-i/r-30: 34:15' Nur-Samas A^u-ur-''UTU: 40:43'; 43:2'; F6:l'(?) Nor-samsi Nu-iir-D'W-si: 25:2' Nur-Tispak [/Vw-«]r-TISPAK:Fl:5' Paniya Fa-m-ya: 23:18'; 39:1(?) Pi-Kubi YJK'-Ku-bi: 19:8 Puzriya(?) PU.SA-i/a: 17:3
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
Rabut-Adad :2'(?); F6:2' Resntum Re-su-tum: 26:2' Ristati 'Salamatum Sa''-la-rna-tum: 39:15 (Gelb 1980, No. 5410) Sanqum Sa-an-\q\wn: 4li:S' Sin-bel-aplim [''EN.ZU'-E]N-ap-;im:41:6' Sin-eri bam ^EN.ZU-e-n-foa-am: 23:8'; F9:3'(?); F26:5' Sin-iddinam ''EN,ZU-i-e/(M-/iam: 3:5'; 20 vi 7'; 41:8' Sin-imguranni '^EKZV'im-gur-an-ni: 9:2; 10:2;F12:3'(?) SO-im-gur-ra-an-ni: 6:2 ^Q-im-gur-an-ni: 7:2, 8:2; 51:5' 30-[: F22:3'(?) \-an-ni:F2A'{?) Sin-iqisam '^ENZV-i-qi'Sa'am: 2:3 Sin-ma ''EN,ZU-ma: 20B:2' Sin-piya 30-pi-yo: 19:1 Sin-remeni •^ENZV-re-me-ni: 38:2 30-/e'-m[e-m]:48:8' Sin-sar ''EN.ZU-sar 3:8' Sukallatum Sii-kal-la'-t[um%S7-A 'Supurutum Sii'-pti-ru-fum: 26:9' (cf.Gelh 1980, No. 5742) Suqatum Su-qd-tum: 48:3 Sabitum S'a-/«-f»m: 26:8'; 35:8'; 39:13 Saruru Sa-ru-rti: 19:5 {see Gelb c/a/.1943: 175,324) Sihti-ruqat ^SHh-ti-ru-qd-at: 55:2'; 56:4'
71
Silli-nisu S(7-/('-n)-sw:20ill' Silii(ya) Siki: 18:5 { Idi) SI'HM/O: 19:3 (DUMU Idi); 20 i 5; 37:9'; F14:2' Suddurum Su-du^-rum: 20 vi 9' Sakiru .Sa-/c/-r«: 41:13' Salmutum Sa-al-mu-tum: 30:15'; 32:8'(!) Samas-kurub '^\]i:\}-ku-ru-ub: F6:3' Samas-nawir Sama.s-semi ''UTU-iV»i(: 19:4,6 Satiya Sa-fi-i/a, 30:13'; 32:3'; 50:3 Semitum Si-mi-i-tum: 41:13' §e-mi-tum: cylinder seal (DUMU.MUNUS Nuriri) Seriktum Se-ri-ik-tum: 26:5'; 40:4 Sibirssa S(-/ji-/?-,sa: 41:11' Sibitatum Si-bv-ta-tum: 36:8' Si mat-Sin Si-ma-at-ZO: 32:4'; 40:40' Simatum Si-ma-tum: 22:4; 24:2; 40:15 Si-umml Si-i-um-mi: 34:1; 43:9'(?); F15:4' Su-Amurium S'w-''MAR.TU: 8:4; 38:4 SK-MARTU: 7:4
Su-ma-abum Sti-ina^''-a-bu-nm: 20 v 2'
Su-Mama Summa-ilum S«m-ma-DINGIR: 20 v 14' Sumu-lisi Su-mu-li-si: 2:1 Sunakku Su-na-ak-ku: 33:9'
72
KAMYAR ABDI AND GARY BECKMAN
Su-Namar Su-'^Na-mar. 19:17 Ta-ra-am-Is^-tdr: 30:18' Tu rah turn Tu-ra-ah'tum: 35:9' Tabiini Ta-a-k(-m: 31:3'; 48:1 Ummi-damqat (7m-mi-SIG5:21:3;F21:2'(?) Ummi-tabat Um-mi-ia-ba-af: 40:26' f/m-mi-DUio-af: 36:11' Ummi-waqrat Um-mi-ua-aq-ra-af: 21:4; 35:12'; 39:18 Uqtiim Uq-fU7n: 36:10' Waqar-beli Wa-qar-be-li: 28:3; 30:4'(?); 34:11(?); 40:18; 41:15' V^a-qar-b[e-J\i: 40:\6
Waqartum V^a-qar-tum: 22:9', 15'; 29:3', 16'; 31:1'; 32:3'; 36:3'; 40:10; 45:5'(!); 54:2; 55:6'; F23;l'(?) Warad-A m u n u m ARAD-''MAR.TU: 19:9, 10, 11, 12; 45:10'; 51:2' Waiad-ili AMD.DINGIR: 38:3 ''\Z\u--tQ-nim-zi: 20 i 6' Zu'-uz-zu. : 182,279) : 20 v 12' (SIPA) DUMUGAL-''S/rfl-a/j-/a-/w;n: 20 v 9'
A-hu-u\m': F7:4' A-/VHI-M/-X: 40:21
AHi-i!i-m-a\g-: F4:3' [''-'': F9:5' Ar'-i: F7:6' A-ta-na:-a\b'- 26:7' At-ta-\: F7:3' A-u:{i-ir-:Fl:V A-\-blma-\: 27:4' A-[o-|/?/ma-s[il:20ii9' A-x-x-'iEN.ZU: 40:35' \A\m--mu-x\: 1:1 Ba-AK^x[:20ii2' Ba-za'-\: 20 iv 10' Be-\: 45:7' Dan-na-at-x-x: 40:41' Da'-)(i(-|: 40:24'
hi-ba-x: 1:18; 2:5 I-ra-h{a-. F7:2' l-sar-'^xl 1:3 Ku-u-\: FS:5' lM'-mi-[: 20 ii 10' Ma'~dM-\: 39:10 Mil-k\i-: 38:7 Mi-nu'ulf»'-- F24:2' Nu-\: 47:2' Nu-ur-%F9:7' Qi-ki-af-x-x: 33:3' Ra-x: 30:23' m-i-su-a-x{-)[: 27:2' Sa-mu-sa-x{-)\: 40:44'
Sf-he---x-l55:r Fragmenfary Here and in the other lists of damaged proper nouns most very mutilated writings where only a common sign or two can be recognized (e.g,, \-tum) have been excluded. A-/;r-[; 41:1' A~da-[ o~\x-ija: 23:5' Arf-KA-x-/v-i': 23:19' A-f/a'-x-[o-lI/O: 23:5'
^A-ha-af-ma-\\ 50:2
'Sa-x\: 11:3 Sa-ma'-l 20 iv 15' Se-pu-u-\o-\x: 20 vi 8' Si-[: 45:3' ra-na-|: 33:6' Ta^-x-^[f'-: 36:5' Te-ija-\\: 40:1 Ta-ab'x-bi: 20B:6' Wa-f/[ar-: 45:5'(!) ARAD-fa-[: 34:14' ARAD-lo-]x: 28:2
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN •'LAMA-x[:48:2 ''EN-ZU-x: 5:2 ''UTU-5i-ma-x(-)i:27:3' Acephalic -«-/«(: 26:5'; 40:2; 42:6(?) -x-am-blma'-tum: F15:3' x-AN-DAH-/a-rum: F3:7' \-an-ki''-rum: 41:7' •a\r-du-gu-\-H\u\: F3:2' Yx-x-ar-hv. 23:3' \-b(i-tum: 53:1'
\-x-ga-mil: Fl:3' -g\i-tja: 25:6' x-hi-fi-fum: 29:5' \-x-ik-ku-fa-nu: 20 i 8' -i\m''-ri-ta: 41:4' x-ka-lja-\: 30:10' \'X'ka-la-at\ 23:13' x-x-ki-i\nr\: 22:5' x-x-x-kr-ifi: 53:3' -KUi-bi: FIO:2' x-x-la^-rum: F3:5' |x-//"/,s-.s«-O(': 23:2'(?)- F^
-m\u^-ur: 20 v 18' \x-na-hi-ta-ti: 25:4' x-x-na-mu: 20 i 3' x-na-x-na-mi-x: 23:9' x-mt-us-ba-su: 20 i 14' x-ra-Ku'bi: 20B:4' X~7'(2~titiH',
r^t.O
\x-um-mi: F3:r ]x-n.s^-.sa-?-»-«/^: 23:4' \-x-we-de-e-ma: 20 v 14' (DUMU Summa-ilum) ]-x-ZI-a-fou:Fl:6' lx-x-ZI-fa:20vi5' x-x-SH'^-wr: 39:19 1-''EN.ZU: 41:9' 1-''ISKUR:F3:4';F1O:3' II. Geographic Names Agade A-ga-de: 18:4 Almati ^"*^'A;-mfl-M:'":4:ll' Amurrum A-mu-u\r-ri\:
18:1
Atusarri A-fu-sV-n'": 18:6 Bari ''^'Ba-ri"^'-. 20 i 7' Der ': 18:2 Elam ^: 17:6 Gulu Haburatu "''^''Ha-bu-ra-t[u^'']: hlO (SeeWafler2001:82) Idamaraz \l-d\a''-ma-m-az: F18:2' Kidan
V- ^/7-/l/J' -fit" • SV "il
Kellan
x-.s'a-i-//': 25:5' Off
Ci tltil
I ou
U/.
an: 23:11' l: 19:18
^yJit
-,sV^ja-aM-Hi: 20 vi 3 '
\x-sa-tum: 52:3 |x-se-BU-U|D(-):20ii4' lx-,S7-h/:20vi4' \x-si-ya: 20 i 4' A A oM I til.
