The Smell of the Cage

CDLJ 2009:4

Cuneiform Digital Library Journal (ISSN: 1540-8779)

Published on 2009-08-21

© Cuneiform Digital Library Initiative

Creative Commons License This work is licensed under a Creative Commons Attribution 4.0 International License except when noted otherwise and in the case of artifact images which follow the CDLI terms of use.

Robert K. Englund

University of California Los Angeles

Mesopotamia, Late Uruk period, proto-cuneiform, slavery, personal names

for Slava Ivanov on the
occasion of his 80th birthday


§1.1. It may seem less than remarkable to many observers of the advancing civil rights movement in the United States that, in November of 2008, citizens of this country elected a black man to the office of President. Barack Obama is not personally descended from African slaves; still, his ascension to the highest elective US office, despite the lingering liability of his skin color, represents a true benchmark in a sordid history of abuse that is intimately related to the European pillage of the New World. The history of European enslavement of Africans for the purpose of forced labor in transatlantic colonies describes a cultural atrocity whose flames burned brightly in the American South, but, we might note, longest in Brazil, where, beginning in the 16th century, hard labor in sugar cane production and mining operations was transferred by the Portuguese from the deteriorating indigenous slave populations into the hands of imported Africans. Here as in other New World colonies, slavery well outlived its abolishment in Europe—in 1761 in Portugal,[2] or with the Slave Trade Act effectively frozen in the British Empire in 1807 until its eventual prohibition in 1834.[3]


§1.2. The US followed Britain in the abolition of the slave trade in the early 19th century,[4] but retained legal ownership of slaves, in the Confederate states until Lincoln’s famous Emancipation Proclamations of 22 September 1862 and 1 January 1863, finally banning all forms of slavery with adoption of the 13th Amendment in December of 1865. Approximately four million black slaves were freed by July of 1865,[5] but, as post-war federalism would play out, freed into the very uncertain future of Reconstruction that eventually failed them, and rewarded the intransigence of secessionist Southern states. By 1877, with the final withdrawal of federal troops in a kowtow by the US president, Hayes, to advocates of “states rights,” all Republican state governments were replaced by Democrats who instituted a system of segregation and poll taxing that effectively disenfranchised recently freed black men. This was, however, as the history of southern paramilitary organizations comprised of former Confederate soldiers demonstrated, not the most pressing existential distress of blacks in the post-war United States; still, poll taxes and other means of intimidating blacks, including the Jim Crow laws passed by the Democratic state legislatures, were an infection of the US body politic that held through the freedom marches of the 1960’s and beyond—the 24th Amendment, ratified in January of 1964, finally abolished poll taxes, and the Civil Rights Act was passed in July over the Senate filibuster led by Southern Democrats, one month before Obama’s third birthday. The best chronicler of the Southern experience with Reconstruction and the succeeding Confederate resurgence is William Faulkner, from whose Go Down, Moses this paper’s title is borrowed:

The Sam Fathers whom the boy knew was already sixty—a man not tall, squat rather, almost sedentary, flabby-looking though he actually was not, with hair like a horse’s mane which even at seventy showed no trace of white and a face which showed no age until he smiled, whose only visible trace of negro blood was a slight dullness of the hair and the fingernails, and something else which you did notice about the eyes, which you noticed because it was not always there, only in repose and not always then—something not in their shape nor pigment but in their expression, and the boy’s cousin McCaslin told him what that was: not the heritage of Ham, not the mark of servitude but of bondage; the knowledge that for a while that part of his blood had been the blood of slaves. “Like an old lion or a bear in a cage,” McCaslin said. “He was born in the cage and has been in it all his life; he knows nothing else. Then he smells something. It might be anything, any breeze blowing past anything and then into his nostrils. But there for a second was the hot sand or the cane-brake that he never even saw himself, might not even know it if he did see it and probably does know he couldn’t hold his own with it if he got back to it. But that’s not what he smells then. It was the cage he smelled. He hadn’t smelled the cage until that minute. Then the hot sand or the brake blew into his nostrils and blew away, and all he could smell was the cage. That’s what makes his eyes look like that.” [6]


§1.3. Many questions still surround Fathers’ almost mystical role in this classic novel. The reader is, though, informed of where he got his name. He was described as part Chickasaw (his biological father), part African and part European (his quadroon mother), but his name derived from “Sam (Had-Two‑)Fathers,” since his mother had been married off to a black slave before his birth. Such personal name etymologies (“anthroponomastics”) can form a vital part of social and linguistic research where source material is scarce. Genealogical research has always enjoyed a high degree of interest among informal learners in the United States, in particular of late among descendents of more recent European immigrants whose family records, though now much better searchable online, often end with the Ellis Island Online Database of New York passenger lists.[7] With increasing digitization and networking of birth, marriage and death records from foreign organizations, including most importantly churches, we may expect in the near future to enjoy the capability of tracing, from our home computers, the lives of ancestors reaching back several centuries, and thus add dimensions to our family histories we had imagined long lost. Onomastic resources that might assist in charting the history of the African slaves imported into the Americas, however, are very meager indeed, and not likely to ever be recovered in substantial form. For another indignity imposed on slaves arriving in the harbors of the New World was the stripping of their names, and the assigning of new ones by their masters. Recent research conducted on ship rosters has shown us that transatlantic slaves’ names were not included, but rather just numbers, age, and gender of individuals, much as we might expect in the stock car transportation of cattle to market.[8]


§1.4. And in no less dehumanizing a fashion, slaves sold into the chattel possession of plantation owners of the South were renamed willy-nilly, with no reference to practice in their African homeland (as fragile as this practice may have already been in African communities, where names often changed following important events in the individuals’ lives). Many black Americans thus today carry the European names of or assigned by their ancestors’ owners, their plantation trades, or of any of a number of other associations from their past in the Americas, including new names chosen by emancipated slaves, but very rarely the names of their African past.[9] Aside from the educational and social value a full reckoning of displaced Africans in the Americas would represent to the descendents of slaves, it is not difficult to imagine the geo-linguistic value such rosters would bring to research on the African diaspora.


§1.5. The destinies of slaves and the recording of slave names can be followed back much further in recorded history than many would suspect. While debates among historians concering the role of slavery in classical Greece and Rome can grow uncomfortably heated (see conveniently the survey of research in McKeown 2007), more sober discussions of early state development in 3rd millennium Mesopotamia led, in particular, by scholars from the former Soviet Union and its Eastern European allies, helped to build a theoretical foundation for a longue durée analysis of this abuse and its effects on social progress.[10] I am honored, as a sign of my gratitude for his intellectual generosity and his genuine personal warmth, to dedicate my paper to a close colleague of those discussants, Vyacheslav Ivanov, whom I discovered at UCLA later than I would have wished, but to whom I have stuck like glue since. Slava celebrates his 80th birthday on the 21st of August of this year. While the two of us have had occasion to discuss Babylonian onomastics, I have never compiled for his consideration a list of designations of slaves from early Mesopotamian texts. I hope that the personal names offered here, while, at least to my understanding, not credibly to be connected to any known Babylonian languages, will serve as a basis for further discussions with him.


§2.1. It is understandable that earlier research on slavery in ancient Mesopotamia has concentrated on those periods best reflected in the inscriptional record. While most popular histories cite references to slaves and slave prices culled from the famous Babylonian law codes, certainly it is the documentation from legal contracts on the one hand, and from administrative accounts on the other, that offers the best evidence of the day-to-day existence of slave populations and their overlords. Historians are not entirely clear as to what constitutes chattel slave property, nor in many cases what the social, political or military environments were within, and beyond Babylonian borders that led to the enslavement of often large numbers of individuals. I would like to present here what little I have been able to gather from recent work on what I believe are personal names of slaves in proto-cuneiform documents dating to the Late Uruk period, ca. 3350-3000 BC, many of which derive from irregular excavations and are thus unprovenienced. Indeed, without the rich resources of the Norwegian Schøyen Collection made readily available for study by its owner, our current harvest of, at minimum, 440 personal names, would be reduced to a statistically insignificant 38.[11]


§2.2. We should be clear that much that has been proposed in the identification of laborers in the eras prior to the fully historical Early Dynastic IIIb period (pre-Sargonic Lagash, ca. 2500-2340 BC) is highly speculative, necessarily based as it is on analogies drawn from later periods. Thus it seemed reasonable, in the absence of countervailing evidence, to attach the semantic field of “slave” or “dependent laborer” to graphic precursors of characters know from Ur III and ED IIIb accounts to represent slaves or dependent laborers. The sign geme2 (“female slave”) appears in ED IIIa texts (Fara period, ca. 2600 BC) in a form slightly different from that known in the pre-Sargonic Lagash texts (“SAL×KUR” vs. SAL with the three Winkelhakens of the KUR sign spread out to its corners; see figure 2), itself the precursor of our conventional form of geme2 composed of the element SAL followed by KUR.[12] This component KUR of the compound sign has in all discussion of geme2 been considered a geographical qualifier, thus literally “mountain-woman,” where, with ample textual justification, the chattel slaves of early Babylonia were believed to have been purchased, or taken, by force or threat of force, from the mountains, or more generally foreign lands, to the east or north of the Mesopotamian alluvium. The corresponding male designation ARAD2 derived from the grapheme representing males (NITA) in combination with the same KUR sign.[13]


Figure 1: Map of Mesopotamia, early settlements


§2.3. Successive publications of excavated text artifacts attesting to earlier and earlier phases of cuneiform led, in the mid-1920’s, to the most ancient examples of the writing system. Conventionally known as “proto-cuneiform,” the sign forms found on texts from Jemdet Nasr and Uruk invited comparison, both graphic and semantic, with characters found on later texts. Included with these earliest cuneiform signs was the sign combination SAL.KURa interpreted by Langdon, and following him all other Assyriologists who dealt with these texts, to represent the precursor of geme2 and thus “female slaves.”


§2.4. As with so much of note in researching early Mesopotamian administration, the first systematic discussion of 4th millennium slave designations was published by the Russian scholar A. Vaiman. In a 1974 article, Vaiman reviewed the then available textual evidence and concluded, correctly, that SAL and KURa (KURb-d are graphic variants of this sign) in the archaic texts in fact represented female and male humans, respectively, and that these were recorded much as were the stock of herding accounts, including, in the case of Uruk IV period texts, the qualification of children with a special numerical sign that was otherwise employed to designate fractions of some whole unit.[14] The next discussion of proto-cuneiform designations of archaic laborers was offered by Englund and Damerow in an edition of proto-Elamite texts from Tepe Yahya,[15] followed by a re-interpretation of texts from the Langdon Jemdet Nasr publications by Englund/Grégoire, and by Nissen, Damerow and Englund in a catalogue prepared for an exhibition in Berlin’s Charlottenburg Palace in 1990.[16] Englund provides an overview of previous research on this matter in a 1998 publication.[17] As this research has shown, the accounting for apparent slaves in the Late Uruk period reflected the same degrading abuse of fellow humans as was the defining flaw of the American South, but it collaterally resulted in lists of personal names, names that, in the tradition of Mesopotamia, should bear much linguistic, or at least orthographic information. With the infusion of large numbers of recently available proto-cuneiform texts, we have been able to add very substantially to the list of clear personal names ascribed to humans in the Late Uruk period, and can begin to investigate these names for elements that may support, or by their absence tend to hamper an identification of the language of our earliest cuneiform scribes.

Figure 2: Paleography of geme2


§2.5. The discussion about the “Sumerian question,” that addresses the linguistic affiliation of these archaic scribes, continues, at least in my mind, and has taken a rough edge of late, the more so with publication of the 2003 Leiden Rencontre volume that made no credible advances in the now fairly stale list of “proofs” that Sumerian phoneticisms, or even number words, were a clear element in Late Uruk documents.[18] The lines of sign analysis that have accompanied this research are fairly straightforward. In the first instance, a rebus use of discrete signs (for instance, the words for “arrow” and “life” are homophones in Sumerian, where as in the example below, if correct, the arrow pictogram is more likely to represent “life” than “arrow” or some other homophonic word). There are precious few proposed pairs in this vein of attack, although we would hope that, with improved access to all Late Uruk texts, interested scholars would perform more systematic searches.[19] Second, we might expect to discover the use of phonetic rather than semantic values of signs (see the instance of “sux-pa” below). Third, and most often seen, specialists will attempt to isolate use of phonetic glosses attached to logograms in some way (best known are instances of such phonetic glosses inscribed within sign frames, but also simply near to the sign of reference). This strategy considers the possible combinations in complex graphemes to include semantic element + semantic element (uninteresting for language identification), semantic element + phonetic element (interesting but difficult to identify), or phonetic element + phonetic element (very interesting, and very difficult to identify). I list below a selection of the multivalency proposals made heretofore on Sumerian phonetic signs,[20] together with possible instances of iteration common to Sumerian orthography, and the proposal of M. Powell[21] that the uniquely sexagesimal structure of Sumerian number words offers proof that Sumerians invented proto-cuneiform, where sexagesimal notations are amply attested in the earliest texts. In this regard, we should note the examples of multivalent sign use cited from the other pristine writing systems, Egyptian (with its key example of proposed b3-st for the place name (per)-bastet, “(house) of the goddess Bastet”[22]), Chinese and Mayan. I have set off in bold those candidates for Sumerian in the archaic texts that appear strong, although of these only the very poorly attested šabu carries real conviction.


