Piecing Together Song of Release - JCS 65

127

Piecing TogeTher The Song of releaSe

Eva von Dassow (University of Minnesota)

1. Introduction

The hurrian poem that bears the ancient title “Song of release” was recorded in a bilingual edition, hurrian with a hittite translation, by scribes working in hattuša circa 1400 bce.1 Several manuscripts of this edition were produced, but not one tablet of any of those multi-tablet manuscripts survives intact. What remains of them exists in the form of many small fragments and a few big ones, which were found, together with pieces of other com-positions, in a disordered assemblage in and near one building (Temple 16) at Boghazköy.2 only a few of these fragments preserve colophons that would help establish the identity and sequence of the dismembered parts of the Song of release, whose title is given in the hittite colophons as ŠÌr para tarnumaš. Since, unlike their later Meso-potamian counterparts, hittite scribes did not standardize the distribution of content across successive tablets of multi-tablet works, the text of a given composition may be divided differently in every manuscript of it; and so it is with the manuscripts of the Song of release. Thus the most fundamental task in studying this composition is to determine which tablets and fragments belong to it, in what order those pieces go, and accordingly, in what order the narrative of the poem goes (so also Wilhelm 2012: 158). it need hardly be said that the arrangement of a literary work is normally essential to its meaning, and reconstructing it in the wrong order will necessarily yield a faulty understanding of the work.

i am grateful to gernot Wilhelm for responding to many queries, especially concerning the analysis of hurrian forms, and providing pho-tographs in support of this inquiry, which originated while studying the Song of release and the parables in his hurrian reading course at the University of Würzburg in the summer semester of 2008. My thanks also go to Mary Bachvarova for discussing and arguing about these texts with me; while we disagree on many points (though fewer than she believes), these conversations have helped me in thinking about the mate-rial, as well as leading me to literature i would otherwise have overlooked. abbreviations used herein accord with those employed in Archiv für Orientforschung.

1. The fragmentary exemplars of this work are dated, as an ensemble, to “around 1400” on palaeographic and linguistic grounds; the evi-dence is summarized by neu (1996: 3–6). The age of the hurrian composition is greater, by a couple of centuries (see further below). all tablet fragments representing the Song of release were published in KBo 32, together with other tablets and fragments found in the oberstadt area of hattuša from 1980–1986; the poem received its editio princeps in StBoT 32 (neu 1996).

2. See neve 1984: 366; 1999: 70–72. The find context is fairly coherent, despite the disarray of the finds. Most of the tablet fragments were found within two of the three excavated rooms of the basement of Temple 16, the rest of which was overlain or obliterated by the Byzantine church (and associated structures) built in part of materials salvaged from the hittite buildings. other fragments were found in the vicinity of Temple 16, presumably having been scattered in the course of its destruction and the construction of the later Byzantine buildings (though neve does not discuss how and when the fragments came to be dispersed); about twenty fragments were found within the area occupied by the neighboring Temple 15, leading neve to postulate that this building contained a library. neve infers from the distribution of the fragments in Temple 16 that the tablets were originally stored in a pithos that lay upturned by the western wall of the western cellar room. There is a disjunction between the mid-thirteenth-century date neve assigns to the group of buildings that includes Temples 15 and 16 (1984: 369–70; 1999: 11–12, with Table 1b) and the date of the tablets (ca. 1400) that was determined already by otten (1984: 375, in his appendix to neve’s preliminary report on the 1983 excavations).

JcS 65 (2013)

128 eva von daSSoW

The reconstruction presented in the editio princeps by erich neu (1996) has been followed without challenge by most other scholars.3 neu reconstructs the sequence of tablets, and of identifiable episodes, as follows (tablets are cited by their KBo 32 numbers). The proem (no. 11) is followed by the parables (nos. 12, 14); then the netherworld banquet (no. 13); next comes the senate scene, in which Zazalla makes a speech (no. 15, partially duplicated by no. 16), followed by the scene in which the god Teššub communicates to Megi, king of ebla, and Megi conveys his message to the senate (no. 19, partially duplicated by no. 20); and the last surviving episode may be represented by a fragment mentioning Purra in captivity and Pizigarra of nineveh (no. 10).4 gernot Wilhelm, however, has argued for a different reconstruction, pointing out flaws in neu’s reading of the evidence.5 The most fundamental differences are these: Wilhelm excludes the parables from the Song of release, and places the senate scene (nos. 15, 16) after, not before, the scene in which Megi communicates Teššub’s message to the senate (no. 19). i have disco-vered that on the latter point the manuscript evidence accords with Wilhelm’s reconstruction, and refutes that of neu. The content of KBo 32, 214, a small fragment of the hurrian column of one manuscript, duplicates the end of no. 19 followed by the beginning of no. 16, confirming that the senate scene follows Megi’s transmission of the divine message to the senate, rather than the other way around. furthermore, the hittite text on the reverse of no. 16 proves to duplicate the hurrian text on the reverse of no. 20, as quoted on the lacunose reverse of no. 19; thus, unless the obverse and reverse of no. 16 have been misidentified, this tablet contains excerpts from two discon-tinuous parts of the poem. if the extant exemplars include not only tablets representing segments of continuous manuscripts, but tablets containing excerpts, that complicates the process of reconstruction while elucidating the process of textualization.

Before detailing these discoveries, i discuss the possible historical background of the Song of release and the context of its textual transmission (§ 2). Then, after discussing the evidentiary basis for neu’s and Wilhelm’s re-constructions of the poem, i demonstrate how fragment no. 214 bridges nos. 19 and 16, and examine how the text of no. 16 relates to the other tablets (§ 3). next i consider the purpose of producing a bilingual edition of this poem, and the relationship of the hittite translation to the hurrian original (§ 4). lastly, i draw out the ramifications of piecing the poem together one way rather than another, focusing on the meaning of its core theme of release (§ 5).

2. The “Song of Release,” Its Historical Background, and Its Transmission in Hatti

The Song of release is a mytho-historical poem about ebla, formerly a powerful city, and igingalliš, a city ebla had subjugated. The main divine characters are Teššub, the storm god, allani, goddess of the netherworld, and išhara, goddess of oaths. The main human characters are Megi, king of ebla; Zazalla, son of fazanigar, the speaker of the senate of ebla; Purra, the leader or representative of the people of igingalliš; and Pizigarra of nineveh, who was sufficiently important that the proem introduces him immediately following the gods, but whose role remains obscure. The narrative revolves around the storm god Teššub’s demand that the city of ebla release the people of igingalliš from subjection. Teššub conveys his will to Megi, promising blessings if the city complies and utter annihilation if it does not. Megi conveys the god’s demand to the senate, where the orator Zazalla argues against effecting the release of the people of igingalliš. at some point in the story Teššub pays a call on allani, goddess of the netherworld, who holds a banquet for her guest. Their conversation is not preserved, but one may guess that the purpose of his visit was to discuss the fate of ebla; after all, allani’s realm would receive a substantial popu-lation influx upon Teššub’s fulfilment of his threat to wipe ebla out. although the end of the poem is missing, it

3. The validity of neu’s reconstruction is assumed by, for example, hoffner (1998: 65–80), otto (1998: 147–51), and Bachvarova (2005). astour considered nos. 15–16 to follow nos. 19–20, a conclusion he presents without arguing against neu and without reference to Wilhelm (see n. 5), but like others he accepted the attribution of the parables to the poem (astour 2002: 146–47).

4. See neu 1996: 16–20. he also associates no. 17, which mentions Zazalla, and no. 18, in which Teššub appears and Megi is addressed by the title “Star of ebla,” with nos. 15–16.

5. See Wilhelm 1997: 279–93; 2001: 82–85; and now 2012.

Piecing TogeTher The Song of releaSe 129

appears that the destruction of ebla must have ensued upon the senate’s refusal to obey the god’s will and let the people of igingalliš go.

The poem was originally composed in hurrian, in a dialect substantially older than that known to have been in use in the fifteenth–fourteenth centuries.6 By itself, the linguistic evidence would suffice to indicate no more precise a date of composition than “long before 1400,” as the textual record of hurrian is too thin to establish a chronology of dialectal development; moreover, epic poetry normally employs a form of language substantially older than current speech (as archi has observed, 2007: 189). Based on the broad historical points of reference the poem contains, however, it was most likely composed in the early sixteenth century bce. its historical reference points consist principally in the conjunction of the two cities, ebla and igingalliš, in a hurrian poem that became known in hatti; this points to the end of the Middle Bronze age as the period when the poem originated. While igingalliš is attested, under the form agagališ, as a tributary (perhaps) of pre-Sargonic ebla, to bridge the centu-ries separating that historical context from the nascent hittite kingdom’s involvement in the area would require evidence that is unavailable or postulates that are unsustainable.7 Middle Bronze age ebla was destroyed around 1600 bce,8 and hattušili i visited igingalliš around the same time. igingalliš was one stop on an itinerary through northern Syria that commenced with the destruction of alalah, according to hattušili’s annals (preserved only in late editions); although the hittite version of the annals indicates that he destroyed igingalliš, the akkadian version does not say what he may have done there.9 The contemporaneous textual and archaeological evidence provides no basis for connecting the destruction of Middle Bronze age ebla with hattušili’s visit to igingalliš, nor can either event be dated with precision adequate to establish or refute a link between them, but the conjunction of both cities in the Song of release suggests the possibility of a correlation.10

none of the human protagonists of the poem can be identified as a historical figure. no other extant source mentions a Zazalla, son of fazanigar, or a Purra of igingalliš, and the very name of Pizigarra the ninevite is other-wise unattested. Three kings are named in the surviving portion of a passage (no. 20, i) that evidently relates the succession of nine kings, three in igingalliš and six in ebla, preceding Megi, the tenth, upon whom Purra now attends (no. 20, iv 16’–20’; no. 19, i/ii 3–10).11 The three whose names are preserved, arib-ebla, Paib-ebla, and

6. neu (1996: 5–6) summarizes the linguistic evidence, on the basis of which he dates the hurrian composition to ca. 1600 bce. The fact that on linguistic criteria the hittite text dates approximately two centuries later (3–5) would suffice to confirm that the hurrian text represents the original composition and the hittite the translation, were there any doubt.

7. attestations of agagališ in the pre-Sargonic archives of ebla are collected in RGTC 12/1. astour (1988) argues strenuously for an ex-pansive vision of ebla’s empire during the “archives” period, and for the view that at that time agagališ was not merely tributary but “a district directly controlled by ebla” (2002: 123); from the Song of release (alone), he infers that agagališ (igingalliš) had however become independent from ebla by the end of the Ur iii period (124), the moment to which he prefers to assign the events that gave rise to the poem. Bachvarova (2005: 54) conjoins the cult of deceased royal ancestors at ebla with that practiced in hatti to support the proposition that the poem concerns the release of the people of igingalliš to serve ebla’s royal dead; see further below, § 5.

8. Paolo Matthiae gave the date for the destruction of Middle Bronze age ebla (Tell Mardīh iiiB) as 1650–1600 (on the “middle chronol-ogy”), attributing the event “with high probability” to hattušili i or Muršili i (RlA 5: 15, s.v. ibla; also Matthiae 1997: 379). he has lately reas-signed responsibility for destroying ebla to Pizigarra of nineveh, postulating an alliance between this character and Muršili i (2006: 46–47), a scenario the scant evidence available cannot support.

9. The akkadian and the hittite versions of the annals of hattušili i diverge at many points. editions of both are given by imparati and Saporetti (1964); the akkadian version has most recently been edited by elena devecchi (2005), and the hittite version by Stefano de Martino (2003). The statement, “from Uršu i went to igingalliš, from igingalliš i went to Tašhiniya,” is followed in the akkadian version by, “Upon my return i destroyed Uršu” (obv. 9), and in the hittite version by the more expansive statement, “i destroyed these lands” (i 19).

10. neu (1996: 11–12) and haas (2006: 177) likewise posit that the poem originated in the sixteenth century, referring to the destruction of Tell Mardīh iiiB and to the campaigns of hattušili i, but they relate the latter events to the poem by way of hattušili’s liberation (amar-gi 4) of enslaved people in hahhum; on this subject see further below, § 5.

11. The text on the obverse of no. 20, of which only the hurrian column survives, is elucidated by Wilhelm 1997: 289–91. almost twenty lines of col. i are sufficiently preserved to show that this passage is part of a series of identical statements about each of a succession of kings. The transition from one king to the next is marked by a statement that “they raised rn to the throne as king” (rn=n evern=a kišh e=n=a ag=id=o), which appears twice in the preserved portion of the text (i 4’, 16’), followed in both instances by a statement of the number of years that rn was enthroned. in the only instance of the sequence that is fully preserved (but for the ends of lines), the text continues with a sentence mentioning Purra and Teššub, then states that “in the nth year” rn did something, upon which the succession to rn2 evidently ensued, for the next fully

130 eva von daSSoW

ešeb-abu, would presumably have ruled ebla, but no external evidence from any period corroborates the exis-tence of these kings. The appellation of ebla’s current king, Megi, recurs in a series of diverse sources during the early second millennium, normally in the form Mekum (or genitive Mekim), and most often as a designation for the ruler of ebla. The distribution of attestations has given rise to the view that in origin Mekum/Megi was not a name but a title.12 on this interpretation, the Megi of the Song of release could simply be the current holder of the office—in whatever period.

The fact of the poem’s transmission, translation, and preservation in hatti is a stronger indicator of the histo-rical context in which it originated than is the internal evidence of its contents or its hurrian dialect, neither of which can be securely pegged to precise external points of reference. it is a matter of conjecture exactly what histo-rical episode may have given rise to this poem.13 it appears most likely, however, that the Song of release presents a mythic aetiology explaining why the city of ebla was destroyed around the time it came within the purview of the hittite realm.

Why and how the poem came to be preserved on clay is another matter. Sometime around 1400 bce, hittite scribes working in hattuša recorded this hurrian poem in writing and translated it into hittite.14 They wrote the original and the translation side by side, like a loeb classical edition in cuneiform. hittite scribes also transcribed much other non-hittite material, but few compositions received a bilingual edition. another one that did was a hurrian collection of parables, parts of which (KBo 32, 12 and 14) were found together with the fragmentary manuscripts of the Song of release (see § 3). The reason for producing bilingual hurro-hittite editions of these particular texts may have been that they were used in language instruction—that is, to train hittite scribes to read and write hurrian. This hypothesis is developed below in § 4, after examining the poem’s contents and their sequence.

hittite scribes made little further use of this poem, on present evidence, after producing the bilingual edition of it that was left in the basement of Temple 16. Until quite recently, not a scrap of the Song of release had been identified in any of the other tablet assemblages found at hattuša. one tiny fragment of what may be a later version of the poem has now been published, while another fragment may refer to it; the original findspot of neither is known. ABoT 2, 247, which features sign forms diagnostic of a thirteenth-century date, mentions Megi in ll. 1’ and 2’ and Zazalla in l. 3’; what little text is preserved does not appear to match any passage of the poem that is extant in the manuscripts from Temple 16. KBo 57, 180, a fragment of a cultic text, may read pa-ra-a tar]-nu-ma-aš ŠÌr

comprehensible statement is again “they raised rn2 to the throne as king” (rn2=n evern=a kišh e=n=a ag=id=o; on the form ag=id=o, an old hurrian 3pl. ergative, see giorgieri 2000: 227, 229–30).

in the hurrian text of the passages that sum up the nine kings Purra has served (no. 20 iv 16’–21’, repeated in no. 19 i/ii 3–10), the cities igingalliš and ebla both appear in the dative: igingallišša (< igingalliš=va) and ebla=va (Wilhelm 1997: 286–87, n. 47; cf. neu 1996: 402–5). The text thus says Purra fed three kings “for” igingalliš and six “for” ebla, not kings “of ” each city; whatever this means regarding the relation of the kings to the cities, one may assume, given the names arib-ebla and Paib-ebla, that the three predecessors of Megi whose names are preserved in no. 20 did rule ebla.

12. The instances so far known have been collected and discussed by Tonietti (1997), who also proposes a plausible derivation from Semitic mal(i)kum. The fact that, in the Song of release, Megi takes the hurrian relator (article) -ne-, as proper names do not, indicates that it is used as a designation or title rather than as an individual’s name; see Wilhelm 1999: 413, with n. 8.

13. among the scenarios that have been proposed, the most elaborate are those offered by astour (2002) and Matthiae (2006), who situate the events of the poem at opposite ends of the Middle Bronze age while distorting the evidence in equal measure to do it (see above, notes 7–8). i will investigate the reconstruction of the poem’s historical context in more detail elsewhere.

14. There is nothing to recommend Matthiae’s idea that “the translation into hittite of the poem was requested by Tudhaliya i” in order to draw the attention of his contemporaries, especially the kings of Mittani, to the conquests that Mursili i had (putatively) achieved in alliance with hurrians (Matthiae 2006: 47, citing Jacques freu and Stefano de Martino for the suggestion that Tudhaliya ordered the translation). for this to be plausible one must suppose that Tudhaliya had the Song of release broadcast to an international audience somehow. Yet the poem was sufficiently unimportant to the hittite royal administration that—with the exception of the two recently published fragments discussed immediately below—it was neither represented among the tablets kept in the palace and main temple, nor mentioned in the tablet catalogues found there (as Wilhelm has pointed out; 2001: 83). The poem was not “jealously kept in the capital of the hittite empire until its destruction” as Matthiae states (47); rather, after its translation ca. 1400, virtually the only copies that were produced were apparently left unattended in a jumble of other material in the basement of Temple 16.


Š[Ìr-RU, “they sing the song of para tarnumar,” in l. 4’; if indeed the line refers to the Song of release, it would attest the recital of this poem in a cultic context.15 it might however refer instead to a “song of release” like the one sung in a festival described in KBo 37, 68, discussed by franca Pecchioli daddi (2001), although the preserved text does not use the phrase para tarnumar. Thus, one fragment attests the poem’s textual survival, and the other may attest its survival in song, but probably in forms different from that represented by the bilingual edition.

