Chapter 3

Literacy, numeracy and the

state in early Mesopotamia
Eleanor Robson

INTRODUCTION
The intertwined origins of numeracy, literacy and the state in fourth-millennium Mesopotamia
have been widely recognised for several decades now. Yet their exact relationship remains a
matter of hot dispute. At one extreme, Denise Schmandt-Besserat (1992; 1996) argues that
early literacy was no more than a bureaucratic accounting tool whose sole precursor was an
increasingly complex and uniform system of clay counters, attested all over the Middle East
from the eighth millennium BC. At the opposite pole, Jean-Jacques Glassner (2003) sees no
role for numeracy in the origins of cuneiform, evidencing private and familial uses as well
as institutional ones.1
In this article I provide an interpretation that gives equal weight to numeracy and
literacy, which I hope amounts to more than a simple compromise between the two sides. I
argue that from about 6000 BC Neolithic villagers used simple geometric counters in clay
and stone to record exchange transactions, funded by agricultural surpluses. As societies
and economies grew in size and complexity, increasing strain was placed on trust and
memory, driving both an increasing numerical sophistication and the invention of written
signs for the commodities, agents, and actions involved in managing them in the late fourth
millennium BC. Here I take the story beyond its usual endpoint, further arguing that this
numerate, literate urbanism in turn pushed both literary and mathematical development,
giving rise to the world’s first supra-utilitarian mathematics and literature in the territorial
kingdoms of southern Iraq over 4,000 years ago. I shall also argue that literature and
mathematics were very closely intertwined in the ideology of Middle Bronze Age statehood
and kingship.
The evidence presented here, collected and analysed by a vast range of scholars over
the last 150 years, comes not only from the documents themselves – written on clay
tablets in the long-dead Sumerian and Akkadian languages with a complex wedge-
shaped script now called cuneiform – but also from other objects and relationships
discovered through archaeological excavation in Syria, Iran, and Iraq. Covering the
 38 ELEANOR ROBSON

period approximately 6000–1600 BC, this is necessarily far from an exhaustive survey;
rather I have tried to exemplify each major development with evidence from one or
two sites.

COUNTING WITHOUT WRITING IN THE NEOLITHIC MIDDLE EAST

There have been people living in the Middle East for at least 500,000 years, when
*This page
reference
Neanderthals occupied caves in the Zagros mountains of northern Iraq (Matthews 2000:
must be 13–6*). Anatomically modern humans have inhabited the region for about 100,000 years
wrong, please(Matthews 2000: 17–22). From 15,000–10,000 BC2 the slow process of sedentary settlement
correct
began, with “new aspects of social structure, such as architecture, burial, ritual paraphernalia
and exchange networks […] already well established features of human communities long
before any fundamental alteration in subsistence strategies had occurred” (Matthews
2000: 52). Over the next few millennia, hunting and gathering gradually and patchily
gave way to agriculture and then animal husbandry, but never entirely died out. By 7000
BC three domesticated cereals – barley, emmer wheat, and einkorn – were widespread.
Hunting dogs were the first animals to be domesticated, probably even before the existence
of agriculture. Sheep, goats, and then pigs, none of which are direct competitors for food
eaten by humans, were exploited for meat and secondary products. Cattle were the last of
the major domesticates to appear in the Middle East, some time after 6500 BC (Matthews
2000: 51–4).
As farming techniques became more efficient and cultivars higher yielding, the protection
and equitable distribution of harvest surpluses also became increasingly necessary. It is in
these circumstances that we first find recorded number, in the form of tiny ‘tokens’ made
of clay or stone and shaped into simple geometrical forms – spheres, cylinders, cones, discs,
and tetrahedra. They are found in villages and settlements all over the Middle East, from
western Syria to central Iran. The very earliest evidence, from five early eighth-millennium
BC sites in northern Syria and western Iran, predates both farming and the use of pottery
in the region.3 Thereafter, Schmandt-Besserat (1992: 36–9) documents tokens at a dozen or
more sites per millennium, from southeast Anatolia and the Levant to eastern Iran, right
up until about 1500 BC, “tapering off in Mesopotamia in the third millennium BC while
lingering longer in Syria and Iran” (Schmandt-Besserat 1992: 39). She argues for a regional
system, in which the meanings of individual tokens were fixed across vast distances and
time-spans (Schmandt-Besserat 1992: 164).
Of course, in many contexts small clay objects are very difficult to interpret: we have no
way of knowing why prehistoric villagers might have made or kept them. For instance, Tell
Abada was an early fifth-millennium site in the foothills of eastern Iraq, 80% excavated
during rescue excavations in the 1970s and 80s preparatory to the construction of the
Hamrin Dam (Jasim 1985; Jasim & Oates 1986). The village comprised some 10 large
tripartite houses, each arranged around a central courtyard. Around 90 tokens were
recovered from two phases of Building A, the largest of the settlement and in the centre of
the site. All were found in seven discrete groups of four to 16 tokens in the western half of
the building, mostly in bowls or jars but also in little heaps on the floor. Around half of the
tokens were spherical, including a few with up to five notches in them; another third were
roughly conical, the remainder discs and rods. Each cluster was composed of a mixture of
shapes, ranging in size from a few millimetres to 5cm (Jasim & Oates 1986: 352–7). Not a
single sealing or other remnant of administrative apparatus was found in the building (see
below). As Jasim and Oates conclude,
“Although the context of the small clay tokens at Abada proves for the first time their mutual
relevance, it does not indicate their actual purpose. The unusual numbers of infant burials
found under Building A could suggest some religious significance, with the implication of a
ritual function for the tokens. But an accounting use as genuine ‘calculi’ is equally plausible,
 LITERACY, NUMERACY AND THE STATE IN EARLY MESOPOTAMIA 39