1) Multivalence?
archaic sign(s) proposed Sumerian interpretation
EN-E2-TI en-lil2-ti, “Enlil (gives) life” (Langdon 1928: VII; Falkenstein 1936: 37-38; etc.)
PA-NAM2-RAD/ZA(A) nam2-sux-pa, /nam-sipa(d)/ (van Dijk 1989: 446)
DARA4/PIRIG+MA alima with MA = /ma/ (Green in Nissen and Green 1987 s.v.)
GA2×AN ama with AN = /am/ (Green, op. cit., and see above, n. 18)
GA2×EN men with EN = /en/ or / men/ (Green, op. cit.)
EN-ME-MU endub, with /en/ of EN (Krebernik 2007: 43)
EN-ME-GI engiz suggests /en/ of EN and / gi/ of GI (Krebernik 2007: 43)
E2-BAHAR2b-NUNUZ zilulu with NUNUZ = /za/ (Krebernik 2007: 43)
GIR2-SU gir2-su (Krebernik 2007: 43)
ZI // SI4 with both = /si/ (Englund 1994: 38, W 9123,a1)
PIRIG+NUNUZ az(a) with NUNUZ = /za/ (Green, op. cit.)
URI3-NA nanna with NA = /na/ (passim)
GI gi (gi4) “return” (Vaiman 1974b: 16)
NUN-ME abgal among “gal-words” in the Lu2 A list, with GAL = /gal/ (see above, n.18)
ŠA3-BU ša3-bu // ED LAK50/ša-bu-nun, OAkk ša-ab-bu-nu-um (Krebernik 2007: 43)[23]
2) Possible Sumerian verbal iteration?
3) Sumerian sexagesimal system?


As is evident from this list, classical graphotactics have played only a minor role in such research, based on strong, though by no means overwhelming evidence that sign sequences in this largely logographic, or even saccades-based[24] ancient orthography were fluid, and not dependable indicators of word or phoneme flow within textual sub-units (“words,” cases or lines).


§3.1. To this discussion I would like to add some material concerning Late Uruk personal names that have often been cited in literature generated by the Berlin based project “Archaic Texts from Uruk,” but never gathered systematically, and that I have only ordered in a preliminary way. The major difficulty in isolating clear instances of personal names, where we must expect that the accounts and perhaps sections of the lexical lists were replete with such designations, is that the text formats do not explicitly identify what is what once you leave the realm of numerical notations, object designations and signs or sign combinations of thematic meaning derived from the lexical lists. Of course, we have been unable to identify, nor should we expect to find, any semantic glosses of personal names—aside from the simple number sign representing “one unit,” these were a millennium off. Frankly, one of the more dissatisfying discussions that I had with Peter Damerow and Hans Nissen in preparation of the Berlin Erlenmeyer exhibition catalogue[25] was in fact having to admit that we could not state whether the sign combination “KU ŠIM,” central though it was to understanding the archival meaning of the core texts in this collection, referred to a human, to a profession, or to a household. We agreed to an individual “human” (brewery foreman), but only as an expedient convention.[26]


§3.2. The same frustrations can be applied down the line to any number of signs or sign combinations that can, due to considerations of tablet format, or as part of a procedure that eliminates from consideration other spatially associated signs whose semantics are identifiable, be isolated. Since we cannot know how many variables are at play in these residual sign combinations, it would be less than prudent to simply assign to them all the role of personal names. There may be though other strategies to increase the likelihood that we are looking at names of specific persons. For instance, you can imagine an automatic text parser that searches all instances of sign combinations from the lexical lists “Professions” (Lu2 A) and “Officials” from all sign strings found in discrete tablet cases (corresponding to “lines”), removes from the resulting list first these lexical notations, then eventual identifiable signs or sign combinations (numerical notations, object designations and so on) from the remainder, and writes a list of all still remaining signs and sign combinations. Aside from possible functional terms, including for instance verbal forms, we would anticipate that these entries represent the personal names of cited household officials. We might also look for parallels in the text formats that isolate distinct personal names for us—for instance, some designation of personnel inventories as was well known in later periods, or, say, a format like later table accounts with some global qualification followed by strings of individual cases, each with signs or sign combinations with no further qualifications.


§3.3. Isolating these names would help to satisfy our curiosity about the conceptual organization of its members that archaic household accountants imposed on their books, but more importantly, since cultural continuity is regularly cited as one of the lynch pins of Sumero-Babylonian civilization, and since personal names as a conservative cultural trait should be discoverable in texts that code, or are coded by Sumerians, this prosopographic material from the Late Uruk texts could play a prominent role in discussions of archaic linguistics. For despite all the caveats offered by specialists in early cuneiform, it has, since my time as a student in Dietz Edzard’s seminars in Munich, reading 3rd millennium texts and examining, as was his wont, earliest sign etymologies, seemed to me curious that if these should be texts written by Sumerians, we did not immediately recognize a substantial number of forms that could at least plausibly be interpreted to represent elements of the Sumerian language—quite aside from the seemingly missing references to the Sumerian pantheon. And in the first instance, I would have expected language, or if you wish, culture-specific patterns to show up in personal names. Still, neither the list Lu2 A, nor the socalled list of officials, gave any clear indication of sign patterns that would comport with later, often predicative formulations in personal names such as “servant of Enlil,” “he is my lord,” or “lady of Inanna.”


§3.4. It turns out that the Late Uruk accounts of herds of animals led us to the sorts of texts that clearly included personal names.[27] Records of such herds, first edited by M. Green,[28] contained data much like that known to specialists working on texts from later periods, including numbers and designations of animals, of their ages and gender, as well of course as identification of their owners, herders, and whereabouts, and the real or anticipated dairy and textile products associated with these animals. As is the case with other types of accounts, these texts detail conceptually important terminological categorizations, for instance qualifying x ewes (sign U8) and y rams (UDUNITA) as x+y small cattle (UDU). Just as with small and large cattle, and, as we are seeing with a substantial recent influx of archaic herding accounts, with donkeys,[29] pig herds were also differentiated according to animal age and use, in the case of cattle also according to gender. The text W 23948[30] records the distribution of animals from a large herd of 95 pigs into two groups of adults associated with large household units in Uruk, and a third comprised of juvenile animals. The juveniles were qualified with a designation borrowed from time accounting metrology to represent animals that had reached the age of one year; one porker, together with ten mature animals, were then, according to this text, possibly slaughtered for the household kitchen.[31]


Figure 3: W 9827 contains an apparent account of a number of groups of male and female laborers, listed individually on the obverse (23+ in the first column, 22+ in the second) and totaled on the reverse (preserved is a notation representing in the sexagesimal system 211+ female and male laborers, in proto-cuneiform SAL KURa).


§3.5.1. During our work on the Uruk III period texts from Jemdet Nasr, Grégoire, Damerow and I noticed that a similar terminology and syntactically motivated text format were visible in accounts of what were, in totals of the texts, qualified as SAL KURa ERIMa and SAL KURa SAG×MA, that is, what we speculated to be “yoked” and “noosed” female and male slaves, following Vaiman’s interpretation of SAL and KURa.[32] With the series of three Jemdet Nasr texts MSVO 1, 212-214, we were able to demonstrate several things. First, that the numbers of individuals qualified as SAL or KURa in archaic texts were not large—at most 211+ recorded on the reverse of the account W 9827, doubtless representing the summation of smaller groups recorded on the obverse (see figure 3).[33] Second, we saw that the accounting procedure of text consolidation, so well attested for later periods of Mesopotamian history, was employed already by household bookkeepers at the dawn of writing. MSVO 1, 213 and 214, were in fact entered, sign for sign, into the larger account MSVO 1, 212. But then third and most significantly, we could see that the accounting format of these texts was very complex, but foresaw the division of individual records into sub-cases with formal differentiations. The first sub-case of one entry contained a numerical notation, an object designation (as we believe, “slave of quality x”) and one or more signs apparently referring to persons or offices. There followed one or more sub-cases, with one exception[34] never with a numerical notation, containing signs that we interpreted to represent the personal names of the designated slaves. Where the initial numerical notation was 1, there were one or two such associated sub-cases; where 2, there were at least two.


§3.5.2. Thus the initial entries of MSVO 1, 212, are (reconstructions according to MSVO 1, 213 obv. i):


1a 1N1 ˹SAL KURa˺ SAG×MA ŠA E2a MUŠEN×2N57
1b1   ˹ZATU751a˺ ERIMa
1b2   [...] X
2a [1N1] ˹SAL KURa SAG×MA ŠA˺ [...]
2b1   ˹DUR2 3N57 ZATU751a˺
2b2   [ABa TUR? N2] KU3a
3a 1N1 KURa E2a ŠA ˹MUŠEN×2N57˺
3b1   SI ˹MA? ENa˺ X
3b2   [GI×KUb1 BAR]
4a 1N1 ˹KURa MUŠEN×2N57˺ [E2a ŠA]
4b1   1N14 ˹UDUa˺
4b2   1N1 [KIŠ KURa]
5a [1N1 SAG×MA GEŠTUc5 MUŠEN×2N57 ]
5b1   [GI ŠA E2a AMAa]
5b2   [TAK4a NIa] ˹SAG ERIMa˺ [MUŠ3a UR2 DUR2]


and the summation of all entries on the reverse:


col. ii
1 1N14 7N1 SAL KURa SAG×MA
2 1N14 SAL KURa ERIMa X [...]
col. iii
1 [2N14] ˹7N1 SAL KURa˺ UB ˹PAa ? SAG×MA SANGAa X ENa˺ N4


§3.5.3. Unfortunately, the complexity of the individual entries in this account makes it very difficult to understand the syntactical relationships among those entities represented by individual sub-cases, and the text would furthermore appear to contradict, with its combination in initial sub-cases of SAL, KURa and 1N1, our belief that SAL denotes a single female, and KURa a single male. I have no convincing explanation for this seeming contradiction. Similar accounts from Uruk with less complex accounting format, however, do help to fill out this picture with terminology more reflective of that known from herding accounts. Where herding texts recorded domesticated animals according to species, gender and age of breeding significance—we expect also qualifying the males as to whether and when they had been castrated—the archaic accounts of groups of humans added new levels of qualification, with clearer differentiation of the terms SAL and KURa, and with designations of slaves that contained greater terminological texture.


§3.6.1 The two Uruk texts in figure 4 are good examples of this accounting procedure. Each has in the left column a total, eight individuals in both texts, corresponding to numerical entries to the right. Clearly enough, the first text[35] lists 1 + 1 + 2 + 2 + 1 + 1 for a total of 8, while the second has (4+1=) 5 + (1+2=) 3 = 8. The latter text demonstrates that SAL and KURa qualify different objects, probably female and male slaves, that are themselves in the accounting terminology further divided into apparent age qualifications. Thus, in the former text we have, viewed syntactically, the qualifications AL, ENa TUR, 1N57×U4 TUR, BULUG3, U2a A and ŠU; in the second text, SAL, KURa and ŠA3a TUR. Several of these designations are terms well known to Sumerologists. TUR (a presumed pictogram of human breasts) representing young children (Sumerian dumu), 1N57×U4 representing “one year,”[36] and AL (picture of a type of hoe) representing “adult” (with later Sumerian reading mah2, this sign usually qualifies sexually mature domestic animals, but is also possibly an element of two personal names in the ED IIIa period, and is even a qualifier of the capacity unit gur [WF 76 rev. x 3]). Finally, ŠU will be associated by some with later šu(-gi4), “old one,” found in many herding accounts and laborer inventories.