3. Reconstructing the Sequence of the Poem

i now turn to the problem of how to ascertain which tablets and fragments belong to the Song of release, and in what sequence to arrange them. for the purposes of this discussion, it must be borne in mind that the compo-sition is not divided into fixed segments represented by numbered tablets. instead the distribution of text across tablets varies from manuscript to manuscript, apparently freely; so does paragraphing and lineation. Therefore, rather than thinking of the composition as consisting of Tablets i, ii, etc., each tablet having specified content, one must think of it as a work whose text may be divided at will, or as needed, depending on factors extrinsic to the composition.

leaving aside physical joins, the evidence for identifying and sequencing the material is of the following kinds, in order by strength:

1. colophons, which give the composition’s title and the number of the tablet in the manuscript at hand;2. duplicate content;3. Similar content, especially the appearance of dramatis personae known from this composition and no

others;4. Similar format, specifically, the bilingual “loeb edition” format: hurrian column on the left, hittite trans-

lation on the right, a double ruling down the middle separating the two, and horizontal paragraph rulings spanning both columns;

5. narrative logic, that is, what should follow what for the story to make sense.

The last two indicia are the weakest, for different reasons. any bilingual hurro-hittite text could have the same format; to attribute every tablet exhibiting this format to the Song of release would require the a priori assumption that this was the only composition to be recorded in a bilingual edition. The appeal to narrative logic entails the interference of the interpreter’s assumptions, and thus invites reading passages of the text to conform to the assu-med framework, in a circular process that tends to reinforce preconceived ideas rather than illuminating the sense of the narrative. hence different indicia must be used in combination; even the evidence of colophons requires supplementation when they are incompletely preserved.

Using these various types of evidence, neu (1996) and Wilhelm (1997, 2001, 2012) have developed differing reconstructions of the Song of release. Table 1 presents both reconstructions synoptically, together with the evi-dence of the surviving colophons. roman numerals are used to number the tablets according to their sequential position in the composition, based either on the evidence of colophons or on the hypothesis of reconstruction; in the latter case they are placed in square brackets. Since the manuscripts differ in their division of the text, a tablet’s number in a given manuscript (i, ii, etc.) bears no fixed relationship to its textual content. Therefore to designate a tablet or fragment as representing “Tablet n” is not to define what “Tablet n” of the text contains, but

15. So Jared Miller observed in the summary of KBo 57’s contents (2007: ix–x). KBo 57, 180 is transliterated in groddek 2011: 94. The transliteration of ABoT 2, 247 given in akdoğan 2010: 124 may be filled out slightly on the basis of the photograph provided on the web site of hethitologie Portal Mainz: 1’) [ … ]⌈x iMe-gi⌉[ … ] 2’) [ … ]x ma x iMe-gi x[ … ] 3’) [ … ]x ŠA iZa-a-za-al-la [ 0 ] 4’) [ … ]x-aš [ 0 ]; following a paragraph ruling are bits of three lines, in which the only complete sign is a new Script da. i am most grateful to JCS’s referee for pointing out both of these fragments to me.

132 eva von daSSoW

rather to indicate the position of that tablet in a particular manuscript, thus the relative position of its textual content in the composition.

Table 1. contents and sequence of the Song of release: two reconstructions (tablets and fragments cited by number in KBo 32; roman numerals indicate sequential position in mss.)

neu’s reconstruction colophon Wilhelm’s reconstructionno. 11 = i(proem)

“Tablet i, ŠÌr para tarnum[ar …,] whi[ch …]” i = no. 11 (proem)

no. 12 = ii(parables)

“Tablet ii, p[a?- … -] ⌈x-x⌉[- … ]”

(not part of para tarnumar; colophon of no. 12 may be restored otherwise; genre and content dissimilar to p.t.)

no. 14 = [iii](parables)

no. 13 = [iv](allani hosts Teššub in netherworld)

“[ … ] para tarnumar [ … ], the singer [ … ], not finished”

“[ … … , not] finished” (no. 10)

[ii] = no. 37 (fragment involving Teššub, išhara, ebla, release); no. 10 (Purra in captivity and Pizigarra of nineveh)[iii] = no. 20 (obv. records kings of ebla; rev. overlaps with no. 19)

“[ … para] tarnuma[r, not fini]shed”

[iv] = no. 19(Teššub’s speech to Megi, Megi’s report to senate)

nos. 16/15 = v(senate scene: Zazalla’s speech, Megi’s reaction)

“Tablet v, para tarnumar, not [finished]” v = nos. 16/15(senate scene: Zazalla’s speech, Megi’s reaction)

(no. 17, mentioning Zazalla, and no. 18, mentioning Megi, belong to [v])

“[ … ] para tarnumar [ … ], the singer [ …], not finished”

[vi?] = no. 13(allani hosts Teššub in netherworld)

nos. 20/19 = [vi](Teššub’s speech to Megi; Megi’s report to senate)

“[ … para] tarnuma[r, not fini]shed”

no. 10 (Purra in captivity and Pizigarra of nineveh)

“[ … … , not] finished”

among the tablets with colophons that survive at least in part, the title (ŠÌr) para tarnumar is wholly or partly preserved in only four (nos. 11, 13, 15, and 19), and of these four, the tablet’s number in the manuscript is preser-ved in only two: the colophon of no. 11 identifies it as “Tablet i, song of para tarnum[ar …,] wh[ich … ],” and the colophon of no. 15 identifies it as “Tablet v of para tarnumar, no[t finished].”

in addition, one of two tablets containing parables (nos. 12 and 14), which were found in the same locus (Temple 16) as most of the fragments of para tarnumar, bears an incompletely preserved colophon that has prompted attri-bution to the Song of release. no. 14 is a bilingual hurro-hittite tablet, almost completely preserved; no. 12 is a fragment of the left half of a tablet, containing part of the hurrian column of what was presumably also a bilingual hurro-hittite tablet, and it has the colophon “Tablet ii of ⌈x⌉[- … ]⌈x-x⌉[- … ].” neu restores this colophon to read “Tablet ii of p[ara tarnuma]r, no[t finished]” (dUB ii.KaM p[a-ra-a tar-nu-ma-a]š Ú!-U[L QA-TI]), and accor-dingly identifies no. 12 as Tablet ii of the Song of release; since no. 12 ends with the standard paragraph intro-


ducing another parable, and no. 14 begins immediately with a parable, no. 14 could then be identified as Tablet iii of the poem.16 however, the surviving traces hardly compel the restoration of para tarnumar in the colophon of no. 12, and the content of nos. 12 and 14 has no organic relationship to any part of this composition. it is on the contrary dissimilar in several respects, as Wilhelm observes, pointing out further that the bilingual format is abandoned on the reverse of no. 14, which has no colophon (2001: 84; 2012: 159). Wilhelm accordingly excludes the tablets of parables from the Song of release. indeed, scholars studying the Song of release have produced fairly forced and speculative arguments to explain why the parables should be part of this composition.17 Such interpre-tive gymnastics are unnecessary if they are not attributed to it.

The surviving colophons of tablets that do belong to the Song of release thus yield two fixed points in the com-position’s sequence: the proem, represented by no. 11, is Tablet i, and the scene featuring Zazalla’s oration in the senate, represented by nos. 15 and 16, is Tablet v. The contents of nos. 15 and 16 overlap: the first seventeen sur-viving lines of no. 15 are duplicated by no. 16 ii 14–31. Between the proem (Tablet i; no. 11) and the senate scene (Tablet v; nos. 16/15) neu placed the tablets of parables, nos. 12 (Tablet ii) and 14 (Tablet [iii]), and the episode of Teššub’s visit to allani in the netherworld narrated in no. 13, which would represent Tablet [iv] (neu 1996: 18, but with reservations). as well as excluding nos. 12 and 14 from the poem, Wilhelm placed no. 13 after the senate scene, as Tablet [vi(?)], rather than before.

as noted at the outset, besides the inclusion or exclusion of the parables, the main point on which neu’s and Wilhelm’s reconstructions differ is the placement of no. 19, the beginning of which is partly duplicated on the reverse of no. 20, in relation to Tablet v (nos. 16/15). it will therefore be useful to describe the contents of these four tablets (nos. 15, 16, 19, and 20) before continuing to explicate the reconstruction process.

• Nos. 16 and 15: no. 16 is a large fragment of a tablet’s top right quadrant; as such, it preserves mostly the hittite translation of the poem, with only scraps of the hurrian column on the obverse. The text on the obverse starts with the introduction of Zazalla as a powerful orator whom none can best in argument. Zazalla speaks to Megi in the senate, rebuking him for counseling submission. The next several para-graphs become increasingly fragmentary until the break, but are duplicated by the first preserved para-graphs on the obverse of no. 15. no. 15 is the bottom half (roughly) of a tablet, and it preserves both the hurrian and hittite columns (with lacunae). following Zazalla’s initial rebuke to Megi (no. 16 ii 12–13), the next paragraph, poorly preserved in both tablets, refers to Teššub and ends with “he responded,” Zazalla probably being the subject (no. 16 i/ii 14–16 = no. 15 i/ii 1’–3’). Then, Zazalla sarcastically asks Megi whether Teššub himself is oppressed, that he should demand release (kirenzi / para tarnumar).18 he goes on to declare that if Teššub were in debt, or hungry, naked, or parched, the entire membership of the Senate would readily aid him; but they will not grant release for the people of igingalliš. for who then would serve them? The people of igingalliš wait on them, cook for them, and do their laundry. Megi can release his own slaves if he likes, even send away his son and his wife, but the senate will not grant the requested release. hearing this, Megi goes weeping before Teššub to report on the outcome of his meeting with the Senate; he purifies himself and casts the sin upon his city … and then the tablet breaks off.

The reverse of no. 16 contains part of three paragraphs of the hittite column followed by, apparently, eight lines’ worth of blank space at the end of the tablet. not a single complete line survives of the preser-ved text, but none of it corresponds to any part of no. 15, as it would be expected to do, given the over-lap between the two tablets. after a passage mentioning a singer’s pay, a millstone, and there not being

16. neu 1996: 17, 72. it should however be noted that neu expresses reservations about including the parable tablets, stating that they may belong to a different series (17, 18).

17. for example, neu (1996: 9–10) shoehorns the didactic exempla of reprehensible behavior into the story revolving around the theme of release from servitude by invoking the context of public performance, while otto (1998: 150–51) interprets the Song of release as a cautionary tale about the consequences of disobedience and ingratitude, these being the theme of the parables.

18. This sentence (no. 16 ii 17–18 = no. 15 i/ii 4’–5’) is a rhetorical question, not a declarative statement, and serves as an ironic opening to the ensuing counterfactual scenario of the god himself being in debt, etc., as Wilhelm has explained (1997: 280–83).

134 eva von daSSoW

release, the very last lines of no. 16 (iii 15’–18’)—of each of which only the last word remains—duplicate either the first four lines of no. 19 (i/ii 1–4) or their quotation 75 lines later (19 iv/iii 26’–29’).

• Nos. 20 and 19: no. 20 is a fragment from the middle of the left side of a tablet, preserving only the hur-rian text. no. 19 is a virtually complete tablet, though full of large lacunae, and it preserves both the hur-rian and the hittite columns for much of the obverse and part of the reverse. at the start of no. 19, Teššub is addressing Megi; his speech was evidently in progress at this point (see below, on no. 20). Teššub tells Megi to release the people of igingalliš and, in particular, the captive Purra, who has served nine kings and who now stands before the tenth, Megi. Teššub promises blessings if ebla does effect the release, and threatens to destroy ebla utterly if it does not. The threatened destruction is described in detail until, toward the end of the obverse, lacunae interrupt the text. at that point, Teššub’s speech seems to have concluded and Megi begins speaking, but this transition is mostly lost, with the loss of the entire hurrian column and most of the hittite column on the top two-fifths of the reverse. Where we can pick up the thread, Megi is speaking to the senate (iii 1’–3’), quoting the speech Teššub addressed to him, which he has not quite finished doing by the end of the tablet.

The first ten lines of no. 19 are duplicated on the reverse of no. 20 (iv 15’–21’); curiously, what little remains of the next lines in no. 20 (iv 22’–24’) does not correspond to the next lines in no. 19. nothing that could be understood as introducing direct speech is found in the preceding lines of no. 20 (so far as preserved), nor in the poorly-preserved recapitulation of these lines in Megi’s quotation of Teššub’s speech farther on in no. 19 (iii/iv 9’–25’), but for an occurrence of the quotative particle in the hittite column (iii 9’, and also in l. 26’). The text preceding the duplicated passage on the reverse of no. 20 should therefore be part of Teššub’s speech to Megi; in the absence of the hittite column (but see below, § 3B), these paragraphs are difficult to understand. The obverse of no. 20 preserves part of a long section relating the series of kings on whose behalf Purra has served, as mentioned above (n. 11): they add up to nine kings, and now Megi is the tenth, as Teššub declares to Megi (no. 20 iv 15’–21’ = no. 19 i/ii 1’–10’). The section relating the series of kings could conceivably belong to Teššub’s speech—which in that case would occupy at least one entire tablet—but if so, Megi did not quote this section to the senate.

Because the last few lines on the reverse of no. 16 duplicate the first four lines of no. 19, neu placed no. 19, in which Teššub communicates his demand to Megi and Megi communicates it to the senate of ebla, after Tablet v (see neu 1996: 19, 283–84, and 286). Since the first ten lines of no. 19 are duplicated by no. 20 iv 15’–21’, neu recognized that the content of no. 20 precedes that of no. 19 (1996: 19, 453–55); given his determination that no. 19 follows Tablet v, both tablets would then represent Tablet [vi] (though he is inexplicit on this point). however, the overlap neu observed between the end of no. 16 and the beginning of no. 19 would effectively disallow the overlap between nos. 20 and 19, as the missing part of no. 16 does not suffice to encompass the entire contents of no. 15 followed by the entire contents of no. 20.

Wilhelm argues that nos. 20 and 19 must instead precede Tablet v (nos. 16/15), primarily on the basis of narra-tive logic (1997: 292–93; 2012: 160–62). in nos. 20/19, Teššub commands Megi to release the people of igingalliš, promising blessings if ebla grants the release and utter destruction if the city refuses, and Megi conveys the god’s message to the senate. in nos. 16/15, Zazalla responds to Megi and declares the senate’s refusal to release the people of igingalliš, whereupon Megi goes weeping before Teššub and explains that not he but his city refuses to grant the release. The story only makes sense this way around.19 neu’s argument that no. 19 follows no. 16 is premised only on the recurrence of the last lines of no. 16 at the beginning of no. 19. This is a passage of Teššub’s speech, “Set the sons of igingalliš verily free! Set free Purra, the captive, who has served nine kings!” which is also repeated

19. The question that is open in no. 19, namely, whether or not ebla would grant release, is decided in no. 15, as Wilhelm emphasizes (2012: 161).


when Megi quotes Teššub’s speech on the reverse of no. 19. in fact, based on comparison of the hittite text on the reverse of no. 16 with the hurrian text on the reverse of no. 20, and comparison of both with the surviving text on the reverse of no. 19, the last lines of no. 16 prove to duplicate not the first lines of no. 19 but their quotation on the reverse of no. 19 (iii/iv 26’–29’). The results of this investigation also indicate that no. 16 records not a single continuous segment of the poem but two different excerpts from it, one on the obverse and one on the reverse (unless perhaps the two sides are to be switched around). Meanwhile, Wilhelm’s argument from the logic of the narrative is validated by the discovery that the content of fragment no. 214, supplemented by no. 107, bridges the end of no. 19 and the beginning of the obverse of no. 16.

i first detail this discovery (a), then the relationships between nos. 16, 19, and 20 (B).

A. Fragments 214 and 107

no. 214 is a fragment from the bottom left corner of a tablet, and as such it contains part of the hurrian column only. The preserved text picks up toward the end of Megi’s quotation of Teššub’s threat to destroy ebla. The first two lines (no. 214 i 1’–2’) duplicate the last three lines of no. 19 (iv 49’–51’), which correspond to the first two lines of a passage of Teššub’s speech as given on the obverse of no. 19 (i 24–30). The next four lines of no. 214 (i 3’–6’) reiterate the next four lines of Teššub’s speech, and they would duplicate the first few lines of the tablet following no. 19. There follows a blank space, then a paragraph ruling. Below the ruling, parts of four lines are preserved, the first three of which are duplicated by no. 107, a tiny fragment from a tablet’s hurrian column. These lines give the hurrian original of the first few lines of the hittite translation found at the top of no. 16, col. ii. This passage, which introduces the orator Zazalla, presumably continues on the reverse of no. 214, of which very little remains.

here follows a composite edition of the passages represented by no. 214.

Composite Edition of KBo 32, 214 (obverse) and Duplicate Manuscripts

Key:plain italics = no. 214underlined italics = no. 19, col. iv (quoting col. i) and col. i (for continuation of quotation)boldface = no. 16 (col. ii, hittite translation)doubly underlined italics = no. 107 (duplicate of no. 214 i 8’–10’ and no. 16, [col. i] 1–3)n.B.: spellings vary both within and among these manuscripts. for example, where no. 214 i 1’ has waa-[h é-tab], no. 19 i 24 (quoted in iv 49) has pa-h é-tab; where no. 214 i 3’ has ka-a-zu-uš, no. 19 i 27 has ka-a-šu-uš, but then ka-a-zu-u-uš two lines on. Most likely the space at the beginning of 214 i 5’ does not permit the plene spelling i-ki-e-ne- of no. 19 i 29. Such variations between manuscripts are not indicated here inasmuch as they do not affect interpreting the text and determining which passages correspond to which.