^'±,\j^
\-tu-ublp-ri{-)\:F2:V
73
Lazawan Me-Turan ^''^'Me-fw-ra'-a'-^i'i: 20 i 6' IVRIi'Mi-ZI*^': 23:26' Nakabum ^'^^/Vfl-i
74
KAMYAR ABDI AND GARY BECKMAN
Nakitta ^'"^'A^[rt--fc|r-iY-/[a^'''], 1:9 Natima "'^^Na-ti-ma'^': 19:6 Nikkum ''^'''Ni-ik-k\um]: 19:2 (see RIA 9, 569f,) Pahuna ^"^^Pa-/iU-W*^''i: 19:14 Palum ^•"^'F«--/um^': 5:5 ^'^^'Pa-Zt'": 4:9' Sigiya ^"'^Si-gi-(/a: 19:16; 25:6'(?) Sabanugi
III. Deities in Theophoric Names Adad: see Isme-Adad, Libbi-Adad, Rabut-Adad; see also legend of cylinder seal Amurrum: see Amat-Amurrum, Ibni-Amurnim, Su-Amunum, Warad-Amurrum Ea: see Bamat-Ea Gula: see '^Gu-(\aIrra: see Irra-dan, Isguni-Irra Isliara: see Nur-Ishara Istar: see Imer-Istar, Istar-imdi, MSrat-Istar, Taram-Istar Kubum: see Amat-Kfibi, Apil-Kubi, Pi-Kubi Lahma: see Iddin-Lahma Lamassum: see ''LAMA-x[ Mama: see Su-Mama Namar: see Su-Namar NE.MUS: see ]-^''NE^MUS'^ Sin: see Amat-Sin, Nur-Sin, Sin-piya, Sin-eribam, Sin-iddinam, Sin-imguranni, Stn-iqisam, Sinma, Sin-rem?ni, Sin-sar, Simat-Sin, ''EN.ZU-x[, A-x-x-''EN.ZU S/Tahlarum: see DUMU,GAL-S/Tahlarum Sama.s: see Ibbi-Samas, Itti-Samas-dini, NabiSamas, Nur-Samas, Samas-kurub, Samas-nawir, Samas-semi, '^UTV-si-ma-x[ Tispak: see Nur-Tispak Tutu: see Aha-Tutu
Zalli Za-al-li : 20 i 16' URUGIBIL: 19:4 Fragmenfary '"'''G\a'-: 33:4' ^•'^^'KH'-I:
19:20
"-'"^'Sa-x*^': 20 i 9' L'R'.'C^.^f xi'_nllK|l, 9C1 ; Q' JU C/[.5 (.)[ . ilU I O
^^•'^^'Sfr-lol'^': 19:19 References Abdi, K. 1999a Archaeological Research in Irans Islamabad Plain. The Journal of the International Institute of the Lhiiversity of Michigan 111: 8-11, 1999b Archaeological Research in the Islamabad Plain, Central Western Ziigros Mountains: Preliminary Results of the First Season, Summer 1998. Iran 37: 33-43. 2000 Islamabad 1999. ban 38: 162-63. 2001a Archaeological Research in the Islamabad Plain: Report on the First and Second Seasons. Iranian journal of Archaeology and Ilistorij 26-27:47-61 (Persian), 2001b Islamahad 2000. Iran 39: 299-300. 2002 Strategies of Herding: Pastorallsm in the Middle Chalcolithic Period of the West Cential Za^ros Mountains. Ph.D. dissertation. University of Michigan.
2003 The Early Development of Pastoralism in the Central Zagros Mountains, journal of World Prehistory \71A: 395-448. forth- Excavations at Operation WKiS ai Chogha coming Gavaneh. Report on tlic First and Second Seasons. 1998-1999. Ahdi, K.; Biglari, F; and Heydari, S. 2002 Islamabad Project 2001: Test Excavations at Wezmeh Cave. Archdologische Mitteilungen auslran und Turan'34: 171-94. .Abdi. K.; .\zadi. A.; Biglari, F; Famiani, D.; Heydari, S.; Nokandeh, C; and Ma.shkour, M. 2002 Tuwah Khoshkeh: A Middle Chalcolithic Pastoralist Campsite in the Islamabad Plain. Iran 40: 43-74. Baqir, T, and Mustafa, M. A. 1945 Irac| Covernment Sounding at Der. Sumer 1: 37-54. (Arabic)
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
75
Battini-Villard, L. Jones, J. F. 1989 L'espace domestique en M^sopotamie de la 1857 Narrative of a Journey to the Frontier of llle dijnastie d'Ur a ri^po(iue paleobabijTurkey and Persia, Through a Part of Kurlonlennc, BAR International Series 767. distan. Selections from the Records of the Oxford: BAR Romixiy GovemmentNS 43: 137-213. Biglari, F, and Abdi, K. Kordevani, M. 1999 Paleolithic Artifacts from Cham-e Souran, 1971 Excavations at Chogha Cavaneh. Bastan the Islamahad Plain, Central Western Zagros Chenassi va Honar-e Iran 7-8: 36-71. Mountains, Iran. Archaologische Mitteilutigen (Persian) aus Iran und Turan 31: 1-8. Levine, L. D. Braidwood. R. J. 1973 Ceographical Studies in the Neo-Assyrian 1961 The Iranian Prehistoric Project, 1959-1969. Zagros, 1, han 11: 1-27. Iranica A)iti(f\ta I: 3-7. 1974 Ceographical Studies in the Neo-Assyrian Cohen, M. E. Zagros, II. Iran 12: 99-124. 1993 The Cultic Calendars of the Ancient Near Mashkour, M., and Abdi, K. East. Bethesda: CDL. 2002 Archaeozoology and the Question of PastorCol Ion, a alist Campsites: Tbe Ca.se of Tuwah Khoshkeh. 1987 First Impressions: Cijlinder Seals in the Pp. 221-27 In Archaeozoology of the Near Ancient Near East Chicago: University of East, ed. V H. Buitenhuis et al. Croningen: Chicago Press. ARC: Publications. Eidem, J. Muhamed, .'\. K. 1992 The Shenish&ra Archives 2: The Adminis1992 Old Babylonian Cuneiform Texts from the trative Texts. Copenhagen: Royal Danish Hamrin Basin, Fdubha I. London: Nabu Academy of Science and Letters. Publications. Eidem, J., and L^ssoe, J. Postgate, J. N. 2001 The Shenishdra Archives 1: The Letters. 1979 The Historical Ceography of the Hamrin Copenhagen; Royal Danish Academy of Basin. Smner 35: 594-91. Science and Letters. Parpola, S., and Porter, M., eds. Frayne, D. R. 2001 The Helsinki Atlas of the Near East in the 1992 The Early Dynastic List of Geographical Neo-Assyrian Period. Helsinki: The NeoNames. New Haven: American Oriental Assyrian Text Corpus Project. Society. Saporetti, C. 1997 On the Lwation of Simurrtim. Pp. 243-69 in 2002 IM rivale di Bahilonia: Storia di Ehinnna ai Crossing, Boundaries and Linking Horizons: tempi di IJamiiiurapi. Rome: Newton & Studies in Honor of Michael C. Astour, ed. Compton Editori. C. D. Young et al. Bethesda: CDL Press. Schmidt, E. F. Casche, H. 1940 Flight Oisr Ancient Cities of Iran Chicago: 1989 Tell ed-Der 1988: Un quartier du 17e siecle Universit}' of Chicago Press. avant notre ere. Northern Akkad Project Stein,A. Rejx)rtsS: 15-18. 1940 Old Routes of Western Iran. London: CeIb,I,J. MacMillan. 1980 Computer-Aided Analysis of Amorite. Chicago: Streck, M. P. The Oriental Institute of the Ihiiversity of 2000 Das (unurritisclw Onomastihm der altlxihyChicago, loni.schen Zeit, AOAT 271/1. Munster: Ugarit Gelb, 1. J.; Purves, P M.; MacRae, A. A. Verlag. 1943 Nuzi Personal Names, OIP 57. Chicago: UniWafler, M. versity of Chicago Press. 2001 TaU al-Hamidiya 3: ?Air historischen GeograCofF, C. L, phie von Idaiuaras zur Zeit der Archive 1966 New Evidence oj Cultural Development in von Mari und SulxitenliUSehna, OBOSA 21. Luristan in tlie iMte Second and Early First Freiburg, Switzerland: Universitatsverlag. Millennia. Ph.D. dissertation. Institute of ArWoolley, C L. chaeolog)', University of London. 1976 The Old Bahylotiian Period, Ur Excavations Heydari, S. Report VII. Oxford: Oxford University Press. 2001 Ceology and Ceomorphology of the IslamaZiegler, N. had Survey Area: Some Preliminary Ohserva1999 Le Harem de Zimri-Um. Paris: SEPOA. tions. Iranian journal of Archaeology and History 26-27: 61-65, (Persian)
76
KAMYAR ABDI AND GARY BECKMAN
obv.