Figure 4: The Uruk texts W 20274,2 and W 23999,1 (reverse surfaces are not inscribed) 


§3.6.2. The most compelling accounting practice that emerged from the analysis of these two proto-cuneiform accounts from Uruk was the clear practice of associating numerical notations and general slave designations with sub-cases of signs and sign combinations that corresponded exactly to the numerical notations. Thus, in the first text of figure 4, 1 AL (i 1a) is followed by one sub-case with nonnumerical signs; 2 1N57×U4 TUR (i 3a) by two subcases, each with non-numerical signs. The case with 4 SAL in the second text (i 1b1a is followed by four subcases, each, again, with non-numerical signs. It appears reasonable to conclude that these sub-cases contain personal names associated with individuals recorded in numerical sub-totals to their left (leaving aside a discussion of the true orientation of the proto-cuneiform texts), and that signs or sign combinations associated with these sub-totals qualified the named individuals in very much the same way as herding and dairy accountants recorded gender and age-specific sub-goups of agricultural units.


§4.1. This format was then the “tracer” to locate further instances of the same phenomenon, which differs from accounting formats of herding accounts chiefly in the inclusion of these non-numerical sub-cases.[37] Due in part to the poor state of preservation of most Uruk texts, only about a dozen comparable accounts have been isolated among the more than 5000 tablets and tablet fragments unearthed there in regular excavations, and some few others from other sites.[38] These numbers have been significantly increased with nearly 40 new reference texts that form part of the Norwegian Schøyen collection.[39] One of these artifacts, first observed in Brussels by Philippe Talon, who recognized its significance and kindly posted to me his carefully done copy and transliteration before it entered the Oslo collection with the manuscript no. MS 3035 (figures 5-6), is of particular note.[40]


§4.2. The large account exhibits the same correspondence between cases with numerical notations and associated sub-cases with non-numerical notations that we have seen in smaller texts above. For instance, the section in the lower left of the tablet’s obverse surface (figure 5) contains a numerical notation representing “12” in the sexagesimal system, qualified by 3N57×U4 TUR, probably “three-year-old children.” Exactly 12 sub-cases follow, each with one or more signs representing as many personal names of the individuals summarized in the left-most case.


Figure 5: The section in the lower left of the obverse of the Schøyen text MS 3035 (figure 6) demonstrates the numerical relationship between the initial notation (sexagesimal “12” qualifying a notation that may be interpreted to mean “three-year-old children”) and the number of sub-cases to the right with ideograms that in all likelihood represent personal names. Note the occurence of the same names in sub-cases 2 and 7 (as well as 1b7 of the same column), and the possibility that sub-case 10 is to be interpreted as (KURx. ZA7=) “ZAGINx” = “Lapis,” “Blue(-eyed one).”


§4.3. The account at a higher structural level employs procedures that are well known from the grain accounting office of Jemdet Nasr.[41] The double dividing line down the middle of the text indicates that it is the compilation of two still quite significant accounts, each beginning with the most valuable objects (here AL, presumably adult slaves) and continuing through numbers of less valuable items. The first sub-account appears to be globally qualified by the sign 2N57 MUNa1, the second 1N57 MUNa1. This MUNa1 is likely to represent some sort of accounting (rationing?) period, possibly connected to the sign combination PAPa SUa discussed below, note 43.



Figure 6: MS 3035, a complex account in the Schøyen collection, contains notations representing numbers of apparent slaves qualified according to age, though not (visibly) gender.


Transliteration of MS 3035:

obverse i      
1a 3N14 2N1 AL 2N57 MUNa1 5a 1N14 2N1 U4×3N1 TUR
1b1 PAa 5b1 SIG7 MUŠEN
1b2 KAŠc DAa 5b2 ENa U2b DU
1b3 SUKUDgunûd NIa 5b3 ZATU659
1b4 U4 ŠU 5b4 GI ŠA3a1
1b5 ZATU659 5b5 SAG×GEŠTUb GIŠ
1b6 BUa ŠEa ŠUBUR 5b6 GU4 SAL ENa
1b7 ENa U2b DU 5b7 ENa U2b DU
1b8 EZINUa ENa 5b8 GIR3gunûc ENa
1b9 ZATU659 5b9 E2a BUa
1b10 ZATU659 5b10 3N57 NUNUZc
1b11 ŠU ŠU 5b11 E2b GIR3c
1b12 BUa GI 5b12 SI ADa? AN
1b13 ŠU2 PAPa obverse ii
1b14 3N57 SAL 1a 1N14 1N1 AL
1b15 3N57 A 1b1 ŠEa MUŠEN
1b16 BUa URa 1b2 INb ENa
1b17 3N57 1b3 HAL PAPa
1b18 DIMa 1b4 TUR3a 5N57
1b19 ENa GIŠ×ŠU2a 1b5 PAPa
1b20 ŠU ŠU 1b6 AN TE KI GALa
1b21 EZENb ENa 1b7 ZIa E2a
1b22 ŠU2 SAL 1b8 ZATU773a MAŠ KURa
1b23 ENa AMAR ŠU 1b9 ENa AN E2a
1b24 E2a BUa 1b10 BUa ŠEa
1b25 DIMa 1b11 MUŠ3a NU11tenû
1b26 PAPa 2a 2N1 ENa TUR
1b27 ŠA3a1 A TAK4a 2b1 BAHAR2a BUa
1b28 NAM2 2b2 BUa DU
1b29 BAHAR2a ENa 3a 3N1 U4×1N57 TUR
1b30 ENa U2b HI 3b1 GA2a1 ENa GU
1b31 BAHAR2a 3N57 3b2 ŠU2 BUa
1b32 6N57 ˹U8˺ 3b3 1N57 A NEa
2a 9N1 ENa TUR 4a 3N1 U4×2N57 TUR
2b1 PAa TUN3a 4b1 GI/GI/GI ENa
2b2 SAG HI NIN 4b2 NU (UDUa×TAR)a
2b3 SI BARA3 4b3 ENa ŠITAgunûa ABa
2b4 NIa GIR3c 5a 4N1 U4×3N57 TUR
2b5 BUa+DU6a GUL 5b1 3N57 DUR2 BUa+DU6a
2b6 DIMa 5b2 ENa A
2b8 KIDe NUNa ENa 5b4 ZATU659
2b9 MAHb×NAa 6 1N57 ˹MUNa1˺
3a 7N1 U4×2N57 TUR reverse i
3b1 ŠU2 SAL 1 4N14 3N1 AL
3b2 ŠUR2b 2 1N14 1N1 ENa TUR
3b3 TI SUa ENa 3 5N1 U4×1N57 TUR
3b4 DARb E2a 4 1N14 U4×2N57 TUR
3b5 GIŠxŠU2a ENa 5 1N14 6N1 U4×3N57 TUR
3b6 SI E2a MEa reverse ii
3b7 3N57 SAL 1 1N34 2N14 5N1 1N57 2N57 MUNa1
4a 2N1 U4×1N57 TUR   SUa PAPa 1N58BADa SI AN ADa
4b1 BUa GI   GIRa
4b2 PAPa HAL


§5.1. Using this, and the 50 other accounts registering numbers of humans in this way, we may compile a list of general qualifications for what we interpret to be archaic slaves:


general terms
  KURa male
  SAL female
  SAG head, human[42]
  SAG×MA noosed head
  ERIMa yoked one
  PAPa SUa ?[43]
  AL of working age (“hoer” ?)
  ENa TUR four years old and older up to AL?
  KURa TUR boy, younger than ENa TUR?
  KURa ŠA3a1 boy, very young?
  SAL TUR girl, younger than ENa TUR?
  SAL ŠA3a1 girl, very young?
  ŠA3a1 TUR = KURa/SAL ŠA3
  3N57×U4 (TUR) three-year-old (or: child in 3rd year)
  2N57×U4 (TUR) two-year-old (or: child in 2nd year)
  1N57×U4 (TUR) one-year-old (or: child in 1st year)


§5.2. These then are the higher-level qualifications of persons in proto-cuneiform accounts, quite possibly chattel slaves, or humans in some form of servitude to Late Uruk households. While I must admit to some doubt about the interpretation of the complex signs including “U4” (“day,” but a general anchor for time metrology notations in this period), it may be relevant to mention the analyses by I. Gelb, H. Waetzoldt and others that children of state-dependent laborers will have been assigned full work loads by the age of six or shortly thereafter. If our designation ENa TUR encompasses a period of several years, AL might indeed qualify workers of an age that would appear young to us, but certainly not to many sweatshop owners around the world, and certainly not to the industrialized West prior to such legislation as the British Factory Act of 1833 aimed at curbing abusive child labor in British textile manufacturing. According to this at the time heralded advance in labor rights, children aged nine to thirteen could not be forced to work more than nine hours a day. Nevertheless, why did archaic accountants so exactingly record the ages of children from their first through their third years? This system of dating bears an uncanny resemblance to herding accounts of large cattle and of pigs of later periods, or even of the initial lines of the so-called archaic Pig List.[44] The age designations of domestic animals employed in those accounts are explicit tools known to any dairy or pig farmer; they track age to know when to wean the young, to judge weight gain, and to prepare sexually mature animals for breeding, or to train oxen for the plough. It is difficult to recognize a comparable need in accounting for young children, aside possibly from the intent of accountants to retain strict control of juveniles as they grew to working age. As slave laborers, after all, they would have represented a substantial chattel asset to ancient households.


§5.3. Doubtless, tagging all proto-cuneiform accounts that contain the format for personal names described above will result in a list that is, for a number of reasons, by no means complete. In the first place, H. J. Nissen and his research collaborators have stated again and again that we must understand the nature of the texts taken from Uruk excavations. To make historical, occasionally just aesthetic points, often the best preserved of those accounts are cited and put in illustrative graphics or on book jackets, but these are the tablets that survived more than 5000 years of deposition in Uruk, after having been rudely gathered and tipped, as detritus of a burgeoning administration, into construction projects of the ancients. Most artifacts could not survive such ill treatment intact.[45] Thus the very fragmentary nature of the great majority of our texts gives fair warning that we are missing much of the original depositions, certainly most of the original text material, and that those exemplars we do have are so incomplete as to make a measured judgment of their contents very difficult. In the second place, the state of decipherment of proto-cuneiform approached a natural barrier with publication, in ATU 2 (1987),[46] of the results of research conducted by H. J. Nissen and M. Green on the interpretation of non-numerical signs in the proto-cuneiform texts, and of research conducted by P. Damerow, R. K. Englund and J. Friberg on the numerical signs and sign systems. Advances in the understanding of Late Uruk texts from Mesopotamia have, since that publication, been modest.[47] Particularly the interpretation of much of the source material that is not directly associated with numerical notations, with counted or measured objects, or with signs or sign combinations found attested in the thematically ordered archaic lexical lists whose uninterrupted history of transmission resulted in sign-for-sign copies well into the 3rd millennium, and even into the Old Babylonian period, remains highly problematic. These remaining sets of signs will include personal names.[48]


§5.4. Nevertheless, the limited method of sign and sign string isolation used here has resulted in a list of ca. 450 discrete entries (see the appendix below), each with fair probability representing the given name of an individual. We may look at these personal names in a number of ways. The resolute decipherer will first just count and rank signs, always aware that the sample may be skewed, given that so much now derives from one private collection of inscriptions of unknown provenience. Persons whose names included the sign ENa, possibly the ruler of archaic communities or even of regions, should not surprise us, and this may be the correspondence to lugal in later Early Dynastic personal names. This sign is attested more than twice as often as the runner-up signs BUa (unclear meaning; pictographically “snake,” but its only contextually derived denotation points toward field surveying) and 3N57 (in some and possibly most instances an abstracted form of the sign KURa, “male slave” or perhaps after all also “mountain,” “foreign land”).


