by Denise Schmandt-Besserat

       Small clay tokens of multiple shapes found in Near Eastern sites of 8000-3000 B.C. constitute the earliest evidence for accounting. 1 (Slide 1) The tokens had two main functions: they served as counters to calculate quantities of goods; they were mnemonic devices used to store data. In this paper I will discuss this latter function: 1. How accounting was practiced with tokens, and 2. how accounting was related to sociopolitical changes.

I. Accounting Before Writing

      The study of tokens may give an insight into several aspects of accounting before the invention of writing and, in particular, what information was kept, where, how, by whom, and for how long.

1. What Information Was Kept

      The hypothesis that there was an evolution, from tokens to the Sumerian pictographic writing, allows the meaning of a number of counters to be established. The tokens matching Sumerian pictographs stood for units of merchandise, suggesting that each counter stood for a specific unit of a commodity such as, for example, a particular measure of grain, a jar of oil or a fleece of wool. Consequently, it appears that the token system was restricted to the recording of exclusively economic data.

      It is noteworthy that the goods recorded with tokens changed over the millennia. When the token system came about, ca. 8000 B.C., the first assemblages of counters consisted mainly of cones, spheres, disks, and cylinders ( See Slide 1) that stood for quantities of grain and livestock. In other words, the earliest tokens probably recorded the most basic staples.

      These plain tokens continued to be used to the very end of the system in the third millennium. About 3500 B.C., however, "complex tokens," characterized by multiple new shapes and markings, appeared in cities. The majority of these stood apparently for finished products, such as bread, oil, perfume, wool and rope, and for items produced in workshops such as metal, bracelets, types of cloths, garments, mats, pieces of furniture, tools and a variety of stone and pottery vessels. The object of accounting varied, therefore, in the course of time. At first it dealt with the products of the farm and the country and, later, with goods manufactured in cities.

2. Where Data Was Stored

      When tokens were recovered within buildings, the structures suggest that, from early on, accounting with tokens belonged to non-secular places. For example, in the sixth millennium B.C. site of Hajji Firuz, Iran, a number of cones were located in a house showing no trace of daily life activities, such as cooking or flint chipping. 2 The building also differed from the usual domestic architecture in other ways. First, it was smaller, consisting of a single room, instead of the normal two-roomed units. Second, unusual features, such as a low platform and two posts, were erected inside. 3 In sum, the structure that yielded most tokens at Hajji Firuz was apparently serving a special non-domestic function.

      In the fourth millennium B.C., public buildings that produced tokens took the form of citadels or temple precincts. It seems revealing, for instance, that among the 778 tokens excavated at Uruk, the first and most important Sumerian city, 685 or 88.1% were excavated in the sacred precinct of the goddess Inanna, the tutelary goddess. From the remaining 93 examples, 43 or 5.5% originated from a second shrine, the Anu temple, and only 50 or 6.4% from the city's private quarters. The tokens from Uruk, therefore, can be said to belong primarily to sanctuaries. 4

      The repeated occurrence of complex tokens in gate rooms, and in particular temple gates, may also be singled out as being of particular significance. In Habuba Kabira, Syria, for instance, the southern city gate, which led to the Temple area, produced not only tokens, but other administrative material such as cylinders and seals. 5 At Uruk, tokens were concentrated at the periphery of the religious precinct, and in particular, in the eastern and western part of the temple area, where the gates to the precinct were presumably located. Now, gateways to cities, temples and palaces have traditionally been associated with administrative centers in the ancient Near East. 6 The Bible also alludes to the city gates as the location where public procedures took place and where records were kept. 7 The significance of the gateway in administration in historical as well as prehistoric times may therefore explain why the temple gates are so often depicted in ancient Near Eastern art. The two bundles of reeds, symbolizing it, appear time and time again on seals 8 and on many fourth millennium ritual stone vessels. 9 In particular, on a famous vase from Uruk, the Priest-King of Uruk was represented leading a procession of offering bearers towards the temple gate, shown filled with goods. 10

3. The Quantity of Data Stored

      Clusters of counters recovered in situ seem to indicate that the accounts kept in archives by the means of tokens dealt with small quantities of different kinds of commodities.

