§1. Introduction

In April 2010 Robert Englund brought to my attention LAUSD: AA 74, 13 (E 135.4-set 1), a tablet in the collection of the Los Angeles Unified School District (henceforth: LAUSD: AA 74).^[2]  At its longest the preserved part of the tablet measures 13cm x 13cm, and is about 2.5cm thick, with writing on both sides.  The right, left, and lower parts of the tablet are broken.  The left side is missing a substantial portion (perhaps one-quarter or more), a small part is missing on the right (enough for a few signs), and an unknown extent of the bottom has broken off.  Internal evidence from the text, namely the mention of the goddesses Ištar and Nanā and two temple officials, reveals that the text belongs to the archive of the Eanna temple of Uruk.^[3]  It probably dates to the second year of Neriglissar (c. 558-557 B.C.).    

The text labels itself an “inspection (amirtu) of sheep and goats, the property of Ištar of Uruk and [Nanā],” and the preserved part assesses the growth of six different herds, under the control of seven different herdsmen, from the first year to the second of Neriglissar's reign. I am unaware of another text like it from first millennium Mesopotamia, although it can be understood in light of other texts from the Eanna archive. While the word amirtu appears in other Eanna texts that assess herds, none of them take the form of this text.^[4]  Closest to it are audits (NÍG.ŠID ép-šu-tu, lit. a “settling of accounts”) of individual herdsmen (NBC 4897, NBC 4847, NBC 4846).^[5]  Those texts help elucidate LAUSD: AA 74, but the differences make LAUSD: AA 74 unique.  For example, those texts assess the growth of one herd of one herdsman each, and they also consider the growth of those herds over multiple years. LAUSD: AA 74, on the other hand, assesses multiple herdsmen and charts the growth of their herds over one year. Unfortunately, the fragmentary nature of LAUSD: AA 74 precludes a detailed analysis, but a general picture and some details emerge from a close reading.

From other inspection texts we know that LAUSD: AA 74 would have had an additional five vertical columns to the left: one for each category of sheep (ram, ewe, male lamb, and female lamb) and then a subtotal of sheep.  Starting from the left of the best-preserved part of the tablet, there are four vertical columns for types of goats (he-goat, she-goat, young he-goat and young she-goat, although most entries for the first two are obliterated) and a subtotal column for goats (called “blacks”).  Following that is a column for the grand total of sheep and goats, and the final column gives the name of the herdsmen, deductions from the herd in animals, and the balance of wool and goat hair.  An edition follows.

§2. Text

reverse

reverse

§3. Text notes

Line 2: Kleber puts Mušēzib-Marduk's career from at least Neriglissar 2/III/15 until Nabonidus acc/IV/22.  LAUSD: AA 74 seems to have been composed at the end of Neriglissar year two or early year three, so this attestation fits within these parameters.  The šatammu in the text, identified only by his first name Bānia, is certainly Bānia//Tabnēa/Bā'iru.  Kleber's earliest attestation of this individual comes from YBC 11647, which dates to an unknown year of Neriglissar.^[6]

Line 3: I know of no other attestations of a Nabû-qātē-ṣabaṭ//Bēl-ahhē-erība.  The fairly unusual name Nabû-qātē(šu)-ṣabaṭ, without patronymic, appears in BIN 1 158: 14^[7] (undated) written ^mdAG-ŠUII-šú-ṣa-bat and BIN 166: 8 (Ner. 02/VII/xx).  Neither of these texts is about sheep or goats.  

Line 15: Bēl-lē'e//Nabû-[…]: a Bēl-lē'e//Nabû-ušallim appears in the Eanna archive associated with sheep and goats in TCL 13 171: 40 (Camb: 05/VI /23) and PTS 2328: 12f (Cyrus: 07/V/21).

Line r. 10': Bēl-epuš//Bēl-uballiṭ: A Bēl-epuš//Bēl-uballiṭ appears in association with sheep and goats in the undated UCP 9/1, 85: 13f.