LoE
ohv. obv.
5' LoE
rev.
rev.
10'
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
obv.
/
-
obv.
•,
'I
rev. uiiinsc rihed but sealed
rev.
obv.
10
H^.
obv. rev.
sealed
rev. sealed
77
78
KAMYAR ABDI AND GARY BECKMAN
obv.
obv.
LeE rev. uninscribed rev.
sealed
^K^ obv. obv.
rev. sealed LoE
rev.
LoE
rev. •,:,/.;
1/^
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
14
15 obv. C^
obv.
rev.
LoE 17
rev.
obv.
h
16 obv.
rev.
7
79
80
KAMYAR ABDI AND GARY BECKMAN
18 obv.
obv,
LoE rev.
LoE
rev.
21 obv.
rev. lost
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
10'
15 15'
unplaced fragment A
LeE
81
KAMYAR ABDI AND GARY BECKMAN
82
22 LeE obv. 23
rev.
24 rev. obv.
^< 5
y
-t=
:
,.
20
/i
r 10 ¥
LoE
T^
V
f
j
15 very worn
rev. lost
r 25'
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
25 obv.
27 obv.?
preserved rev. uninseribed
obv.?
rev.?
obv.
rev. obv. V
5' 15'
rev. uninscribed
83
84
KAMYAR ABDI AND GARY BECKMAN
30 obv. obv.
, ..
5'
. '.- L-»»
LoE
^m^:^ . V !.:•:>
rev.
I
/pff 37
rev.
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
34 obv.
35 obv.
85
86
KAMYAR ABDI AND GARY BECKMAN
36
39
obv.
obv.
5 I
rev.
15
38 obv.
•A rev. f
f
rev. uninscribed 10
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
obv. '
5'
87
KAMYAR ABDI AND GARY BECKMAN
42
43
obv.
obv.?
I \-
"I
^m^:
Tfl
-
H^J HI
AT
^ r nrrr
•^
TT
_
5'
rev.?
-^
rev.
10'
46 44
.^^<
a otber side
1^ obv.?
r side 47
48 50 obv.
rev. rev. uninscribed
other side lost
89
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
51
49
55 obv. r\
obv.? (
M^j^
J
rev.?
LoE otber side lost
rev.
very flakev
obv.
UE 54 obv.
rev. uninscribed
rev. uninscribed obv.
very worn
56
rev.
rev.? - a few illegible traces
90
KAMYAR ABDI AND GARY BECKMAN
Fl
FIO Fll
F12
ohv.?
MA
F13 edge
rev.?
A CUNEIFORM ARCHIVE FROM CHOGHA GAVANEH, WESTERN IRAN
F14
F16
F15
F17 ob\.?
obv.
••#F
HI
-*^I^ rev,
F20
F19 obv.?
rev.?
rev.?
F24
5'
F28
F21
a.
91
F18
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOGIETY Mogens Trolle Larsen (University of Copenhagen) For Bob Adams in gratitude and admiration
During more than two centuries in the early second millennium B.C., Assyrian traders, with their base in the city Assur on the Tigris, conducted an intensive commercial operation that took them to central Anatolia more than a thousand kilometers away. Here, on the plateau of Asia Minor, they huilt up a system of colonies and trading stations in or adjacent to urban centers belonging to the numerous small pohtical entities in the region. The largest and most important of these settlements was at Kanesh, modern Kiihepe, some twenty kilometers from Kayseri, and the administration of the entire colonial system was located here. Excavations at this site have uncovered a considerable number of houses in the lower city, where Assyrian merchants and their local families lived, and where their archives have heen found. At present we know of ahout 23,000 texts from here, approximately 4,000 from the antiquities market and the rest from the Turkish excavations that have been pursued under the direction of professor Tahsin Ozgiig since 1948.' No more than about 25 percent of the texts have heen published so far, and many of those that have heen available in cuneiform copies for
a long time have never been edited properly or placed in meaningful contexts such as reconstructed archives or dossiers. The excavations at Kanesh have revealed four different occupation levels in the lower city. The two earliest ones, levels 4 and 3, are almost entirely unknown, and the hulk of the textual material comes from level 2; the houses here were destroyed at some point during the second half of the nineteenth century B.C., and the residents could only remove their most valuable possessions, including the most recent documents, before disaster struck. The dated texts refer to a period lasting about ninety years, but most of the years (the early ones) are either unattested or only rarely found in dated documents, so the archives cover in reality a phase of some fifty to sixty years, that is, two generations that lived in the houses before the destruction.^ The phase referred to as level lh has yielded only a few hundred tablets, most of these unpublished. The Assyrian commercial system has been analyzed and described in some detail.^ It was hased on a system of long-distance trade with an import into Anatolia from Assur of large quantities of tin
1. 1 wish to thank Klaa.s Veenhof, Bob Adams, and Thomas Hei tel for their constructive comments to this paper. At a crucial point in the writing of this article a conversation with professor John F. Padgett in Chicago gave new ideas and information concerning the conditions in Renaissance Florence.
2. See Veenhof (2003a: 57), destruction of level 2 in 1836 B.C.; and Kryszat (2004). 3. See Larsen (1967); Veenhof (1972); and a collection of letters in Michel (2001).
93
JCS 59 (2007)
94
MOGENS TROLLE LARSEN
and textiles; the traders more-or-less permanently hased in Kanesh took care of the sale of these commodities, either directly on the market or hy way of commission arrangements with traveling merchants, and sums of silver (plus some gold) were shipped hack to Assur in order to he reinvested in new caravans carrying tin and textiles. A very lively trade in copper and wool took place within Anatolia itself, and that too was in the hands of the Assyrian merchants.* The Old Assyrian trade in Anatolia was emhedded in a much larger, interregional network that is either poorly or not at all attested in either archaeological or textual sources: Assur itself produced only some of the textiles sent to Anatolia and imported others from the cities of the alluvial Bahylonian plain in the south; the tin also came from here, though ultimately it must have heen hrought from mines much further away, presumahly in central Asia. Despite very real advances in the study of the Old Assyrian trade and society during the last decades, starting with the puhlication of Paul Garelli's Les assyriens en Cappadoce in 1963, we are still a long way from an adequate understanding of many central issues. Some of the limitations are of course due to the fact that only a very small group of scholars are working on this material, and, in comparison with the situation in other fields of study, one is struck hy the contrast: a handful of scholars are faced with a mountain of evidence and only a very limited foundation in the form of indepth analyses of vital aspects exists. A further complication hampering our understanding of the historical, political, and social structures stems from the fact that all of our evidence comes from a colonial context, whereas the archives from the main offices and governmental hureaus in the capital are unavailahle to us.^ The excavations conducted at Assur hy a German team under Walter Andrae during the first decade of the twentieth century did not result in the discovery
4. See Dercksen (1996). The wool trade has not yet been studied. 5. For a discussion of the sy.stem of commercial colonies see Larsen (2000: 77-87); for a general discussion of Old Assyrian colonies see Stein (2004: 143-71).
