The industrial park of Girsu in the year 2042 B.C. interpretation of an archive assembled by P. Mander.

• Author: Heimpel, Wolfgang

G. BUCCELLATI WAS THE FIRST to recognize an archive characterized by registration of expenditures of bread for a group of Amorite women. He collected a corpus of thirty-one texts and considered them a variety of messenger texts.(1) M. Sigrist, in his article "Courriers de Lagas," discussed the three texts from HSS 4 and the then unpublished text BM 20799, He understood the texts as expense vouchers for travelers and permanent personnel of a "caravanserail," and called them caravan texts.(2) He did not refer to Buccellati. In SAT of 1993 he published BM 20799 and a number of texts of the same general type. G. Pettinato published in the same year three additional texts in MVN XVII. A. Uchitel, referring to Buccellati but not to Sigrist, identified additional unpublished texts in CBT 1, published five of these in copy, and drew up a new list of forty documents. He came to the same conclusion as Sigrist that the texts were expense vouchers for permanent personnel and patrons of a resthouse (e-[kas.sub.4]).(3) In the monograph under review Mander has published a considerable number of additional texts, adding the texts already identified by Buccellati and published later by Sigrist and Pettinato. Mander did not become aware of Uchitel's contribution until he had finished his manuscript and refers to it in an addendum. Mander drew up a list of seventy-four documents, all of which he transliterated. He also collated the texts which Sigrist and Pettinato had published. In NABU 1997, no. 8 (p. 8) he added a text published by M. Yoshikawa in ASJ 10: 19 and the texts which Sigrist transliterated in the introduction to CBT 2. I can add S. J. Levy and P. Artzi, Sumerian and Akkadian Documents . . . in Israel (Atiqot 4 [1965]), nos. 4 and 13; SAT I 449; Watson Umma, fragment f; and Uchitel's BM 14873. CBT 2 contains many more texts which Sigrist identified as "caravan texts." A number of these will turn out to belong to Mander's archive.

The texts recorded expenditures of bread, occasionally also mutton, and exceptionally beer. Some recipients were "kennelmen," as Figulla in CBT I, and after him Mander, translated sipa ur-[gi.sub.7](-ra). The "other workers" of the title of Mander's monograph were Amorite women, shipbuilders, a bird keeper, apprentice scribes, and various workers of problematic identity.

Mander focused on assembling the archive and on distinguishing text types. He also compiled extensive, full, and accurate indices, and thus put at our disposal the means to identify and evaluate precious new information. I am very grateful to make use of it and, basing myself on his work, I will consider the questions of the place where the texts were written, the identity of the recipients of bread and mutton, and the connection of these recipients with that place. I have drawn on most, but not all, of the published texts from Girsu.(4) Especially relevant among these are the texts of what will be called the governor's food allowance archive.(5) I hope to apply an interpretative angle to the texts of Mander's archive as new texts are published by him and others.

The one-year hypothesis. Most texts were dated by month and day, the texts of the first month also by year,(6) which was AS 5 or according to the middle chronology the period from spring 2042 to spring 2041 B.C. Only one text recorded expenditures for more than one day. Buccellati already proposed that they all belonged to one and the same year, and Mander elaborated and solidified the proposal. He also quoted H. Waetzoldt's suggestion that antiquities' robbers found the locus of a basket in which the tablets of this year were stored. Crucial for the validity of this one-year hypothesis is the question of overlap. The corpus includes two, or possibly three, pairs of texts from one and the same day. Texts 12 and 23 are dated to VI 3. Their information is complementary, so that they may have recorded the expenditures of the same day in two installments. The second pair is formed by texts 22 and 61 of IX 2. Both list Amorite women as recipients of bread while the remainder of the information is complementary. If these were the same women, then there is a problem with assigning the texts to the same date. The problem is addressed below in the section on redundant complementarity. The third pair consists of texts 15 and 16. Their information is identical. Mander suggested that one was a copy of the other, a feature that "we cannot presently comprehend." He mentioned also that the reading of the number 6 of text 16 is "not clear" and considered the possibility of reading 7.