      Tokens are frequently found in clusters varying in size from two to about one hundred counters. In one case, for instance, thirty-five cones and one sphere were nestled together; 11 Other Iranian sites produced small hoards of 2 to 37 tokens of mixed types tightly packed together. Yet another Iraqi excavation yielded a dozen clusters of 4 to 16 tokens. 12 Finally in Syria, a room of Sabi Abyad produced 31 tokens and small groups of at most 9 tokens were recovered in five different rooms of a large building of Habuba Kabira. Furthermore, the token clusters were always composed of several types of counters. 13 For example, Uruk produced a cache of 75 tokens, composed of spheres, large spheres, cones, tetrahedrons and cylinders. 14

4. The Method of Accounting

      The system worked according to the most simple and basic principle of one-to-one correspondence which consisted in matching each unit of a set to be recorded with a token. Accordingly, one jar of oil was shown by one ovoid, six jars of oil by six ovoids, etc... (Fig. 1) There were seemingly only a few tokens that stood for a collection of items, like the lentoid disk which probably meant "a flock" (- perhaps 10 animals). The token system did not allow one to express numbers abstractly. There was no token for "one," "two" and "three" independently from the commodity counted.

      The evolution of the token system seems to reflect an ever increasing need for accuracy. This is exemplified, for example, by tokens dealing with livestock: the early plain cylinders and lentoid disks apparently stood for "heads of livestock" whereas the fourth millennium complex tokens indicated the species, "fat-tail sheep;" the sex, "ewe;" and the age, "lamb." The quantum jump in the number of token types and subtypes, which occurred in large cities about 3500 B.C., seems therefore to indicate a concern for more precise data.

5. Who Practiced Accounting

      In rare, but significant instances, tokens were recovered in a funerary setting. These findings may reveal who handled the tokens and practiced accounting.

      Among the thousands of graves dating from 8000 to 3000 B.C. excavated in the Near East, only a dozen are known to have yielded tokens. In these instances the counters were visibly restricted to the tombs of individuals of high status, some richly furnished. In one case the interments held a variety of alabaster vessels and ornaments of dentalia shells and carnelian. 15 Three of the burials of the North Mesopotamian site of Tepe Gawra seem particularly diagnostic. They were among the richest of the site, holding gold rosettes and beads, precious vessels of obsidian, serpentine, electrum. These particular tombs also included symbols of power such as maceheads and lapis lazuli seals. Tomb 107 included special architectural features: a shrine was erected above the burial of an obviously very prestigious individual who held six stone spheres as his only funerary gift. 16

      The rarity of funerary tokens, their association with luxurious burial deposits, with artifacts symbolizing power and with special architecture, seems to indicate that the counters did not belong to the masses, but were the privilege of an elite.

6. How Long the Accounts Were Kept

      An ancient rubbish pit of the fourth millennium B.C., excavated in Iran may give an insight on the time span during which the accounts were kept. The layers of trash that could be distinguished according to seasons revealed that tokens were most often associated with early summer deposits. The excavator noted, therefore, that the counters were discarded in the traditional season for plenty, after the harvest and threshing, when the crops would be stored. 17 This in turn, might suggest that transactions were made in the course of the year to be completed at the time of the harvest. If this was so, the usual length of keeping accounts in archives was less than a year.

II. Tokens and Socioeconomic Development

      The fact that the first stage of the token system (plain tokens) coincided with agriculture and the second (complex tokens) with urban formation, seems of great significance. It clearly indicates that the evolution of accounting coincided with major socioeconomic developments.