of the houses where the main offices of the Old Assyrian merchant firms were located, where the main financiers lived, and where their archives were kept. Consequently, practically all of the textual material availahle for study in the field of Old Assyrian stems from archives discovered in private houses in the trading colony at Kanesh, and this fact naturally colors our view of hoth the social and commercial structures of this society. The excavated houses were inhahited hy families who tapped into the overland trade hetween Assur and Anatolia, and their activities took place within a wider framework of representation and agency, linking the people in the colonial system in Anatolia with each other and with the community in Assur. Here, in the home city that was almost invariahly referred to simply as "The Gity," we find not only the political institutions governing the entire Old Assyrian society, hut also the persons who directed the activities of the men in Anatolia. Only very few of the men who lived in Assur had presumahly ever heen in Anatolia, hut through their investments in joint stock companies and partnership contracts with traveling traders, and as heads of the families or clans or companies of which the men in Anatolia were junior memhers, they exercised control over the larger issues and decisions that shaped the trade. Einally, an equally critical ohstacle is the fact that most of the puhlished texts have no archaeological context, having heen dug up hy local villagers in the years hetween ahout 1890 and 1948. This means that the archival association of each of the unexcavated texts has to he reconstructed on the hasis of an analysis of internal criteria, and a comparison with the excavated texts shows us that this task is virtually impossihle to carry out in a satisfactory way. An analysis of the texts found in a house excavated in 1994*' shows that as much as 45 percent of them would he implacahle without the evidence of the excavator, either hecause no names are preserved in them (apart from wit-
6. Entrusted to me for publication. See further discussion of this archive below.
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOCIETY nesses), or hecause they do not mention any individual who can he related clearly to the family in whose house the texts were discovered. These difficulties notwithstanding, the existing studies and analyses have shown that the Old Assyrian archives constitute the earliest known documentation for an elahorate long-distance trading system, reaching presumahly from the tin mines in central Asia via the cities on the Mesopotamian plain, where many of the textiles were produced, to Syria and Anatolia—in fact as far as the Black Sea coast and the Aegean region. Moreover, the commercial organization was in many respects comparable to what we find in the Geniza texts from Old Cairo (eleventh to fourteenth century), and even the Italian cities of the Renaissance. The classic description hy de Roover of the desk-hound commercial manager who conducted his business by way of agents,/actors, who were stationed in hranch offices far away, the so-called "Commercial Revolution of the Thirteenth Century" in Europe, has a close parallel in the Old Assyrian period (see de Roover 1942: 34-39). And there is no comparable evidence available to us from the many centuries, even millennia, that separate these two historical situations, a fact that underscores the comparative potential of the Old Assyrian archives. This article is intended to point to a few preliminary observations concerning the relationship between individual and family in Old Assyrian society, and it will be based on my study of two family archives: a reconstructed "archive" reflecting the activities of the family of Assur-nada and an unpublished archive found in 1994 that was excavated in the house belonging to the family of a certain Salim-Assur. The mere existence of these archives must of necessity structure our view of the social and commercial patterns: family archives reflect the activities of what we have usually called a family firm. But despite the obviously crucial role of the family, whose importance is evident to anyone, we have to admit that very basic questions about the nature of the family remain without adequate answers. The term bet ahini, "our fathers house," is used by members of the same kinship group to
95
refer to the institution that defines their common origin and interests, but we have no study to tell us who shared this institution, whether the strict nuclear family or a wider kinship-hased groupperhaps even the individuals who seem to function as a kind of client. The fact that the Old Assyrians did not use a system of surnames, and used patronymics sparingly, means that it is sometimes extremely difficult to reconstruct family patterns. The religious significance of the ancestral spirits, etammu, and the rituals that must have been associated with them is unclear.*^ Extremely common references in the letters to crisis situations, where people are admonished to protect and shield the paternal house from evildoers,^ show the central position of the family as a force binding a group together. However, the interface between the kinship structure and the commercial one determining the concrete function of the family as a socio-economic unit needs to be clarified. Property The study of the archive of Assur-nada (Larsen 2002) dug up in a house in the lower city at Kanesh by the local villagers before excavations started, made me realize that although it seems reasonable to describe the relations existing between the members of the nuclear family, whose activities are reflected in the texts, as a "family firm," there was a strong emphasis on individual responsibility and rights. It seems an inescapable conclusion that these men did not share ownership of either goods or money, and it is my contention that the evidence shows that ownership to funds of any kind was individual, not familial. The texts in the reconstructed archive are concerned with the activities of a small group of
7. See for instance kt n/k 1192: 20-21: "Save my fathers house, my own, and the ancestral spirits of my fathers house" (E a-bi-a i-a-la\m' ii e-td-me E a-hi-a I e-te-er). See Hirsch (1961). 8. kt 94/k 1146: "Clear the affairs of your fathers house! No evildoer should harm you!" (E a-bi-ku-nu za-ki-a ina-nia-an .sa-li-um i-na se-n-ku-nu Id i-sa-la). The verbs used in such contexts are typically etarum, "save," zaka^um, "clear," ebabum D, "cleanse," esdrum S, "bring in order."
96
MOGENS TROLLE LARSEN
persons: the father Assur-idi, who lived in Assur and sent instructions to his sons in Anatolia; Assurnada, who lived in Kanesh, and who, as the oldest son, was in charge of the activities of the family in Anatolia; and his two hrothers Ili-alum and Assur-taklaku, hoth of whom functioned as caravan leaders, transporters, and traveling agents. This was clearly a family-run husiness where orders and instructions came from the father, who was also responsihle for securing a steady flow of commodities from Assur to Kanesh. The importance of the family ties for these men is easy to demonstrate, and it is clear that the nuclear family served to create and maintain a kind of solidarity among its memhers. The authority of the pater familias was never in douht and his instructions were central to all decisions and activities in Anatolia, although he often expressed deep anxiety or frustration hecause of what he perceived as the unacceptahle hehavior of his oldest son. The solidarity existing in the group manifested itself in a variety of ways, and sons and father acted in each other's hehalf in all kinds of commercial transactions. This meant, for instance, that Assur-nada could send a shipment of silver to his father in Assur, where he would take care of the purchase of tin, textiles and donkeys and see to it that these commodities were sent hack to Kanesh;® and Assur-idi could ask or order his sons in Anatolia to sell goods that helonged to himself;'" or he could make use of their services as caravan leaders, transporting his own shipments." The affairs of the family did not always go smoothly, and the father's letters contain regular complaints caused hy Assur-nada's unwillingness (or inahility) to send silver promptly to Assur. In one particularly excited letter, he claims that this is now the third time it has happened, and he has heen placed in a difficult situation in Assur where, he says, "for every mina or even half a mina they (i.e., his creditors) prowl around me." He is in fact so angry with the son that he swears hy divine Assur 9. An example is Larsen (2002:117). 10. An example is Larsen (2002: |2). 11. See for instance Larsen (2002: |33 or 34).
that with respect to a certain loan of no less than 40 minas of silver "I shall not make any concessions to you regarding interest," hut that he intends to charge the full amount.'^ Incidentally, this illustrates the apparently quite regular occurrence of interest-hearing loans hetween the memhers of the nuclear family. In difficult situations the sons would appeal to their father for help, hut his assistance never seems to have taken the form of direct gifts of money or goods without some quid pro quo. There was clearly no common family fund on which they could draw, hut each memher of the family had his own funds. I was therefore led to the conclusion that this Old Assyrian family firm should he characterized as "a vehicle for collahoration" rather than as a true company."' The father's position, representing the prestige and status of the family's husiness, was of great importance for his sons. Assur-nada in Kanesh at one point wrote a letter to his father in which he implored him to help in a difficult situation. He seems to have heen ready to emhark on a husiness trip around Anatolia with a quantity of goods entrusted to him hy a third party, hut he was afraid that his own status would not he quite reassuring for this investor. He was undouhtedly involved here in a commission sale of tin and textiles that had arrived from Assur, and we know from many texts that in such situations the reliahility of the agent was deemed of the utmost importance.'"* Apparently Assur-nada was afraid that he might he seen as not quite sound, so he wrote to his father for help. In the letter he was not asking for a loan or a gift, hut instead he suggested that the father should send off a shipment of merchandise that would reach Kanesh, and whose arrival there
12. Larsen (2002:1[41). It seems that otherwise friends or colleagues could reckon with a certain degree of flexibility when it came to the payment of interest on loans, "like one brother pays another"; see Larsen (2002:1163: 31-33). 13. Larsen (2002: xxvi), in a discussion of the letter 1[45 (CCT3,8b). 14. The agent has to be trustworthy {keniim), "as reliable as you yourself" (.sa ki-ma qdrqi-di-ku-nu, AKTS, 88: 45-46), or "a reliable agent who gives no reason to worry" {a-na DUMU um-ml-a-nim ke-nim Id sa sa-ha-tim, CCTS, 49a: 18-19).
