In March 2013, the cuneiform collection of the Musée d’Art et d’Histoire (MAH), Geneva, Switzerland, was reviewed for tablets with numerical content. In addition to multiplication tables, a table of squares and the famous geometrical tablet MAH 16055, all in “portrait” format, two unpublished documents in tabular “landscape” format, MAH 16038 and MAH 15886 + 16295 were noted. The former tablet is poorly preserved and only numbers are visible, even at the heads of columns. The latter, however, is a full-fledged account of the grain rations of four gangs of workers for one month during the reign of king Rim-Sin of Larsa. Although the type of tablet is well known from previous publications (Riftin 1937, Seri 2002, Robson 2004), the features described below make it worthy of note.
The document is of pale clay, dimensions are 10.0 × 5.5 × ca.2 cm. The MAH catalog shows that it came into the collection in November 1938 from the personal collection of Genevan assyriologist Alfred Boissier. The artifact bears a strong resemblance to a number of Rim-Sin-era accounts in the Plimpton collection at Columbia University, the origins of which have been described by Robson (2002). Boissier’s European location suggests acquisition from one of the Paris dealers mentioned in the on-line historical notes to the Hermitage collection (Anonymous 2003) which houses the very similar Riftin 114 and 116, perhaps Géjou, who had other Larsa tablets on the market between 1910 and 1915 (Dyckhoff 1999, p. 56 note 177).
Recto (Fig. 1, Tables 1a and 1b)
Although the rightmost column of the face of the tablet has been seriously damaged, enough remains on two lines to enable us to read the names of Rim-Sin-rappašunu and Nergal-lamassašu. These names also occur in the “mu-bi-im” column of Riftin 114 and 116, The third name, on line 4 of our tablet, has suffered more, but seems to end in LAM (compare the end of line 12 of tablet HMA 9-01847). This leads us to propose Rim-Sin-atpalam as the name in that cell of the table. Our tablet can thus be assigned to the corpus of texts involving these persons discussed by Veldhuis (2008).
Unfortunately, the traces of the contents of the fifth cell in the column do not seem to fit any of the names associated with Rim-Sin-rappašunu and (Rim-Sin)Nergal-lamassašu on other tablets. The corresponding row of the tables in Riftin 114 and 116 contains ša ša-ru-um i-ša-ka-nu, “whom the king will assign”, and shows the smallest contingent of workers there, as does our tablet, but the surviving wedges in MAH 15886 + 16295 do not seem to fit the phrase. One would like to imagine that the insertion in our tablet that begins ŠU+LAGAB ša dri-im ... continued with dSUEN i-ša-ka-nu. The tails of vertical wedges projecting into the sixth cell of the last column are too numerous for the signs of this phrase, but some of those tails may go with the signs of the proper contents of the fifth cell, not the insertion.
A prospective account of corvée labor?
Šakānum was the term used for assigning people to corvée labor, and one month, the period covered by our tablet, was the standard length of corvée service (Stol 1995: 294 and 299). The iparras form of the verb in the Riftin tablets is interesting in that it suggests that these accounts are not records of expenses already incurred, but rather prospective estimates of the resources to be deployed in the projects named at the bottom of the table. The similarity of the numbers in our tablet and in Riftin 114 and 116 leads one to wonder if 2000 is the approximate number of workers available to the Larsa administration in any given month. We see the same sized labor force proposed for major canal work at Mari under Rim-Sin’s contemporary Zimri-Lim (ARM 6 7; Kupper 1954, Durand 1998). Largescale maintenance or construction work (at the bottom of Riftin 116 the scribe noted eren2 si2-hi-ir-ti E2.GAL, “workers of the surroundings of the palace”) is a typical use of corvée labor. The modest contingent under the king’s direct control in Riftin 114 and 116 and perhaps in our tablet may represent palace staff: at contemporary Mari, labor forces were recruited from “the district, Palace and (city of) Mari” (Evans 1963, p. 68).
Like Riftin 114 and 116, MAH 15886 + 16295 records the quantities of grain allocated to workers who fall into two categories, one receiving much more per day than the other. In the case of Riftin 114 and 116, the more favoured receive 6 2/3 sila, approximately 6 2/3 liters, per day while the less favored receive only 2. Our tablet shows one category receiving 3 bariga 2 ban2, about 200 liters, and the other 1 bariga or 60 liters. These are 30 times the daily amounts given in the headings of the first two columns of Riftin 114 and 116, and indeed our tablet gives a monthly total consumption by each gang of workers in column 4, while the other tablets give both daily and monthly figures. In both our tablet and Riftin 114 and 116, the daily rates are identical, and the larger quantity is 3 1/3 times the smaller one. As noted by Joannès (2000), the lesser ration, about 2 liters or 1 kg per day, would be barely adequate to maintain a physically active man and would leave nothing to share with a dependent person or to trade for anything to fill out the diet. Perhaps the marginal ration reflects the mobilization of residents for day labor, a regime that would ensure the workers access to a fuller and more varied diet than their allowance alone would provide. The naming of “foremen-of-ten” (UGULA NAM 1(u)) among those recruited by corvée would account for their receiving no more than the common laborers, whereas the real leaders of the operation enumerated in the first column are entitled to compensation at the market rate.