§5.5. For comparison, it may be helpful to list the number of attestations of highest frequency signs used in all discovered personal names (left), and the most frequent signs in the proto-cuneiform texts generally (right; excluding lexical list attestations):


ENa 91 ENa 1470
BUa 43 AN 811
3N57 40 GALa 783
PAPa 33 SAL 683
AN 31 GI 679
ŠU 31 BA 662
E2a 24 PAPa 623
DU 21 SANGAa 545
ŠUBUR 21 NUNa 519
MUŠEN 19 ŠU 505
A 17 E2a 463
HI 17
SAL 17
GI 16
KAŠc 16
SAG 14
SI 14
U2b 14
GIR3c 12
ZATU659 12


§5.6. Although I cannot make out a meaningful pattern in these numbers, at least we now have a basis for comparing the frequency of signs used in personal names versus those used in the texts as a whole; such frequency tables can serve, for instance, to test in Babylonian texts the hypothesis of Meriggi, Vallat and Dahl that proto-Elamite scribes developed a syllabary used exclusively to record proper nouns.[49] It might here be more instructive to consider the signs and sign combinations that are most often found in our list as those representing true names of individuals, and to compare these entries with the most frequently attested names in the texts from the “historical” ED IIIb (ca. 2400-2350 BC) and the Ur III (ca. 2050-2000 BC) periods.[50]


Late Uruk, ca. 3200 BC
  names times attested
  ZATU659 10
  PAPa 7
  BUa GI 6
  DIMa 5
  ENa PAPa 4
  ENa U2b DU 4
  EZENb ENa 4
  NIa GIR3c 4
  ŠU ŠU 4
  3N57 SAL 3
  E2a DAH 3
  ENa GIŠ×ŠU2a 3
  UB ZIa 3
ED IIIb, ca. 2400 BC
  names, men names with this element
  dDN-… (in any position) 210
  lugal-… 190
  ur-… 170
  en-… (excluding den-ki/den-lil2) 82
  e2-… 81
  a-… 68
  amar-… 32
  lu2-… 27
  me-… 24
  nam-… 23
  sag-… 20
  names, women names with this element
  nin-… 141
  geme2-… 24
  ama-… 24
Ur III, ca. 2000 BC
  names, men names with this element
  dDN-… (in any position) 1664
  ur-… 683
  lu2-… 589
  lugal-… 585
  …-mu (some = muhaldim) 368
  e2-… 290
  du11/inim-… 197
  dingir-… 157
  ha/he2/hu-… 150
  (en-… 94)
  (amar-… 32)
  names, women names with this element
  nin-… 320
  geme2-… 201
  ama-… 85


§5.7. Comparing the list of proto-cuneiform personal names with those of the most common personal names or name elements in the Early Dynastic and Ur III periods, we see quite substantial differences. First is, our archaic personal names contain no obvious theophoric elements. Indeed, in this list, there is not one instance of a name that might plausibly be interpreted to include a Sumerian divine element, whereas such names outnumber all other examples in both ED IIIb and Ur III texts. Then also, the common elements ur, amar, a (seed) are nearly unknown in the archaic texts, and those instances of ENa (in bold) that we might consider archaic correspondences to later lugal contain other elements that make no sense if interpreted to be Sumerian. Finally, the Sumerian names of women from later periods find no counterparts in the archaic texts.


§6.1. I have stated elsewhere[51] that this search for personal names among slaves might be skewed in another telling way. We might suspect that as in later periods, and as the designations SAG+MA and ERIMa, as well as seeming prisoner scenes on many Late Uruk seals might tend to support, the chattel slaves were above all taken from foreign populations, their names thus in some non-Babylonian language. But frankly, it would surprise me if the Uruk overlords did not rename their foreign slaves with terms comprehensible to the local population, much as did the buyers of African slaves shipped to the Americas, since it is difficult to imagine that those engaged in the exchange and exploitation of humans, of whole families judged as little better than local livestock, would have made an effort to retain their native names. I can offer only indirect evidence that this may have been true. Contracts of the sale of chattel slaves in the Ur III period followed a standard format that included the name of sold persons in the form “one (slave type), PN his/her name, his/her price n shekels of silver ... .”


§6.2. A quick search of available documents, restricting myself for the present to only those contracts and related court records that included the phrase “PN mu-ni-im,” “PN is his/her name,”[52] demonstrates that some of these names are clearly of foreign origin, or are Akkadian, but that the majority carried a plausible Sumerian pedigree.


In Nippur:
  sag nita nam-dumu mu-ni-im
    ur-lugal mu-ni-im
    lugal-ur2-ra-ni mu-ni-im
    ad-da-[…] mu-ni-im
    lu2-den-lil2-la2 mu-ni-im
    šar-ru-a mu-ni-im
    nu-hi-dingir mu-ni-im
    lu2-dsuen mu-ni-im
  guruš i-din-dda-gan mu-ni-im dumu a-bi-ša-ru-um
  sag munus maš-da2-gu-la mu-ni-im
    en-ni-dla-az mu-ni-im
    ni-za-ti-a mu-ni-im
    a-za-za mu-ni-im
    nin-mu-ba-zi-ge mu-ni-im
    geme2-e2-zi-da mu-ni-im
In Ur:
  sag nita2 šu-gu-bu-um mu-ni-im
    en-um-diškur mu-ni-im
    dnin-gir2-su-ka-i3-sa6 mu-ni-im
    dingir-ma-lik mu-ni-im
  sag munus ta-re-ša-am3 mu-ni-im
    i3-li2-bad3-re mu-ni-im
In Wilayah?:
  sag nita2 |PU3.ŠA|-ha-ia3 mu-ni-im
  sag munus na-an-na-a mu-ni-im
    a-ga-ti-ma mu-ni-im
    18-dar-um-mi mu-ni-im
In Umma:
  sag nita a-ba-in-da-an-e3 mu-ni-im
  dumu nita2 a-ba-a-in-da-an-e3 mu-ni-im
  sag munus dba-ba6-lu2-sa6-sa6 mu-ni-im
    nin-mu-ušurx(LAL2.TUG2)-mu mu-ni-im
In Girsu:
  sag nita a-lu2-du10 mu-ni-im
  sag munus geme2-aga mu-ni-im



§6.3. Isolating personal names in the proto-cuneiform texts represents an important beginning in our efforts to lemmatize all proto-cuneiform transliterations with an eye toward identifying the signs that we do understand, or that we believe we understand, and toward more broadly defining what the sign combinations represent that do not correspond to common entries in our lexical lists. I put these data up to underscore the lingering problems in determining the linguistic affiliation of the earliest Babylonian scribes. It may be debated whether the rough translation “male slave” and “female slave” are correct renderings of the proto-cuneiform signs SAL and KURa, but I think the unbiased observer will not reasonably doubt that most, perhaps all of the sign combinations discussed above in selection, and listed in the appendix below, do in fact represent personal names. They are directly, or by association categorized by Late Uruk scribes using terminology that ultimately points to SAL and KURa; they are found in a distinct text format that removes them from the realm of simple object designations; and they do not correspond to entries in the thematic lexical lists.


§6.4. The list of presumed slave names is by no means definitive, but I think a good indication of problems inherent in the archaic Sumerian postulate. Even under the assumption that the personal names in our texts were those of prisoners of war, or of slaves imported into Babylonian bondage from regions surrounding Mesopotamia, and thus were not of the “Uruk core,” sharing the language and culture of their overseers, it remains difficult to understand the absence of, among other linguistic clues, theophoric elements, Sumerian or otherwise. This reminds us of the fact that we have found no lexical god lists of the pantheistic form well attested in the ED IIIa period—it is in fact difficult to point to any clear evidence of anthropomorphic deities in the Late Uruk period at all, once the presumed depiction of Inanna on the Uruk Vase is put in doubt—and that such theophoric elements have not been identified in any other sign combinations that would be credible candidates for personal names. That would leave us with the common elements for males, lu2, lugal, nin, ur, and ARAD2, and for females nin, geme2 and ama—all exceedingly rare, or missing here. If we exchange SAL for geme2, and KURa or 3N57 or, for skeptics, even ŠUBUR for, say, ur, then the corresponding names in our list are not more reflective of expected early Sumerian forms. How much more agreeable this discussion would be if Langdon, now eighty years ago, had been right and not just en-lil2-ti, but other names in this vein had been uncovered in the proto-cuneiform archives!



*  The following is the revised version of a paper presented at the conference Origins of Early Writing Systems, held in Beijing, PRC, in October of 2007. Origins was funded by the CAENO Foundation, New York, and organized by the Department of Near Eastern Languages and Civilizations of Peking University. I am grateful to Henry Zemel, Yushu Gong and Yiyi Chen for their kind support before and during that meeting. Otherwise unpublished (proto-)cuneiform texts will be cited in the article according to persistent URL’s assigned the texts upon entry to the Cuneiform Digital Library Initiative, in the short form <>. Publication of the texts will thus not alter this pathway. Abbreviations of text publications follow < doku.php/abbreviations_for_assyriology>.

2   12 February 1761, signed by ‘Minister of the Kingdom’ Sebastião José de Carvalho e Melo. Slavery was abolished in Brazil with adoption of the Lei Áurea (“Golden Law”) signed in 1888 by Princess Isabel. As elsewhere, a strong incentive to commit to this act of manumission was that slavery was simply not profitable compared to the depressed wages paid poor European immigrants whose labor resulted in no collateral costs—housing, clothing, rationing while sick or during off seasons— whatsoever. Cf. conveniently Schwartz 1996; Pang 1979; Conrad 1972. In an act of “national reconciliation,” many of Brazil’s slavery records were burnt following a 14 December 1890 order of the then Minister of Finance, Rui Barbosa.

3  The parliamentary “Act for the Abolition of the Slave Trade” prohibited slave trade in the British Empire, but not slavery, that would remain legal for another 27 years, in some parts of the kingdom longer. The act levied fines of £100 for each offence, that is, for each slave found to be in transport by British-owned ships. Ingenious captains did not simply transfer their flags to those of Spain, but, when cornered by the Royal Navy, were reported to have dumped their “cargo” at sea (P. S. Foner 1975: 120-122).

4  The law passed on 2 March 1807 in the US went into effect on 1 January 1808, but was rarely enforced (cf. Franklin and Moss 1994: 90-92). It has been conjectured that the prohibition of the slave trade by the UK, and then other European nations and the US, led to the institution of slave “breeding stations” in Virginia and elsewhere in the South. The breeding of slaves, however, was already attested in the late 18th century, due to the rapid expansion of slavery in southern plantations, and to the limited stocks of African slaves entering American ports (Franklin and Moss 1994: 114-120). This chapter of abuse is not well understood and based for the most part on anecdotal histories. But certainly the rapid expansion of slave populations in the US, easily seen in the US census reports beginning in 1790, demonstrate that owners were not repressing pregnancies, and were probably actively promoting them.

5  The 1860 census counted 3,953,760 slaves in the Union. At this time, the slave populations of Mississippi and South Carolina easily surpassed those of free men (434,696 vs. 354,699 and 402,541 vs. 301,271, respectively), though with Virginia in the lead throughout the 19th century in total numbers (1860: 490,887 slaves). Though an abbreviated report due to political turmoil, the 1860 cartographic representations of the Census bureau did serve Union commanders with vital information concerning the populations—white and black— they would expect to encounter, the location of transportation routes, and even the crops they could count on to feed invading troops. See the historical resources of the US Census Bureau at < prod/www/abs/decennial/>.

6  “The Old People,” in: Go Down, Moses (Modern Library 1942, pp. 166-167).

7  <>. Online genealogical resources are growing, with (<>), GenealogyBank (<>), and the Mormon site Family Search (<>) among the better known current services.

8  Curtin 1969 is the first attempt at a more systematic compilation of data documenting this trade from both East and West Africa via European ships to the Americas (“triangular trade”). Curtin concludes that the bulk of the trade went to the tropical Americas (from Brazil up through the Caribbean) and that relatively few slaves (ca. 5% of the total from Africa) entered North America. The ambitious Trans-Atlantic Slave Trade Database sponsored by Emory University and directed by David Eltis and Martin Halbert (<>) combines ship manifests with historical annotation and the market accounts, often anecdotal, available to earlier historians, and will fill in many of the gaps noted by commentators on Curtin, including testing Curtin’s hypothesis that slaves in the US enjoyed a much higher rate of survival than did their counterparts to the south, given census numbers of the mid-20th century. African names of these slaves remain hard to come by. Only in the case of repatriation or legal challenge following the British Slave Trade Act of 1807 were slave cargoes recorded according to African names. These name rosters are the subject of further research by the Emory-led team and others (see <> and Nwokeji & Eltis 2002).