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOCIETY would reassure the son's investors of the financial soundness of the family. Otherwise, he feared that he would he forced to turn to a stranger with a request for a guarantee, and that he would he met with the response: "Go to your father! Who knows whether the father is good for his name?" (Larsen 2002,145: 22-25). In other words, the reputation of the family had a direct impact on the son's activities in Anatolia. He was dependent on other people's perception of the status and viahility of the family, especially of his father in Assur. This is at the very least an indication of the importance of the close links tying family memhers to each other, hut the frantic tone of the letter also shows that although expected, the father's solidarity could not simply he assumed. And at any rate, the son in Anatolia was not in a position to rely on the family's common funds, merely at its status as a financially sound firm. In another situation the father explained in a letter to Assur-nada how his help to the son was to he understood:
97
He would then huy tin for the silver and let the son enjoy whatever profit he was ahle to realize on that. The "gift" was, in other words, a kind of interest-free loan. This type of procedure is known from many other contexts in Old Assyrian trade. For instance, it was the way in which caravan personnel were paid: they would receive a be^ulatum, a "working capital," usually consisting of Vi mina of silver This was a loan that had to he repaid at the end of the journey hetween Assur and Kanesh. During the journey the employee could use this amount as he saw fit; he was expected to make a profit which, in effect, constituted his salary. The point can he further illuminated hy another text from this archive, where Assur-nada at a later stage hailed out his own son Iddin-Istar from various dehts, perhaps simply hy paying off the son's creditors. His payment for this was a special loan of 10 minas of silver that the son received against a deht-note recording an ohligation of 29 minas. Within one year Iddin-Istar had to travel twice to Assur with the silver to make purchases, and he was to guarantee his father a profit of 190 Why have you written to me, saying: "I shall pay my percent on this transaction. If he could make more, accounts with your silver. You gave me your word" I did he could keep the surplus. In other words, Assurnot give you my word! However, you begged me, so with Assur-kasid as witness I gave you a gift of 20 kutdnu- nada's help to his son had to he repaid with his gratis work throughout a whole year, including textiles. Since our affairs are concluded you must send me at least 10 minas of silver, both from the proceeds of twice going to Assur and hack, a journey of some the textiles and that of the donkeys, the very first silver five thousand kilometers (Larsen 2002:1142). you obtain, and then I shall buy 2 talents of tin and Examples such as these lead to the conclusion send that to you so you can enjoy the profit. Do not take that each individual had control over, and was reany of my silver! (Larsen 2002,
98
MOGENS TROLLE LARSEN
archive, and the texts reflect the activities of a much larger group of persons. Obviously, we are dealing with a much richer and more influential family, and the letters show that a fairly large group of persons whose relationship to the central figures is not clear, whether family, in-laws or clients, were engaged in the affairs of the family. The pater familias here was a certain Issu-arik; he had moved to Assur late in life and left his two sons Salim-Assur and Iddin-ahum in Anatolia, while a third son, Assur-hel-awatim, appears to have lived in Assur where he held a high post as laputtaum in the city government (see Dercksen 2004: 65-72). At a time when the father was still alive he had ordered Salim-Assur to lend his brother a large sum of money to help him pay his debts. This request is contained in a quote found in a letter:
As a logical consequence of the fact that all property was individually owned, no firm or family {bet PN) ever appears in a debt note as either creditor or debtor,'^ and writers never refer to money owned by the family, only to funds that may be stored and available in the paternal household.'^ The exact meaning of phrases such as "on behalf of" or "in the name of the paternal house," referring to situations where an individual collects debts, entrusts money to others or discusses with customers, has to be established on the basis of an analysis of the context.'^ These references point to the importance of the family in the Old Assyrian commercial system without challenging the fundamental observation that all capital was individually owned.
Yes, Issu-arik spoke as follows to him in our presence: "Pay the debt of^ your brother, both here and abroad, and then go and take it again out of your brother's silver. If there is no silver belonging to your brother, then pay out of your joint-stock capital."'^
The naruqqum, or joint-stock partnership, formed the basis for the individual Assyrian
Years later, after the brother Iddin-abum s death, this led to a complex lawsuit where Salim-Assur attempted to retrieve his money from his brothers heirs. The fact that he did not collect the debt, but left the large sum of money advanced to his brother in his hands, and the fact that it was apparently ignored in daily transactions between the members of the family over a period of several years, again indicates the close relations that could exist between brothers. It is even more striking given that the archive does not indicate any commercial collahoration between the two brothers. It is doubtful whether Salim-Assur could have refused his father's order, for even though the brothers functioned as independent traders in Anatolia, administering their private funds, the authority of the pater familias was presumably overriding. 16. kt 94/k 840: 2-23: ke-na I-sii-ri-ik i-na ma-ah-ri-ni ^e'ta-wu-su-um um-ma- su-ut-ma ha-bu-ui; a-hi-ka: lu a-na-kam: lu i-na GANA-lim su-qil-id-ma ii a-Uk-ma : i-na KU.BABBARdp a-hi-ka ; le-qe : su-ma [KU.BABBAR]-dp a-hi-ka Id-su i-na na-ru-qi-ka: su-qill.
Investments
17. With the exception of ownership of investments in joint-stock funds; see further below. 18. See for instance AKT 2, 57: 5-14, a letter to a lady in Assur from her brother in Anatolia, where he writes: "My dear sister, my dear lady, if there is any silver there of our fathers house, then satisfy Idaya's son. If you do not wish to (do that), then let Pilah-Assur sell my house there to pay off my creditor." [a-ha-U a-ti be-el-ti a-ti a-ma-kam i-na KU.BABBAR so E a-bi^-ni su-ma i-ba-si DUMU I-da-a-a ta-i-bi^ su-ma. Id li-bi^-ki a-[ma\-kam E-ti-a Fi-ld-ah-A-siir l^i-dfma. tdin-ka-ri iu-sa-bi^). The situation seems to indicate that the father was dead, and the request to spend money available in his house may have been somewhat irregular, since noone could know, presumably, to whom the money belonged. The very rare references to debts of the paternal house (see for instance BIN 4, 83: 33-37) are probably in all instances to be placed in the context of inheritance problems. 19. See Larsen (2002,1168: 3-7), where Assur-nada writes to the lady Abaya: "In accordance with the missive I sent to you Innaya has here discussed with the customers in your name and in the name of your father's house" {a-ma-ld na-dspe-er-tim sa ds-pu-ra-ki-ni a-na-kam }-na-a a-su-mhki ii a-na. su-m\ E a-bi-ki tdm-ka-ri e-ta-wu; kt 94/k 1742: 26-29: "I also entrusted 40 minas of refined silver that was in the name of my father's house to Pilah-Istar" (;/ a-ha-ma 40 ma-na KU.BABBAR sa-ru-pd-am sa a-na su-mi E a-bi^-a a-na Fi-'ldah-I.^\tar] djrqi-id); or kt aA 1030: 1-5: "Out of Iddin-Suen's copper Tab-Assur received 13 talents 20 minas of poor copper on behalf of his father's house and Alili" {i-na URUDU .5o I-diSu-in 13 GU 20 ma-na URUDU la-mu-nam DViQ-A-sur ki-ma E a-bi-su ii [A\-li-li il^-qe).
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOCIETY
99
trader's possihihty to function independently in the comnnercial system. It was the contracttaal foundation for an arrangement where a group of financiers came together to invest in a capital fund that was entrusted to an individual to function as the hasis for "carrying out trade" {makdruiri) during a specified numher of years, prohahly always a period of at least ten years.^° The creation of such a partnership is illustrated hy a small group of texts in the Salim-Assur archive. They are special loan contracts recording funds advanced to the hrother Iddin-ahum, clearly at a very early date in his career when he was a young man with amhitions to hecome an independent husinessman. It appears that he had approached a numher of persons in Anatolia with the suggestion that they should invest in his joint-stock fund, since he was planning to go to Assur to try to convince his hosses there that he should he given such a fund. One of the contracts states this clearly: "Iddin-ahum owes 8 minas of refined silver to Kula. Reckoned from the week of Hapi-ahsu and Sukallia he must go to the City within 25 weeks and together with his investors he must write Kula's name for 2 minas of gold."^' Other texts in the group say simply: "If they write his tahlet (i.e., the joint-stock contract), then he must hook x in the name of PN," indicating that it was not at all a foregone conclusion that the men in Assur would feel comfortahle giving him this new status. These texts show how a young trader would collect funds among family and friends in Anatolia, money that would allow him to approach the financiers in Assur. All such contracts had to he set up in the capital, and it seems that all major accounting operations relating to this partnership fund were regulated there. His hrother Assur-helawatim contrihuted funds (24 minas of silver), and among his Anatolian investors we find also Ikun-pl-Assur (2 minas of gold), and of course the
Kula mentioned in the contract cited. We do not know who his investors in Assur were. Once the contract had heen set up to run for a numher of years, the manager would have to keep the joint-stock capital separate from the funds he owned outright. We do not have any account texts that give us information ahout how this was handled, at least no such texts have heen identified, and it may he that they would have to he found in Assur, for it was here that the manager or his representatives each year would have to settle accounts with his investors in accordance with a set of rules that are only poorly understood. It must have heen highly unusual, perhaps even illegal, to spend naruqqu-iunds on personal dehts, for instance, as we can see Salim-Assur heing forced to do hy his father when he had to hail out Iddin-ahum from his deht ohligations as referred to earlier. It seems ohvious that this partnership transaction had a profound impact on the way interpersonal relations were structured. It created a network of commercial contacts, where the manager of a joint-stock fund would hecome the Anatolian agent for investors located in Assur, a man to whom they could turn with a variety of tasks. At the same time it meant that he had senior partners close to the center of power in the Old Assyrian system, men who could help him in all kinds of situations. And although some of the central investors would certainly he memhers of his own family, many would helong to other firms and could ensure the setting up of a much more complex set of relationships. A letter in the correspondence of the great Kanesh trader Pusu-ken shows that the tractator, or manager of the fund, had a particularly close relationship to the firm that organized the estahlishment of the naruqqu and presumahly contrihuted the largest sums to it:
20. See Michel (2001: 303-57) for a commented selection of texts concerned with this contract; cf. also Larsen (1977: 11945), and Dercksen (2004: 83-89). 21. kt 94/k 1016: 1-12: S ma-na KU sa-ru-pd-\am\ i-se-er I-di-a-bi^-im Ku-la. i-Sii is-tii. ha-mu-us-tim sa Ha-pi-ah-'su^ il Sii-ka-lfa a-na 25 ha-am-sa-ltim a-na A-lini'' i-ld-ak-ma isti um-mi-a-ni-su 2 ma-na KU.GI su'-mi. Ku-ld i-ld-pd-at.