      The appearance of the first token assemblages in 8000-7500 B.C. suggests that the system of counting and record keeping of goods became necessary when survival depended on domestication of grains and accumulation of the products of harvests. In other words, the token system fulfilled new needs for accounting brought about by agriculture, and data storage can be considered as directly related to the "Neolithic Revolution."

      The multiplication of token types and subtypes in the fourth millennium B.C. was also not by chance. These complex tokens, which included many new forms and which were characterized by bearing incised lines and punctations, corresponded to the creation of cities. It can be presumed that the creation of workshops, and the more diversified urban economy that followed, required more accounting techniques.

      It is likely, however, that the two major changes in the token system were not mere corollaries of farming and cities but were determined by new sociopolitical structures that derived from agriculture and urban settlements. I propose that accounting may be related to the rise of an elite. This idea is supported by three sets of data.

      First, the creation of the token system coincided with a new settlement pattern characterized by larger communities. For example, in an early community in Syria, Mureybet, there is no evidence for the use of counters in the two earliest phases of the site, about 8500-8000 B.C., when it was a small compound of 0.5 ha. Tokens occur in the third phase, ca. 8000-7500 B.C. when the hamlet had grown to become a village covering 2 or 3 ha. It is difficult to evaluate the size of the population of Mureybet III. It is estimated, however, that the community then exceeded the number of individuals manageable in an egalitarian system. In other words, in the case of Mureybet, and probably elsewhere, the first token assemblage coincided with the advent of rank society characterized by a new type of leadership overseeing the community resources. 18

      Second, as discussed above, the tokens recovered in the tombs of prestigious individuals suggest that, from the Neolithic to the Bronze Age (6000-3500 B.C.), the use of counters served as an instrument of power for an elite controlling real goods.

      Third, the fact that the complex tokens phenomenon took place at the time of state formation was no coincidence. In all the major ancient Near Eastern cities such as Uruk, Susa, Chogha Mish and Habuba Kabira the complex counters occur in levels characterized by 1. seals and sealings featuring the ruler; 19 2. public buildings built according to an identical plan and decorated with typical mosaics; 20 and 3. potteries which probably served as grain measures. 21 In other words, the administrative centers that yield complex tokens were the seats of a large bureaucracy, housed in similar buildings, using the same administrative devices--complex tokens, seals and grain measures--and, most importantly, were headed by the same powerful ruler. The complex tokens can be considered, therefore, as being part and parcel of the bureaucracy used by the Sumerian king to govern and control real goods in the first city-states.

      The appearance of tokens in the earliest rank societies, their inclusion in rich burials, and the place of the complex tokens in the state bureaucracy, suggest that, from the beginning, accounting was the privilege of an elite and that the more the system became efficient and precise, the more power it wielded.


     A faint sketch of the places and people associated with ancient accounting begins to emerge. Counting and data storage with tokens started in open air compounds where subsistence was based on cultivating or, at least, hoarding cereals. Their first purpose was to record quantities of the traditional Near Eastern staples: grain and small stock. In the fourth millennium B.C. complex tokens kept track of manufactured goods in large centers. Starting in the sixth millennium, tokens are recurrently found in public buildings. The clusters of tokens found in situ range usually between a dozen to 75 artifacts, showing that the counters were never kept in large quantities. There is some evidence that the counters were mostly discarded during the summer, after the harvest. Tokens, together with other status symbols, are sometimes included in the burials of prestigious individuals, suggesting that they were used by the elite who controlled the economy of redistribution.