and … perhaps the most persuasive – though the two functions are not necessarily as distinct
as our terminology implies.”
(Jasim & Oates 1986: 355)
Evidence for the function of tokens comes through particular types of archaeological
context. A particularly clear example comes from Tell Sabi Abyad, a late Neolithic village
on the Balikh river in north-central Syria which was destroyed by fire around 6000 BC
(Akkermans & Verhoeven 1995; Akkermans 1996). Five densely clustered rectangular
houses, composed of many small rooms, and four round ones, known anachronistically as
tholoi, were uncovered in Dutch excavations at the so-called Burnt Village in 1991–92. Two
houses and one tholos yielded clay tokens, carefully fashioned into simple geometric shapes.
Buildings II and V were primarily used for storing and processing cereals, as witnessed by
knee-high deposits of grain in some rooms and large numbers of grinding slabs, mortars and
pestles in others. In each case, the tokens were found in just one or two rooms, where they
had been carefully stored along with hundreds of clay stamp-sealings from bags, baskets,
jars, and doors as well as a variety of other small stone, ceramic, and bone artefacts. Tholos
VI was used for both storage cloth manufacture. None of the three constructions seems to
have been inhabited domestically.
These sealings, amongst the earliest attested anywhere in the Middle East, were
designed to stop unauthorised entry into whatever they marked, thus showing that people
were storing and controlling access goods of high value. Although over 60 different designs
of seal impressions are known at Sabi Abyad, suggesting that a large number of individuals
were involved in the process, not a single actual stamp seal has been recovered. The
excavators conclude that either the seals were made of perishable material such as wood or
bone – or that the objects were sealed at some other place and came to Sabi Abyad through
trade or exchange (Akkermans & Verhoeven 1995: 25). We should probably understand the
tokens, so carefully stored with the sealings, as records of the transactions, but exactly which
quantities and commodities they represented it is impossible to say. It is likely, however,
that systematisation was local at best (Friberg 1994: 484). Particular shapes and sizes must
have had particular meanings to the individuals involved in trade and storage at Sabi Abyad
but, although the general signification of tokens was well understood across the region,
the clusters of tokens could themselves be meaningfully deciphered only by someone who
remembered the transaction they invoked.