From Su-Hubur to Ptlsu-ken: Concerning the matter you wrote about, saying: "La-qep's two sons are grown up; they are now fit to administer a joint-stock fund"—do not let them free (to other investors)! Here Assur-samsi [one of the sons] promised me to take up a Joint-stock fund, and the fund will enter our house.
100
MOGENS TROLLE LARSEN
Concerning his brother Assur-malik—a letter from me and from Assur-samsi went to him, (and) Assur-samsi wrote: "I have promised Su-Hubur a joint-stock fund. Do not promise anything to anyone else there lest you put me to shame. Do come here and have Su-Hubur clear your affairs!" My dear brother, put in a word on my behalf to Assurmalik, saying: "In this house where your father and grandfather have been clothed, you too should take a joint-stock fund and be clothed yourself !"^^
La-qep was a close associate of Pusu-ken and like him he had intense commercial contacts with a group of men in Assur who counted Salim-ahum and Su-Huhur, the writer of this letter; he in turn was an uncle of Imdi-ilum, another Kanesh merchant who was very active in this circle of traders. The letter makes it clear that the trader who administered a joint-stock fund had a special relationship with one particular house in Assur, and that such partnership arrangements could tie people together in tight husiness contacts over several generations. Like all other funds, the joint-stock capital was tied to the individual trader, not to the family,^'^ and it seems clear that in many families every adult memher would manage such a fund as their private husiness, and at least some of them would also have investments in other peoples naruqqum. A letter from Assur-idi in Assur to Assur-nada contains the passage: "people less important than you administer joint-stock investments, and people less
22. Adana 237D, an unpublished text partly communicated by K. Hecker (1978: 150 n. 39): [um-ma Su]-Hu-bur-ma a-na [Pu-su-ki-\ii^ qi-bi-ma \a-su-m\i sa ta-ds-pti-ra-ni \um-m\a ata-ma me-er-u (5) [U-jqe-ep : ra-bi-a-an a-na \n\a-ru-qi-im : na-su-i-im \w\a-as-mu la th-wa-sa-ar-su-nu a-na-kam : A-siir''UTU-sJ pi-su ama na-ru-qi-im narsa-i-im (10) i-d(-nain it naqum- a-na be-ti-ni: e-rardb a-su-mi A-siir-ma-lik a-hi-su t:up-pu-um sa i-a-ti ii A-siVr-UTU-si 'a'-na A-siir-ma-lik i-talka-sum (15) [um-ma] A-sur-VTV-Si-ma a-na-kam [a-n\a SuHu-bur a-na na-ru-qi-i\m] \pd\-i Hu'' a-ti-di-in (rev.) [a-makam] : a-ta: pd-ka a-ma-ma-an [la t]a-da-an-ma la tii-ba-sa-ni (20) [a-ta-[al-kam-ma: Su-Hu-bur a-na-kam u-zu-ka-kd: a-putum .su-ma a-hi a-ta : a-ma-kam : a-wa-tdm is-te-et a-na ^A^-sur-ma-lik ki-ma i-a-ti \q\d-ri-ib um-ma a-ta-ma a-sar (25) [a-bu]-kd it a-bu a-bi-kd i-na \E-t\ini a-«mi-»nim : TUG.HI.A [ilr,-t[a-ab-su-ni : ii a-ta [a-m]a-kam : na-ru-qam na-an-si-ma |TUG.HI.A] li-itrba-d[s]. 23. There is one example of a father and son jointly administering a joint-stock fund: EL 327 and 328.
important than myself have invested 10 minas of gold" (Larsen 2002, 1119; 15-19). The father advises his son to get hold of all the investments he can lay his hands on in Anatolia and to come to Assur to have his status as manager of such a fund regulated. The passages clearly show how important it was hoth for the son and for the father to he memhers of the group of traders who operated with a joint-stock fund. Such a partnership would he dissolved at the death of its manager, hut on the other hand, investments in a naruqqum could he inherited after the death of the investor. We have a few texts that tell us that after the death of the important trader Pusu-ken, his sons inherited shares of his investments in other persons' joint-stock funds, and the oldest heir, Assur-muttahhil, hought these from two of his hrothers.^^ A document in the Salim-Assur archive contains a list of investments in the joint-stock fund of one of the memhers of the family, prohahly SalimAssur himself. 22 minas: the house of Assur-imitti and Hunniya; 20 Vi minas: Ill-alum; 10 minas less 15 shekels: Su-Kubum; 13 minas: the house of Su-IUil; 13 minas the son of Amur-Istar; 13 2/3 minas 5 shekels belonging to Laqip son of Kayatum; 6 Vi minas: Assur-malik son of Ennam-Adad: in all: Wi talent 8'/2 minas of silver from my joint-stock capital, including (money) from their tadmiqtu-\oans that my investors have entrusted to me. For half of the silver their terms have expired. For half of the silver he must pay within 4
We have here a list of at least some of the men who had invested in Salim-Assur's joint-stock fund. None of them helong to his immediate family: his
24. Cf. EL texts 310 and 311 {CCT 5, 22a and Dalley 11) plus ATHE 33 and AKT 1,11. 25. kt 94/k 1458: 1-21: 22 ma-na E A-.suM-mi-ti u Hu-ni-a 20 1/2 ma-na 1-li-a-lum 10 ma-na LA 15 GIN Su-Ku-bu-um 13 ma-na E Su-<'EN.LIL 13 ma-na DUMU A-mur-Jstar 13 2/3 ma-na 5 GIN sa Ld-qe-pi-im DUMU Kh-a-ti-im 6 1/2 ma-na A-siir-ma-lik DUMU En-um-'^lM SU.NIGIN 1 1/2 GU 8 1/2 ma-na KU.BABBAR sa na-ru-qi-a qd-dum sa ta-ad-mi-liqti-su-nu sa um-me-a-nu-a i-na se-ri-a i-di-ii-ni mi-sa-al KU.BABBAR u^-mti-su-nu ma-al-u mi-sa-al KU.BABBAR a-na M U 4 . S E i-sa-qal
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOCIETY hrothers do not appear, nor does his father, hut it is unclear whether the text lists all of the persons who had contrihuted to his naruqqum. Of those who do appear, we can point to Assur-imitti, prohahly the son of Amur-ili, who is said in other contexts to he Salim-Assurs representative;^® Hunniya, his son, does not play a role in the archive. The men Ill-alum, Su-Kuhum, and Assur-malik son of Amur-Istar appear in the letter 94/k 1257 as SalimAssurs investors in Kanesh. It is therefore prohahle that the men mentioned in this text were all in Anatolia, and that his investors in Assur are ahsent. Salim-Assur may have heen one of the investors in his hrother Assur-bel-awatim's naruqqumF A certain Assuris-takil is known to have invested 3 '/3 minas of gold in Salim-Assur's joint-stock capital,^^ and his father Issu-arik had an investment of no less than 10 minas of gold running for only two oq
years. The references in the text cited above to investments that helong to a "house" rather than to an individual indicate that investments that would run for many years could be viewed as the property of the paternal estate, presumably in the same way as real estate. When a house or family of an investor is mentioned rather than just his name, the explanation must be that such investments were long-term ones that were expected to remain active even after the death of the investor, and as has been mentioned, such investments could be inherited. Inheritance Under normal circumstances, all brothers and sisters were entitled to a share of the paternal
26. kt 94/k 1051. See Larsen 1982, with genealogical table on page 219, for the family of Assur-imitti. See also Dercksen (1996: 140 and table on page 92). Assur-imitti had four sons and it is not clear why only Hunniya appears here. Assur-imitti was a prominent man, the uncle of Imdi-ilum; see further below. 27. kt 94/k 1757, a text that refers to a payment of 12 minas of silver to be made as the next disbursement for na-ru-uq A-sitr-be-el-a-iva-tim. 28. kt c/k 1460. 29. kt 94/k 964, a text that records an accounting between Iddin-abum and Salim-Assur.