A look at Mander's chronological arrangement in charts 1-3 on pages 86-88 gives the impression of a natural flow of increase and reduction.(7) Note especially that the amount of bread allocated for apprentice scribes rose from 20 to 24 liters during the fifth month, and that it was consistently 20 liters before that time and 24 after it.(8)

Text types. The texts of the archive can be identified easily because they start with a line recording a modest amount of bread for Amorite women or other often-recurring recipients. A closer look reveals many variations in notation and continuous change in kind and number of recipients. The notational variations seem quite unnecessary and demonstrate a curious lack of consistency, Mander brought some order into the inconsistencies by distinguishing twelve text types, several subtypes, and some exceptions. A chronological arrangement of the texts by types can be found in his chart on pages 55-57.(9)

The texts fall into two broad categories as Uchitel already saw. A certain Urebadu appears in one, dogs and their keepers, shipbuilders, and apprentice scribes in the other. Amorite women were listed in texts of both categories. Text 51 may serve as an example for the first category: "5 liters bread, female Amorites; 2 liters Urebadu, 2 liters Ur-Damu - they are work-impaired; 2 liters beer, 2 liters bread, Lugal-Ezen bird keeper; expenditure; day 7; month III." Text 6 may serve as example for the second category: "12 liters bread, dogs; 4 adult males (gurus), 2 liters each, dog keepers; 20 liters, apprentice scribes, 5 liters, Amorite females; 1/2 sheep Elamites of Sabum, intermediary PN; 10 liters bread, 1 leg, Abbamu, soldier of the mirror holder; day 11; expenditure; month III." The first category corresponds to Mander's text types C, D, and E; the second to A and B.

The categories were distinguished from the second month on. In the first month Urebadu and shipbuilders were recorded together. The relationship between the two categories is curious because Amorite women were recorded in both. If we assume that these were the same Amorite women in all texts we arrive at a most improbable scenario.(10) Whenever the recipients of one category were present, those of the other category were absent. For example, in III 5 dogs, their keepers, and apprentice scribes were present. Two days later they were absent and the work-impaired Urebadu and Ur-Damu as well as the bird keeper Lugal-Ezen were present. Four days later these were gone, and dogs, keepers, and apprentice scribes were back. We might be tempted by Sigrist's idea that the dogs and their keepers were guarding caravans en route.(11) Their coming and going would then be explained by the movement of caravans. But the watchfulness of dogs lessens in proportion to the distance from their home territory, so that these animals are not well suited for guarding a caravan en route. Also, why were apprentice scribes and shipbuilders absent when the dogs were absent? The situation would be perfectly understandable if the texts of the two categories were complementary and the full archive consisted of pairs of texts for each day. There is indeed the pair of texts 12 and 23 of VI 3, mentioned above. Text 23 registered bread expenditures for Urebadu and Ur-Damu, for Lu-Ningirsu, and the UN-il of the sheep house, while text 12 registered the expenditures for dogs, dog keepers, apprentice scribes, Amorite women, and three Makkanite shipbuilders. Clearly, dogs and scribes did not drive Urebadu, UrDamu, and Lu-Ningirsu away. All groups of recipients were compatible. But why were the expenditures for the Amorite women recorded elsewhere in texts of both categories?

One possible answer is that there were two groups of Amorite women, both of which received 5 liters of bread daily. We might assume that there was only one such group in the first month, when the two text categories were not yet distinguished, and two groups in subsequent months. We would then have to assume that only one group was present on the dates of texts 23, 26, and 70. The second possible answer is that there was only one group of Amorite women, and that the expenditures for them were often recorded in texts of both categories for one and the same day, giving the false impression that they received their daily rations twice. Such a practice of double notation is not tolerated by modern administrative procedure, but it was so in the provincial administration of Ur III Girsu, as the following example demonstrates.

Why were the expenditures for Ur-Ninsun, Usgina, and Kala registered on two tablets? What kind of administrative procedure causes such redundancy? Were different scribes under instruction to register any expenditure that came to their attention regardless of repetition? Was there no clear demarcation in responsibility? Or did one scribe register expenditures twice a day, perhaps once before siesta and another time in the evening, repeating in the evening some entries from the morning? While I have no explanation for the phenomenon, we may have to consider that a particular expenditure could be registered on two daily records. That such redundancy has affected only the Amorite women in Mander's archive is an enigma.(12)

Where the texts were written. The texts recorded the expenditure of bread and mutton, indicating that they were written in...