No. 214 I, with restorations based on duplicates

1’/iv 49’ [a-ar-ti-ma-a-an Ur]U⌈E-eb⌉-la waa-[h é-tab] 2’/iv 50’–51’ [na-ah -h u-pa]-⌈a⌉-du-uš-ša-a še-ri-tab i[n4-na-an-na-am] 3’/i 27 [a-ta-aš-ši-]i ka-a-zu-uš h u-wuú-u[š-tab] 4’/i 28 [ki-ir-h é-]ma-a du-ú-ti-ku-u-uš ti-lu[-lu-u-uš-tab] 5’/i 29 [i-ki-(e-)n]e-ma-a ma-h i-ir-re ka-a-zu[-u-uš]

136 eva von daSSoW

6’/i 30 [sí-i]l-ma-ni-tab du-ru-uh -h é 7’ (blank) _________________________________________

8’/107 1’ [ú-ru-uk-ku-un]-na al-zi-ia-al-li š[u-u-uk-ka-ni?] 16 II 1 [NU.GÁL-m]a-aš me-ma-iš-ši ku-iš me-na-ah -⌈h a-an-da⌉

9›/107 2’ [iš-t]a-⌈ni⌉ ki-wii-ra-a-ša-an ⌈ú⌉[-ru-uk-ku(-un-na)] 16 II 2 [(ŠA?) UKKI]N iš-tar-na LÚ.MEŠ ŠU.GI-aš NU.GÁL-ma-aš

10’/107 3’ [a-a]l-zi-ia-al-li du-[ … ú-ru-uk-ku?] 16 II 3–4 [me-ma-i]š-ši ku-iš me-na-ah -h a-an-ta [ … ] ar-ku-ar-ši ku-iš i-e-ez-zi

11’ ti-pa ku-ri-i[a-a-al-li?] 16 II 5 [Ú-UL] ku-i[š-ki] te-ez-zi

Translation

1’/iv 49’ and the city of ebla i shall destroy, 2’/iv 50’–51’ like a place of no habitation i shall make it, 3’/i 27 The lower town i shall smash like a cup, 4’/i 28 the upper town i shall trample in the dump, 5’/i 29 the agora inside it, like a cup, 6’/i 30 i shall crush underfoot. ___________________________ 8’/107: 1’/16 II 1 There is no speaker opposing him, 9’/107: 2’/16 II 2 Within the senate, there is no 10’/107: 3’/16 II 3–4 speaker opposing him, [no one] who 11’/16 II 5 makes an argument against him.

Notes

214 i 1’–7’: neu (1996: 490) suggests that no. 214 could be an indirect join to no. 24+216, the obverse of which bears a passage of the hittite column that duplicates either no. 19 ii 14–33 (from Teššub’s speech) or no. 19 iii 39’–51’ (from Megi’s quotation of it), followed by a break; the reverse bears a passage that does not match any part of no. 19 (nor of no. 16), or none that is intact enough for comparison. The same paragraph that occurs in hurrian in no. 214 i 1’–7’ occurs in hittite in no. 24+216 ii 10’–19’ (at which point the fragment breaks off). however, the mismatch in the number of lines this paragraph occupies in no. 214 (seven including the blank line) versus no. 24+216 (ten or even eleven) excludes the possibility that the two fragments belong to the same tablet.

214 i 5’–6’: This pair of lines appears twice in no. 19, col. i. having written the sentence in ll. 29–30, the scribe repeated it immediately below, varying his spelling, in ll. 32–33 (following a blank line and a ruling), apparently in order to fill the space resulting from the disparity in length between the hurrian text and the wordier hittite translation. The hittite translation of this paragraph (ii 27–34) requires double the number of lines of the hurrian


original (i 27–30), so the scribe probably committed dittography on purpose in order to maintain the corres-pondence in content between the two columns.

214 i 8’–11’: neu (1996: 544–45, 550–51) recognized that these lines are duplicated by no. 107, and moreover recognized a similar passage in the Kešše epic. With restorations based on the Kešše epic and no. 214, the text of no. 107 reads as follows:

1’ [ú-r]u-uk-ku-un-na a-al-zi-i[a-al-li šu-u-uk-ka-ni? iš-ta-ni] 2’ [ki-b]i-ra-a-ša-an ú-ru-u[k-ku-un-na] 3’ [a-al-]ši-a-al-le-e d[u- … … ti-pa ku-ri-ia-a-al-li?]

(given that there is blank space between l. 3’ and the break, the sentence probably concludes in l. 3’.) The similar passage in the Kešše epic (ChS i/6 no. 30, col. iv 13–15) runs as follows:

iv 13 ú-ru-u[k-k]u a-al-ši-ia-al-li šu-u-u[k-ka-ni? … ] iv 14 ⌈ke-e-wee⌉[-er-r]a-bi ú-ru-uk-⌈ku⌉ a-al-š[i-i?-i]a?-⌈a-al-li⌉ iv 15 [ … … … ]x te-i-wa⌈a-a⌉ ku-ú-li-i[a-a-a]l-li

here follows an analysis of the words of the passage in no. 107 = no. 214 i 8’–11’. all of the elements and mor-phophonological processes adduced are readily found in grammatical descriptions of hurrian (giorgieri 2000; Wegner 2007; Wilhelm 2004), with a few exceptions for which references are provided.

ur=o=kko=nna (ú-ru-uk-ku-un-na, 214 i 8’, 9’), “he does not exist,” corresponds to nU.gÁl-ma-aš in the hittite translation. it is composed of the verb stem ur-, “to exist,” with theme vowel -o- and negative suffix -kko, followed by 3s. enclitic pronoun -nna (indicating the subject in the absolute). i suggest that the word ur=o=kko (without the enclitic) may recur at the end of no. 214 i 10’, since the final sentence of the paragraph in hittite is another statement that an entity that does a specified thing does not exist.

al=ž=i=a=lli/e ((a-)al-zi/ši-ia-al-le(-e), 214 i 8’, 10’; 107 1’, 3’) corresponds to the hittite phrase memaišši kuiš menah h anda, “one who speaks against him.” it appears to be a nominalized form of a 3s. transitive verb: the stem al-, “to speak,” bearing an augment -ž- (of unknown function), is conjugated with the transitive vowel -i- and the 3s. ergative marker -a; this finite form is somehow nominalized by -lli (or -lle), perhaps a by-form of the deriva-tional suffix -nni that forms occupational designations.20 The whole word is a noun in the absolute serving as the subject of ur=o=kko=nna. hittite menah h anda may render the augment -ž-, if the augmented stem alž- carries the nuance “oppose.”

šukk=a=ni (šu-u-uk-ka-ni, 214 i 8’) is “a single one,” composed of šukk-, “one,” the theme vowel -a-, and the adjec-tival suffix -ni. if this is indeed the word to be restored, it may be understood to modify alžialli (“there is not one single speaker …”), but it does not have an exact counterpart in the hittite translation. one might seek a hurrian word corresponding to hittite menah h anda, “against,” unless this meaning is contained in the hurrian stem alž- as suggested above.

ištani (iš-ta-ni), “middle” (214 i 9’), is rendered by hittite ištarna; here it functions as a preposition to the next word (see Wegner 2007: 115).

20. i owe this suggestion to gernot Wilhelm, who adduces the parallel h avurul=le=da < h avurun=ne=da (ChS i/5, no. 78 rev. iii 18’) in support of -lli < -nni. The normal suffix used for nominalizing a finite verb form is -šše (Wilhelm 2004, §4.5.13.2; giorgieri 2000: 239–42).

138 eva von daSSoW

kivir=ra=až=a=an (ki-bi/wii-ra-a-ša-an, 214 i 9’), “the senate” (dative), is composed of kivir- (or kever-), “elder” (see neu 1996: 550; Salvini and Wegner 2004: 174) with the plural relator (article) -na (> ra, by assimilation), the pluralizing suffix -až-, the dative -va (> a, by elision), and the conjunctive enclitic -an. Semantically it corresponds directly to lÚ.MeŠ ŠU.gi-aš in the hittite translation, but hittite ištarna should govern a preceding noun (hoff-ner and Melchert 2008, §20.19). a preceding noun is here restored as UKKin, “assembly” (conjecturally prefaced by ŠA, indicating the genitive), which would supplementally explicate kivir=ra=až=a=an; that is, where the hur-rian says “amid the elders there is none” the hittite says “amid the assembly, of the elders there is none.”

du-[ … ] (214 i 10’) must begin a word corresponding to some part of the phrase that includes hittite arkuarši kuiš iezzi, “who makes an argument against him.”

tiv(e)=a (ti-pa, 214 i 11’), is tive-, “a word,” with theme vowel -a, corresponding to hittite arkuar-, “argument.”

kur=i=a=lli (? ku-ri-i[a-a-al-li?], 214 i 11’) is built on the verb stem kur-, “to answer.”21 if the restoration (based on the comparandum in the Kešše epic) is correct, the form is the same as that of alžialli, which according to the analysis suggested above is a nominalized finite verb; thus kur=i=a=lli would be “one who answers, rebuts.” The phrase combining this word with tiv(e)=a seems to be rendered in hittite both by arkuarši kuiš iezzi, “who makes an argument against him,” and by (natta) kuiški tezzi, “(no one) who speaks”; this would be another instance of the hittite translation over-explicating the hurrian wording.

No. 214 IV

on the reverse of no. 214 there remain only a few signs in the middle of the first three lines. These lines may be assumed to correspond to no. 16 i 6–8, the first three lines of the hurrian text of the second paragraph of no. 16, which is rendered into hittite in col. ii 6–10. only the ends of no. 16 i 6–8 are preserved; using neu’s restorations, plus my own suggestions, they might be combined with the remnants of no. 214 iv 1–3 as follows:

i. hurrian ii. hittite1/6 [a-a-i? a-ar-di-]ne-w[a e-gi-da waa-zu-ra]-⌈an-ni⌉ [(nu) ma-]a-an [Ur]U-ri-ma me-ek-ki ⌈me⌉-mi-iš-kà-tal-la-aš 2/7 [ x x x x x]⌈x⌉-li ⌈x⌉[ x Za-a-za-al-la] waa-zu-ra-an-ni [ … -]kán [ud-]da-a-ar a-ap-pa Ú-UL ku-iš-ki3/8 [ x x x x x-]mi-ra l[i? … ú-ru-u]k-ku [wa-ah -nu-]⌈zi⌉ [i]Za-a-za-al-la-aš-ma me-ek-ki me-mi-iš-kat-tal-la-aš9 [nu-uš-ši tu-l]i-ia-aš pé-e-di ud-da-ar-še-et10 [Ú-UL ku-i]š-ki tar-ah -zi

[if] in the c[ity] there is a powerful speaker, [one whose] word no one contra[dict]s, Zazalla is (such) a powerful speaker, [against whom,] in the place of assembly, [there is none w]ho overcomes his argument.

This hypothetical reconstruction of the hurrian text assumes, based on measuring signs and spacing, lines that generally accommodate 12–16 signs; it also assumes that the hurrian column of no. 16 had lines of similar length,

21. The meaning “answer” for a stem kur- is suggested by Wegner (2007: 186, and glossary, p. 265), based on her analysis of a form kuru=ve or kur=uva in the Mittani letter; the same stem may yield the adverb kuru/o, “again.”


with the text similarly distributed, as no. 214. The break at the beginning of no. 214 iv 1–3 would accommodate five or six signs, and again five or six signs would be missing between the preserved bits of those lines and the ends of the corresponding lines 6–8 of no. 16 col. i.

neu assumes that waa-zu-ra-an-ni must correspond to mekki memiškatallaš, “powerful speaker,” and if this is correct his restoration of the word in no. 16 i 6 must also be correct. (Would the word, analyzed as paž(i)=ur=a=nni, perhaps be based on paž(i)-, “mouth”?) it would then follow that Zazalla’s name must occur prior to the second occurrence of waa-zu-ra-an-ni, in the break following the sign li in 214 iv 2. if neu’s restoration of [Ur]U-ri, “in the city” (dative), in the hittite version of the first line of this paragraph is correct, the hurrian version must also have “in the city”; in hurrian, this would be the dative ardi=ne=va followed by the postposition eg(i)=i=da, “inside.”22 and if neu’s reconstruction of the preceding word as mān, “if,” is correct, its hurrian counterpart could be āi, as in no. 19 i/ii 11. given that the paragraph apparently concludes with a statement of nonexistence, and that no. 16 i 8 ends with -uk]-ku, no. 214 iv 3 may conclude with the hurrian word ur=o=kko—which occurs in the preceding paragraph—or else another intransitive verb negated with -kko.

no. 214 omits quoting the last few lines of Teššub’s speech, represented by no. 19 i/ii 35–42, and proceeds straight to the introduction of Zazalla, which corresponds to no. 16 ii 1ff. This omission may be accounted for by haplography: the next sentence of Teššub’s speech begins with ú-úr-ru-ki-wuú-ú-i (urru=g(i)=iffu=we, “(that) of my urrugi,” meaning unknown; 19 i 35), and the first paragraph following Teššub’s speech begins with ú-ru-uk-ku-un-na (ur=o=kko=nna; 214 i 8’ = 107: 1’). The first few syllables of both words sound the same and could be spelled the same way, so a scribe’s ear or eye could easily have jumped ahead one passage. notwithstanding the omission of those lines, the text of no. 214 shows the order in which the narrative proceeds: on the obverse of no. 214, the last passage to appear on no. 19 is succeeded by the first passage to appear on no. 16; thus, Megi’s report to the senate of Teššub’s message is succeeded by the response Megi receives in the senate from Zazalla.

B. No. 16 reverse = no. 19 IV/III 7’(?)–29’, quoting no. 20: 3’(?)–17’

as mentioned above, parts of three paragraphs of col. iii are preserved on the reverse of no. 16, followed by blank space at the end of the column, and nothing whatsoever of col. iv. on the reverse of no. 20, three paragraphs of col. iv are more or less preserved, plus bits of the preceding and following paragraphs, and nothing of col. iii. a series of correspondences between no. 20, col. iv and no. 16, col. iii indicates that they represent respectively the hurrian and the hittite versions of the same passage, right through the last few lines of no. 16 (iii 15’–18’). Those lines correspond to the first few lines of the first paragraph of no. 19 (i/ii 1–10), which duplicates no. 20 iv 15’–21’. This paragraph is part of Teššub’s speech to Megi, which Megi then quotes to the senate, as recorded on the reverse of no. 19; accordingly, no. 16 iii 15’–18’ also corresponds to no. 19 iv/iii 26’–29’. The quotative particle -wa that marks two successive sentences in the preceding paragraph of no. 16 (col. iii 4’, 5’) indicates that this passage represents quoted speech. Thus, no. 16 more plausibly duplicates Megi’s quotation of Teššub than Teššub’s speech to Megi.23 on the reverse of no. 19, the preserved bits of the first 20 lines of col. iii indicate that Megi, addressing the senate, has begun his quotation of Teššub’s speech by l. 9’: nothing remains of l. 8’, but l. 7’ ends with a 2pl. verb that may be restored as [iš-ta-m]a-aš-tén, “listen,” which Megi would be saying to the senate, and l. 9’ ends with a word containing the quotative particle ([ … ]⌈x⌉-wa-ra-aš-ta). Unfortunately for the purpose of restoring the text, what is preserved of the following lines shows that Megi—or the poet—does not always quote Teššub precisely.

22. The hypothetically reconstructed spacing would just barely accommodate e-gi-da, and i had initially wanted to restore only ardi=ne=va as an exact equivalent of hittite h appiri. however, gernot Wilhelm points out (personal communication) that “the hurrian dative has no loca-tive meaning” and to carry this meaning it should be followed by eg(i)=i=da. for such postpositions (composed of words for the body or body parts in the directive case), which follow nouns in the dative, see giorgieri 2000: 245; Wegner 2007: 114–15.

23. This was suggested already by neu (1996: 284; see especially the first paragraph of his comment on iii 4’). neu observed many of the correspondences between nos. 16, 19, and 20 that i discuss, but his observations (cited below) did not lead him to the conclusions that i draw.

140 eva von daSSoW

Meanwhile, what remains of the paragraphs preceding 20 iv 15’ = 16 iii 15’ hardly suffices either to interpret the hurrian text of no. 20 col. iv fully or to restore the hittite text of no. 16 col. iii, nor yet to restore the text on the reverse of no. 19 that inexactly quotes the former and duplicates the latter.

on the facing page i present the texts of nos. 20 iv and 16 iii side by side (lined up artificially to compensate for discrepancies in paragraph length). i now explicate the correspondences between them; thereafter i compare what survives of no. 19 iv/iii 7’(?)–29’ to these manuscripts.

n.B.: assuming that col. iii of no. 16 has the same width as col. ii (about 7.5 cm, based on photographs that include a scale), very little can be missing from the beginnings of the preserved lines, but for the first and last few. at the beginning of ll. 3’ and 4’, probably no more than two signs are missing. Thus neu’s suggested restoration of [pa-ra-a tar-nu-mar] at the beginning of l. 5’ cannot be accommodated (neu 1996: 278). as to the end of col. iii, the lacuna continues past the ruling beneath no. 16 iii 18’, but there is sufficient blank clay to the right of the lacuna to indicate that the text did not continue beyond this ruling.24

20 iv 6’ / 16 iii 2’: hittite h aššeš(-a), “hearth” (nom. pl. with enclitic -a) corresponds to hurrian h um=ni (h u-[u-um]-ni); h umni is the hurrian equivalent of hittite h ašša-, “hearth,” in 19 i/ii 36 (see neu 1996: 283 on h aššeš=a, and pp. 429–30 for discussion of h um=ni and h ašša-, with alternative interpretations for both words). not enough of the hittite text is preserved to discern the context of the intact words ši’e, “eye” (or šī(y)e, “water”), and šuġuri, “life,” preceding h u-[u-um]-ni in the hurrian text, or the partly preserved word that follows, h a-a[p-ša?- …], perhaps a form of the verb h apš-, “to turn (one’s eyes) toward.”