The Geneva and St. Petersburg accounts differ in the terms used for the more- and less-favored workers. The expression in the heading of column 1 of our tablet has previously been attested only in TCL 10, 133: 88 (see CAD R, 393 s.v. rubbû; also D. Arnaud, RA 70 88), which seems to be a list of rations for various individuals and groups of persons. TCL 10, 133 is also of interest for including two entries (ll. 58 and 83) for šuku eren2.sag; eren2.sag appears as a heading in the tabular account that is tablet 43 in Seri 2007.
A scribal short-cut
An error in column 4 of our text shows that although the scribe was forced to compute the amounts of grain for each gang of workers by multiplying their rates of pay by their numbers, he arrived at the grand total at the bottom of column 4 by simply adding the column of figures and failed to cross-check his calculations. If he had multiplied the totals of each category by their total entitlements, he would have arrived at 412 gur 1 bariga, which would be the sum of the preceding numbers if one corrected the monthly ration of gang 3 to 120 gur 2 bariga 2 ban2.
A striking difference between the Geneva and St. Petersburg tablets is the decimal representation of the numbers of workers in Riftin 114 and 116 versus the sexagesimal representation in MAH 15886 + 16295. The consistent format of these three tablets and the many Rim-Sin 31 tablets tabulated by Robson (2004: 131) implies that they come from the same chancery, so the use of different numeral notations is striking. Although one’s first reaction is to consider the latter the “normal” Old Babylonian system, Friberg (2007: 5 and 98) reminds us that the number words of Akkadian are decimal, and that ordinary people in second-millennium Mesopotamia probably used decimal counting and arithmetic in daily life. The recording of numbers by an additive system using hundreds was normal practice in Assyria (Michel 2006) and is frequently encountered in Mari texts (Proust 2002) but its use in Riftin 114 and 116, presumably accounts from Southern Babylonia, seems to be a rare deviation from the long-established sexagesimal positional notation current in that region. Another case of non-sexagesimal counting south of Babylon is found in a tabular account from Nippur noted by Robson (2004: 127; PBS 8/1 26), dated Warad-Sin 11.
The object of the account
The note at the bottom of the table in MAH 15886 + 16295 suggests that the workers were engaged in construction work at two places, one effaced and the other a fortified place named for Rim-Sin’s father. Kudur-mabuk’s name appears three times in the Old Babylonian volume of the Répertoire géographique des Textes Cunéiformes (Groneberg 1980): in E2 ku-du-ur-ma-bu-uk in Riftin (1937: No. 68), dated to the reign of Hammurabi; in URU KI ku-du-ur-ma-bu-uk in contract TCL 10, 118; and in a letter of Hammurabi to an official named Šamaš-hāzir regarding the contested possession of a field alongside a watercourse at Dūr (BAD3) ku-du-ur-ma-bu-uk (Kraus 1968: No. 87). The fact that our tablet predates the Babylonian conquest of Larsa and reflects the deployment of a large workforce makes it a possible witness to the construction of the fortress, and in any case constitutes the earliest attestation of the place-name.
Verso (Fig. 2, table 2)
The colophon is almost identical to that of tablet TCL 10, 79, date transcribed in Kraus (1959: 157) and text collated by Arnaud (1976: 86), a contract to which Rim-Sin-rapašunu, Nergal-lamassašu and perhaps Rim-Sin-atpalam were parties. While identifying the year in the colophon seems straightforward at first, determining the correspondence of the month and day with a Gregorian date is not.
Rim-Sin conquered Isin in the 30th year of his reign (van de Mieroop 1993, Fitzgerald 2002) and thereafter counted years from that event. Although there does not seem to be enough space between ŠU and IN in line 7 for NAM, TI and LA2, there are similar lacunae in the many attestations of the year formula compiled by Sigrist (1990: 59-60), so it is justified to take the year to be the one after Rim-Sin’s year 30. Rim-Sin introduced not only a novel approach to year-naming but also a complex system for indicating dates (described by Kraus 1959, with references to earlier literature). As for the month, KIN.dINANNA/elūlu was basically the 6th month of the year (Michel 2001), but the added KI and NAM terms in the date confuse the issue. Kraus proposed (1959, p. 160) that the date format seen in TCL 10, 79 combines two different ways of designating the month: traditional (kin dInana ki 6) and numerical (10 ki 3), linked by NAM. Our tablet conforms to that format, but differs in the day number. Numbers beyond 30 (34 in the case of TCL 10, 79) led Kraus to see in these “ungewöhnliche Datierungen” a system of superimposed cycles of named and numbered months that repeated themselves over periods longer than one solar year. The KI numbers related to the position of a date in a given cycle. One consequence of this interpretation (Kraus 1959: 163-164) is that “Isin 2” is not necessarily Rim-Sin 31. Kraus was unable to discover a unifying system that could account for all the attestations he collected, and the latest study to take note of this chronological complexity (Fitzgerald 2002) notes (p. 146) that in spite of the supply of new data by Robertson (1983) a satisfying explanation remains elusive.
The months during which the three tablets discussed here were composed (Riftin 114: KIN.dINANNA = Elūlu, month 6; Riftin 116: GAN.GAN.E3 = Kislimu, month 9; MAH 15886 + 16925: KIN.dINANNA = Elūlu, month 6) fall at appropriate times of the year for public works, between harvest and planting of the next crop. This suggests that, whatever cycles of numbered months may have been in operation, the named months retained their traditional positions in the solar year on account of their correspondence with the economically essential agricultural cycle.