Line r. 12': iš-mi-tu-ma id-da-áš-šú(?): the reading of the second verb is uncertain.  [link 1 to signs]           

§4. General Observations

This text has a number of unusual features.  It charts the growth of multiple temple herds under individual herdsmen from one year to the next, including the deductions taken by the temple over the course of that year.  As mentioned above, other tabulated texts chart the growth of herds for individual herdsmen, often over multiple years.  Texts that list multiple herdsmen either do so just as raw numbers (i.e., the number of animals that those herdsmen had that year), or they list the running balances of the herdsmen for a particular year.^[8]   

We do not know what precipitated this inspection, although we might connect it to an interest in temple livestock audits by various royal administrations in Babylon.  For example, there is the following part of a letter to the Assyrian king Assurbanipal, referring to the Ezida temple of Borsippa:

SAA 10 353: 14-27^[9]

[To the king], my [lord]: your servant [Mar-Issar]. [Good health] to the king, my lord! May Nabû and Marduk bless [the king], my [lord]! 

 . . . In the sp[ringtime the king, my lord], sent a [bo]dyguard to the commandant and the [prelate of] Borsippa (with the following orders): "Make an account of the bull[s and sh]eep belonging to Nabû (NÍG.ŠID ša GU₄.NÍTA[MEŠ] ⸢UDU.⸣ḪI.A.MEŠ ša ^dPA ep-šá). Supply, as [in ol]den times, the regular ram offerings from the estates of the citizens of Borsippa! The fattest rams should be [deliv]ered to Nabû!"

The shepherds have (however) bribed both the commandant [and the] pr[elate]: up till now no account [of] the bulls and sheep has been made, nor have they supplied the regular ram offerings. They have not (even) sacrificed the king's offerings, the ungelded bulls, in the month Nisan (I), (but) do the bull herdsmen's bidding. They turn the blesser's ungelded bulls back from the (temple) gate, and have served (such bulls) on the table of Nabû.

 . . . Perhaps they are telling the king, my lord: "In olden times, [no accounts] (of oxen and sheep) were made." (If so), they are lying. [In] an ordeal proverb attributed to [Burnabur]iaš, king of B[aby]lon, it is s[aid: "The (time of) accounting is] the ordeal of the sheph[erds]." I am dispatching [the relevant tablet to the king], my lord.

The writer of this letter, Mar-Issar, was “Esarhaddon's roving envoy and troubleshooter in Babylonia.”^[10] This letter, interesting for a variety of reasons, shows a new royal administration at Babylon taking an interest in temple accounts in animals, sending various representatives to “do an accounting” (NÍG.ŠID . . . epēšu).  We find a new royal administration taking a similar approach in the early Achaemenid period Eanna as well.^[11]  Moreover, the letter clearly evinces local pushback to that accounting—the fact that a royal order prompted the accounting should be an indication of its irregularity (or at least its onerousness).  Indeed, the “ordeal proverb” attributed to the Kassite king Buraburiaš understands the audit as problematic for the shepherds.  If nothing else, I am unaware of evidence to suggest that yearly audits were commonplace.  In fact, as texts like SAA 10 353 indicate, they were often done under considerable pressure from outside and with resistance from the audited.

Perhaps, then, LAUSD: AA 74 belongs in this context: it is the result of a general inspection of the temple's herdsmen, by order of the royal administration, run through temple officials with royal backing.  This is obviously speculative, so there is no need to push the analysis too far, but LAUSD: AA 74 may prove useful should more evidence of such an inspection come to light.  

§5. Analysis

If we start at line 3, and put aside for now the rightmost vertical column, the text divides into ruled horizontal sections of six lines for most entries, and eight lines for the final entry.^[12]  Each section of six or eight horizontal rows charts the growth of a herd from the first year of Neriglissar to the second. Schematically, each six-line section contains the following information:^[13]

Table §5.1: Schema of each six-line horizontal section

Line 1 of each section, then, gives the actual size of each herd in Neriglissar year 1.  Since most of the left half of the tablet is broken it is best to concentrate on the “grand total”—that is, the total number of sheep and goats—at the start of this assessment period.  Table §5.2 gives the grand total of each herd in Neriglissar year 1 (i.e., the grand total of each herd reported in line one of each horizontal section).  