NEW MANAGEMENT MEDIA IN CHALCOLITHIC TOWNS

The dramatic social and economic developments that took place across the Middle East
during the fourth millennium BC have been collectively dubbed the “Uruk phenomenon” by
the region’s archaeologists, after the enormous city of that name in southern Iraq (Collins
2000; Rothman 2001; Postgate 2002; Algaze 2004). We shall consider Uruk itself in the
following section, but first let us examine the management of increasing social complexity
in the Chalcolithic towns of northern Mesopotamia.
Hacınebi is situated in the foothills of the Taurus mountains in Anatolia, on a key crossing
of the Euphrates river. Here the riverine trade which brought copper and timber to southern
Iraq intersected with the overland caravan route which linked the eastern Mediterranean
with the rest of the Middle East. It was excavated for six seasons in the 1990s by a team led
by Gil Stein, whose aim was to explore the impact of the Uruk expansion on Chalcolithic
culture (Stein (ed.) 1999; Stein 1999; 2001; 2002). In the first four centuries of its occupation
(c.4100–3700 BC), Hacınebi already exhibited many key features of indigenous social
complexity:
“What we see is the regional center for a small-scale, complex polity, perhaps something we
could call a simple chiefdom, with hereditary élites, a complex administrative technology, advanced
metallurgy, socioeconomic differentiation, long-distance exchange of raw materials and/or
prestige goods, and a craft economy that combined household production with small-scale
40 ELEANOR ROBSON

specialisation by independent producers of ceramics and copper ornaments. In general,

the economy seems to have been largely geared towards local consumption, although some
surplus production must have been taking place to support the limited importation of exotic
raw materials for use as prestige goods legitimating the existing social hierarchy.”
(Stein 2001: 279; my italics)
However, that ‘complex administrative technology’ was not, according to the evidence,
a numerate one. Elegantly made, élite stamp seals were in use alongside cruder domestic
ones to seal a wide range of portable objects with local clays, often in purpose-built store
rooms (Pittman 1999; Blackman 1999; Stein 2001), but never in conjunction with tokens as
at Sabi Abyad.
During the period c.3700–3300 BC life and society continued relatively unchanged
across most of Hacınebi. However, the northeast corner of the site displays a radically
different material culture, typical rather of Uruk and other settlements in the south of
Iraq, over 700km away. Diagnostic finds include a wide range of locally made Uruk-style
pottery, architectural decorative elements typical of Urukean public buildings, and personal
ornaments (Stein 2001: 285–9; 2002: 151–2). And while the rest of Hacınebi continued, for
the most part, to seal its surplus goods with local stamp seals, the Urukean colonists rolled
cylindrical stone seals over both local and southern clays and stored tokens in sealed clay
envelopes (Pittman 1999; Blackman 1999).
At Hacınebi, southerners came north to dwell alongside the local population in an
economically and socially autonomous enclave, presumably set up to exploit local copper
resources (Stein 2001: 299). But they also founded new settlements on the upper reaches
of the Euphrates. Habuba Kabira South, for instance, was a walled town of about 10ha on
a strategic bend of the Syrian Euphrates about 150km due south of Hacınebi, excavated
in 1971–76 (Strommenger 1980). There is no doubt that communal living at Habuba was
organised on a completely different scale to the Neolithic village of Sabi Abyad and even to
the Urukean colony at Hacınebi. Inhabited for about 120 years some time between 3400 and
3000 BC, it was centrally planned and managed, with a grid system of paved and drained
streets, and very regular architecture. Rations were distributed, here as at Uruk, the
Hacınebi enclave and elsewhere, by means of mass-produced, standardised bowls. Urukean
administrative technologies – cylinder seals and clay sealings, tokens and clay envelopes
– were also ubiquitous. The 200-odd baked clay tokens display an unprecedented variety of
shapes; over half were incised or otherwise marked before firing (Schmandt-Besserat 1992:
88–91). Two broken clay envelopes containing two and six tokens respectively were found
in ‘the largest and most imposing house of the northern sector of the city’, along with other
administrative debris including flat pieces clay impressed with both seals and token-like
marks (Strommenger 1980: 64 fig. 58; Schmandt-Besserat 1992: 116, 136).
At both Habuba and Hacınebi we thus see the merging of counting and sealing
technologies, with the invention of the sealed clay envelope in which clusters of tokens
could be stored. Impressions of the tokens to be put inside were first made on the outside
of the envelope, and the whole thing sealed – now with cylindrical seals that could cover
the whole surface. But the redundancy in this system rapidly became apparent: why go to
all the trouble of storing tokens when one has made tamper-proof records of them already?
Thus recording quantities of objects took another step towards abstraction, as empty
envelopes were flattened into flat tablets of clay, whose surfaces bore impressions of tokens
and seals.
Even before writing, then, people from the southern alluvium were recording numbers
of stored commodities to protect them from theft or to document transactions. Yet
we probably should not call this activity mathematics, or even arithmetic, as there is
no evidence yet of consistent number bases (Englund 1998: 51). Further, it was just as
dependent on memorising the correlations between individual transactions and individual
tablets as the Neolithic token system of Sabi Abyad. It was still not possible to record
the objects of accounting, or the agents, circumstances, or types of activity. All that had
 LITERACY, NUMERACY AND THE STATE IN EARLY MESOPOTAMIA 41