20 iv 7’ / 16 iii 3’: in the next line, hittite appa išh iši piandu (egir-pa BE-LÍ-ŠU ⌈pí⌉-an-du) corresponds to hurrian pend=o=n evr(i)=i=da ([pé-en-du-un] e-eb-ri-i-ta, as restored based on no. 26; neu 1996: 439 n. 4 and 493–94), “be it returned to its lord.” The verb pend=o=n is an imperative of pend-, “return,” with the valence vowel -o- indicating passive or resultative meaning, and the patient (subject) indicated by the suffixed 3s. enclitic pro-noun -n; it is rendered in hittite by the 3pl. imperative piandu and the 3n. enclitic pronoun -at- (in n=at=za). The hurrian noun evri-, “lord,” is provided with the 3s. possessive suffix -i- and the directive case marker -da, perfectly matching BE-LÍ-ŠU in the hittite version.

20 iv 9’ / 16 iii 4’: in the first line of the next paragraph, ŠA lÚnar in the hittite text corresponds to h almi=ne=ve (h al-mi-ni-bi) in the hurrian text. in one of the polyglot vocabularies from Ugarit, hurrian h almi is equated with (eZen =) ŠÌr = akkadian za-am-ma-rù = Ugaritic ši-i-ru, “song” (Ugaritica v no. 137 iii 7), or perhaps “singer” if zammāru is meant by the akkadian spelling.25 here h almi is provided with the singular relator -ne, followed by the genitive case ending -ve, “of a singer,” ŠA lÚnar. hittite kuššan, “hire,” in the dative-locative case, must cor-respond to the word preceding or following h almi=ne=ve, and lē ešzi must correspond to a word lost in the break at the end of 20 iv 9’.

This sentence, “[ … ] for a singer’s hire is there no [ … ],” and the next are probably rhetorical questions. They are parallel in structure (“for x is there no y”) but not in content.

20 iv 10’ / 16 iii 5’–7’: The first word of this sentence in hurrian is h i-ša-ar-tu, built from h iže-, “millstone,” which is represented by na4.ara5 (with the conjunction -ya) in the hittite text. The sentence ends with para tarnumar lē ešzi in hittite (16 iii 7’), which should render hurrian kirenzi followed by a statement positing its nonexistence (20 iv 10’). What comes in between is rather difficult to figure out in either language.

24. The space left free could have been partly occupied by a colophon, as neu suggests (1996: 286), or the colophon could have been writ-ten on the lower edge.

25. cf. huehnergard 1987: 97 (on Sa 197.2), “eZen = ŠÌr = zamāru, ‘song,’ here incorrectly with double -mm-.” Perhaps it is not incorrect, although the Ugaritic spelling šīru surely means “song” (not “singer”), if polyvalence was admissible in columns to the right of the logographic one. hurrian h almi occurs in the ergative case, thus as agent of a transitive verb, in the citations from Boghazköy tablets given in Ugaritica v by laroche (1968: 455).


KBo

32,

20

IV

KBo

32,

16

III

1’

[

-]⌈r

u-u-

ub⌉[

]

2’

[

]

__

____

____

____

__ 3

’ [

d i]M

-ub

i-ti ⌈

zu-ú

? ⌉-zi

-⌈x-

x-an⌉ k

a-ab

-[ …

] 4

’ [

]⌈x⌉ t

u-um

-ni-i

a-am

⌈d iM

? -ub? x

⌉[ …

]

5’

[

]⌈x⌉-

ri u

-ul-l

u-h a

-i-in

-na

⌈x-x⌉-

ti

1’

[

] ⌈x⌉ [

… a

]n-⌈

ta-r

u⌉ x

[

]⌈x⌉

6’

[

-i]n

-ni š

i-i-e

šu-u

-h u

-ri h

u[-

um]-

ni h

a-⌈a

p? ⌉-[š

a? -x]

2

’ [

]⌈x⌉ h

a-a-

aš-š

e-ša

ku-

i-e-e

[š

]

x-an

-ni

7’

[pé-

en-d

u-un

] e-

eb-r

i-i-ta

3

’ [ x

x ]

na-a

t-za

egir

-pa

BE-L

Í-ŠU

⌈pí⌉-

an-d

u 8

’ (b

lank

line

)___

____

____

____

____

____

____

____

____

_

__

____

____

____

____

____

____

____

____

____

____

__ 9

’ [ x

x x

-]li-

ni-n

i-bi

h al-m

i-ni-b

i ki-⌈

x-x⌉

-bi ⌈

x x

x x

⌉[ …

]

4’

[ x x

]⌈x⌉-

wa

ŠA

lÚn

ar

ku-

uš-š

a-ni

10’

h i-š

a-ar

-tu h

u-ul

-le-e

-bi ú

-ul-m

[i-n]

i-bi k

i-⌈re

-en-

zi? ⌉[

…

]

5’

[

] l

e-e

e-eš

-zi

na

4.ara

5-ia-w

a k

u-iš

11’

h u-u

-um

-pa-

a-ar

-ši m

a-a-

am-m

u-ri-

né-e

h u-

um-p

a-a-⌈a

r-⌉[

ši …

] 6

’ [ x

x ]⌈

x⌉-lu

-uš-

ki-z

i nu

a-p

é-e-

da-n

i-ia

12’

d iM-a

š pa

-li-ia

ši-i

a a

-ki-r

i-ni-b

i ⌈x

x x

a x⌉[

…

]

7

’ [

]pa-

ra-a

tar

-nu-

mar

le-

e e-

eš-z

i13

’ i-š

i ši-

ia-n

e-ta

8’

[

]⌈

x⌉ k

u-iš

Íd-a

z h a

-a-n

i

9

’ [

]⌈x⌉

ku-

uš-š

a-an

Ú-U

L p

a[- …

]14

’ (b

lank

line

)

10’

[

]⌈x⌉ a

-pé-

el h

u-un

-h u

-ul[-

…

-i]š

11’

[

-]ú

-i-iš-

ki-iz

-zi

__

____

____

____

____

____

____

____

____

____

____

__

12

’ [

-]⌈ia

-tal-l

a⌉-a

š Si

g5-in

13’

[

]⌈

x⌉ a

-pé-

el

____

____

____

____

____

____

____

____

____

____

____

_

14

’ [

]⌈x⌉

15’

I-ki

-in-k

al-iš

-h e-

na-a

-ma

na-a

k-ki

pu-⌈u

t⌉-ki

-⌈x

ke-e

l⌉[-ta

-a-i]

15’

[dU

MU

.MeŠ

UrU

I-ki

-in-k

al-w

a ar

-h a

aš-š

u-]⌈

li⌉ ta

r-na

16’

na-a

k-ki

-ma

i Pur-

ra-a

n a-

az-z

i-i-r

i ta

-am

-ra

[ (0

)

]

16’

[ar-

h a-m

a-an

-wa

tar-

na i Pu

r-ra

-an-

pát? e

gi]

r-pa

17’

e-bi

-ir-n

a za

-a-z

u-u-

lu-u

š-te

-ri {

u} I-

ki-in

-kal

-i-iš-

ša [

(0)

]

17’

[pí-i

a-an

-ta-a

n A-

NA

9 l

Ug

al.

MeŠ

-wa

ku-iš

]⌈a⌉

-da-

an-n

a18

’ še

-e-e

r-ri

ki-i

-ke-

e za

-a-š

u-lu

-u-š

a e-

eb-⌈

ri⌉[-

in-n

a

(0)

]

18’

[pí-i

š-ki

-iz-z

i

]19

’ U

rUE-

eb-la

-wa a š

e-e-

er-r

i še

-e-š

i-e z

a-⌈a⌉[

-šu-

lu-u

-ša

(0)

]

____

____

____

____

____

____

____

____

____

____

____

20’

e-eb

-ri-i

n-na

h i

-in-n

e-e-

ma

[e-

em-m

a-an

-sí-n

e-w

a a (0)

]

(r

est a

ppar

ently

bla

nk)

21’

Me-

e-ki

-ni-w

a a ⌈ú-

um⌉-

mi-e

-h a

a-b[

i-wa a

]

__

____

____

____

____

____

____

____

____

____

____

___

22’

[x-]

ra-b

i-na-

a[ …

]

23’

[x x

] ta-⌈a

? -ni? ⌉[

…

]

24’

[x x

]⌈w

a? ⌉[

…

]

(res

t miss

ing)

142 eva von daSSoW

The word h i-ša-ar-tu may be analyzed as h iž(e)- with the root augment -ar-, denoting iterativity, followed by the verbal morpheme (of unknown function) -t- (here > d because adjacent to a liquid) and the ergative valence mar-ker -o. To the hurrian root augment -ar- corresponds the hittite suffix -ške-, denoting iterativity, in the verb [ x x ]⌈x⌉-lu-uš-ki-zi (16 iii 6’). This verb must denote what one does, iteratively, with a millstone. Thus the entire hittite phrase “who repeatedly x-es the millstone” corresponds to hurrian h iž(e)=ar=(t >)d=o. The next hurrian word, h u-ul-le-e-bi, whose derivation is obscure, may contain some part of this phrase’s meaning or relate it to the next. in the hittite text, the clause na4.ara5-ia-wa ku-iš [ x x ]⌈x⌉-lu-uš-ki-zi serves as the antecedent of apēdani=ya, “for that one, too,” and refers to the potential beneficiary of para tarnumar.26

The following word in the hurrian text, ú-ul-mi-ni-bi, is readily analyzed as ulmi-, “maidservant,” with the singular relator -ne, followed by the genitive case ending -ve, thus “of the maidservant.” in the hittite text there is space for the maidservant, gÉMe, at the beginning of 16 iii 7’, and she would be the beneficiary of para tarnumar. Thus the whole sentence would say, “and for the one who repeatedly x-es the millstone, shall there be no release (on behalf) of that one, too, of the maidservant?”

20 iv 11’–13’: even though half the vocabulary remains unknown, the remainder of this paragraph is more com-prehensible in the hurrian text of no. 20 than in its fragmentary hittite counterpart in no. 16. no. 20 iv 11’ reads h umb(a)=ar=ži mamm=o/u=ri=ne h umb(a)=ar=ži, which can be subjected to morphological analysis but hardly interpreted: h umb(a)=ar=ži may be formed from the (unknown) root h umb(a)- with iterative root augment -ar- and derivational suffix -ži; in mamm=o/u=ri=ne, the (unknown) root mamm- bears the derivational vowel -o/u- followed by the suffix -ri (used to form nouns designating professions), then the ablative-instrumental case ending -ne.27 no word in the surviving hittite text is repeated, making it hard to see what might correspond to h umb(a)=ar=ži, except that the iterative root augment -ar- might again be rendered by the hittite iterative suffix -ške- in [ … -]ú-i-iš-ki-iz-zi (16 iii 11’; see also below, on no. 19 iii 24’). The next two lines in the hurrian text read as follows (20 iv 12’–13’): Teššob=až pal=i=a ši=i=a ag=i=ri=ne=ve ⌈ … ⌉ [ … ] eže(?) šiya=ne=da, “Teššub knows, he sees; of the one who raises … … earth(?) into the water.” The name Teššub is marked with the ergative case ending -až; the verb pal-, “know,” has the transitive vowel -i- followed by the 3s. ergative marker -a; the root šī-, “eye,” is construed likewise as a 3s. transitive verb; and the next word is a transitive participle formed of the root ag-, “rise,” with the valence vowel -i- and the suffix -ri. after the break the sentence ends with the noun eže, “earth,” in the absolutive, followed by “water,” šiye-, with the relator -ne and the directive case ending -da.28

The only part of the hittite text of no. 16, col. iii that might be identified with any part of this passage is ku-iš Íd-az h a-a-ni, “who draws from the river” (16 iii 8’), which may correspond to “the one who raises [from the water?].” although there is a ruling below no. 16 iii 11’, the next three lines should belong to the same passage and render some part of the hurrian text given in no. 20 iv 11’–13’ (the placement of paragraph rulings in manuscripts of the poem varies, and occasionally violates sense boundaries).

20 iv 15’–21’ / 16 iii 15’–18’: The last four lines of the hittite text before the final paragraph ruling correspond to the first two sentences of the next paragraph in hurrian (20 iv 15’–21’), which is in turn replicated by no. 19 i/ii 1–10. Why the scribe of no. 16 stopped there, when there was plenty of room left on the tablet, one can only guess; had the bell rung to announce the end of the examination period?

regardless of that question, the placement of the paragraph ruling below no. 16 iii 18’ is not an anomaly but replicates the placement of a ruling at the conclusion of the same sentences in no. 19 iv/iii 29’ (quoted above):

26. neu restored the verb as [wa-a]l-lu-uš-ki-zi, “(s)he praises,” i.e., praises the millstone, and understood the clause to connote slave labor (1996: 279, 285; he also suggested [tatt]aluškizi, “sells, gives away”). The copy and photograph do not compel reading -a]l-, and if this subordinate clause indeed refers to the maidservant in the main clause, the verb can hardly describe her as praising the tool of her subjection.

27. if the noun is in the absolutive it should not bear the relator, therefore the suffix -ne should instead be a case ending, as gernot Wilhelm points out to me (personal communication).

28. The spelling ši-ia-ne-ta, with -ia-, would seem to contain a form of the stem šiy- with a theme vowel -a- instead of -e-, but perhaps it merely represents an ad hoc graphic or phonetic variation.


“Set the sons of igingalliš verily free! Set free Purra, the captive, who has served nine kings!” This confirms that no. 16 iii 15’–18’ duplicates no. 19 iv/iii 26’–29’, rather than 19 i/ii 1–4; accordingly, i have restored the quotative particle -wa in no. 16 iii 15’–18’ where it appears in no. 19 iv/iii 26’–29’. Working backwards from that point to where Megi’s quotation of Teššub’s speech must begin, it is possible to match some of the few surviving elements of the preceding text of no. 19 either to the hurrian text of the passage it quotes, preserved on the reverse of no. 20, or to the hittite text of the passage duplicated on the reverse of no. 16.

Below i present what remains of no. 19 iv/iii 7’–29’, followed by notes identifying correspondences with no. 20 iv or no. 16 iii.

KBo 32 19 IV/III 7’–29’

iv. hurrian iii. hittite 7’ [ … ] [ … iš-ta-m]a-aš-tén 8’ [ … ] [ … ] 9’ [ … ] [ … -]⌈x⌉-wa-ra-aš-ta10’ [ … ] [ … ]11’ [ … ] [ … ]12’ [ … ] [ … egir?-p]a13’ [ … ] [BE-LÍ-ŠU pí-ia-an-du ] _______________________________________________________________________________14’ [ … ] [ … ]⌈x⌉-pa-aš-ša-a-aš15’ [ … ] [ … Ú]-UL wa-ša-a-an-za16’ [ … ] [ … ]17’ [ … ] [ … -]e ku-uš-ša-an18’ [ … ] [ … ]19’ [ … ] [ … -]zi20’ [ … ] [ … ]21’ ⌈iš?-x⌉[ ]⌈a?-x⌉[ … a-ki-ri-ni-bi? … ] [ … ]22’ i-ši š[i-a-ne-ta … ] [ … -]an-zi23’ (blank) [ … ] [ … ] _______________________________________________________________________________24’ ma-am-mu-u-r[i-né-e h u-um-pa-a-ar-ši … ] [ … -]iš-h a an-ni-iš-ki-zi25’ diM-aš pa-⌈li-ia⌉[ ši-ia I-ki-in-kal-iš-h e-na-a-ma] [ … sa-a-a]k-ki26’ na-ak-⌈ki⌉-il-la-⌈a⌉[-an pu-ut-ki-na [dUMU.MeŠ UrUI-]⌈ki-kal-wa⌉ ar-h [a aš-šu-] li ke-el-ta-a-i ] tar-na 27’ na-a[k-ki i]⌈Pur-ra⌉[-an a-az-zi-i-ri ] [ar-h a-ma-an-w]a tar-na iPur-ra-an[-pát]28’ tá[m-ra e-bi-ir-na za-a-zu-lu-u-uš-te-ri ] [ap-pa-an-t]a-an A-NA 9 lUgal.MeŠ-wa ku-iš29’ (bl[ank) ] [a-da-an-na] pí-iš-ki-it _______________________________________________________________________________

19 iv/iii 7’: if Megi introduces his quotation of Teššub’s speech at l. 7’ with the exhortation [iš-ta-m]a-aš-tén, “listen,” and then begins quoting Teššub by l. 9’, this passage should correspond roughly to the first preserved lines on the reverse of no. 20.

19 iv/iii 9’: neu suggests restoring ták-]ku-wa-ra-aš-ta, “if within” (1996: 390–91, 434).

144 eva von daSSoW

19 iii 12’–13’ = 16 iii 3’: The half-preserved sign at the end of 19 iii 12’ appears to be -p]a, in which case egir-p]a / [BE-LÍ-ŠU pí-ia-an-du] may be restored on the basis of 16 iii 3’.

19 iii 17’ = 16 iii 9’(?): neu observed that the word kuššan, “hire,” in 19 iii indicated some relationship with 16 iii, where the same word appears twice (ll. 4’, 9’). if the match suggested above between 19 iii 12’–13’ and 16 iii 3’ is correct, the occurrence of kuššan at the end of 19 iii 17’ should more likely be correlated with the second occur-rence in 16 iii 9’. What remains of both manuscripts indicates, however, that they contain somewhat different versions of this passage and the next.

19 iv 22’ / 20 iv 13’: The restoration of i-ši š[i-a-ne-ta was proposed by neu on the basis of comparison with 20 iv 13’ (1996: 388, with n. 10). Since the following lines in no. 19 correlate with the preceding lines in no. 20, this would mean that Megi quoted the phrases of Teššub’s speech in a different order than Teššub uttered them, at least in one manuscript of this passage.

19 iv/iii 24’ / 20 iv 11’: neu pointed out the correspondence between nos. 19 and 20, on the basis of ma-am-mu-r[i-né-e (1996: 388, with n. 11; 434). no. 19 deviates from no. 20 in not repeating h u-um-pa-a-ar-ši, which must follow in the break. if this word contains the iterative root augment -ar-, rendered into hittite by the iterative suffix -ške-, then the hittite phrase rendering h umb(a)=ar=ži in 19 iii 24’ does so using a different iterative verb than that which conjecturally corresponds to it in 16 iii 11’ (see comment on 20 iv 11’–13’, above). neu restores the verb in 19 iii 24’ as dam-mi-]iš-h a an-ni-iš-ki-zi, “he keeps on treating harmfully” (ibid. 392–93, 435).