101
estate, although the oldest brother seems to have received a larger share. We have a limited number of testaments, a fact that might lead to the conclusion that such documents would be drawn up only when a special situation existed in the family. It seems from the few examples that are extant that they were set up to safeguard the interests of one or more of the heirs, and, in one case, we hear of the disinheritance of a son. However, as Veenhof points out, "in nearly all cases where a division of a paternal estate {bet abim) is at issue, the existence of a testament is mentioned or implied" (Veenhof 2003b), which I take to mean that important, wealthy traders would nearly always leave a will. A proper testament would apparently often be set up in Assur, but in other cases the last will may have been merely a statement made by the dying person. We hear in some instances that a man has died without speaking his will: "Sadly, our father has died. He did not speak his will. If it is good for you, then set out and come. Bring our father's house in order."^° It seems clear that in this case the absence of a last will was creating difficulties. It is in fact obvious that a great many inheritance situations gave rise to prolonged, and highly complex lawsuits. One reason for this was the dispersed nature of the Assyrian community, which meant not only that distances of up to 1500 kilometers separated the individual members of the family, but, even more importantly, that the merchant who died in many instances would be involved in a large number of commercial relationships with people who were scattered across Anatolia as well as in Assur itself. Documents had to be collected, funds evaluated, and outstanding claims collected. We see time and again that customers, both investors, creditors and debtors of the deceased, would attempt at once to lay their hands on the available evidence in the form of tablets that would be stored in tbe houses of the dead merchant.
30. CCT2, 33: 2-6. It is a letter from Su-Kubum in Assur to Ikun-pi-Assur in Kanesh; he also wrote about the same affair to other contacts in Kanesh in CCT A, 41b.
102
MOGENS TROLLE LARSEN
Another reason for conflict was disagreement among the heirs. Because of the way in which the financial relations hetween the members of the family were organized, the structures that were in place and secured the orderly running of the familys business while the pater familias was alive tended to dissolve, or at the very least to come under severe stress, when he died. It seems that the normal result would be dissolution of the existing family operation, so that the surviving sons would start over as independent traders, or in partnership with others outside the original family group. This is perhaps the most striking consequence of the lack of a common, familybased fund. The Salim-Assur archive provides vivid illumination of this point. When Issu-arik died, his three sons continued on separate paths and had only limited, if any, commercial contacts with each other And, in this family, deaths regularly gave rise to complex lawsuits. I have already referred to the situation that arose after the death of Iddinabum, when Salim-Assur had to engage in protracted lawsuits in order to reach a settlement with his nephews over the money he had spent on Iddin-abum's debt. Salim-Assur had been appointed as the executor of Iddin-abums estate,^' an example of the collaboration and solidarity that must have existed hetween the brothers, but it is characteristic that immediately after Iddin-abums death, the third brother, Assur-bel-awatim, lodged a claim to no less than 24 minas of silver, his investment in Iddin-abum's joint-stock capital, which moved his son to take Salim-Assur's house, his slave-girl and other servants as security, and only after this large sum had been paid could SalimAssur reclaim his own property.^^ When Salim-Assur then died, his two sons became involved in a lawsuit over the inheritance that lasted at least two years, and which is directly referred to in more than one hundred texts of all types, i.e., close to 10 percent of all the texts in the archive. Of the five large legal dossiers in this archive, four are concerned with the aftermath of a death, while one originates in a conflict with a 31. kt 94/k 1006. 32. kt94/k950andll44B.
local official. It may be estimated that at least fifty percent of all legal texts in the corpus (published and unpublished as far as known to me) were drawn up in connection with complications arising from contested inheritances. Some of these, such as the complex dossier referring to the death of Puzur-Assur (cf. Matous 1969: 156-80) have been studied in some detail, and it can safely be concluded that this particular situation when an important, active trader died was of crucial importance. Often the legal disputes came to involve the city assembly in Assur, and in several cases verdicts were issued commanding customers to return tablets and other valuables that they had removed from the strong rooms and magazines of the deceased, and special "attorneys" were sent to Anatolia in order to clear up the tangled affairs of the dead man's estate. The breaking up of collaboration among the heirs was not a foregone conclusion, of course. Immediately after the death of Salim-Assur the younger son Al-ahum, who had been with the father in the city Durhumit when he died, wrote to his brother Ennam-Assur: Sadly, our father has died. It is not Salim-Assur who is our father, it is you who are our father Take care there of our father's instructions and clear the affair.^'
Very quickly the two brothers fell out, however, and their conflict came to involve not only their closest family, but also the many other persons appearing in the archive as employees, partners, agents, and so on. It is interesting that the conflict was caused in large measure by the fact that Salim-Assur had left two different wills, namely, a proper one set up in Assur where it was in the care of his bosses, and a document that contained a statement made on his deathbed in the Anatolian city of Durhumit, where only Al-ahum had been present. This seems to have referred only to the obligations of the two surviving sons, and Ennam-Assur refused to accept this document, claiming that Al-ahum had tricked the dying father into making special concessions. 33. kt 94/k 747: 5-12: Id li-bi i-li-ma a-bu-ni: me-et Id Sdlma-A-sur a-bu-ni a-ta a-bu-ni a-ma-kam a-na ti-ir-tia-bi^-ni i-hi-Hd^-ma za-ki-i
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOCIETY Ennam-Assur was murdered before the lawsuit had heen brought to an end (we do not know by whom), and it is interesting to see that the fairly extensive correspondence between Al-ahum and his agents and representatives appears to show that the entire group of persons who had been associated with the Salim-Assur family split up, and several of the men who appeared prominently in the letters to be dated before Ennam-Assur's death disappear completely. Apparently they had sided with Ennam-Assur in his conflict over the inheritance. A family firm should accordingly be seen as a much wider circle of individuals, reaching beyond the nuclear family. In most instances, we have no precise knowledge of the relationship that binds a group of persons to a firm in a pattern that may be referred to as clientage (such as the family of Amur-Istar and his sons, who were all closely related to the Salim-Assur family), but it is surely likely that marriage bonds and perhaps more distant blood relationship could play a vital role.'''* Obviously, some brothers continued to get along well with each other after the death of their father, but I do not know of a single clear example in our documentation of continuity across the generation gap, where brothers in Anatolia continued to work together with the eldest as the new boss, the situation envisaged in the letter from Al-ahum quoted above.*^^ For Assur we have an example in the family of Imdi-ilum; his father Su-Laban was in charge of the family affairs until his death, but then the second-eldest brother Assur-imitti moved 34. It is presumably in this context that we must understand the many references where a man is said to belong to "the house" of another man; AKT 3, 1; 1-7: Assur-taklaku son of Ikuppia, (grand)son of Assur-taklaku of the house of Ukubturus, owes 2 minas of iin-silver to Iddin-Assurson of Assur-malik" (2 ma-na KU.BABBAR ti-ri i-se-er A-siir-tdk-ld-ku DUMU l-kupi-a DUMU A-siir-tdk-ld-ku sa E U-ku-ub-tit-ru-us E-di-inA-Siir DUMU A-siti^maMk irsu); for other examples see BIN 6, 64: 4-5, 23-24; CCT4, 48b: 7-8. Members of the family often refer to "our own PN" in order to distinguish them from namesakes. 35. This point can be illustrated much more elaborately by way of the unpublished archives. Veenhof points out (personal communication) that the two brothers Elamma and Al-ahum, whose archives were discovered in 1991 and 1993 respectively, and who lived in neighboring houses, had little or no commercial contact after their father's death.
103
from Anatolia to Assur and became the new pater familias, and the relationship is attested in numerous letters (see Larsen 1982: 221). In view of the observation presented earlier that all ownership of property was individual, not familial, such a lack of continuity is logical. The idea of the family firm as an incorporated legal entity was not yet invented, and instead the emphasis was on individual ownership, responsibility, and accountability. Goidthwaite has described a development that can be traced in the documents from Renaissance Florence and that may have relevance for our understanding also of the Old Assyrian situation. In Florence there was a change from the early-fourteenth century, when a family's common ownership of property was the norm (extending also to business organization), to the fifteenth century, where brothers regularly split up and maintained their individual property, a development that is marked by what Goidthwaite calls the "disintegration" of the family.^'' The main reason for this change was the need to create funds large enough to secure a profitable participation in the international trade by the introduction of external partners into what was otherwise a family partnership. It is not possible to trace a similar development in the Assyrian evidence, moving from communal ownership to individual property. The practice of common ownership to land is of course known from later periods in the north of Mesopotamia, for instance at Nuzi, and the Middle Assyrian Laws refer to brothers living in undivided households after the death of their father. However, it is unknown whether such was the norm in Assur before the expansion of the trade. Nevertheless, the point to make is that the practices represented in the Old Assyrian texts may well have been driven by
36. Goidthwaite (1968: 253-58). Goidthwaite (1980: 63) wrote: "Not until the sixteenth century did the notion that wealth somehow had its own integrity as something inviolable and permanent, to belong forever to a family, gain currency, and only then did inalienability and fideicommissa begin to complicate the inheritance of property." The Babylonian companies, such as Egibi and MurasO, that operated in the NeoBabylonian and Persian periods seem to have maintained themselves across generations, sometimes for hundreds of years, but the precise organization of these firms is not known.