9  A helpful general overview of naming practices, as is to be anticipated highly dependent on the particular language and culture of the naming owners, is offered by Miller and Smith 1997 s.v. “Names;” see, further, the illuminating description of Jamaican slave onomastics in Burnard 2001. Thus, slaves imported to the US from Spanish or Portuguese speaking colonies in the Caribbean often retained (first, but seldom sur-)names drawn from those languages, where slaves from Jamaica or Barbados carried common English names. In many cases, owners drew names from ancient history or the Bible, evidently trying to keep individuals identifiable. See Berlin 2003: 73; he cites, pp. 57-58, Chesapeake plantation owner Robert Carter, writing to his overseer in 1727: “I name’d them here & by their names we can always know what sizes they are of & I am sure we repeated them so often to them that every one knew their names & would readily answer to them.” The correspondence and papers of “King” Carter dating from 1701-1732, including transcribed inventories of slaves, have been made available by the University of Virginia at <http://etext.lib.virginia. edu/users/berkeley/>. For instance, Falls Quarter, located in King George County, listed 24 slaves, among them “Negroes: Sam Foreman, Grace his Wife, Gowin a boy, about 7 years old, Tomboy, about 3, ditto; Bristo a Man, Beck his Wife, Robin, about 6 Ditto, Ben, about 3 Ditto,” etc., going on to record horses, hogs and cattle in precisely the same format, though without personal names. In similar fashion, Ball 1999: 98 describes the 18th century purchase of three slaves in Charleston, South Carolina, with succinct records: “1721 – Bought: Fatima, Hampshire, Plymouth.” While the motivation for naming one of them “Fatima” is open to discussion, the names of the second and third slaves in this record surely derived from favored place names of locales (county, city) near the native Devon of the buyer, Elias Ball. This is not the place for a full discussion of terminology employed by slave owners in the South to qualify their chattel work force according to labor capacity; but I mention in passing that we have ample description of the “hand” terms applied to African slaves. As F. L. Olmstead 1862: 246 has described this system, “The field-hands are all divided into four classes, according to their physical capacities. The children beginning as “quarter-hands,” advancing to “half-hands,” and then to “three-quarter hands;” and, finally, when mature, and able-bodied, healthy and strong, to “full hands.” As they decline in strength, from age, sickness, or other cause, they retrograde in the scale, and proportionately less labor is required of them. Many, of naturally weak frame, never are put among the full hands. Finally, the aged are left out at the annual classification, and no more regular field-work is required of them, although they are generally provided with some light, sedentary occupation” (cf. further Blackburn 1997: 467). Olmstead goes on to describe labor production norms employed, in plantations of eastern Georgia and South Carolina, to chart tasks of field gangs, for instance foreseeing the excavation of 1000 cubic feet of clear meadow soil per full-hand workday, etc.—in uncommon parallel to worker categories and workday norms already recorded in ancient Babylonian labor accounts (the Gullah scholar L. Turner 1949: 283 offers the following: “They have three class: whole hand, and three-quarter, and half hand. The taskrow length is thirty-five feet long. That’s thirty-five feet long-task-row length. The breadth of the task—that the widest of the task cross and cross—is twenty-four bed. This carry twelve row each side. [They] call that one task. Now, these whole hand have to do two task of that one day for day’s work. That’s the whole hand, now. Not a row must [be] left. The three-quarter hand must do one of those whole task and a half. That’s his day’s work. The half hand shall do one of those whole task, and that is his day’s work. That was the way they had them fix” [and see there Appendix H for a Gullah transcription of the Wadmalaw Island, South Carolina, informant’s text]; cp. the parallels to 21st century BC accounts described, for instance, in Englund 1991). How slaves named their own children, so far as they retained some control of them in an American market heated by increasing values, is often unclear, but was also customarily tied to the naming practices of previous owners, or of the owners of their ancestors. Creoles did retain some vestiges of their African past, though as a rule in names reserved for private, not public and thus not documented use. See generally Turner 1949, and Burnard 2001: 329-338 and the more recent literature cited there, p. 328 fn. 7. As has been amply noted, the name “Barack Obama” bears clear witness to the Kenyan Luo heritage of his father.

10  Dandamaev 1984: 30-35 and 67-80, offers a review of the history of philological and social-historical research of Babylonian slavery. The difficult terminology of slave trade and exploitation played a central role in debates conducted mostly in the 1930s and 1960s, debates as to the social status of dependent laborers known in 3rd millennium cuneiform texts as guruš (males) and geme2 (females), and organized in labor troops under the strict control of state foremen. See Struve 1947 and 1969 (engl. translation of a 1949 article). In the 1960s, I. M. Diakonoff and I. J. Gelb opposed the more stringently ideological views of Struve in his application of Marxist formation theory to the particularly Mesopotamian variant of state and empire evolution (“Asiatic mode of production”), including his presumption that Ur III laborers were chattel slaves. In a series of articles, they proposed a more pluralistic model of late 3rd millennium social structure in Babylonia, with only slightly varying opinions about the status of the large numbers of laborers organized in Ur III labor gangs. See, for instance, Diakonoff 1969 and 1976; and Gelb 1965 (particularly pp. 238-241), 1967, 1971, 1972, 1973, 1979, 1982a. Further, Pecírková 1979; V. Afanasieva et al. 1968; Melekišvili 1974; Komoróczy 1978; Brentjes 1987, 175-180; and Westbrook 1995. Englund 1990: 63-68, basing his argument above all on accounting practice, comes down on the side of Diakonoff that there was little difference in practice between the state-organized system of labor (characterized by the terms guruš and geme2) and household chattel slavery, in which male slaves were designated with the sign ARAD2 (in lead lines of contracts of sale often sag nita2, literally “male head”), female slaves with the same geme2 (in contracts of sale often sag munus, “female head”). The chief difference would be that chattel slaves in 3rd millennium Mesopotamia were freely marketable, while laborers in state servitude were not. See more recently B. Studevent-Hickman 2006; Koslova 2008.

11  Assyriologists have taken a lot of flak recently, above all from members of the archaeological community, for their determination to publish and discuss all ancient cuneiform texts, with no regard to their immediate provenience. Thus the American Schools of Oriental Research, and the German Archaeological Institute, are currently restricting the publication of inscriptions that derive from recent antiquities market activity. Despite these roadblocks in scholarly communication and very possibly worse, most will agree that it is incumbent upon researchers to seek and exploit all avenues of evidence relevant to their work, but to condition the information derived from sources of varying reliability (s. for instance Owen 2009: 125-142). Regardless of the irregular origin of many, indeed most cuneiform tablets in public and private collections, specialists are, based on a number of factors, well able to date, and even place in rough geographical locale, these unprovenienced documents, and are therefore able to judge their value in their own research. In the matter of the decipherment, or we should say the description and interpretation of proto-cuneiform, archival locus of text artifacts has in fact played no more than a passing role, insofar as the great bulk of texts derive from regular excavations of Uruk, and as these texts came exclusively from secondary, even tertiary ancient context. They had been discarded in antiquity and, together with the other detritus of administrative households, used to level depressions in underfloors, to fill mud-brick-faced walls, and so on. The private Schøyen cuneiform collection consists of a very substantial number of artifacts, with an over-representation of Old Babylonian and of Late Uruk period texts. The owner was fairly decided in his purchases in acquiring high-impact texts, with a representation of literary, epistolary and mathematical documents that far outweighs their percentage of a normal set of excavated texts. Four volumes of these texts have appeared as of August 2009 (Friberg 2007; Alster 2007; Dalley 2009; George 2009). Together with a small number of Ur III administrative texts published in Owen and Mayr 2007 (nos. 1514-1526), two Gilgamesh witnesses published in George 2003 (vol. 2, p. 7, MS 2652/5 and pp. 8-9, MS 3025) and various other texts published before they were purchased by Schøyen, these editions amount to over 700 published exemplars, a growing fraction of the full collection. The remainder, including my own volume of the Late Uruk collection, are being prepared for publication under the general editorial supervision of Andrew George of the University of London. There can be little doubt but that the historical and linguistic content of this collection rivals that of most national collections on earth. But even if it consisted entirely of mundane copies of long-known literary compositions, it seems to me the ethical imperative of specialists to fully document the texts’ content, and to communicate their findings to the scholarly community as well as to the general public. Those who are not prepared to utilize all sources in their research, including texts available to us through private collections, and certainly those who would presume to limit the access or use in scholarly communications of unprovenienced sources, as has begun to happen with submissions even to such politically neutral editorial boards as those that oversee the publication of papers on the history of mathematics, may want to reconsider the professional choices they have made in their lives.

12  Cf. the forms a-c in the paleographical table compiled by Gelb 1982a: 98. Only the text WF 93 obv. ii 1 attests the sign in clear semantic relationship with the male counterparts guruš in the ED IIIa period. This ED IIIa period sign form was retained in Nippur into the Old Akkadian period (see, for instance, TMH 5, 28 i 7-8 and rev. i 2; 44 rev. ii 4; OSP 1, 23 vii 5; 1, 139 ii; but also the conventional form of other Old Akkadian sites, with exceptions in Nippur [cp. OSP 1, 41 obv. ii 1, and s. OSP 1, 25, 26, 27; OSP 2, 84 [onion archive] i 2), in Isin (BIN 8, 39 obv. ii 9 [and 66 obv. 8?]) and Adab (OIP 14, 56 obv. ii 7’) through ED IIIb. The ED IIIb form cited here (figure 2) is a peculiarity of Girsu.

13  The earliest clear attestations of both ARAD2 and GEME2 are found in the ED I-II (ca. 2800 BC) text UET 2, 259 (with possibly contemporaneous OIP 104, no. 7 obv. i 1; a search for “IR11” in CDLI will list instances of ARAD2, of unclear meaning, in the proto-cuneiform texts). Though this text is beyond the scope of the current paper, it should be noted that it contains on its obverse lists of 23 male and then 12 female personal names, totaled in two cases on the reverse that are qualified with UŠ.KUR and SAL.KUR, respectively. The clear break of the latter sign form from the highly standardized use of its individual components to represent female and male laborers, respectively, in the preceding Uruk phases is another indication of the disruption in proto-cuneiform brought on by the break between Uruk III/Jemdet Nasr and ED I.

14  Vaiman 1974a, in Russian; German translation available in Vaiman 1989. See also Vaiman 1981 (Russian) = Vaiman 1990 (German). The interpretation of the numerical sign N8 as a sign qualifying young animals and children also goes back to the two works by Vaiman.

15  Damerow and Englund 1989, 32003: 24 and 53-57.

16  Englund and Grégoire 1991; Nissen, Damerow, and Englund 2004: 111-120 (English translation published by the University of Chicago Press as Nissen, Damerow, and Englund 1993).