19 iv/iii 25’ / 20 iv 12’: here too neu pointed out the correspondence between nos. 19 and 20, and on the basis of diM-aš pa-li-ia he restored ša-a-a]k-ki, “he knows,” in the hittite column of no. 19.

19 iv 26’ / 20 i: 15’: The hurrian 2s. imperative nakk=i in 19 i 1 = 20 iv:15’ is quoted in 19 iv 26’ as nakk=i=ll(a)=an, 2s. imperative with 3pl. enclitic pronoun -lla- followed by the conjunctive particle -an. This is one unambiguous instance in which the quotation does not reproduce the original utterance exactly.

19 iii 28’: The space seems insufficient to accommodate appa piyant]an, which is found in both 19 ii 3 and (based on egi]r-pa) in 16 iii 16’–17’, and which neu suggests (1996: 400) the hittite translator used in the sense of appantan, “captive, prisoner (of war)” to render hurrian aziri (loanword from akkadian asīru). So i suggest res-toring ap-pa-an-t]a-an here instead.

despite all the gaps, and discrepancies indicating that the scribe of no. 16 was working from a somewhat dif-ferent (oral or written) text than the scribe of no. 19, the evidence examined above allows the conclusion that the text of no. 16 [iv/]iii duplicates that of no. 19 iv/iii, up to l. 29’, and accordingly quotes the text of no. 20 iv[/iii], up to the first lines of no. 19 i/ii. Meanwhile, the text of no. 16 [i/]ii begins just after the point where the text of no. 19 ends, as shown by no. 214, which duplicates the last lines of the latter followed by the first lines of the former; thereafter no. 16 duplicates no. 15. The diagram in fig. 1 illustrates the correspondences linking nos. 15, 16, 19, 20, and 214 (for simplicity fragments that, like no. 107, duplicate a few lines of one or more of the aforementioned exemplars are omitted).

assuming that the manuscripts do not form a clay Möbius strip, no. 16 must be a tablet containing two separate excerpts rather than a single continuous segment of the poem. So far as preserved, the obverse of no. 16 contains part of the scene with Zazalla in the senate (but it cannot accommodate the entire scene as found in no. 15), and the reverse contains an excerpt from Megi’s quotation of Teššub’s speech in his address to the senate. Since these two scenes go the other way around in the narrative, perhaps the “obverse” and “reverse” of no. 16 should also be read the other way around, so that the order of the two excerpts matches the narrative order of the poem. it


fig.

1. d

iagr

am sh

owin

g re

latio

nshi

ps li

nkin

g K

Bo 3

2, n

os. 1

5, 1

6, 1

9, 2

0, a

nd 2

14.

dot

ted

lines

indi

cate

whe

re th

e ta

blet

or i

ts e

dge

is no

t pre

serv

ed.

146 eva von daSSoW

matters rather little for a tablet containing excerpts, but nothing prohibits reversing the identification of obverse and reverse given in the editio princeps.29

The foregoing inquiry yields other results, ones that are not altogether encouraging for the project of recons-tructing the poem. first, if no. 16 contains excerpts, so may other fragments, large as well as small. at an extreme, there may never have been a complete manuscript of the poem, though the two colophons that indicate the tablet’s number in a sequence (nos. 11 = i and 15 = v) suggest that indeed there was. Second, not only may quotation be inexact, manuscripts of the same passage may diverge rather considerably. This is demonstrated incontrovertibly by no. 19’s inexact quotation of itself and its divergence from no. 20, in the passage where Megi quotes Teššub’s speech. in duplicating the same passage, no. 16 likewise diverges from no. 19, as well as from no. 20 (which gives part of the speech no. 16 quotes). This indicates that there was no fixed text in the sense we might prefer to posit, even if the extant manuscripts do manifestly “duplicate” each other. Perhaps it is more accurate to say that they replicate a more or less stabilized text, the contents and arrangement of which could still vary (at least up to the level of the sentence); but did that text exist in oral or written form? Were the scribes who produced the extant manuscripts copying their text from a master text, or taking it down from dictation? even if paragraph rulings fall in the same places as often as not, the observable variation points to the latter. The lack of homogeneity among manuscripts indicates that they were each produced in response to a distinct recitation (like an oral exam), with the possible exception of the manuscript that included tablets i (= no. 11), [iv] (= no. 19), and v (= no. 15).

4. Poetry and Translation

The composition under discussion is designated a song, indicating its character as poetry. does this charac-terization apply equally to the original hurrian poem and to its hittite translation? This question has been asked of other “songs” rendered into hittite from hurrian, such as those constituting the Kumarbi cycle.30 it may also be asked of imported Mesopotamian literary works, such as the story of gilgamesh, of which different akkadian, hurrian, and hittite versions were produced in hatti; some such works also carry the designation “song.”31 in none of these other cases, however—on present evidence—was the composition recorded in a bilingual edition, as were the Song of release and the series of parables. Several further questions therefore present themselves. first, why were bilingual editions of these two compositions produced? other non-hittite literary works were written down in distinct monolingual editions, whether in their original language or in hittite. What different purposes account for these different textual outcomes? The answer affects the nature of the hittite translation of the Song of release: was it meant to be poetry, like its hurrian original, or was it meant to accomplish some other purpose? it should be noted that this last question does not apply to the series of parables, which is not (according to the one extant colophon) called a song. further, addressing that same question to other poems that were rendered into hittite, but not in a translation set down side-by-side with the original, would yield answers that do not automatically apply to a bilingual edition.

let us take it as given that compositions designated “song” really were sung, in their original form; it does not follow that their translations were likewise. nevertheless, in the case that significantly different versions of a composition existed in its original language and in hittite (or in yet other languages, as in the case of the hurrian gilgamesh and the akkadian Kešše), it may reasonably be inferred that the literary creativity that accounts for

29. regarding this suggestion, gernot Wilhelm (personal communication) observes that the reverse (as identified in the editio princeps) seems to bulge somewhat, and it is mostly lines of the obverse that run onto the right edge; on the other hand, there is no unambiguous instance of avoiding a line (that is, one encroaching from the other side).

30. it was first asked of the Song of Ullikummi, in the same terms, by güterbock (1951: 141–44).31. lorenz and rieken distinguish “songs,” based on their colophons, as a category consisting of imported literary works (2010: 219); not

all such imports, however, are designated “song,” notable exceptions apparently being the compositions featuring Sargon and naram-Sin of akkad. an overview of the diverse versions of gilgamesh produced in hatti is given by Beckman (2003: 42; see also archi 2007: 186–88, 197).


the differences was not confined to text production but reflects refashioning the composition prior to its textua-lization.32 That is, the songs and stories of Kumarbi, gilgamesh, Kešše, and so on had a life outside their fixation in text. But what life did the texts themselves have? a singer of tales needed no text to inform him. Scribes, not poets, produced the texts we have, for purposes that need have had nothing to do with producing poetry.33 a singer (lÚnar) is indeed mentioned in the partly preserved colophon of one tablet of the Song of release (no. 13), as well as in one fragment that preserves nothing but the bit of a colophon (no. 66, [ … par]a tarnuma[š … ] lÚnar). depending on what these colophons recorded in their entirety, they may well attest that at least one manuscript of this poem was recorded from a singer’s dictation. But even if this were shown to be so, it would constitute too slender a basis for extrapolating the claim that all texts recording poetic compositions were dictated to scribes by their authors.34 and even if they were, that would say nothing about whether or how the resulting tablets may have been used by those who composed and transmitted poetry orally.

in other words, the existence of poems does not account for the existence of written records of them, and the recording of a poem in writing cannot be assumed to have had any effect on its existence outside writing. all the more so in the case of a translation that had no evident existence apart from the composition translated: no doubt the Song of release was sung in hurrian (or it would not have existed to be written down), but the mere fact that it was translated into hittite on clay, right alongside the hurrian original, does not mean that it was ever sung in hittite. if it was, its life as a poem in hittite did not depend on its textualization as a translation from hurrian.

Since asking what such compositions were written down for ineluctably means asking what scribes wrote them down for, the question is usually answered by reference to scribal instruction or, more broadly, to the texts' use by hittite literati in the service of the court.35 The case for the use of imported Mesopotamian literature in scribal training is secure; stories of gilgamesh, for example, were part of the cuneiform curriculum that was acquired in hatti along with the script.36 it is harder to see how recording several different versions of a story in different languages would have served the narrow purpose of instruction in reading and writing, while on the other hand to gesture toward the possibility of performance at court or at banquets runs up against the problem of showing that performers required written texts.37 But the instructional purpose of bilingual editions is practically self-evi-

32. Pace Beckman (2003: 49), in whose estimation it is “not credible” that the hittite version of gilgamesh represents “a direct reflection of oral tradition,” this version is so different from the akkadian epic (in its successive forms) known from Mesopotamia that it seems impos-sible that it was the product only of written redaction, without the intervention of story-telling. on the other hand, better support is required for archi’s assertion, predicated primarily on the presence of multiple versions of the gilgamesh story at hattuša, that scribes regularly took dictation from bards (2007: 197); see immediately below. Moving beyond gilgamesh, dijkstra has compared the hurrian and hittite versions of the Song of Kešše, leading to the conclusion that the hittite version is the outcome of retelling and rewriting, not simply adaptation, of the hurrian original (2008: 215); the fragment of an akkadian version imported to egypt from hatti and found at Tell el-amarna is too small to assess its relationship with the hurrian story. for giorgieri’s (2001) comparison of the hurrian fragment of the Song of Ullikummi with the hittite version, see further below.

33. These generalizations would hold regardless of the likely existence of occasional exceptions, i.e., scribes who could sing poetry or singers who could write, such as Mary Bachvarova postulates (in press; i thank her for providing me with a draft of this article in advance of publication). The professions of singer (lÚnar) and scribe (lÚdUB.Sar) were distinct, and the proposition that the same individuals did both would require evidence.

34. as certain compositions of other kinds purportedly were, according to explicit indications in the texts themselves: for example, Kik-kuli’s instructions in horse training (cTh 284) are presented as having been spoken by Kikkuli, and the tales of illuyanka embedded within the text of the Purulli festival (cTh 321), as recorded, are presented as the words of a priest.

35. See most recently lorenz and rieken (2010), with references there. They examine several lines of evidence—colophons, tablet cata-logues, the findspots of different types of tablets in hattuša, and comparative material found elsewhere—to determine how mythological texts were employed by hittite scribes, and they conclude that imported mythological literature was used in scribal education while anatolian myths were used in ritual. Beckman similarly situates the gilgamesh material exclusively in the context of scribal instruction (2003: 37), although in discussing the hittite interest in legends of Sargon and naram-Sin, he dwells on the appeal this material had for kings and the court, rather than its curricular utility (Beckman 2001: 89–91).

36. for a detailed overview, see Beckman 1983, updated by the more recent references in the preceding notes.37. cf. Beckman 2003: 37; haas 2006: 128. Thus, on the one hand, lorenz and rieken’s conclusion that “importliteratur” functioned exclu-

sively in scribal education (2010: 229) seems too restrictive, while on the other hand, archi’s postulate that such texts were written as supports for recitation (2007: 198) requires evidence that they were indeed so used.

148 eva von daSSoW

dent. The series of parables and the Song of release may readily be understood to represent two successive stages of instruction in hurrian, for hittite-speaking scribes who already knew how to write.38 The parables represent a more elementary stage, the stage that proverbs did in the Mesopotamian scribal curriculum borrowed into hatti. Just as proverbs served to teach Sumerian grammar in the Mesopotamian curriculum,39 the parables, together with their translation into hittite, would have served the purpose of teaching hurrian grammar—as they may do again nowadays. The Song of release represents a more advanced stage of instruction, comparable to that represented by the stories of gilgamesh in the Mesopotamian curriculum; notably, fragments of an akkadian version of gil-gamesh were found in the same archaeological context as the parables and the manuscripts of the Song of release.

The hypothesis that the hurro-hittite bilingual edition of the Song of release was produced in the course of scribal education would neatly explain certain facts about the extant sources. first, there are several distinct manuscripts that exhibit a relatively standardized text but substantial variation in format, as if students were asked to reproduce the poem as a sort of final examination in hurrian. Second, the hittite translation tends to be gram-matically over-explicit, unbundling clusters of agglutinated hurrian morphemes into whole hittite clauses whose syntax can seem rather unnatural and overwrought. for example, from the passages discussed above, hurrian alžiyalli, “debater,” is “analyzed” in hittite as memaišši kuiš menah h anda (no. 214 i 8’/no. 16 ii 1; above, § 3a), and hurrian h ižardo is “analyzed” as na4.ara5-ya(=wa) kuiš [ … ]-luškizzi, “who repeatedly x-es the millstone” (20 iv 10’/16 iii 5’–6’; § 3B). hittite often takes more words than hurrian to say the same thing, but instances such as these suggest a deliberate effort to represent every element in hurrian with a corresponding element in hittite. The resulting hittite translation may sometimes be verbose and awkward, but grammatically exact, as would be appro-priate in the context of language instruction. and if instruction in hurrian was the primary purpose of translating this poem into hittite, the translation need not have been intended to be poetry.

The relationship of the hittite translation of the Song of release to the hurrian original was the subject of a study by Stefano de Martino (1999). although most of his examples derive from the parables, rather than from the poem, his investigation yields significant observations regarding differences in formulation in the hittite vis-à-vis the hurrian of these texts. as he emphasized, the two languages are dissimilar in structure, and each requires the use of elements the other doesn’t, so the match between a statement in one and its counterpart in the other can seldom be exact.40 Moreover, it may be hard to tell whether differences in formulation should be attributed to grammatical or literary considerations. features de Martino noted that are salient for such an inquiry include the following: the syntax of the hurrian text is simple and tends to employ parataxis, where the hittite instead employs hypotaxis (1999: 9); hurrian words that, presumably, had no ready equivalent in hittite were provided with a periphrastic translation that served to explain the original (10); the hittite translation tends toward sim-plification, for instance, dismembering chiasmus in hurrian into two parallel syntagms in hittite (11–12); some-times the same hurrian formulation is rendered in diverse ways in hittite (12–13); while occasionally an element was omitted in translation, sometimes an element was added, serving to clarify the meaning (15–16). altogether, in de Martino’s estimation, the translator(s) exhibited “scarsa cura nel cercare di riprodurre gli elementi stilis-tici della composizione hurrica” (17). in the case of the Song of release, if the translator(s) aimed to produce a poem in translation, it could have been hittite style, rather than reproducing the hurrian style, that was the goal. regardless, most of the features itemized above would serve the purpose of language instruction, inasmuch as they explicate and analyze the target language through the translation; thus it stands to reason that they would abound in the parables, the more elementary text.41

38. This idea originated from reading the Song of release in class with gernot Wilhelm at the University of Würzburg. Wilhelm observed that the assemblage of tablets that includes the bilingual texts reflects an instructional context, and i suggested that the parables would logically precede the Song of release in a curricular sequence.

39. This is concisely explained by veldhuis (2000).40. for example, as de Martino noted (1999: 10), hittite makes much more promiscuous use of coordinating conjunctions than hurrian,

a feature that is correlated with differences in syntax.41. Studies of hittite poetics that have examined the bilingual texts have sometimes focused on the parables (a recent example is francia


To carry forward this inquiry with regard to the poem, let us examine the well-preserved passage of no. 19 in which Teššub threatens to annihilate ebla (featured above in the study of fragment no. 214 [§ 3a], which pre-serves part of Megi’s quotation of Teššub). The hurrian and the hittite text are given clause by clause, along with their grammatical analysis, to facilitate diagramming correspondences or differences between elements of each.42 instances of the kinds of difference that de Martino observed in other passages are present in this one, too.

19 i/ii 24 city (abs.)-conj. gn (abs.) destroy-fut.-1s. erg.hurrian: ardi=ma Ebla paġ=ed=aw

hittite: Eblan h appiran h arnikmi gn (acc.) city (acc.) destroy-1s. pres.-fut.

in this clause the grammatical correspondences between one language and the other are straightforward. The hurrian absolutive, patient of an ergative verb, is rendered in hittite by an accusative, object of a transitive verb. neu observes that the hittite translator has ignored the hurrian enclitic conjunction -ma and, regarding the change in word order (“the city ebla” becomes “ebla the city”), he remarks that, “even this simple sentence with its hittite translation … illustrates very clearly an essential structural difference between the hurrian and hittite form of speech” (1996: 419). This structural difference, however, presented no grammatical problem, nor did the transformation from ergative to accusative.

19 i/ii 25–26 settle-neg.-equative liken-fut.-erg. so-3s. pron.-conj.hurrian: nah h =obad(e)=ož šer=ed=aw inna=nna=m

hittite: n=aš mān natta kuššanga ašanza n=an apiniššan iyami &-3s. nom. as not ever settled &-3s. acc. thus make-1s. pres.-fut.

here the translator chose to unpack the hurrian equative nah h =obad(e)=ož into a separate hittite clause, “as if it (were) never settled.” in the second hittite clause, “thus i shall make it,” the hurrian ergative šer=ed=aw, “i shall make resemble,” is rendered by hittite iyami, “i shall make,” complemented by the adverb apeniššan; the adverb both carries the sense of the hurrian root šer- and corresponds grammatically to inna-, to which is affixed the hurrian 3s. enclitic pronoun -nna (absolutive), which has its exact counterpart in the hittite 3s. accusative enclitic pronoun -an. The connective nu that opens each of the hittite clauses is required by the grammar of the sentence, as it is formulated, but may at the same time correspond to the hurrian enclitic conjunction -m(a) that is affixed both to the first word of the preceding verse (ardi=ma) and to the last word of this one.