Table §5.2: The initial grand-total size of each inspected herd

As we know from audit texts, the temple had a mathematical formula for modeling the growth of a herd from year to year.  In general terms, the model follows stipulations laid out in YOS 6 155: that for every 100 viable ewes, the herd grew at a rate of 66 2/3rds lambs per year.^[14]  Before applying the model, though, the temple subtracted a standard set of herd deductions.^[15]  The second and third horizontal lines of each section list these deductions from each herd.  Respectively, these are hides (KUŠ.ME) and “wages” (idu): the former is 10% of every animal, except young males, and represents natural deaths in the herd.  The latter seems to be a consistent mathematical deduction, yet it is unclear exactly how the temple arrived at that number or what it represents.

Because the text is broken, we cannot discuss the model and deductions on a case-by-case basis here; the part of the tablet where the scribes would have charted the growth and deductions of each category of animal is mostly lost.  However, with the grand totals in the preserved section, we see the results of the modeled herd growth less the two deductions in the fourth horizontal line of each section.  In other words, each horizontal section begins, in the first line, with the size of each herd at the beginning of the inspection period, broken down by animal, sex, and age (young vs old); the next two horizontal lines of each set subtract standard deductions from each animal from first line, and then the fourth horizontal line records the modeled growth of the herd, less the deductions of lines two and three.  

The broken state of the text limits how far we can work with these numbers, but the preserved part of the text does allow us to chart (a) the initial size of the herd and (b) the number of animals the temple anticipated the herdsmen having one year later, using its model and standard deductions:

Table §5.3: Initial herd size compared to the anticipated herd size of the next year

In other words, Table §5.3 shows the actual size of the herd of each herdsman in Neriglissar year 1 and then the number of animals the temple calculated the herdsmen to have in Neriglissar year 2, using its model and deductions.  

Of the two remaining lines in each six-line section, the fifth horizontal line lists the number of animals that the herdsman actually had at the inspection, the “seen” (amrā) animals.  Table §5.4, then, lists the initial size, anticipated size, and then this actual size of each herd:

Table §5.4: initial herd size, anticipated herd size, and actual herd size for each herdsman

Finally, the sixth horizontal line of each section records the difference between the anticipated number of animals and the actual (“seen”) number of animals in the herd.  Some of these entries are broken.

Table §5.5: initial herd size, anticipated herd size, and actual herd size

Were this all the text recorded, the anticipated size of the herd less the actual (“seen”) size of the herd would equal the difference recorded in the text.  This clearly is not the case in Table §5.5.  For example, in the table's first entry 808 - 720 is 88, but the text records 62; in the second, 1343 less 1120 is 223, but the text's broken entry is between 100 and 199; in the table's final entry 1068 – 686 is 382, but the text records 364.    

At this point the rightmost vertical column of the text, which marks extractions in actual animals removed from the herd, comes into consideration.  Since a Mesopotamian balanced account contrasts the number of animals that the temple anticipated the herdsman having (the “debits”) against the number of animals that the herdsman actually had (the “credits”),^[16] the temple had to factor in animals that it took from the herdsman between audits.  These, as much as the actual animals in the herdsman's possession at the time of the inspection, counted as “credits” to his account.

These extractions fall into two categories: the first is those animals that the temple removed from the herd at the time of the inspection; the second is those animals that the temple removed before the inspection.  For the first, which this tablet lists after the word amrā (“inspected,” see, e.g, the rightmost entry in lines obv. 7 and rev. 16'), we know from audits of single herdsmen that the accountants subtracted the number of actual animals taken out at the time of the inspection from the temple's modeling of each type of each animal for the next year.^[17]   So, for example, in line 7 the 2 parru-lambs and 9 he-goats would have counted as deductions from the appropriate column in an account for the third year of Neriglissar, which is not in the purview of this text.  