104
MOGENS TROLLE LARSEN
similar forces, i.e., the need for capital accumulation, in the estahlishment of businesses based on partnerships between individuals who were not related to each other Conclusions Old Assyrian society was divided into two distinct groups, the capital and the colonies, and the hierarchical relationship governing their interactions may be seen in several features: the creation of a joint-stock fund always took place in Assur and it seems that it could only be created with the acceptance of a group of financiers based there; the city assembly, whose members counted at least some of the correspondents known from our archives in Kanesh, had the final say in a great many cases that directly involved the colonists, and it also laid down the rules and guidelines for the trade by way of verdicts;'^^ in most instances it seems that also testaments were set up in Assur; and even though some men in the colonies could reach a certain degree of autonomy and amass great wealth, they were fundamentally dependent on the bosses in the capital.''^ In the case of many families, we can see that the father, sitting in Assur, ran the affairs of the firm—Assur-idi is an example, but other arrangements were possible. As pointed out earlier, the important trader Imdi-ilum accepted the leadership of his uncle in Assur after his father had died. After the death of Issu-arik, Salim-Assur's directors in the capital consisted of a group of persons who were presumably relatives, although we cannot establish the ties that bound them to him. Another influential trader in Kanesh, Pusuken, did a very large part of his business as an agent for a man called Salim-ahum who seems not
37. See, for instance, the verdict where Assyrians in Anatolia were forbidden to trade in gold with Hurrians, Amorites, and Babylonians (kt 79/k 101); or the case where a group of Assyrians working in Anatolia was fined for having traded in a local textile type that could compete with the export articles from Assur (VAT 9290; see Veenhof 1972: 126-28). 38. Status relations that are apparent from the introductory formula of letters (see Larsen 2002: xxxviii-xl) almost invariably show the men in Assur to be in a situation of superiority.
to have been a member of his immediate family. Salim-ahum's sons Dan-Assur and Ennam-Assur, whose archive was found in 1970, did not handle his commercial affairs in Anatolia, although they were in regular contact with their father as evidenced by the many letters in the archive.''*^ We also have several examples of letters written to men in Anatolia by a group of persons who refer to themselves as "your investors.'"*" Relations between the men at the two ends of the system could accordingly be based on close family ties, but in some cases the association appears rather to have been based on economic bonds such as the joint-stock funds. The lack of documentation from Assur makes it impossible to determine whether the features that can be demonstrated for the Old Assyrian colonial society were as important in the capital. The institution of the bet abini, the paternal house, seems to belong in Assur at any rate, and it is quite possible that even the breakdown of the family on the death of the principal in Anatolia would have had scant relevance for the men in Assur, who may have constituted a continuity on a different level, maintaining the reality of the paternal house. This seems to be the case of the Imdi-ilum family referred to earlier. In many cases, the cohesion of the family was limited, and to some extent counteracted, by the emphasis on the individual as the only legally constituted owner of funds. This meant that a constant tension was introduced between family and individual, tending to dissociate individuals from each other, breaking up established patterns of authority at regular intervals as generations succeeded each other, and creating extremely stressful relationships between close relatives. The individual trader in Anatolia was defined first as a member of a family, and in this role he would be constrained in his actions by the authority of the paterfamilias. At the same time he would be the manager of a joint-stock fund, and on the basis of that he would be in a relation39. Partly published in AKT 3. 40. See Adana 237B; BIN 4, 32; BIN 6, 105; Prague 486; TC 2, 41; and kt91/k 350.
INDIVIDUAL AND FAMILY IN OLD ASSYRIAN SOCIETY ship to a number of financiers, who were not members of his family but who nevertheless exercised some degree of authority over him. It is highly likely that the creation of a joint-stock fund that gave him a large capital was a prerequisite for a trader's ability to function with success in the basic commercial procedures. It may well be that the Assur-nada family's constant financial worries had their roots in their too-limited capital basis, whereas the family of Salim-Assur appears to have been so wealthy that inevitable losses never created a critical situation. This narwf/c/u-institution must have been invented by the Old Assyrian merchants at some point during the first decades of their commercial penetration of Anatolia,"*' at a time when we must assume that the trade expanded geographically as well as in terms of the size of individual transactions, and it seems clear that apart from making large funds available to younger members of the commercial community, it also introduced new interpersonal relationships that cut across family
41. The ones we know of may be dated to the period between KEL 75 and 80. 42. Veenhof notes (private correspondence), as another reason for the introduction of this new type of transaction, that in this kind of partnership the investors ran limited risks, in contrast to the situation in families, where every member could be made responsible for the failure of the father's business.
105
Some of the colonists undoubtedly came to establish a degree of independence from Assur'"' also in commercial affairs, for instance, by becoming heavily involved in the trade in copper that took place within the frontiers of Anatolia. In fact, Salim-Assur spent much of his time in the copper-producing region, and died in the main center there, Durhumit, and his family was clearly a major player in the trade in copper and wool. Nevertheless, it was in Assur that the economic basis for the entire system was rooted; here we would have found a large number of merchant houses who operated both production of textiles and the transit trade in tin and imported textiles, and the men in Kanesh could not maintain their existence without constant contact with Assur or without regular supplies of goods arriving from there. Changes would undoubtedly happen over time, as the colonial society in general developed its own practices, including the new financial partnership contract, and especially as certain individuals in Kanesh and the other colonies in Anatolia became rich, and established families of their own abroad. These developments may well have put a strain on traditional patterns of relationships, leading to a fragmentation of families and even to new concepts of property and individual responsibility.
43. Note that the second generation of colonists in our documentation regularly established an Anatolian family, marrying local women.
References de Roover, R. 1942 The Commercial Revolution of the Thirteenth Century. Bulletin of the Business Historical Society 16: 34-39. Dercksen, J. G. 1996 The Old Assyrian Copper Trade in Anatolia, PIHANS 75. Leiden: Neederlands Instituut voor het Nabije Oosten. 2004 Old Assyrian Institutions, MOS Studies 4. Leiden: Neederlands Instituut voor het Nabije Oosten. Garelli,E 1963 Les assyriens en Cappadoce, Bibliotheque archeologique et historique de l'lnstitut fran-
gais darcheologie d'Istanbul 19. Paris: Adrien Maisonneuve. Goidthwaite, R. A. 1968 Private Wealth in Renaissance Florence. Princeton: Princeton University Press. 1980 The Building of Reiiaissance Florence. Baltimore: The Johns Hopkins University Press. Hecker, K. 1978 Zu den Verwandtschaftsbezeichnungen in den Kiiltepe-Texten. Pp. 137-55 in Festschrift Luhor Matous, Vol. 1, eds. B. Hruska and G. Komoroczy. Budapest: Eotvos Lorand Tudomanyegyetem, Okori Torteneti Tanszekek.
106
MOGENS TROLLE LARSEN
Hirsch, H. 1961 Untersuchungen zur altassyrischen Religion, Archiv fur Orientforschung Beiheft 13/14. Graz: Biblio-Verlag. Kryszat, G. 2004 Zur Chronologie der Kaufmannsarchive aus der Schicht 2 des Karum Kanes. PIHANS 99. Leiden: Neederlands Instituut voor het Nabije Oosten. Larsen, M. T. 1967 Old Assyrian Caravan Procedures. PIHANS 22. Leiden: Nederlands Instituut voor het Nabije Oosten. 1977 "Partnerships in the Old Assyrian Trade." Iraq 39: 119-45. 1982 Your Money or Your Life! A Portrait of an Assyrian Businessman. Pp. 214-45 in Studies and languages of the Ancient Near East in Honour of I. M. Diakonoff, eds. M. Dandamayev et al. Warminster: Aris & Phillips. 2000 The Old Assyrian City-State. Pp. 77-87 in A Comparative Study of Thirty City-State Cultures, ed. Mogens Herman Hansen. Copenhagen: The Royal Academy of Sciences. 2002 The Assur-nada Archive, Old Assyrian Archives 1. PIHANS 46. Leiden: Nederlands Instituut voor het Nabije Oosten.
Matous, L. 1969 Der Streit um den Nachlass des Puzur-Assur ArOr 37: 156-80. Michel, C. 2001 Correspondance des marchands de Kanish Litteratures anciennes du Proche-Orient 19. Paris: Editions du Cerf. Stein,C. 2004 The Political Economy of Mesopotamian Colonial Encounters. Pp. 143-71 in The Archaeology of Colonial Encounters: Comparative Perspectives, ed. G. Stein. Santa Fe, NM: School of American Research Press. Veenhof, K. R. 1972 Aspects of Old Assyrian Trade and Its Terminology. Leiden: Brill. 2003a The Old Assyrian List of Year Eponyms From Karum Kanish and Its Chronological Implications. Ankara: Tiirk Tarih Kurumu. 2003b The Old Assyrian Period. Pp. 457-58 in A History of Ancient Near Eastern Law, ed. R. Westbrook. HdO 1.72. Leiden: Brill.