19 i 27/ii 27–28 lower town (abs.) cup-equative smash-ošt-1s. erg.hurrian: adašši kaž=ož h ub=ošt=aw

2010, a study flawed by claiming to examine “figures of speech” or “figurative language” when the features actually discussed—rhyme, asso-nance, etc.—are neither). Since parables, proverbs, and the like are normally composed in verse, or in a form similar to verse, it is reasonable to treat the bilingual edition of hurrian parables as a source for such an inquiry, provided this composition is not misunderstood as a song.

42. restorations of the hittite text are based on the fragment no. 24. lexical and morphological elements of hurrian may be found in one or more of the following: giorgieri 2000; Wilhelm 2004; and Wegner 2007.

150 eva von daSSoW

hittite: Eblaš h appiraš katteraš wah nueššar arh a teššumiyaš iwar duwarnah h i gn (gen.) city (gen.) lower (gen.) circuit (preverb) cup (gen.) like smash-1s. pres.-fut.

This time the translator handled the equative in a more straightforward way, as if to elucidate its form and meaning through presenting a simpler solution to the same problem: the phrase teššumiyaš iwar represents as close to a one-to-one correspondence as possible with hurrian kaž=ož, “cup-like.” The translation of a hurrian erga-tive into a hittite transitive is again unproblematic; not so the translation of its object. for hurrian adašši, which forms a pair with kirġe, the subject of the next verse, the translator chose what would appear to be a roundabout rendering: he renders adašši into hittite as “the circuit of the lower town,” adding the specification “of ebla” (as if clarity required the exegetical genitive),43 a wordy construction that enables him, in the next verse, to use only a nominalized adjective governed by wah nueššar to render kirġe as “the circuit of the upper (town).” This expansive translation does not simply represent variation in the hittite vis-à-vis the hurrian poem. it may represent a choice to explain the hurrian word pair adašši and kirġe, rather than simply find approximate hittite equivalents for them (perhaps there were none?). it may also represent a revision for the purpose of achieving the standard hittite verse structure (see further below).

19 i 28/ii 29–31 upper town (abs.) dump-equative trample-ol-ošt-1s. erg.hurrian: kirġe tudi=g(i)=ož til=ol=ošt=aw

hittite: šarazziyaš=a wah nueššar arh a h uššiliyaš iwar šakkuriemi upper (gen.)-conj. circuit (preverb) dump (gen.) like trample-1s. pres.-fut.

The noun kirġe having been dealt with in the manner explained above, for the rest of this verse the hittite translation corresponds straightforwardly to the hurrian: a noun in the genitive governed by iwar again renders the hurrian equative, and a transitive verb renders the hurrian ergative. in both this and the preceding verse the hurrian verb form contains the morpheme -ošt- (of uncertain function), which may be represented by the pre-verb arh a that accompanies the corresponding hittite verb in each case. neu observes (1996: 425) that the clitic conjunction -a (affixed to šarazziyaš) lacks any counterpart in the hurrian text of no. 19, but has one in no. 214 i 4’, where the same verse begins [kirġe]=ma.

19 i 29–30/ii 32–34 inside-3s. poss.-rel.-dir.-conj. marketplace-rel.-dir. cup-equative crush?-fut.-1s. erg. nether-deriv.-adj.hurrian: egi=i?=n(e)=e?=ma maġir=r(e)=e? kaž=ož silm=an=ed=aw tur(i)=o=h h e

hittite: Ki.laM-ni=ma=kan ištarna pedi Ebla[n? h appiran teššu]miyaš iwar du[ … ] marketplace-conj.-ptcl. inside place (dat.) gn (acc.?) [city (acc.)? c]up like c[rush-1s. pres.-fut.]

This is the longest hurrian verse in the passage under consideration; it is written twice (being repeated in 19 i 32–33), the hittite translation having taken up so many more lines than the hurrian that the two texts got out of alignment (see above, § 3a).

in the hittite rendering of the phrase “(at) the marketplace within (it),” Ki.laM obviously corresponds to hurrian maġiri, ištarna to egi, and the enclitic conjunction -ma to its hurrian counterpart in the same position.

43. de Martino (1999: 15–16) suggested that the city’s name was added for clarification. if so, the reason for clarification can hardly have been ebla’s obscurity by the time the translation was made (as he suggests), but rather clarity of expression in hittite.


While no equivalent to the hurrian relator -ne exists in hittite, if the hurrian directive case -e is present, it may be represented by the dative pedi as well as the “local” particle -kan.44 in the second half of this verse, the transla-tor has again rendered kaž=ož as teššumiyaš iwar, and the lacunae prevent certainty about how he dealt with the rest. for the unknown hurrian verb silm=an- (containing the causative morpheme -an-), neu suggested hittite du[warnah h i], as in ii 28 (since the verb must denote something one can do to a cup), and he looked for an object to correspond to tur(i)=o=h h e, an adjectival form45 apparently serving as the patient (in the absolutive) of the erga-tive verb. Perhaps to provide an unambiguous object for the corresponding hittite transitive verb, the translator has again inserted “ebla,” perhaps followed again by “city” (as in ii 27).46

The hittite rendition of the next verse is fairly straightforward albeit only partly preserved and partly compre-hensible: ammedaz=ma=an[ … -a]ššurami translates hurrian urru=g(i)=iffu=we tal=ašt=aw, “i shall take away (that) of my urru=gi” (19 i/ii 35). The hittite rendition of the next few lines, however, expands them to more than double their length in hurrian, adding verbs as well as other words, as if to parse out the meaning of the original:

19 i 36–39 19 ii 36–42kirġe=we adašši=n(e)=e h umni šarazziy[aš=a wah nuešnaš] h aššan n=an katta katt[eraš wah nuešni] arnumi

The hearth of the acropolis into the lower town, The hearth of the circuit of the upper (town), i shall remove it down to the circuit of the lower,

adaššī=we šiye=n(e)=e katteraš=kan w[ah nuešnaš h ]aššan kattanda h api [arnumi?]

(that) of the lower town into the water, the hearth of the circuit of the lower (town), down to the river [i shall remove],

kirġe=we adašše=n(e)=e h apš=[aw] šarazziyaš=ma[=kan wah nu]ešnaš [h ]aššan kattand[a katter]i [wah nu]ešni išh u[h h i]

(that) of the acropolis into the lower town i cast. the hearth of the circuit of the upper (town), down to the circuit of the lower i cast.

The concise hurrian verses tumble forth like a sudden demolition. The hittite translation spells out the process at pedantic length.

is it also poetry? it may indeed be: for the most part, the text can be analyzed into verses having four beats each, as hittite poetry is supposed to have.47 Since the shorter hurrian verses do not exhibit such a four-beat rhythm, one could argue that the expansions in the hittite version were intended to achieve this poetic form, rather than—or

44. That is, if the hurrian phrase egi=i=n(e)=e=ma maġir=(n>)r(e)=e is correctly analyzed above. The -e- in the spelling i-gi-e-ne-ma, in l. 29, may represent the 3s. possessive suffix -i, while the geminated /n/ in the spelling i-ki-en-n[i-m]a in the repeat in l. 32 is presumably an error.

45. not “manhood,” nor “foundation,” as neu suggested (1996: 426–27), rather, “pertaining to (-h h e) (what is) beneath (tur(i), with deri-vational vowel -o).”

46. neu restored a genitive, ebla[š, and suggested it might be followed by a noun meaning “foundation” (in his translation) or else by UrU-aš (in his commentary; neu 1996: 385, 425–27). What remains of the sign before the break in l. 33 is the head of a single horizontal, which could be an as well as aŠ.

47. This structure was discerned first by güterbock (1951: 142); its characteristics were given more precise form in a study by Mcneill (1963), and have subsequently been elaborated by durnford (1971), Melchert (1998), and most recently Kloekhorst (2011), each of whom on

152 eva von daSSoW

as well as—to explicate the grammar and meaning of the hurrian poem; these are not mutually exclusive options. relevant here are the observations güterbock made upon endeavoring to assess whether the hittite version of the Song of Ullikummi was poetry. it was indeed in verse, he determined, but “not written in very good verse,” whether because of the difficulties attending metrical translation or because it was not meant as verse and “only reflects the structure of the hurrian original” (1951: 144). his assessment may need modification in light of giorgieri’s com-parison of the hurrian fragment of Ullikummi with the corresponding passage in the later hittite version, from which giorgieri concluded that the latter was a revision, rather than a direct translation, of the hurrian version.48 But in the case of the Song of release, the hittite text is clearly intended not as an independent recension but as a translation of the hurrian text, set down side by side on clay as they are, paragraphs neatly lined up with each other. regardless of whether the translator’s intent was also to render the hurrian poem into a hittite poem, to the extent that he sought to replicate the form of the original, its poetic structure would be retained in translation, as güterbock remarked in discussing Ullikummi.49 an experiment could be done to help decide the matter: someone could produce both a more literal hittite translation and a more poetic one, and see how each differs from the text we have. (Unfortunately, not knowing hittite, i myself cannot do this experiment.) notwithstanding the outcome, while efforts to analyze the poetics of the hittite text of the Song of release are not misguided, they should proceed on the premise that it is translation first and poetry second.

5. Subjection and Liberation

The central issue of the poem is the divine demand that the city of ebla release the people of igingalliš; it is for this that it is called the song of para tarnumar, “release,” in hittite. Teššub conveys the demand to Megi, ruler of ebla, in the following terms:

19 i/ii 1–23Igingalliš=h e=na=ma nakk=i release the sons of igingalliš fut=ki=na kel=d(i)=ae in well-being,nakk=i=ma Purra=n assīr=i50 release the captive, Purra,tamra ever=n(i)=a51 saž=ol=oš=t=iri who has served nine kings._______________________________Igingalliš=ša(< va) šēr(i)=ri(< ni)52 for igingalliš three kings

different grounds corrected Mcneill’s assumption (1963: 240, 242) that hittite poetic meter was derived from akkadian through the mediation of hurrian (see durnford 1971: 69, n. 3; Melchert 1998: 483–84, 493; Kloekhorst 2011: 157).

48. giorgieri 2001, esp. pp. 134, 152–53. for the most part, however, the differences he identifies, based on what little of the hurrian text is preserved, are of the same order as the differences between the hurrian and hittite texts of the bilingual.

49. curiously, the assessment of haas (2006: 297) is somewhat the inverse: he avers that verse structure tends to be neglected in transla-tions of Babylonian or hurrian poetry, that this is especially clear in the hittite version of the Song of release, and that “gebundene Sprache eines hethitischen literaturwerkes ist am besten in dem ‘gesang von Ullikummi’ zu erkennen.” Wilhelm, too, considers that “gebundene Sprache” has not been identified in the Song of release (2012: 156, n. 3).

50. This word is akkadian asīru, “captive, prisoner of war,” borrowed into hurrian and here spelled a-az-zi-i-ri; it is translated into hittite as appa piyantan, “the one delivered up” (for references see de Martino and giorgieri 2008: 116). i take the double z in the hurrian spelling to represent the allophone /ss/ (not /z/), since only by graphic gemination would the phoneme /s/ of the akkadian word be preserved.

51. The form e-bi-ir-na is not the plural of ever=ni, “king” (evri, “lord,” with the individualizing suffix -ni), which takes the form evre=n(i)=na (as in ll. 6 and 8 below), unless a variant or error is assumed. instead, e-bi-ir-na may be analyzed as ever=ni with the essive case marker -a. The “nine as king” are recipient of the action of saž=ol=oš=t=iri (“having served”), participial form of saž=ol-, literally “feed,” which may contain not the morpheme -ošt- but the past tense marker -ož- plus the morpheme -t- of uncertain function that sometimes follows it. Both of the foregoing suggestions derive from gernot Wilhelm.

52. i suggest that še-e-er-ri in ll. 5 and 7 may be analyzed as širi, “number,” with the individualizing suffix -ni, and that it may function as a measure-word preceding numbers that count individuated entities (in this case, the three and six kings). There being no counterpart to such a word in hittite, the translation inserts in its place “city of the throne” (restored from the quotation of this passage, 19 iv/ii 30’–33’). in both


kig=e saž=ol=ož=a evre=n(i)=na he has served,Ebla=va šēr(i)=ri(< ni) šež=e for ebla six kingssaž=ol=ož=a evre=n(i)=na he has served,h ennē=ma emanzi=ne=va and now, for the tenth,Megi fe=va abi=va meġ=a Megi, before you he stands.53

_______________________________āi nakk=ed=aššo kir=en=zi if you (pl.) decree release,Ebla=va šerže=ni54 mann=a for ebla the fate is (this):nakk=ed=aššo kir=en=zi you (pl.) decree release,en=ar=ġ(e)=a=lla55 h ud=ed=aw to god-like (power) i shall exaltōlmi=šši your weaponry,_______________________________ōlmi=šši nan=ed=i ad(i)=ir=ġ(e)=a56 Your weaponry will beat the opponent,h el(i)=a=ma er=ol=eva av(v)=undi=šši57 gloriously shall your field(s) thrive._______________________________nakk=i=o=(a)ššo=nna kir=en=zi if you (pl.) do not decree release,Ebla=va šerže=ni mann=a the fate for ebla is (this):šindi=šš(i)=uva=d šō=uva on the seventh dayun=eva=tta ede=ž=uda fē=ž=(< v)a i shall come upon you.

Then Teššub’s threat of destruction continues as quoted above (§ 3).What kind of subjection did the eblaites impose on the people of igingalliš, and to what status or condition

were they to be released?The infinitive para tarnumar renders into hittite the hurrian noun kirenzi, an abstract formed with the suf-

fix -ži (> zi) from the root kir-, “release.”58 The equivalence of hittite para tarnumar and hurrian kirenzi with akkadian andurārum and Sumerian ama(r)-g i 4 is well established, as is the basic meaning shared by all four terms: the restoration of persons (or real property) to their original status.59 The specific content of these terms depended on the context and the subjects to which they were applied. The andurārum that released free persons from debt servitude released slaves from the jurisdiction of creditors to their original owners (not from slavery);

occurences of širi, the vowel is lowered and lengthened (i > ē), perhaps for emphasis or metrical reasons, as also happens with the final vowel of h enni, “now,” before the enclitic conjunction -ma, in l. 9, and elsewhere in the poem.

53. Why neu (1996: 378–79, 407–8) makes Teššub enter before Megi here is mystifying. Purra is the logical subject of all the verbs from the participle in l. 4 to meġ=a in l. 10, there is no basis for restoring d[iM-aš in the hittite text of this passage, and gods do not enter before humans.

54. Since the noun šerže refers to both possible outcomes, either Teššub’s blessing (following l. 12) or his annihilation (following l. 21) of ebla, i suggest that it means something like “fate.” The form ma-a-na may be analyzed as composed of mann-, which is the base of both the verb “to be” and the 3s. independent pronoun, with the thematic vowel -a, which functions as the marker of intransitive valence or as the essive suffix, either way yielding the meaning “it is” (the stative verb mann- normally bears the marker -e/i; giorgieri 2000: 227–29, Wilhelm 2004: 112; here i thank JCS’s referee for adumbrating the identity of (pro)nominal and verbal morphemes). again the hittite translator has replaced the phrase with “city of the throne” instead of rendering it into hittite; likewise when šerže=ni appears in no. 15 iv/iii 7.

55. if this word is analyzed as an adjectival (-ġe) form of a noun en(e), “god,” augmented by -ar- (though this augment is normally applied to stems from which verbs are formed), in the essive case (-a), it remains to account for the 3pl. enclitic pronoun -lla. if ōlmi, “weapon,” the object of the ergative h ud=ed=au, is treated as a collective plural, “weaponry,” it may be the antecedent of -lla. on the 2pl. possessive forms -ž-/-šši, see giorgieri 2000: 216, with n. 139; Wilhelm 2004: 107.

56. This word, an adjectival form (-ġe) of a noun derived from a stem ad-, in the essive case (-a), must mean “opponent” or the like; it is translated into hittite with h arpanalliš, “enemy.”

57. av- is the root of words for “(arable) field” (such as avari) and a measure thereof (avih ari); the word here, avundi, is translated into hittite as a.ŠÀh aršawar.

58. The morpheme between root and abstract suffix may be the individualizing suffix -ni, metathesized between kir- and -ži (> zi after -n), following the analysis suggested by giorgieri (2000: 203): kir=i/en=zi < *kir=ni+-ži.

59. for an overview, with citations of illustrative examples and references to the pertinent literature, see von dassow 2011: 208–11.