17  Bauer, Englund, and Krebernik 1998 (=OBO 160/1): 176-181.

18  The RAI section organized by G. Whittaker in Leiden and published in van Soldt 2005 (Ethnicity in Ancient Mesopotamia; s. the Tuesday, July 2nd program, p. 452), was ostensibly devoted to the debate concerning phonetic glosses and other language clues in Late Uruk texts (thus not to be confused with the “Sumerian problem” debate that, at the turn of the 20th century, addressed the question of whether Sumerian represented a real language at all). Two papers, one by the organizer (van Soldt 2005: 409-429) and one by G. Steiner (van Soldt 2005: 340-355; Steiner’s statement p. 345 that “all words transmitted in a “Sumerian” context are, independent of their structure, to be understood as “Sumerian” until they have been unambiguously assigned another language” [translation mine], does place skeptics at a distinct disadvantage!) were informed, and informative. (An important third paper offered by J. C. Johnson [“Complex graphemes in the proto-cuneiform corpus and the problem of phonological reconstruction”] unfortunately did not make it to press in this volume, and will be published elsewhere.) However, the papers by G. Rubio (van Soldt 2005: 316-332) and C. Wilcke (van Soldt 2005: 430-445) remind us that contributions to conference volumes are often not subject to the scrutiny of peer review. To be clear, and since both authors expended some effort in responding to points I and others have made in the past concerning the all too marked willingness of Assyriologists to declare the question of the linguistic affiliation of Late Uruk scribes resolved in favor of Sumerian, I have always professed simple agnosticism in the matter and have attempted to keep a running tally of lines of evidence that may be cited on one side or the other. To satisfy Rubio’s uncommon sensibilities, I am happy to retract my modest spoof equating Sumerian culture with Early Dynastic plano-convex bricks (van Soldt 2005: 321-322 and 325; I have otherwise restricted mention of this matter to my classes, where I make clear to those who do not know their history of cuneiform studies that the butt of the half-jest is the long-deceased Stephen Langdon, who, in Langdon 1931: 595, remarked that plano-convex builders of the ED periods may have represented the “recrudescence of the indigenous [=pre-Indo-Sumerian] civilization” of Mesopotamia). Even a passing remark in Bauer, Englund, and Krebernik 1998: 81 n. 170 about qualifier-noun sequences in archaic lexical lists that seemed inconsonant with Sumerian led to an extended discussion by Rubio of ambivalent word order in a list, the pig list, that may be no lexical list at all—with no mention whatsoever of the pertinent compositions I was referring to, especially “Animals” (Englund and Nissen 1993: 89-93) and “Vessels” (Englund and Nissen 1993: 123-134) with a high level of consistency in the use of qualifier-noun sequences. Rubio states that I argue “the so-called “Pig List” constitutes the best example of this word order” (van Soldt 2005: 322), and directs the reader to n. 350 (about color qualifications in archaic lists) of my publication instead of n. 349, which is the only reference I make to a possibly qualifier-noun word order in Late Uruk texts, citing specifically textile entries of the “Vessels” list. But that comment was only offered as a footnote remark recommending a possibly rewarding review of sign sequence in pre-ED IIIb texts that has in my opinion too facilely been described as “unordered.” The apparently consistent order GAL-NOUN and NOUN-TUR in both scholastic and administrative archaic texts (for instance, Lu2 A ll. 35-36 [Englund and Nissen 1993: 76], and Nissen, Damerow, and Englund 2004: 74 to nos. 6 and 11), quite aside from a number of other considerations about Uruk order of ideograms and numerical signs, might further interest those who are curious about such things. Such research as is demonstrated by Rubio in this volume is not rigorous, is in part misleading, and added nothing new to the debate. In his contribution to Ethnicity, Wilcke, on the other hand, appears to want to enter a discussion, in this case of numerical notations and number words (“das Sexagesimalsystem als sprachliches Phänomen,” roughly van Soldt 2005: 431-439), that he enlivens in a fashion that may be entertaining to some, but bothersome to others, and that in no way contributes to the question of Sumerian origins. We may leave aside the fact that he demonstrates limited command of the terminology of numeracy, to give a kind turn to some of his comments; and that he adds little to, and may rather subtract from previous analyses of the numerical notations in the 3rd millennium texts he cites (to his unique reference of an n-final reading of 7(geš2) in Ukg 4 vi 6 etc., we add the multiple instances of 2(geš2)-am2 from administrative Ur III texts, and we note such potential anomalies as 1(geš’u) = /nur/ or even /šar’u/ in MVN 13, 343 obv. 3). For instance, the ED IIIb royal inscription Ent 35 iv 4 (cited by Wilcke in van Soldt 2005: 436) is of unclear, possibly brick metrology, certainly followed by bitumen capacity (//Ukg 7 ii' 3-4; what is geš2.d’ušu?); and his interpretations of Ent 28-29 A ii 25 and iv 11 are conventional and certainly incorrect (p. 436, and including the Lagash II text Gudea Stat B [p. 437, corrected in addendum, p. 444]) and best viewed as simple šar2 gur = guru7 on the one hand, as 4 šar’u gur = 40 guru7 on the other. He should, further, withdraw most of the comments dealing with early numerical sign paleography, for instance van Soldt 2005: 437, n. 23 and n. 25, that are either wrong or hackneyed; an article by an expert on the subject of sexagesimal notations, J. Friberg (Friberg 2005 with very substantial literature), should be substituted for his remarks, van Soldt 2005: 438-439, on ED IIIa–Old Akkadian mathematical texts. When in all of this the author gathers up a bundle of large 3rd millennium numerical notations, and assiduously assigns Sumerian readings to each, thus “proving” their Sumerian origins, we are left to wonder what lines of logic are being proposed. Such reasoning is, in the end, no more credible than is the now standard means of demonstrating phonetic glosses in proto-cuneiform by attaching Sumerian readings to elements in complex signs, derivatively assigning semantic meanings to the base sign, and then citing the semantic root to justify use of the gloss. The prime example of this practice is the ubiquitously cited ama < GA2×AN (AN = am6), for which no evidence whatsoever has been cited from texts that this complex sign refers to “mother,” Sumerian ama. We would most expect this use to show up in personal names, but the sign’s rare occurences in the appendix below (IM 134762 i 2': AMAa ZATU628b N4, <> obv. ii 1.b9: AMAa AN ENa ; MSVO 1, 212 obv. i 4.b3: ˹AMAa˺ ERIMa MUŠEN MAŠ, ii 1.b: ˹AMAa MUŠEN MAŠ KI ZATU694c GI˺) give no indication of meaning “mother,” nor is the sign AMAa the variant (AMAb = GIŠ×AN) that does appear to represent “mother” in the succeeding ED I and later periods (a search through CDLI files will demonstrate that these are independent syntactical entities, and not just orthographic variants, with a significant shift in context and frequency across the period from Uruk III to ED I-II; for the record, I note one potential instance of AMAa = “adult woman” in <> obv. 1b1a; collation needed of a notation that appears to read 2(N14) GI6 AMAa, “20 black AMA’s”[?]). Instead of citing elsewhere in the paper various correct interpretations, or justifiable speculations by Friberg, Wilcke should rather defer to him entirely. It is difficult to locate anything in the rest that deserves our attention, perhaps excepting the fanciful notion that we might attach number words to Uruk V period clay tokens (van Soldt 2005: 439; the author, pp. 441-443, trumps all earlier speculation by transporting Akkadian glosses back to the Uruk IV period Lu2 A list, and in a short excursus pp. 434-436 resolves, to his own satisfaction, a half century of theoretical discussions among historians of science on what constitutes abstract number in Mesopotamia). We must leave to Wilcke and M. Krebernick the determination of the ultimate source of Late Uruk GAL = /gal/ referred to in our list below (under NUN.ME = abgal), for which see van Soldt 2005: 444, with n. 56 citing Krebernik in Gerber, Ehlich and Müller 2002: 64 n. 4 (and cp. Krebernik in Streck and Weninger 2002: 1-2, n. 1; Krebernik 2007: 43 n. 19). In a startling sign of polygenesis, this identification even landed in Glassner 2000 (s. Englund 2005: 114).

19  I have been thinking about the apparent use of the SLEDGE sign GURUŠ to represent workmen (opposed to SAL) in the text MSVO 1, 1, with which one of the participants of the University of Peking conference, Jerry Cooper, has confronted me in past, and, as we shall see, of the sign AL to represent apparent adult humans, consonant with later Sumerian AL = mah2 (it should be noted that the sign MAH in the archaic texts was identified in Green and Nissen 1987 only according to graphic similarity with the sign mah of later periods, following Falkenstein 1936: sign no. 649, and that the sign mah is attested first in the ED IIIa period with both readings mah and al6. MAH has not been identified in texts from the periods ED I-II, and AL in those texts does not occur in the same context as in the archaic texts). We might imagine a language in both cases with homonym pairs SLEDGE = FIELD HAND and HOE = ADULT SLAVE (unless this means simply “hoer”). The remarks of Steinkeller 1990: 22, based on the differentiation of KAL/GURUŠ in the ED IIIa corpus (GURUŠ a strict rectangle, KAL a rectangle with an angled line at the right, thus more graphically similar to the rounding of archaic GURUŠ and the graphic precursor of later kal/ guruš), may not have accounted for the application field of GURUŠ in Uruk III, where it combines with SAL in parallel to KURa (cp. in particular MSVO 1, 1, and ATU 5, pl. 66, W 9579,ac), thus demonstrating a good fit with later GURUŠ/GEME2 and ARAD2/GEME2). Since “KALa” occurs only in the archaic Tribute List as a qualifier of cows, and given its graphic similarity to archaic GURUŠ, it may be that this “KALa” is in fact GURUŠ, that the ED IIIa correspondence of the lexical line (see the images provided at <> of SF 12 and cf. the duplicates SF 13 and MVN 3, 15) is to be read ab2 GURUŠ in the Fara period, and thus that the second sign is to be interpreted as a failed attempt by Fara scribes to understand the original “sledge cow.”

20  See Bauer, Englund, and Krebernik 1998: 77 n. 158, with reference in particular to the reviews of Green and Nissen 1987 (the revised Uruk sign list) by M. Krebernik and P. Steinkeller. The most powerful example of this list would have been the first, en-lil2-ti; it was, however, already shown in Englund 1988: 131-132 n. 9, to be fallacious.

21  Powell 1972: 172.

22  Dreyer 1998: nos. 103-104.

23  Note the potential correspondence of the personal names A ŠA TAK4a and A ŠA3a1 TAK4a in the appendix below (MS 3887 obv. i 4 // MS 3035 obv. i 1b27, MS 2436 obv. i 4b1 and MS 2431 obv. i 4b2?; cp. MSVO 1, 212 obv. ii 8., MS 2998 obv. ii 6, and IM 134954 rev. ii 4b2).

24  J. C. Johnson and A. Johnson (private communication) are investigating the sign clustering of selected ED IIIa period UD.GAL.NUN texts with an eye to understanding how scribes were overcoming the challenges they faced in representing texts through syntactical rather than formally text structural means as was the case in the preceding ED I-II and Late Uruk periods. Their working hypothesis is that a cognitive reading strategy of harvesting sign clusters for interpretation, rather than a strict linearization, is not only at work in early cuneiform orthography, but is a more natural and efficient means of reading. The “saccade” refers to a rapid movement of both eyes in the same direction, the natural way that humans gather visual information; “saccade generation” to such movements in lexical processing. See for instance Rayner 1998; Reichle et al. 1998; Engbert, Longtin and Kliegl 2002.

25  Nissen, Damerow, and Englund 2004.

26  Nissen, Damerow, and Englund 2004: 66-70.

27  Bauer, Englund, and Krebernik 1998: 143-175.

28  Green 1980; cf. Nissen, Damerow, and Englund 2004: 131-138, with further reference to contemporary herding texts from neighboring Iran.

29  There are currently 68 administrative attestations of “KIŠ” in the CDLI corpus (excluding attestations from the Tribute List that exhibit a different sign form, and appear to refer to a different object). See, for instance, the numerous donkey texts edited by Monaco 2007 (CUSAS 1): nos. 31-40, with examples of complex qualifications of animals divided into sub-totals and subsubtotals. A number of archaic Schøyen texts contain comparable accounts, but including records of donkeys qualified SAL and KURa, that is, as jennies and jacks (cf. the CDLI entries to MS 2863/9, 2963 and 4494). CUSAS 1, 40, lists groups of animals qualified as one and two year-olds; as we might expect, the one-year-old animals are further qualified as AMAR—though specifically referring to “calves,” this sign acted as a general designation of young animals in later cuneiform tradition.

30  Cavigneaux 1991: 57; Englund 1995: 125-128.

31  This is a provisional interpretation of numerical signs from the derived system S' where it is employed to qualify herded animals, and possibly humans. See Green and Nissen 1987: p. 131.

32  Above, n. 14. The justification of MA = “noose” in SAG+MA was based on the associated yoke pictogram ERIMa, on the combination of this sign with animal head signs (and thus in those instances not to be understood as a phonetic gloss), and on a consideration of the pictographic referent of MA. This sign, later peš1, is interpreted to reflect the “string of fruit” that Gelb 1982b convincingly explained, and thus “tied-back cord” generally—in our case, tied round the neck of the slaves, thus qualifying them in some way other than the pictographic ERIMa, “yoke.”

33  ATU 5, pl. 118, W 9827; cf. Falkenstein 1936: no. 577 (and see p. 22); Vaiman 1974a: 141, no. 24; Nissen, Damerow, and Englund 2004: 112, no. 13.2; Bauer, Englund, and Krebernik 1998: 178 fig. 66.

34  And this exception, MSVO 1, 212 obv. i 4b1-2 = MSVO 1, 213 obv. i 4.b1-2, recorded ten sheep and one male donkey, KIŠ KURa, probably purchased together with the recorded slave AMAa MUŠEN MAŠ.

35  Note that “LUGAL” in W 20274,2 obv. i 3b1 probably refers to a one-year-old slave child, and thus is not likely to represent anything like “king” of later tradition. The sign combination LU2 GAL is attested 10 times in Uruk texts [from a total of 36,448 lines], never in a context of any social consequence, based on the value of commodities registered in proximate tablet cases, and 55 times in ED I-II texts [from a total of 4004 lines] in personal names of a form that is largely consonant with later usage. These figures would reflect a level of usage of “LUGAL” in the ED I-II period about 50 times that of Uruk IV-III, of course given these numbers to be understood with a grain of salt.

36  Englund 1988: 121-185, especially 156-160.

37  Vaiman 1974a: 140 (=Vaiman 1989: 123), to no. 20, drew attention to the likelihood that ATU 1, 92 (=ATU 5, pl. 81, W 9655,t) with its notation obv. 1: 3N1 2N8, referred to three adult slaves and two slave children, parallel to the use of N8 (N1 rotated 90º clockwise) to designate young animals (cp. ATU 5, pl. 66, W 9579,ai, pl. 92, W 9656,ba, and pl. 109, W 9656,fx).