changed was the move from ephemeral clusters of tokens to permanent impressions of
them in clay. Indisputably, though, it was the precursor of a system that became incredibly
sophisticated and flexible when writing was thrown into the mix, somewhere in southern
Iraq at about the time that Habuba Kabira was abandoned in the very late fourth
millennium BC.

ACCOUNTING AND TRAINING IN THE FIRST CITY STATES

Throughout the fourth millennium increasing surplus wealth, managed through token
technology, had driven a massive population growth in the region of southern Iraq, which
was then a mix of fertile river plain and abundantly verdant marshes. Permanently dry land,
fit for habitation, was hard to come by, so large settlements grew in single spots, gradually
rising above the extraordinarily flat landscape on the accumulated rubble and rubbish of
centuries. These increasingly large cities, with populations in the tens of thousands, were
far too huge and impersonal to be run along the old lines of family, trust, and memory.
Uruk itself had grown to 250ha in size by 3000 BC, with at least 280 further hectares of
inhabited hinterland (Algaze 2004: 141). German excavators, working regularly at Uruk
since 1928, have shown that its central administration, based in the monumental temple
complex of some 30 hectares at the heart of the city, was supported by a strong religious
ideology, expressed through imagery on objects large and small, focussed on the chief
priest and the anthropomorphic deity he worshipped (Nissen 1986). Just as in the northern
colonies, at Uruk excavators found tokens, sealings, clay envelopes and ration bowls all
around the central religious precinct – along with some 5,000 proto-cuneiform tablets
which had been thrown away after their useful life was over. Frustratingly, though:
“Archival information which might have been derived from the excavated Uruk [tablets]
was in great part lost, due both to the record method of the excavators, but also and
fundamentally to the fact that the archaic texts formed – seemingly without exception – part
of the general debris of pottery shards, animal remains, etc, removed from administrative
units of the central [temple] district Eanna and either deposited in trash holes or used as fill
in constructions of walls and floors.”
(Englund 1998: 41)
Nevertheless, on internal grounds it is possible to isolate three stages of the tablets’
development. “Numero-ideographic tablets” (Englund 1998: 51-5) are almost identical to
those found at Hacinebi and Habuba that bear impressions of cylinder seals and tokens.
But they are also marked with one or two signs, incised with a pointed tool, which ‘represent
discrete objects (sheep, jugs of beer and dairy fats, strings of dried fruits, textile products)’
(Englund 1998: 53). These notations, so far attested only in the big cities of the south,
are thus the first explicit record of the material objects of accounting. Many are clearly
pictographic, schematically representing containers and the like; it is thus probable that
the remainder are too, even if we cannot currently identify the objects signified or the
conventions used to represent them.4
4
In the following stage, known conventionally as ‘Uruk IV’, is hitherto known
only in Uruk itself. These tablets, mostly unbaked and unsealed, exhibit a range
of some 900 incised word signs and five different numerical systems for counting
different types of commodity. They may record one or more transactions of a single
type, whether income, expenditure, or internal transfer. The front side of a tablet
may be divided into box-like cases, each of which records a separate item of account,
typically with a total or other derived figure on the reverse (Englund 1998: 61).
Analysis of the individual entries on the front, combined with summations on the
back, has established that the same numerical signs could take different quantitative
values, depending on what was being counted (Nissen et al. 1993: 25–9).
42 ELEANOR ROBSON