154 eva von daSSoW

the andurārum that freed people from foreign rule restored them to native rule; and the andurārum that remitted taxes or duties restored people to a condition prior to their subjection to the state. These words thus acquired the connotation “liberation,” even if it was not their primary meaning. Their basic meaning could moreover allow seemingly paradoxical usages, for persons could be released from one kind of subjection only to be brought under another. Thus hattušili i states in the hittite version of his annals that, upon capturing hahhum, he freed (arawe-) its people from šah h an and luzzi duties and released them (appan tarna-) to the sun goddess of arinna; in the akkadian version, he says he “placed them in the house of the sun goddess of arinna, and beneath the heavens i established their andurārum.”60 This andurārum evidently consisted in exempting the people from duties to the state in order to dedicate them to the service of the goddess. in an analogous instance, Queen ašmunikkal decreed that personnel dedicated to the service of a royal mausoleum were free (arawe-) from šah h an and luzzi and that they may not be released (para tarna-), i.e., released from service to the mausoleum.61 in this case, para tarna- would have restored the dedicated personnel to secular jurisdiction, so that they could be levied for duties to the state. While the phrasing in hattušili’s annals rhetorically evokes formulae that express emancipation as cleansing before the sun god (aligning emancipation with consecration), in asmunikkal’s decree para tarna- is used in a narrowly legalistic sense to refer to the release of persons from a particular authority’s control. neither instance provides a template for the general meaning and usage of para tarna- or its semantic equivalents.

in the Song of release, the verb nakk-, “release, send forth,” is used with the noun kirenzi as patient. The phrase nakk- kirenzi is rendered into hittite by para tarnumar iya-, “to do (a) para tarnumar” (as in 19 i/ii 11, 13, 20–21, and 15 i/ii 19’–20’). hurrian nakk- alone is rendered by hittite arh a tarna- (as in 19 i/ii 1–4 and 15 i/ii 26’), while para tarna- translates kirenzi alone (as in 15 iv 2/iii 3 and iv/iii 18), or kir- alone (as in 15 iv 2–3/iii 4, duplica-ted by 216 iii 11’–12’). in the last passage cited, kir- is used as a verb in the 3s. imperative, kir=o=n(na), “be (s)he released,” with the patient “slave, slavewoman” (see below, p. 158, with n. 78), indicating that this root does carry the meaning “(set) free”—rather than only “release” of another kind, such as from one jurisdiction to another. The hurrian locution nakk- kirenzi sounds redundant, being composed of two different words that both mean “release” and that may each be used in that meaning on their own: the eblaites must choose whether to nakk- the people of igingalliš, which is the same as to grant them kirenzi, this being the act of kir-, to enact which is to nakk- kirenzi. Yet this locution has its exact akkadian equivalent in the phrase andurāram wuššurum, attested in the Mari corpus in reference to the matter of whether the king of Yamhad should decree an andurārum that would cover the ter-ritories of alahtum (alalah).62 While hurrian kirenzi is the equivalent of akkadian andurārum, hurrian nakk- is equivalent to akkadian wuššurum, “release, send forth.”63 Whichever phrase may be a calque of the other, they mutually illuminate each other by illustrating that, though their semantic ranges overlap, verb and noun function in different senses, and by indicating where the limits of the overlap between them lie; that is, the phrase does not mean “release a release” (vel sim.).64 The noun signifying “release” in the sense of “restoration to original state”

60. KBo 10, 2 iii 15–20 (hittite), and KBo 10, 1 rev. 11–14 (akkadian); see references in n. 9, above. This passage of hattuŠili’s annals is adduced by neu (1996: 11–12) as well as haas (2006: 177) in support of identifying hattuŠili’s campaigns in northern Syria as the historical context for the events behind the Song of release and its transmission to hatti. The hittite and akkadian passages are not fully parallel: both say the king took the hands of slavewomen and slaves off the tools of their labor, but the statement in hittite that he released people from šah h an and luzzi (duties imposed on subjects, not slaves) has no counterpart in akkadian, which instead says he placed them in the house of the goddess; also, the phrase “i loosened their belts” occupies a different position in each.

61. KUB 13, 8: 1–9; for an edition of the passage see otten 1958: 106–7. This text is cited along with hattuŠili i’s annals by Bachvarova (2005: 52–53), in support of her argument about the nature of the “release” Teššub demands in the Song of release (on which see further below).

62. See in particular fM vii 47: 16–28, with durand’s discussion (2002: 80–82 and 155–56). Zimri-lim’s bid to purchase alahtum was be-set by various difficulties, one of which was the issue of andurārum, and the matter is as difficult to sort out now as it must have been at the time.

63. for the equation between nakk- and wuššurum, see Wilhelm 1989: 131 (with references there) and 1992: 132.64. nor yet durand’s “donner libre cours à l’andurārum,” which requires introducing assumptions about the limited scope of a previous

andurārum (2002: 81–82, 160–61, esp. note b).


is put into effect by the action of the verb signifying “release” in the sense of “send forth.” Thus nakk- kirenzi and andurāram wuššurum may best be translated “effect a restoration” or “decree (an edict of) release.”

Because the term andurārum and its equivalents are typically used to refer to remission of noncommercial debts by royal edict, with concomitant release of persons enslaved for debt, the kirenzi at issue in the Song of release was initially interpreted as debt remission. This is the interpretation enunciated by neu in the editio princeps (1996: 9, 479) and by hoffner in his english translation of the poem (1998: 66, 75–76).65 The people of igingalliš would have been enslaved for debt to the eblaites, and Teššub would have been demanding that the city of ebla release them by instituting debt remission. This interpretation was bolstered by references to biblical derōr, the hebrew reflex of akkadian andurārum, which Yahweh required of israel and Judah occasionally, according to the prophets (e.g., Jeremiah 34), or on a regular schedule, according to the Torah (esp. leviticus 25; see neu 1993: 33 and 1996: 9, 480, n. 6). Meanwhile, internal support was drawn from the passage of the poem in which Teššub himself is described as oppressed by poverty, hunger, and perhaps debt (15 i/ii 4’–16’), apparently in an extended simile for the situa-tion of the people whose release the god demands.66 otto, however, has pointed out that the text indicates that the form of bondage to which the people of igingalliš and their leader Purra were subjected was not debt servitude but captivity in war, for Purra is called a “prisoner (of war)” (19 i/ii 3; above, with n. 50); further, otto argues, Teššub’s promise to grant ebla military success if the city released the people of igingalliš, and to destroy ebla if it did not, points to armed conflict as the background of the story.67 To these arguments it should be added that individuals and families, not entire communities, enter into debt servitude. it is apparently the whole community of igingalliš that is held captive by the city of ebla, a situation that military defeat or political subjugation would produce. Andurārum denotes release from this kind of subjection, too: examples include Sargon’s “liberation” (andurārum) of the people of Kish, consequent on defeating Uruk; naram-Sin’s “restoration of freedom” through victory over rebel forces; and Ur-namma’s “liberation” (ama(r)-g i 4) of Mesopotamian cities from the rule of anshan.68

Why did the gods desire that ebla release the people of igingalliš? Bachvarova (2005) has argued that the object of release was to dedicate its beneficiaries to the service of Teššub, whose cult she supposes to have been neglected, or to the service of ebla’s royal dead. To support this interpretation she adduces a hattic song involving exemption from šah h an and luzzi duties; an old hittite instruction text that enjoins the king to take care of his subjects; a Middle hittite prayer featuring claims that the gods had been neglected while people were oppressed with šah h an and luzzi; the passage of hattuŠili’s annals according to which he established the andurārum of people of hahhum and dedicated them to the sun goddess of arinna; and Queen ašmunikkal’s decree establishing that personnel of the royal mausoleum are exempt from šah h an and luzzi and may not be released. That part of Bachvarova’s argu-ment that relies on the latter two passages (discussed above in this section) is undermined by misapprehending andurārum (ama(r)-g i 4) to refer only to release from debt slavery (48, 53), as well as by misconstruing para tarna- to mean “change (someone’s) status” (53, with n. 32). To throw a bridge between ašmunikkal’s decree and the Song of release, she adduces “the archaeological evidence for a royal funerary cult at ebla” (54) and points to the passage of the poem that relates the succession of kings Purra has served (partly preserved in no. 20, obverse;

65. Before studying the poem in the original, i too followed this intepretation, in my entry on hurrian for the new Encyclopedia Judaica (von dassow 2006: 630), which no one should read because the editor neglected to do so, causing the first draft of the entry to be printed with diacritics mutated and without the corrections submitted one week later.

66. neu (1989: 346–49, 353–54) develops this element of the interpretation at some length. Yet nothing in the text explicitly describes Teššub as a debtor.

67. otto 1998: 149–50. The validity of the latter argument is debatable, for Teššub also promises to make ebla’s fields fertile, from which one would not infer an agricultural background for the subjugation of igingalliš; moreover, the storm god similarly adduces his grant of military success in enjoining Zimri-lim of Mari to do justice and hear his subjects’ legal cases (fM vii, 38; durand 2002: 134–35), but not because those cases arose from warfare Zimri-lim waged. in addition, it is not clear why otto considers it inappropriate that debt slaves should wait on kings.

68. These examples are drawn from the legend of the great revolt against naram-Sin (ll. 16–18; Westenholz 1997: 240–43); naram-Sin’s inscription celebrating his suppression of the great revolt, preserved in old Babylonian copies (foster 2005: 59–62, i.3c §9); and the laws of Ur-namma (roth 1997: 16).

156 eva von daSSoW

see above, § 2, with n. 11, and § 3). She asserts that Purra would be released to serve this cult of ebla’s royal dead, while the rest of the people of igingalliš would be dedicated either to that cult or to the service of Teššub.69

But according to Bachvarova’s interpretation Purra was already “tending to the funerary cult of a series of kings of ebla” (2005: 54); he did not need release to keep on doing that. further, the text provides no basis for differenti-ating the release of the people of igingalliš from that of Purra, their leader or preeminent representative, and attri-buting them different destinies. Bachvarova proceeds to adduce a passage of the Song of h edammu, as well as the story of atrahasis, to prove that the gods do indeed suffer in the absence of human toil, therefore it must be that the people of igingalliš are released in order to toil for the gods (54–55, with n. 33). But what is the basis for asserting that Teššub’s cult has been neglected? This idea is premised on interpreting Zazalla’s response to Megi—“if Teššub is under duress, that he requires release” (15 i/ii 4’–5’; see further below)—as a true description of Teššub’s actual circumstances. (he may be able to crush ebla underfoot, but he is hungry and needs to be fed by the very people he proposes to obliterate and do without.) if one understands Zazalla’s speech to contain a series of conditional statements, as Bachvarova does (50), it follows that these are not presented as statements of fact; and no part of the text speaks of Teššub’s cult, its neglect, or its need for staff.

finally, the clinching metaphor adduced by Bachvarova and others to sustain the interpretation of this passage as a description of Teššub’s actual state of want, “the god is man,” is simply not present in the text. Where the hurrian text has ene, “god,” in the absolutive case, at the conclusion of a statement of what Zazalla and his collea-gues would do for Teššub (15 i/ii 12’ and [restored] 16’), the hittite text has the sign sequence an UŠ Un. This sequence has been read as dingir-uš (šiuš, “god,” nominative) Un (antuh šaš, “man,” nominative) and interpreted as a sentence, “the god is (a) man”; in order to make the hurrian text say the same thing, neu has filled in all-too-small breaks at the ends of no. 15 i 12’ and 16’ with e-ne [(ma-a-an-ni) tar-šu-wa-a-ni] (neu 1993: 349, with n. 67; 1996: 315–36). But there is no room for that in the hurrian column (in the first instance, ene is followed by blank clay!), and no warrant for that interpretation of the hittite. The sign sequence at issue is more simply understood as a spelling that uses one phonetic complement, -uš, to indicate that the logogram an is to be read as šiuš, “god,” and another, -un, to indicate that the word is in the accusative.70 Thus, where the hurrian has an ergative verb with its patient in the absolutive, itt=i=l=eva=ž alal(i)=ae ene, “we shall clothe with a garment the god,” the hittite has (as usual) a transitive verb with its object in the accusative: n=an kuišša TÚgkušišiyaz waššaweni šiun, representing the object with the enclitic -an while replicating the noun’s position at the end of the sentence; this unusual syntax may account for the unusual spelling. While the solution i propose may be ad hoc, it has the virtue of yielding a real correspondence between the hittite and the hurrian text. Moreover, albeit my graphic explanation of the odd sign sequence rests on a thin foundation, the connotation-laden reading of it proposed by neu rests on none, while carrying weighty implications not only for the interpretation of this text but for theology in the hittite and hurrian world.

casting the storm god as a suffering human has been a key element of most interpretations of the Song of re-lease to date. The one that departs farthest from the poem’s actual text is perhaps that of haas (2006: 177–85), who restores the colophon of Tablet i (no. 11) as “first tablet: Song of the release o[f Teššub]” (179). he would have this title refer to Teššub’s liberation from the netherworld, where he was held captive as a consequence of partaking in the banquet allani prepared for him. The fact that nothing in the surviving text suggests that Teššub was captive in or freed from the netherworld presents no obstacle to haas, for other works of literature, from the descent of inana to ovid’s Metamorphosis, prove that violating the prohibition on eating in the netherworld results in getting

69. By “funerary cult” she means mortuary cult—hardly unique to ebla (or to royalty). Several sites in Syria have yielded archaeological evidence for mortuary cults, and some, like Ugarit, have also yielded archives that include ritual texts, but do not attest the provision of large numbers of personnel dedicated to the service of such mortuary cults.

70. Both an and UŠ together would effectively form a logogram, dingiruš. combining a sign having logographic value and one having phonetic value to make a logogram is common in cuneiform writing, though not to my knowledge otherwise attested with this word. cf. neu 1999: 301, n. 10 (with references there).


stuck there (181). adding this element to the story would explain why, after somehow being liberated from the land of the dead, Teššub emerges in the world of men as a beggar, the way he is described in the assembly scene (182–83). in this scene haas places half of Zazalla’s speech in the mouth of Megi, who tries to persuade the senate of ebla to help Teššub by pointing out to them the necessity of granting release from debt servitude (as if this were relevant to the situation of an escapee from the netherworld); he is apparently half successful, for Zazalla answers that they will help Teššub, but not grant release to his creditor (according to haas’s translation of no. 15 i/ii 18’–20’; p. 183). That creditors are not the beneficiaries of debt-release edicts seems to have escaped haas’s attention.71

Understanding the poem is not facilitated by inserting elements that are not found in the extant text, introdu-cing assumptions that cannot be grounded in it, and failing to examine the evidentiary basis for reconstructing and interpreting it. The notions that Teššub is assimilated to suffering humanity, that he demands the release of certain people on account of suffering want himself, and that he suffers imprisonment in the netherworld are to be discarded along with the poorly-premised attribution of the parables to the poem. The idea of debt remission is not absent, but it is present only by association: the release that kirenzi denotes typically means release from debt and its consequences, so Zazalla evokes this meaning by implication; but never is Teššub or anyone described as indebted or enslaved for debt. This is what Zazalla says:

15 i 4’–8’:72

[ā]i h enni Teššub h enz=a=du/o73 is it that Teššub is under duress,kirenz=a=mma šar=i=b that he requests release?74

h enz=a ižuġn=ae Teššub Should Teššub be pressed for silver,šigl=adē=mma ižuġni ar=r=eva=ž we shall give silver, a shekel apiece,75

___________________________šaġad=n=adi h iyar=o=h h e a half-shekel gold apiece,šigl=adē=mma ižuġni ar=r=eva=ž a shekel silver apiece we shall give.

There follows a series of statements that if Teššub is hungry, naked, or parched, the senate will provide food, clo-thing, and ointment. in sum:

i 18’–20’:eġel=l=eva=šša Teššub We shall save him, Teššub,h enzi=da h amaz=i=a=šše=dan76 from the one who constrains him oppressively;

71. The text provides no support for distributing parts of KBo 32, 15 i/ii 4’–iv/iii 7 between different speakers, and less than none for placing Teššub’s threat to destroy ebla in the mouth of Purra, as haas does (2006: 184–85); this and other incoherencies have been addressed by Wilhelm (2012: 160–62, with nn. 13–15, and references there).

72. in presenting the following passage i segment the text according to sense divisions, i.e., apparent verses, rather than by lines on the tablet.

73. The verb form here and in the next line has the intransitive valence marker -a-; it is not clear what the suffix(es) spelled -du may be. The stem h enz-, “oppress” (or the like), may be a variant of h emz-, “bind.”

74. assuming the sentence indeed begins with the conditional [a-a]-i, “if,” the protasis consists of two clauses, of which the first reads “if (āi) now (h enni) Teššub (abs.) is under duress,” and the dependence of the second clause is marked by the conjunction -ma affixed to kirenz(i)=a (essive), thus, “that he requests (šar=i=b) release” (with the conjunction lengthened to -mma for emphasis, syntactic or metrical). But then there is no apodosis—or the apodosis is that of the following sentence—hence this protasis seems to have the force of an interrogative that introduces all the unmarked conditional statements that follow.

75. The morpheme -adi is used to form words for measures, and distributive numbers are formed with this morpheme followed by the instrumental marker -ae. Though the instrumental is absent and “shekel” is a measure, not a number, šigl=adi and the measures that follow were evidently understood as distributive by the hittite translator, who expressed it with kuišša, “each,” in every instance: nu kuišša Tarh unni … pāi (or piweni, in subsequent clauses; see neu 1996: 306).

76. The second of these two words is a nominalized form (-šše) of a 3s. ergative verb (h amaz=i=a), in the ablative (-dan); the stem h amaz- may be a variant of h emz-, “bind” (this was suggested by ilgi evrim [now gercek], in gernot Wilhelm’s 2008 seminar). The first word is a noun

158 eva von daSSoW

nakk=i=offo=šša77 kir=en=zi (but) we shall not decree a release.Megi fe=ve tiža=v an=ašt=i=kki Megi, your heart will not rejoice.__________________________

nor will Purra’s heart rejoice, for they will not release the sons of igingalliš. if they did, Zazalla asks, who would serve them (saž=ol-, “feed,” as in no. 19 i/ii 4, 6, and 8)? These people are their waiters, cooks, and launderers. Zazalla concludes his speech by telling Megi:

iv 2–7:āi ur=i=o kir=en=zi if you want a release,kir=o=nna pura=mme=b be your (own) slave released;kir=o=n ulme=b78 be your (own) slavewoman released!__________________________fut=ki=b h ar=a ar=ol=i=b give your son away, ašti=p=pa atta=i=ve=n(e)=e pend=o=n and be your wife returned to her father’s.anz=a=mma ed(i)=iff=až=kež79 (and) … (2s. imp.) … ourselves,Ebla=va šerže=ni Megi the fate for ebla, Megi!__________________________

one may discern a pun on the name of Purra in the injunction “let your own pura=mme (slave) be released” (l. 2). Upon hearing Zazalla’s speech, Megi goes weeping before Teššub, and explains as follows:

iv 15–19:iž=až ar=ol=aw=nna “i myself grant it,ard=iffu=šša kir=en=zi ar=i=a=mma (but) my city does not grant release.fut=ki=ž Fazanigar=w(e)=až Zazalla=ž The son of fazanigar, Zazalla,kir=en=zi ar=i=a=mma does not grant release.”ed(i)=i Megi=ne=ž tal=ah h =o=m himself Megi removed,Eb[la- … ] [upon] ebla [he cast the fault.]

a little more remains of the hittite text, which says that Megi purified the sin from himself and cast it upon the city of ebla (see Wilhelm 2012: 161–62, n. 15). he did what was in his power, limited by the senate as it was.