Slide 1. Plain tokens, Mesopotamia, present day Iraq, ca. 4000 B.C. The cone, spheres and disk represented various grain measures; the tetrahedron stood for a unit of labor. Courtesy Denise Schmandt-Besserat, The University of Texas at Austin. Text

Click for larger image

Slide 2. Complex tokens representing (from right to left above:) 1 sheep, 1 unit of a particular oil (?), 1 unit of metal, 1 type of garment, (from right to left below), 1 type of garment, (?) and one measure of honey. from Susa, Iran, ca. 3300 B.C. Courtesy Musée du Louvre, Départment des Antiquités Orientales, Paris. Text

Click for larger image

Slide 3. Six ovoid tokens representing an account of six units of oil. The tokens were held in an envelope which bore the seals of the administrators involved. The artifacts were excavated in Uruk, Iraq. Courtesy Archaeologiches Institut, Abteilung Baghdad. Text

Click for larger image

Slide 4. Envelope and its content of tokens representing 7 units of oil, from Uruk, present day Iraq, ca. 3300 B.C. Courtesy German Archaeology Institute, Division Baghdad. Text

Click for larger image

Slide 5. Envelope and its content of tokens and the corresponding markings, from Susa, Iran, ca. 3300 B.C. Courtesy Musé du Louvre, Département des Antiquités Orientales, Paris.

Click for larger image

Fig. 6. Impressed tablet featuring an account of grain, from Susa, Iran, ca. 3100 B.C. Courtesy Musée du Louvre, Départment des Antiquités Orientales, Paris. Text

Click for larger image

Slide 7. Pictographic tablet featuring an account of 33 measures of oil, from Godin Tepe, Iran, ca. 3100 B.C. Permission to reproduce to be requested from Dr. T. Cuyler Young, Royal Ontario Museum, Toronto, Ontairo M5S 2C6 CANADA      tel. 416 978 8669

Click for larger image

1. Denise Schmandt-Besserat, How Writing Came About. The University of Texas Press, Austin, Texas, 1996. Text

2. Mary M. Voigt, Hajji Firuz Tepe, Iran: The Neolithic Settlement. University Monograph No. 50, Hasanlu Excavations Reports, Vol. I. Philadelphia 1983, P. 87 and 181-184, H.F. 68-122, 68-170-172, 68-189 and 68-195. Text

3. Mary M. Voigt, 1983, P. 47-49. Text

4. Denise Schmandt-Besserat, Before Writing. The University of Texas Press, Austin, Texas, 1992, p. 49-73. Text

5. Denise Schmandt-Besserat, "Tokens, Envelopes and Impressed Tablets at Habuba Kabira." In Eva Strommenger & Kay Kohlmeyer eds., Habuba Kabira S¸d - Die Kleine Funde. Wissenschaftliche Veröffentlichung des Deutschen Orient Gesellschaft, (Forthcoming). Text

6. Jean-Marie Durand, "L'Organisation de l'Espace Dans le Palais de Mari: Le Témoignage des Textes." In E. Levy, ed. Le Système Palatial en Orient, en Grèce et ý Rome. Actes du Colloque de Strasbourg 19-22 juin 1985. Travaux du Centre de Recherche sur le Proche-Orient et la Grèce Antique, Vol. 9, Strasbourg 1987, pp. 44-48. Text

7. J. Andrew Dearman, "On Record Keeping and the Preservation of Documents in Ancient Israel (1000 to 587 BCE)" Libraries and Culture Vol. 24, No 3, 1989, P. 346. Text

8. Anton Moortgat, Vorderasiatische Rollsiegel. Gebr. Mann Verlag, Berlin 1988, P. 7, Pl. 2: 9; 5: 29b; 6: 30, 31. Text

9. Pierre Amiet, L'Art Antique du Proche Orient. Editions d'Art Lucien Mazenod, Paris 1977, pp. 354-355, Figs. 231 and 236. Text

10. E. Lindemeyer & L. Martin, Uruk, Kleinfunde III. Philipp von Zabern, Mainz am Rhein, 1993, No. 226, Plate 25 i. Text

11. Martha Prickett, Man, Land and Water: Settlement Distribution and the Development of Irrigation Agriculture in the Upper Rud-i Gushk Drainage, Southeastern Iran. Phd Dissertation, Harvard 1985, P. 539. Text

12. Sabah Abboud Jasim & Joan Oates, "Early Tokens and Tablets in Mesopotamia: New Information from Tell Abada and Tell Brak." World Archaeology 1986, Vol. 17, No. 3, P. 352-355. Text