38  Aside from MSVO 1, 212-214, see, for instance, ATU 6, pl. 64, W 15772,p; pl. 65, W 15772,z; pl. 74, W 15860,a4; ATU 7, pl. 86, W 22104,3; BagM 22, 60, W 23972,2; W 17729,bp+bx, W 20593,11, <> and <> (unpub.); MSVO 1, 217-222; MSVO 4, 58; CUSAS 1, 36 and 174. We might wonder, further, whether the archaic “tags” discussed in Bauer, Englund, and Krebernik 1998: 57-60, as well as a large number of recent additions to CDLI (nos. P387483-P387593, P387698-P387725), recorded names of persons.

39  Above, fn. 11.

40  See <>. A second, wholly parallel text has not reemerged since it went through Belgium, but was copied by Talon and posted to CDLI under <>. A third, though poorly preserved parallel text is MS 2863/18 (<>). We may note that many of these texts give clear indication of gender distinctions in names, for instance the young girls named SAL SAL and TUR3a BALAb vs. young boys named ENa GALa AKa, U4 NIMa and ŠU TUR in <>, obv. ll. 3.b1-2 and 4.b1-3.

41  See Englund 2001, especially pp. 26-27 to MSVO 1, 95-96.

42  See the SAG inventory MS 2437, comprising columns of lines, each with one sub-case containing a numerical notation and sign combinations representing presumable personal names, followed by a second sub-case with only counted SAL. The text, including particularly the summation rev. col. iii, is unclear to me.

43  The total of the account MS 3035 (figs. 5-6 and cf. <> and MS 2863/18, bottom of second column) contains this sign combination where we might expect a general designation of the humans recorded in the text; MS 2498 would tend to support the notion that PAPa SUa qualifies slaves in some general way, with the first cases containing numerical notations qualified with PAPa SUa in parallel to AL on our larger accounts. Cp. in particular MS 2439.

44  Englund and Nissen 1993: 22-23, 100-103; Englund 1995; Bauer, Englund, and Krebernik 1998: 169-175.

45  The attractive state of preservation of many archaic collections gathered from the antiquities markets notwithstanding, since these tablets are what remained after a rigorous sifting process that selected “preserved” and left behind “fragmentary” at the site of plunder, and this sifting continues through the markets down to endbuyer. Though now exposed to the elements, we may hope that future regular excavations will gather in the many thousands of fragments of texts that must well litter the edges of illegal excavations of post-Kuwait war Iraq.

46  Green and Nissen 1987.

47  Research conducted above all by the Oxford Sumerologist J. L. Dahl on the approximately contemporaneous, proto-Elamite accounts from ancient Iran has led to substantive gains in accessing that related writing system. See Dahl 2005a and 2005b.

48  Still, public access to proto-cuneiform texts has moved to an entirely new level since the establishment of the Cuneiform Digital Library Initiative (<>), dedicated to the digital capture and dissemination of all cuneiform sources, but in its initial phases focusing on corpora of the 4th and 3rd millennia. No phase of cuneiform is so well documented online currently as is the Late Uruk period, including image and text representations of nearly all available text artifacts, both edited and unedited. Thus, digital facsimiles of nearly all proto-cuneiform texts are available for free use by all networked researchers, and are being profitably exploited by specialists in their work and publications; one successful recent example is the edition of the Cornell proto-cuneiform collection (Monaco 2007). Further, the field may expect in the next years to avail itself of a federated and persistent website that will facilitate wholesale downloads of data packages and accompanying open source software to better interpret locally the descriptions of early cuneiform texts posted by Assyriologists, by linguists and scholars from other related fields, and by informal learners alike. We may therefore be confident that in the near future the resources for study of onomastics in the archaic texts will steadily improve.

49  Meriggi 1975; F. Vallat 1986: 338-339; Dahl 2005a: §5.5, and nd.

50  The numbers of ED IIIb and Ur III names are to be understood as very preliminary, and more relative than absolute; they are based on a count of attestations in the transliterations available to CDLI. Our files contain ca. 8500 names in the Ur III period.

51  Bauer, Englund, and Krebernik 1998: 176 n. 407.

52  Searching for instances of PN1 ARAD2 PN2 (“PN1, male slave of PN2”), PN1 sag nita2 PN2 (“male ‘head’ of”), PN1 sag munus (“female ‘head’ of”) and PN1 dumu nita2/munus PN2 (“male/female child of”) in our files results in a list of more than 300 occurrences, indicating the range of numbers we might expect in a full set of chattel slave names. My perusal of the names of PN1’s indicated no deviation from the general pattern observed in our list of mu-ni-im names, although the terminological differentiation of slave designations in lead lines of sale contracts (sag nita2/munus and dumu nita2/munus) vs. ARAD2 and geme2 in legal case records (di-til-la) and related legal and administrative references is notable. Such texts as TUT 164-12 indicate that, as is generally understood, the more formal designation of ARAD2 and geme2 in the context of chattel slaves is in fact sag (nita2/munus).


Appendix. List of personal names in “slave” accounts
(signs of individual names have been force-sorted without regard to potential language-revealing sequences; an annotated archaic name glossary will appear in due time in the pages of the CDLI)


A AL MUNa1 TE E2a SI MEa KURa.E2a 3N57
A ENa E2a ŠUBUR LA2? NA2a? X
A KI NEa [...] E2a ZIa LA2 SUG5
A NAR E2a [...] LA2 SUMb
A NEa 1N57 E2b BAR 3N57 LA2 TE
A? SANGAa? [...] E2b KALAMa MA MA
A ŠA3a1 TAK4a E2b 3N57 MAHb×NAa
A TAK4a E2a 3N57 [...] MAŠ MUŠEN
A U2b E2b 3N58 MAŠ2 1N57
A 3N57 ENa EN2.E2b MEa? SAL SAL ZATU751a X
ABa KAKa? ENa GA2a1 GU MEa U8
ABa KU6a ENa GA2a1? NUNUZa0? MEa X X
ABa 5N57 ENa GA2a2 MEa? [...]
ADa? AN SI ENa HI MUD [...]
ADa X ENa HI KAŠc MUŠ3a NU11tenû
MUŠEN×2N57 N24? ENa NIGIN TI NAM2 X [...]
AN DU ZATU735a? ENa NIMb1 1N57? NAR
AN Ea? MEa? [...] ENa PAPa NEa ZATU778
AN ENa DU ENa PAPa [...] NIa SAG TAK4a
AN ENa [...] ENa SAL TE 3N57 NIa ŠU2 U4
AN KI ENa TUa [...] NIRa ZATU773a
AN TAK4a U8? ENa ZATU697c PAa
AN ZIa? ENa X [...] PAPa SAL N2
ANŠEe DU DUR2 7N57 ENa [...] PAPa SUa 3N57
APINb EN2.E2b 3N57 PAPa ŠU
BAHAR2a ENa ERIMa SAG [...] PAPa 3N57
BAHAR2a? ENa AN ERIMa ZATU751a PAPa X [...]
BAHAR2a 3N57 ERIMa [...] PAPa [...]
BANb PAPa GA2a1×ENa NUNUZa1 PIRIGb1 3N57 [...]
BAR X [...] GA2a2×3N57 RU NAR
BU3 A GALa SILA3a×NIa RU [...]
BUa ENa KALb2 MAŠ GAR IGb LAL3a? SAG [...]
BUa LAL2a GI NAa [...] SARa URUa1
BUa MAŠ GI ŠA3a1 SARa 3N57 [...]
BUa MUD NAa GI ŠU? [...] SI4a U4 X
BUa PAPa [...] GI×KUb1 SI TUR3a ZATU773a
BUa SAL GI4a ŠA3a1 SI N1
BUa ŠA3a1 GI4a ŠA3a1 [...] SI 5N57
BUa U2a GIR3c׊E3 NUNa [...] ŠU U4
BUa+DU6a GIR3gunûb 3N57 ŠU 3N57
BUa+DU6a BUa+DU6a EZINUa PAPa GIR3gunûc ENa ŠU X
BUa+DU6a DI NAB GIR3gunûc SUKKAL ŠU [...]
DAH [...] HAL MEa ŠUR2b
DARb E2a HI MUŠEN SAL UR5a ZATU628a TI ZIa? [...]
DARb E2b ŠA HI MUŠEN 1N57 TItenû GIR3c
DIMa HI×1N57/HI×1N57 ENa U2b [...]
DIMa DAa HIgunûb U8 6N57
DIN E2a b KAŠc UD5a
DU E2a PIRIGb1 3N57 KA2×LAM GAa [...] UNUGa [...]
DU KI 3N57 KAŠc KAŠc URI3a [...]
DU TUR3a U2b? KAŠc TAK4a ZATU811 3N57
DU? URI3a [...] KAŠc ZATU823 ZATU819? X
DU N1? X KAŠc X N1 [...]
DU7 ZATU686b KI NU U4 3N57 X
DUGa? SI X X KIa ZATU629a 3N57 X [...]
DUR2 DUR2 KI X X 3N57 [...]
DUR2 ZATU751a 3N57 [...] KISALb1 X [...]
E2a BUa KISIMa KUb1 KU6a
E2a ENa AN KITI 3N57
E2a LAMb MUD KU3a [...]
E2a NEa PAPa KU6a RADa URa
E2a PIRIGb1? UDUa KU6a RADa 3N57
E2a SAG 3N57 KU6a? TUMc X