The inference remains that the people of igingalliš entered servitude with the eblaites as a result of military conquest and consequent political subordination of their city. This inference is predicated primarily on two obser-vations: that Purra is characterized as a prisoner of war, and that it is apparently the community of igingalliš that is subject to ebla—or in any case a substantial part of the community, not just individual members of it. The fact that

formed from the same (related?) stem h enz- that was used to form verbs expressing duress in ll. 4’–5’ (above); it is marked with the directive -da, perhaps carrying adverbial force.

77. The verb is a negated 1pl. suffixed with the 3s. enclitic pronoun -nna: nakk- (“decree”) -i- (transitive valence), followed by -aw- (first person ergative) + -va- (negative) > -uffu- (or -offo-), then -ž (plural), plus the 3s. enclitic (ž + -nna > šša) (for the analysis of the verb form see giorgieri 2000: 230–33; Wilhelm 2004: 111–12). The pronoun, with its antecedent kirenzi, serves as the patient of the verb. Without the enclitic pronoun, the same form occurs in l. 24’; the spelling of the contracted morphemes -aw + -va- is -ú-wuu- in the first instance and -u-ub-wuú in the second, yielding ambiguity regarding the vowel quality (/u/ or /o/).

78. The second of these two clauses is written with sandhi, ki-i-ru-nu-ul-mi-ib, as explained by neu (1996: 346). for analysis of the verb forms, with the class-marker -o- conveying resultative action in the imperative, see giorgieri 2000: 234–35; Wilhelm 2004: 113.

79. anz=a=mma may be analyzed as a 2s. intransitive imperative, suffixed with either the 2s. enclitic -mma or the conjunction -ma (con-sonant lengthened for emphasis); the stem anz- is attested elsewhere but its meaning is uncertain. The next word, “ourselves,” is suffixed with an unknown morpheme, -kež. Perhaps Zazalla is telling Megi to deflect the fate threatening ebla.


Zazalla describes their subjection in terms of the labor of cooking and doing laundry and so on suggests that they were taken captive in war and enslaved to households of ebla. Meanwhile, the account of Purra having served nine kings in succession indicates that the subjection of igingalliš lasted for several generations,80 yet without their dis-solving as a distinct community, and this points rather to political subjugation—as of Kish to Uruk in the example of the great revolt against naram-Sin (above, with n. 68), or of the Judaeans in Babylonia after nebuchadnezzar’s conquest, or perhaps of the helots to Sparta. in any case the gods demand an end to this. Why? not because the gods, or any particular god, require oblates to staff their temples, but because the gods in general and the storm god in particular require justice among men—as addu of aleppo informs Zimri-lim of Mari in a famous prophecy (fM vii 38; see above, n. 67)—and permanent subjection of free men is unjust. in essence this is the theme of the prophet amos, whose setting was eighth-century israel, and whose rhetoric inverts that of Zazalla: amos informs the israelites that their god takes no pleasure in their offerings—he does not need them to feed him—and desires instead that they do justice (4:4–5; 5:21–24). This means doing right by their fellow man, not hoarding what they gain by violence and exploiting those they subject (2:6–7; 3:10; 4:1, etc.), else Yahweh will wreck their mansions and strongholds (e.g., 3:11, 15; 5:9; 6:11)—often employing the same kind of stepwise series of parallel terms as Teššub does (in KBo 32, 19 i 24–39). Jeremiah, prophesying in early sixth-century Judah, again communicates the same message, concerning a more specific occasion of injustice (in a passage adduced by neu; see above, p. 153): When the king, Zedekiah, had accomplished what Megi did not and obtained the agreement of Jerusalem’s citizenry to effect a release (derōr), they first did release enslaved members of the Judaean community, but then forced them back into servitude (Jer 34:8–11). This perfidy brought upon them Yahweh’s decision to destroy Judah by the hand of the Babylonians and to make its towns desolate of habitation (34:22), just as Teššub will do to ebla (KBo 32, 19 i/ii 24–25). and the same idea is again present in the poetry of Solon, Jeremiah’s near-contemporary in greece. Solon warned that it was not Zeus who would destroy athens but its own citizens, who by their rapacity and injustice would bring divine retribution upon themselves and cause their city to fall into slavery (West 1992, poems 4, 9, 11, and 13). from this calamity, however, Solon himself saved his city, by proclaiming the seisachtheia that released the earth herself, groaning under the weight of foreclosed mortgages, from her bondage, as well as releasing her citizens from the servitude into which they had been sold for debt (poem 36).

These three comparanda specifically concern injustice committed by financial and legal means, typically in-volving debt servitude, through which powerful members of one polity—israel, Judah, athens—subjugate and exploit weak or disadvantaged members of their own community. That is not the situation at issue in the Song of release, where, rather, one polity has subjugated and exploited another. all these works are nonetheless animated by a shared principle: that the gods do not permit some people to hold others in permanent subjection, for all are originally subjects of the gods alone.81

it cannot be taken for granted that a composition such as this would take written form. i have argued that the Song of release, along with the parables, was recorded in writing and translated into hittite for the purpose of scribal instruction in hurrian. Why the Song of release was chosen for this purpose is another question. The issue of political subjection versus autonomy was one that hatti increasingly confronted as it acquired an empire, which may explain the interest of hittite literati in a poem that dealt with exactly this theme.

The manuscripts of the poem that we have found had no known successors. They were abandoned in the base-ment of Temple 16, along with pieces of other polyglot literature, sometime in the early fourteenth century. But for the fragment ABoT 2, 247 (see above, § 2, with n. 15), no other exemplar of this composition has turned up in the libraries of hattuša, nor yet an entry in one of the tablets that catalogued those libraries’ holdings. Textualization was a virtual dead end for this poem, even as other hurrian compositions continued to be recorded and translated

80. Possibly this indicates mythologizing events and their protagonists, so that the individual Purra lives through the reigns of nine kings, as Wilhelm suggests (2012: 163), or possibly Purra, like the community of igingalliš itself, represents the generations of his forebears going back to the moment of ebla’s subjugation of his city.

81. in this respect Bachvarova’s insight about the poem is right (see above, pp. 153–54).

160 eva von daSSoW

over the next two centuries. Yet that did not inhibit its transmission, for poetry lived on lips and ears, not on clay. Whether it was sung in hurrian, hittite, or yet other tongues, some version of the Song of release or its poetic kin was transmitted from land to land and language to language until its themes and its very turns of phrase found new expression in classical israelite prophecy and archaic greek poetry. indeed, its core theme was reincarnated in the foundation myth of ancient israel, whose archetypal leader conveyed to an implacable oppressor the storm god’s demand to let his people go—or else.

References

akdoğan, r. 2007 Hethitische Texte in Transcription. Ankara Arkeoloji Müzesinde Bulunan Boğazköy Tabletleri ii (= ABoT 2). dresd-

ner Beiträge zur hethitologie 32. Wiesbaden: harrassowitz.archi, a. 2007 Transmission of recitative literature by the hittites. AoF 34: 185–203.astour, M. c. 1988 The geographical and Political Structure of the ebla empire. Pp. 139–58 in Wirtschaft und Gesellschaft von Ebla,

ed. h. Waetzoldt and h. hauptmann. hSao 2. heidelberg: heidelberger orientverlag. 2002 a reconstruction of the history of ebla (Part 2). Pp. 57–195 in Eblaitica: Essays on the Ebla Archives and Eblaite

Language, vol. 4, ed. c. h. gordon and g. a. rendsburg. Winona lake, in: eisenbrauns.Bachvarova, M. 2005 relations between god and Man in the hurro-hittite Song of Release. JAOS 125: 45–58. in press hurro-hittite narrative Song as a Bilingual oral-derived genre. in Proceedings of the Eighth International Con-

ference of Hittitology in Warsaw, Poland, Sept. 5–9, 2011, ed. M. Kapelus and P. Taracha. Warsaw: agade.Beckman, g. 1983 Mesopotamians and Mesopotamian learning at hattuša. JCS 35: 97–114. 2001 Sargon and naram-Sin in hatti: reflections of Mesopotamian antiquity among the hittites. Pp. 85–91 in Die

Gegenwart des Altertums: Formen und Funktionen des Altertumsbezugs in den Hochkulturen der Alten Welt, ed. d. Kuhn and h. Stahl. heidelberg: forum.

2003 gilgamesh in hatti. Pp. 37–57 in Hittite Studies in Honor of Harry A. Hoffner Jr., ed. g. Beckman, r. Beal, and g. McMahon. Winona lake, in: eisenbrauns.

de Martino, S. 1999 Problemi di traduzione per antichi scribi ittiti: la redazione bilingue del “canto della liberazione.” BCILL 100 /

Hethitica 14: 7–18. 2003 Annali e Res Gestae antico-hittiti. Studi Mediterranea 12; Series hethaea 2. Pavia: italian University Press.de Martino, S., and M. giorgieri 2008 Literatur zum hurritischen Lexikon, Bd. 1: A. florence: logisma.devecchi, e. 2005 Gli annali di H attušili I nella versione accadica. Studia Mediterranea 16; Series hethaea 4. Pavia: italian University

Press.dijkstra, M. 2008 new Joins in the hurrian epic of Kešši and Their ramifications. UF 40: 205–23.durand, J.-M. 2002 Le Culte d’Addu d’Alep et l’affaire d’Alahtum. fM vii. Paris: SePoa.durnford, S. P. B. 1971 Some evidence for Syntactic Stress in hittite. AnSt 21: 69–75.foster, B. 2005 Before the Muses: An Anthology of Akkadian Literature, 3rd ed. Bethesda, Md: cdl.francia, r. 2010 The Poetic Style of the direct Speeches in the hittite “Parables” of the “epos der freilassung.” Pp. 65–71 in Inves-

tigationes Anatolicae: Gedenkschrift für Erich Neu, ed. J. Klinger. StBoT 32. Wiesbaden: harrassowitz.giorgieri, M. 2000 Schizzo grammaticale della lingua hurrica. Pp. 171–277 in La civiltá dei Hurriti. la Parola del Passato 55. naples:

Macchiaroli.


2001 die hurritische fassung des Ullikummi-lieds und ihre hethitische Parallele. Pp. 134–55 in Akten des IV. Interna-tionalen Kongresses für Hethitologie, Würzburg, 4.–8. Oktober 1999, ed. gernot Wilhelm. StBoT 45. Wiesbaden: harrassowitz.

groddek, d. 2011 Hethitische Texte in Transcription. KBo 57. dresdner Beiträge zur hethitologie 36. Wiesbaden: harrassowitz.güterbock, h. g. 1951 The Song of Ullikummi. revised Text of the hittite version of a hurrian Myth. JCS 5: 135–61.haas, v. 2006 Die hethitische Literatur: Texte, Stilistik, Motive. Berlin: de gruyter.hoffner, h. a., Jr. 1998 Hittite Myths. 2nd ed. SBl Writings from the ancient World, 2. atlanta: Scholars Press.hoffner, h. a., Jr. and h. c. Melchert 2008 A Grammar of the Hittite Language. Winona lake, in: eisenbrauns. huehnergard, J. 1987 Ugaritic Vocabulary in Syllabic Transcription. hSS 32. atlanta: Scholars Press.imparati, f., and c. Saporetti 1964 l’autobiografia di h attusili i. Studi Classici e Orientali 14: 40–85.Kloekhorst, a. 2011 accentuation and Poetic Meter in hittite. Pp. 157–76 in Hethitische Literatur: Überlieferungsprozesse, Textstruk-

turen, Ausdrucksformen und Nachwirken, ed. M. hutter and S. hutter-Braunsar. aoaT 391. Münster: Ugarit-verlag.

laroche, e. 1968 documents en langue hourrite provenant de ras Shamra. Pp. 447–544 in Jean nougayrol, et al., Ugaritica v. Mis-

sion de ras Shamra Xvi. Paris: geuthner.lorenz, J., and rieken, e. 2010 Überlegungen zur verwendung mythologischer Texte bei den hethitern. Pp. 217–34 in Festschrift für Gernot

Wilhelm, ed. Jeanette fincke. dresden: iSleT.Matthiae, P. 1997 ebla and Syria in the Middle Bronze age. Pp. 379–414 in The Hyksos: New Historical and Archaeological Perspec-

tives. Philadelphia: The University Museum. 2006 archaeology of a destruction: The end of MB ii ebla in the light of Myth and history. Pp. 39–51 in Timelines:

Studies in Honour of Manfred Bietak, vol. 3, ed. e. czerny, et al. leuven: Peeters.Mcneill, i. 1963 The Metre of the hittite epic. AnSt 13: 237–42.Melchert, h. c. 1998 Poetic Meter and Phrasal Stress in hittite. Pp. 483–93 in Mír Curad. Studies in Honor of Calvert Watkins, ed. J.

Jasanoff, et al. innsbruck: institut für Sprachwissenschaft.Miller, J. 2007 Texte aus dem Bezirk des Großen Tempels VIII. KBo 57. Berlin: Mann.neu, e. 1993 Knechtschaft und freiheit: Betrachtungen über ein hurritisch-hethitisches Textensemble aus hattuša. Pp. 329–61

in Religionsgeschichliche Beziehungen zwischen dem Alten Testament, Nordsyrien und Anatolien, ed. B. Janowski, et al. oBo 129. freiburg: Universitätsverlag.

1996 Das hurritische Epos der Freilassung i. Untersuchung zu einem hurritisch-hethitischen Textensemble aus Hattuša. StBoT 32. Wiesbaden: harrassowitz.

1999 ausgewählte Kapitel zur hurritisch-hethitischen Bilingue. Pp. 293–303 in Nuzi at Seventy-Five, ed. david i. owen and gernot Wilhelm. Sccnh 10. Bethesda, Md: cdl.

neve, P. 1984 die ausgrabungen in Boğazköy-hattuša 1983. Archäologischer Anzeiger 1983/3: 329–72. 1999 Die Oberstadt von Hattuša. Die Bauwerke, i. Die Bebauung im zentralen Tempelviertel. Boğazköy-h attuša, ergeb-

nisse der ausgrabungen Xvi. Berlin: Mann.otten, h. 1958 Hethitische Totenrituale. Berlin: akademie verlag. 1984 die Tontafelfunde aus haus 16. Archäologischer Anzeiger 1983/3: 372–75.

162 eva von dassow

otto, e. 1998 soziale Restitution und vertragsrecht: mīšaru(m), (an)-durāru(m), kirenzi, parā tarnumar, šemitta und derôr in

Mesopotamien, syrien, in der hebräischen Bibel und die Frage des Rechtstransfers im alten orient. RA 92: 125–60.Pecchioli daddi, F. 2001 a song of Release from Hattic Tradition. Pp. 552–60 in Akten des IV. Internationalen Kongreßes für Hethitologie,

Würzburg, 4.–8. Oktober 1999, ed. G. wilhelm. stBoT 45. wiesbaden: Harrassowitz.Roth, M. 1997 Law Collections from Mesopotamia and Asia Minor. waw 6. 2nd ed. atlanta: scholars Press.salvini, M., and wegner, I. 2004 Die Texte aus Boğazköy, Bd. 6: Die mythologischen Texte. Corpus der hurritischen sprachdenkmäler, I. abteilung.

Rome: Istituto di studi sulle civiltà dell’egeo e del vicino oriente.Tonietti, M. v. 1997 Le cas de Mekum: Continuité ou innovation dans la tradition éblaïte entre IIIe et IIe millenaires? MARI 8: 225–42.veldhuis, n. 2000 sumerian Proverbs in Their Curricular Context. JAOS 120: 383–99.von dassow, e. 2006 Hurrian. Pp. 623–30 in Encyclopedia Judaica, vol. 9, ed. F. skolnik. 2nd ed. detroit: MacMillan. 2011 Freedom in ancient near eastern societies. Pp. 205–24 in The Oxford Handbook of Cuneiform Culture, ed. K.

Radner and e. Robson. oxford: oxford University Press, 2011.wegner, I. 2007 Einführung in die hurritische Sprache, 2nd ed. wiesbaden: Harrassowitz.west, M. L. 1992 Iambi et Elegi Graeci ante Alexandrum Cantati. vol. II. oxford: Clarendon.westenholz, J. G. 1997 Legends of the Kings of Akkade. winona Lake, In: eisenbrauns.wilhelm, G. 1989 Review of Kaniššuwar: A Tribute to Hans Güterbock on His 75th Birthday, ed. H. a. Hoffner, Jr. and G. M. Beck-

man. ZA 79: 130–32. 1992 Hurritische Lexicographie und Grammatik: die hurritisch-hethitische Bilingue aus Boğazköy. Or 61: 122–40. 1997 die Könige von ebla nach der hurritisch-hethitischen serie “Freilassung.” AoF 24: 277–93. 1999 Kešše. Pp. 411–13 in Nuzi at Seventy-Five, ed. d. I. owen and G. wilhelm. sCCnH 10. Bethesda, Md: CdL. 2001 das hurritisch-hethitische “Lied der Freilassung.” Pp. 82–91 in TUAT Ergänzungslieferung, ed. o. Kaiser. Güter-

sloh: Gütersloher verlagshaus. 2004 Hurrian. Pp. 95–118 in the Cambridge Encyclopedia of the World’s Ancient Languages, ed. R. d. woodard. Cam-

bridge: Cambridge University Press. 2012 ein Konflikt zwischen König und Ältestenversammlung in ebla. Pp. 155–66 in Organization, Representation, and

Symbols of Power in the Ancient Near East. Proceedings of the 54th Rencontre Assyriologique Internationale, Würz-burg, 20–25 July 2008, ed. Gernot wilhelm. winona Lake, In: eisenbrauns.

vonda001

Typewritten Text

vonda001

Typewritten Text

vonda001

Typewritten Text

vonda001

Typewritten Text

©

vonda001

Typewritten Text

American Schools of Oriental Research. All rights reserved. Republished here by permission of the American Schools of Oriental Research.

vonda001

Typewritten Text

vonda001

Typewritten Text

Piecing Together Song of Release - JCS 65

Documents