But for the first time it was not only the objects and quantities of account that were
recorded, but also the human and institutional agents and the type of transaction involved
(Englund 1998: 215). And unlike at Neolithic Sabi Abyad, where it seems that tokens were
used to account for traded goods, here in Uruk the accountants dealt only in the domestic
economy: fields and their crops, livestock and their products, and the temple personnel
themselves. And once again accounting technology had undergone a concomitantly large
shift in the invention of incised symbols, mostly pictograms, to represent the commodities,
land and labour under central administrative control. Numbers, however, continued to be
represented by impressions of tokens – and, crucially, this is how we can infer so confidently
the function of tokens in prehistoric periods.
The mature stage of literate accounting at Uruk – also attested in smaller numbers in
settlements across southern Iraq and southwest Iran – is conventionally known as ‘Uruk III’
and probably dates to a century or so around 3000 BC. The forms of the incised ideograms
show much greater standardisation than the preceding Uruk IV stage, and have lost much of
their pictorial quality. Document formats can be much more complex too, with ‘secondary’
accounts summarising and consolidating the contents of many ‘primary’ records, often over
several years (Englund 1998: 61–4).
Take, for example, the brewing of beer, an essential staple for many societies with
no reliable access to clean drinking water. In southern Iraq brewers used barley which
had been processed in two different ways: malt, which involves soaking, germinating, and
drying the barley grains; and groats, which are hulled and crushed barley grains. On the
front of a typical ‘primary’ tablet four different types of beer, counted by the jug, might
be allocated to several different officials, with the total for each beer given on the back.
While individual objects were counted in groups of 10 and then 60, barley was accounted
for with a special set of four capacity units related to each other by the ratios 1 : 6 : 60
: 180, with the smallest unit equivalent to about 24 litres and the largest to some 4,320
litres.5
A typical ‘secondary’ beer-account might calculate the malt and groats necessary to
brew each of the four beer types in the ‘primary account’, for the officials listed on that
tablet as well as several more. Malt and groats were notated with the same number
system as barley, but marked with tiny dots and diagonal dashes respectively. First the
quantities of groats and malt needed for each of the four types of beer were calculated
separately, then the total number of beer jars summed; finally, the overall amounts of
groats and malt for all beer types were totalled (Nissen et al. 1993: 43–6). We cannot tell
whether these were theoretical calculations, made in advance of production in order to
requisition the necessary raw ingredients, or an accounting after the fact. Either way,
tablets like these comprised a tiny part of the Uruk bureaucracy that traced production
right from sowing the field with barley, to harvest, to the final production and consumption
of the beer.
As the production of accounts entailed complex multi-base calculations, not surprisingly
the accountants sometimes made errors. In the example just discussed, three small
units of grain were accidentally omitted from the total on the ‘primary’ tablet, while on
the ‘secondary’ tablet the malt and groats were totalled to different amounts when they
should have been equal. Trainee scribes thus had to practice both writing and calculating,
and they did so increasingly systematically: whereas less than 1% of Uruk IV tablets have
been identified as scribal exercises some 20% of the much larger, mature Uruk III corpus is
thought to be the product of bureaucratic training (Englund 2004b: 28). The exercises take
two forms: on the one hand, highly standardised lists of the objects of account; and on the
other, apparently ad hoc arithmetical problems.
For instance, a group of about eight small tablets from a single find-spot in Uruk
give the lengths and widths (in conspicuously round numbers) of irregular quadilateral
shapes, with no indication of scribe, institutional authority or administrative context.
Their areas, when calculated, turn out to be “a multiple of the large area unit 5 ešè
 LITERACY, NUMERACY AND THE STATE IN EARLY MESOPOTAMIA 43