Afanasieva, Veronika K., et al.
  1968 Fifty Years of Soviet Oriental Studies: Cuneiform Studies. Moscow: Nauka Publ. House
Alster, Bend
  2007 Sumerian Proverbs in the Schøyen Collection. CUSAS 2. MSCCT 2. Bethesda, MD: CDL Press
Ball, Edward
  1999 Slaves in the Family. New York: Ballantine Books
Bauer, Josef, Robert K. Englund, and Manfred Krebernik
  1998 Mesopotamien: Späturuk-Zeit und Frühdynastische Zeit. OBO 160/1. Freiburg, Switzerland: Universitätsverlag Freiburg Schweiz
Berlin, Ira
  2003 Generations of Captivity. Cambridge, MA: Harvard University Press
Blackburn, Robin
  1997 The Making of New World Slavery: From the Baroque to the Modern, 1492-1800. London & New York: Verso
Brentjes, Burchard
  1987 “Ein Nachwort zur ‘asiatischen Produktionsweise’.” Jahrbuch für Wirtschaftsgeschichte Sb. 1987, 175-180
Burnard, Trevor
  2001 “Slave Naming Patterns: Onomastics and the Taxonomy of Race in Eighteenth-Century Jamaica.” Journal of Interdisciplinary History 31/3, 325-346
Cavigneaux, Antoine
  1991 “Die Texte der 33. Kampagne,” BagM 22, 33-123
Conrad, Robert E.
  1972 The Destruction of Brazilian Slavery, 1850-1888. Berkeley: University of California Press
Curtin, Philip D.
  1969 The Atlantic Slave Trade: a Census. Madison, Wisconsin: University of Wisconsin Press
Dahl, Jacob L.
  2005a “Complex Graphemes in Proto-Elamite.” CDLJ 2005/3
  2005b “Animal Husbandry in Susa During the Proto-Elamite Period.” SMEA 47, 81-134
  nd “Early Writing in Iran: a Reappraisal.” forthcoming
Dalley, Stephanie
  2009 Babylonian Tablets from the First Sealand Dynasty in the Schøyen Collection. CUSAS 9. Bethesda, MD: CDL Press
Damerow, Peter and Robert K. Englund
  2003 The Proto-Elamite Texts from Tepe Yahya. American School of Prehistoric Research Bulletin 39. 3rd edition. Cambridge, MA: Harvard University Press
Dandamaev, Muhammad A.
  1984 Slavery in Babylonia. From Nabopolassar to Alexander the Great (626-331 B.C.) (translated from the Russian by M. A. Powell). DeKalb, Illinois: Northern Illinois University Press
Diakonoff, Igor M.
  1969 “Main Features of the Economy in the Monarchies of Ancient Western Asia.” In J. Meuvret, ed, Troisième conférence internationale d’histoire économique: Munich 1965. Paris: Mouton, pp. 13-32
  1976 “Slaves, Helots and Serfs in Early Antiquity.” ActAnt 22 (1974), 45-78
Dreyer, Günter
  1998 Umm el-Qaab I, Das prädynastische Königsgrab U-j und seine frühen Schriftzeugnisse. AV 86. Mainz: Verlag Philipp von Zabern
Engbert, Ralf, André Longtin and Reinhold Kliegl
  2002 “A Dynamical Model of Saccade Generation in Reading based on Spatially Distributed Lexical Processing.” Vision Research 42/5, 621-636
Englund, Robert K.
  1988 “Administrative Timekeeping in Ancient Mesopotamia.” JESHO 31, 121-185
  1990 Organisation und Verwaltung der Ur III-Fischerei. BBVO 10. Berlin: Dietrich Reimer Verlag
  1991 “Hard Work: Where Will It Get You? Labor Management in Ur III Mesopotamia.” JNES 50, 255-280
  1994 Archaic Administrative Documents from Uruk: The Early Campaigns. ATU 5. Berlin: Gebrüder Mann Verlag
  1995 “Late Uruk Pigs and Other Herded Animals.” In U. Finkbeiner, R. Dittmann and H. Hauptmann, eds., Beiträge zur Kulturgeschichte Vorderasiens: Festschrift für Rainer Michael Boehmer. Mainz: Philipp von Zabern, pp. 121-133
  2001 “Grain Accounting Practices in Archaic Mesopotamia.” In J. Høyrup and P. Damerow, eds., Changing Views on Ancient Near Eastern Mathematics. BBVO 19. Berlin: Dietrich Reimer Verlag, pp. 1-35
  2005 Review of J.-J. Glassner, The Invention of Cuneiform: Writing in Sumer. JAOS 125, 113-116
Englund, Robert K. and Jean-Pierre Grégoire
  1991 The Proto-Cuneiform Texts from Jemdet Nasr. MSVO 1. Berlin: Gebrüder Mann Verlag
Englund, Robert K. and Hans J. Nissen
  1993 Die lexikalischen Listen der archaischen Texte aus Uruk. ATU 3. Berlin: Gebrüder Mann Verlag
Falkenstein, Adam
  1936 Archaische Texte aus Uruk. Berlin: Gebrüder Mann Verlag
Faulkner, William
  1955 Go Down, Moses. New York: The Modern Library (copyright William Faulkner 1940, Curtis Publishing 1942; renewed 1968-1970 by Estelle Faulkner and Jill Faulkner Summers)
Foner, Philip S.
  1975 History of Black Americans: From the Compromise of 1850 to the End of the Civil War. Westport, Connecticut: Greenwood Publishing Group
Franklin, John H. and Alfred A. Moss
  1994 From Slavery to Freedom: a History of African Americans. 7th edition. New York: McGraw-Hill
Friberg, Jöran
  2005 “On the Alleged Counting with Sexagesimal Place Value Numbers in Mathematical Cuneiform Texts from the Third Millennium BC.” CDLJ 2005/2
  2007 A Remarkable Collection of Babylonian Mathematical Texts. MSCCT 1. New York & London: Springer
Gelb, Ignaz J.
  1965 “The Ancient Mesopotamian Ration System.” JNES 24, 230-243
  1967 “Approaches to the Study of Ancient Society.” JAOS 87. 1-8
  1969 “On the Alleged Temple and State Economies in Ancient Mesopotamia.” In Studi in Onore di Edoardo Volterra VI. Milan: A. Giuffrè, 137-154
  1972 “From Freedom to Slavery.” In D. O. Edzard, ed., Geschellschafts-klassen in Alten Zweistromland und in den angrenzenden Gebieten: XVIII Rencontre assyriologique internationale. Munich: Beck, pp. 81-92
  1973 “Prisoners of War in Early Mesopotamia.” JNES 32, 70-98
  1979 “Definition and Discussion of Slavery and Serfdom.” Ugarit-Forschungen 11, 283-297
  1982a “Terms for Slaves in Ancient Mesopotamia.” In M. A. Dandamayev, ed., Societies and Languages of the Ancient Near East: Studies in Honour of I. M. Diakonoff. Warminster: Aris and Phillips, pp. 81-98
  1982b “Sumerian and Akkadian Words for ‘String of Fruit’.” In G. van Driel, et al., eds., zikir šumim: Assyriological Studies Presented to F. R. Kraus. Leiden: EJ Brill, 67-82
George, Andrew
  2003 The Babylonian Gilgamesh Epic: Introduction, Critical Edition and Cuneiform Texts. Oxford: Oxford University Press
  2009 Babylonian Literary Texts in the Schøyen Collection. CUSAS 10. Bethesda, MD: CDL Press
Gerber, Erika, Konrad Ehlich and Jan-Dirk Müller, eds
  2002 Materialität und Medialität von Schrift. Bielefeld: Aisthesis Verlag
Glassner, Jean-Jacques
  2000 Ecrire à Sumer: l’invention du cunéiforme. Paris: Seuil
Green, Margaret W.
  1980 “Animal Husbandry at Uruk in the Archaic Period.” JNES 39. 1-35
Green, Margaret W. and Hans J. Nissen
  1987 Zeichenliste der Archaischen Texte aus Uruk. ATU 2. Berlin. Gebrüder Mann Verlag
Komoróczy, Géza
  1978 “Landed Property in Ancient Mesopotamia and the Theory of the So-called Asiatic Mode of Production.” Oikumene 2, 9-26
Koslova, Natalya
  2008 “Bezeichnungen der Arbeitskräfte in Umma der Ur-III-Zeit.” In S. J. Garfinkle and J. C. Johnson, eds., The Growth of an Early State in Mesopotamia: Studies in Ur III Administration. BPOA 5. Madrid: Consejo Superior de Investigaciones Científicas, pp. 149-206
Krebernik, Manfred
  2007 “Zur Entwicklung des Sprachbewusst seins im Alten Orient.” In C. Wilcke, ed., Das geistige Erfassen der Welt im Alten Orient. Wiesbaden: Harrassowitz, pp. 39-61
Langdon, Stephen H.
  1928 The Herbert Weld Collection in the Ashmolean Museum. OECT 7. Oxford: Clarendon Press
  1931 “A New Factor in the Problem of Sumerian Origins.” JRAS 1931, 593-596
McKeown, Niall
  2007 The Invention of Ancient Slavery? Duckworth Classical Essays. London: Duckworth
Melekišvili, Georgij A.
  1974 “Esclavage, féodalisme et mode de production asiatique dans l’Orient ancien.” In Sur le “Mode de production asiatique.” Paris: CERM, pp. 257-277
Meriggi, Piero
  1975 “Der Stand der Erforschung des Proto-Elamischen.” JRAS 1975, 105
Miller, Randall M. and John D. Smith, eds.
  1997 Dictionary of Afro-American Slavery. Westport, Connecticut: Greenwood Publishing Group
Monaco, Salvatore
  2007 The Cornell University Archaic Tablets. CUSAS 1. Bethesda, MD: CDL Press
Nissen, Hans J., Peter Damerow, and Robert K. Englund
  1993 Archaic Bookkeeping. Chicago: University of Chicago Press
  2004 Frühe Schrift und Techniken der Wirtschaftsverwaltung im alten Vorderen Orient. 3rd edition. Hildesheim: Franzbecker
Nwokeji, G. Ugo and David Eltis
  2002 “The Roots of the African Diaspora: Methodological Considerations in the Analysis of Names in the Liberated African Registers of Sierra Leone and Havana.”History in Africa 29, 365-378
Olmstead, Frederick L.
  1862 The Cotton Kingdom: A Traveller’s Observations on Cotton & Slavery in the Slave States. New York: Mason Brothers
Owen, David I.
  2009 “Censoring Knowledge: The Case for the Publication of Unprovenanced Cuneiform Tablets.” In J. Cuno, ed., Whose Culture? The Promise of Museums and the Debate over Antiquities. Princeton: Princeton University Press, pp. 125-142
Owen, David I. and Rudi H. Mayr
  2007 The Garšana Archives. CUSAS 3. Bethesda, MD: CDL Press
Pang, Eul-Soo
  1979 “Modernization and Slavocracy in Nineteenth-Century Brazil.” Journal of Interdisciplinary History 9/4, 667–688
Pecírková, Jana
  1979 “Social and Economic Aspects of Mesopotamian History in the Work of Soviet Historians (Mesopotamia in the First Millennium B.C.).” ArOr 47, 111-122
Powell, Marvin A.
  1972 “Sumerian Area Measures and the Alleged Decimal Substratum.” ZA 62 (1972) 165-221
Rayner, Keith
  1998 “Eye Movements in Reading and Information Processing: 20 Years of Research.” Psychological Bulletin 124/3, 372-422
Reichle, Erik D., et al.
  1998 “Toward a model of eye-movement control in reading,” Psychological Review 105, 125-157
Schwartz, Stuart B.
  1996 Slaves, Peasants, and Rebels: Reconsidering Brazilian Slavery. Urbana & Chicago: The University of Illinois Press
Steinkeller, Piotr
  1990 “Threshing Implements in Ancient Mesopotamia, Cuneiform Sources.” Iraq 52, 19-24
Streck, Michael P. and Stefan Weninger, eds.
  2002 Altorientalische und semitische Onomastik. AOAT 296. Münster: Ugarit-Verlag
Struve, Vasilii V.
  1947 “Social Structure in Southern Mesopotamia during the 3rd Dynasty of Ur” (in Russian). Yubileinyi sbornik ... II. Moscow & Leningrad: Nauka, pp. 720-742
  1969 “Some New Data on the Organization of Labour and on Social Structure in Sumer during the Reign of the IIIrd Dynasty of Ur.” In I. M. Diakonoff, ed., Ancient Mesopotamia: Socio-Economic History. Moscow: Nauka, pp. 127-172
Studevent-Hickman, Benjamin
  2006 The Organization of Manual Labor in Ur III Babylonia. PhD thesis, Harvard University
Turner, Lorenzo Dow
  1949 Africanisms in the Gullah Dialect. Chicago: University of Chicago Press (republished Columbia, South Carolina: University of South Carolina Press, 2002)
Vaiman, Aizik A.
  1974a “The Designations of Male and Female Slaves in the Proto-Sumerian Writing System” (in Russian). VDI 1974/2, 138-148
  1974b “Über die protosumerische Schrift.” ActAnt 22, 15-27
  1981 “On Deciphering the Proto-Sumerian Writing System” (in Russian). VDI 1981/4, 81-87
  1989 “Die Bezeichnung von Sklaven und Sklavinnen in der protosumerischen Schrift.” BagM 20, 121-133
  1990 “Zur Entzifferung der proto-sumerischen Schrift (vorläufige Mitteilung).” BagM 21, 116-123
Vallat, François
  1986 “The Most Ancient Scripts of Iran: the Current Situation.” World Archaeology 17/3, 335-347
van Dijk, Jan
  1989 “Ein spätaltbabylonischer Katalog einer Sammlung sumerischer Briefe.” OrNS 58, 441-452
van Soldt, Wilfred, ed.
  2005 Ethnicity in Ancient Mesopotamia: Papers Read at the 48th Rencontre Assyriologique Internationale, Leiden, 1-4 July 2002. Leiden: Nederlands Instituut voor het Nabije Oosten
Westbrook, Raymond
  1995 “Slave and Master in Ancient Near Eastern Law.” Chicago-Kent Law Review 70, 1631-1676


Version: 21 August 2009  

Cite this Article

Englund, Robert K. 2009. “The Smell of the Cage.” Cuneiform Digital Library Journal 2009 (4).

Englund, Robert K. (2009). The Smell of the Cage. Cuneiform Digital Library Journal, 2009(4).

Englund, Robert K. (2009) “The Smell of the Cage,” Cuneiform Digital Library Journal, 2009(4). Available at: (Accessed: March 26, 2023).

  • note = {[Online; accessed 2023-03-26]},
  • address = {Oxford; Berlin; Los Angeles},
  • author = {Englund, Robert K.},
  • journal = {Cuneiform Digital Library Journal},
  • number = {4},
  • year = {2009},
  • month = {aug 21},
  • publisher = {Cuneiform Digital Library Initiative},
  • title = {The {Smell} of the {Cage}},
  • volume = {2009},

AU  - Englund,  Robert K.
DA  - 2009/8/21/
ID  - temp_id_237872003222
IS  - 4
J2  - CDLJ
SN  - 1540-8779
T2  - Cuneiform Digital Library Journal
TI  - The Smell of the Cage
UR  -
VL  - 2009
Y2  - 2023/3/26/
ER  -