The second type of extraction is a bit more complicated.  We know that herdsmen routinely delivered animals to the temple throughout the year.^[18]  Texts categorize such between-audit deliveries as happening ina rēhi, “against the balance,” which in effect means that the animals counted as credits at the next inspection.  For example, obverse lines 3-8 of the text under consideration records an anticipated herd size of 808 animals, an actual herd size of 720, and a balance of 62.  The difference between 808 and 720 is 88 animals, which overshoots the recorded balance by 26 animals.  However, we can easily locate those 26 animals by adding up those taken out over the course of the year, between the two inspections:

Table §5.6: Between audit extractions in lines obv. 3-8

Hence, the 26 animals taken over the course of the year were counted against the herd's audited balance.  Table §5.7 shows this in schematic form:

Table §5.7: Schema of between audit extractions in lines obv. 3-8

The next entry with multiple between-audit extractions (obv. 15-20) is broken, so it does not give us exact numbers or dates.  However, the numbers are clear enough to set parameters to show that the accountant dealt with the deductions in the same manner as that just discussed above.  In this case, the temple anticipated 1343 animals and the herdsmen presented 1120, which gives a balance of 223.  Although the written balance is broken, it must be between 100 and 199 (i.e., the tablet reads 1 ME ⸢…⸣).  If, like above, we subtract from the anticipated balance the 34 parru-lambs and 23 other (text broken) animals that were extracted before the audit, we arrive at a balance of 166, which falls within the parameters of what is in the preserved part of the text.

After a series of entries too broken to work with, the final entry (rev. 10'-17') presents some problems I cannot solve.  Line 15' records an anticipated herd size of 1068, line 16' gives the actual herd size of 686 and line 18' records the difference between the two as 364, which is 18 more than the actual difference between 1068 and 686.  As above, one would normally find those 18 animals in the actual extractions in the rightmost column, but that column lists extractions that total at least 49 animals (12 U₈.SAL.MAHX + 12 ⸢ÙZ⸣ + 1 [+ x …] + 10 BAR.GAL + 14 SAL.ÁŠ.GÀR.ME).  In at least one case the scribe seems to have factored an actual extraction into the sub-calculations.  The model anticipated the herdsmen having 82 young she-goats and records a total of 30 inspected young she-goats.  The difference between those two numbers is 52, but the text records a remainder of 38, which is off by 14.  The rightmost column records an actual extraction of 14 young she-goats in month XII, which would seem to account for the difference between the anticipated and inspected numbers.  Yet this does not work for other sub-calculations.  The text gives an anticipated number of 319 she-goats, an inspected number of 163, and records the difference between them as 155.  This is one off from the actual difference between those two numbers (319-163 = 156), but the rightmost column records an actual extraction of 12 she-goats.  Perhaps the between-audit addition to the herd listed in line 11 affected the calculations in ways that would make more sense if the tablet was better preserved. 

 There is one more important observation to make.  We can use the raw data in the text for information about actual herd growth.  Given that each section of the text records the initial size of the herd (in Neriglissar year 1), the actual size of the herd one year later (in Neriglissar year 2), and, finally, all of the animals taken from the herd over the course of a year, these numbers present an opportunity to determine the actual growth of the herds in aggregate size from one year to the next.   Table §5.8 gives the results:

Table §5.8: Actual herd growth from Neriglissar year 1 to 2

I stress that not one of these herds reached the model increase for herd size from one year to the next—that is, none of them met the temple's anticipated numbers—but if we compare the initial size in Neriglissar year 1 to the actual animals inspected in Neriglissar year 2, we see two herdsmen with significant herd growth, one who broke even, and one with a significant loss of herd size.  

It is difficult to say why the final herdsman saw herd losses.  It is possible that an epidemic killed off a significant part of his herd.  There is, however, another possibility.^[20]  Given that the herdsmen always fell short of the temple's model, the more years a herdsman went between audits, the bigger the difference between the temple's expectations and the number of actual animals in a herd.  Since most of the herdsmen in this text saw actual herd growth, I assume that the initial size of their herds is an actual count of animals, generated by an inspection in Neriglissar year 1.  It is possible, though, that the final herdsman failed to have his herd inspected in the first year of Neriglissar; if so, the initial size of the herd for that year might be a modeled number based on a previous year.  In effect, what looks like a significant herd loss might instead reflect a comparison of actual animals to two (or more) years of herd models.