13. Peter M.M.G. Akkermans & Marc Verhoeven, "An Image of Complexity: The Burnt Village at Late Neolithic Sabi Abyad, Syria." American Journal of Archaeology Vol. 99, No. 1, 1995, p. 99, fig. 14; Denise Schmandt-Besserat, "Tokens, Envelopes and Impressed Tablets at Habuba Kabira." In Eva Strommenger & Kay Kohlmeyer eds., Habuba Kabira S¸d - Die Kleine Funde. Wissenschaftliche Veröffentlichung des Deutschen Orient Gesellschaft, (Forthcoming). Text

14. Heinrich J. Lenzen, Vorläufiger Bericht über die von dem Deutschen Arch”ologischen Institut und der Deutschen Orientgesellschaft aus Mitteln der Deutschen Forschungsgemeinschaft unternommenen Ausgrabungen in Uruk-Warka. Berlin, Vol. XXV, 1974, P. 40. (Da XV 4/5 is to be read Oa XV 4/5). Text

15. F. El-Wailly & B. Abu Es-Soof, "Excavations at Tell es-Sawwan, First Preliminary Report (1964)." Sumer Vol. 21, No. 1/2, 1965, pp. 26-28. Text

16. Arthur J. Tobler, Excavations at Tepe Gawra. Vol. II. University Museum Monographs, Philadelphia 1950, pp. 110-111. Text

17. Henry T. Wright, Naomi Miller & Richard Redding, "Time and Process in an Uruk Rural Center." In L'Archéologie de L'Iraq du Début de l'Epoque Néolithique ý 333 avant notre ère. Paris 1980, p. 277. Text

18. Jacques Cauvin, Les Premiers Villages de Syrie-Palestine du IXème au VIIème Millénaire Avant J.C. Collection de la Maison de l'Orient méditerranéen ancien, No 4. Série Archéologique Vol. 3. Lyon 1978, pp. 74-75. Text

19. The Priest-king is represented on an oblong bulla from Habuba Kabira, (72 Hb 102). I am grateful to Eva Strommenger for this communication. Pierre Amiet, Glyptique Susienne, Mémoires de la Délégation Archéologique en Iran Vol. 43, 1972, Pl. 18: 695. Text

20. Julius Jordan, Vorläufiger Bericht über die von der Deutschen Forschungsgemeinschaft in Uruk-Warka unternommenen Ausgrabungen. Abhandlungen der Preussischen Akademie der Wissenschaften, Phil.-hist. Klasse, Berlin, Vol.II, 1931, fig. 16-17; M.J. Steve & H. Gasche, L'Acropole de Suse, Mémoires de la Délégation Archéologique en Iran, Mission de Susiane. Vol. 46, Paris 1971, P. 151 & Pl. 89. Pinhas P. Delougaz & Helene J. Kantor, "New Evidence for the Prehistoric and Protoliterate Culture Development of Khuzestan." Vth International Congress of Iranian Art and Archaeology. Vol. 1, Tehran 1972, P. 27; André Finet, "Bilan Provisoire des Fouilles Belges du Tell Kannas," Annual of the American Schools of Oriental Research Vol. 44, 1979, P. 93. Text

21. Eva Strommenger, "The Chronological Division of the Archaic Levels of Uruk-Eanna VI-III/II: Past and Present." American Journal of Archaeology Vol. 84, No 4. 1980, P. 485-486; Louis le Breton, "The Early Periods at Susa, Mesopotamian Relations." Iraq Vol. 19, No 2, 1957, P. 97-113; Pinhas P. Delougaz & Helene J. Kantor, "New Evidence for the Prehistoric and Protoliterate Culture Development of Khuzestan." Vth International Congress of Iranian Art and Archaeology Vol. 1, Tehran 1972, pp. 26-33. Text

Return to the top of the page

Return to Numerals Project Papers