[c.11ha] … times two extra factors of the type (1 + 1/n)” (Friberg 1997–98: 11–12). In
other words, they appear to be carefully prepared mathematical exercises rather than
genuine surveyors’ records. The calculations themselves were presumably not written
down but performed by the manipulation of counters, as in many other ancient societies
(Powell 1995). Scribal students learned how to write numbers in context by means of
‘Word list C’, an elementary exercise in enumerating the basic commodities of the Uruk
administration known from nearly 60 fragmentary manuscripts (Veldhuis 2005). For
instance, lines 18–25 read:
10 (vessels of?) milk
1 (unit of) cream
10 cows
1 bull
10 ewes
1 ram
10 nanny-goats
1 billy-goat
Trainee scribes also copied and memorised thematic lists of nouns related to the
accounting practices of Uruk: commodities such as containers, textiles, metal and wooden
objects; fish, birds, and animals; professional and administrative titles; names of cities
and other geographical terms. The list of professions now known as Lu A, for instance, is
attested in over 160 exemplars while others, such as a list of bovines, are apparently unique
manuscripts that were rarely used for training. While the thematic range of these word lists
is co-extensive with the bureaucrats’ remit – there are no lists of gods, for instance, or
wild animals – their lexical range is much wider: they record many words that were never
in practice used by the administrators, so far as we know. The point was, though, that they
could have been: the new writing system had to accommodate all the possible uses that the
accountants of early cities might want to put it to (Veldhuis 2005).
In short, by the end of the fourth millennium BC, southern Mesopotamian city states
had implemented an extensible and powerful literate technology for the quantitative
control and management of their assets and labour force. In doing so, they had created
in parallel a new social class – the scribe – who was neither economically productive nor
politically powerful, but whose role was to manage the primary producers on the élite’s
behalf.

ENUMERATING THE LITERARY IN THIRD-MILLENNIUM POLITIES

For much of the third millennium BC the political infrastructure of southern Iraq remained
fairly stable. Individual city states and their immediate hinterlands tended to retain strong
local identities while sharing a common culture, religion, and management structure.
They regularly formed economic and military alliances with their neighbours for the sake
of prosperity and security, while disputes over water and land resources were settled by
warfare or arbitration by a third polity.
During this period, writing gradually accrued new powers as it was adopted for a variety
of institutional purposes, from recording legal contracts to issuing administrative orders
and documenting political events (Cooper 2004). Paranomasia, or punning, enabled scribes
to record the sounds of all words of the language, not only those that could be represented
pictorially. At the same time, visual changes were made: cases became lines of writing;
the script became ever more abstract and ‘cuneiform’ or wedge-shaped. Numerals that
imitated the form of accounting tokens were gradually replaced by cuneiform notation. Yet
literacy and numeracy remained overwhelmingly in the hands of institutional bureaucrats
managing land, labour, and livestock.
In the southern Iraqi city of Shuruppag in around 2600 BC, for instance, over 100 named
temple and palace employees were allocated land and seed grain (at about 35 litres per ha),
44 ELEANOR ROBSON

each precisely recorded and totalled, with which they were expected too feed themselves
and their dependents. Identical quantities of seed grain are recorded as leaving the central
depot - but written by a different scribe, who preferred new-style number notation where
the first preferred the old-style, impressed numeration. (Nissen et al. 1993: 58–9). As this
small example shows, while many individuals and families doubtless earned their livings
through personal enterprise based on trust and memory, never coming into contact with the
written record, institutional economies continued to run on centrally managed quantitative
models, planned and carried out by literate and numerate professionals.
Similarly, scribal training continued to rely on much the same types of exercise as before,
though with the addition of new genres. Trainee bureaucrats still copied and memorised
highly standardised lists of words – often the same exercises that had been created by the
Uruk accountants in the late fourth millennium, even when (as in the case of the professions
list Lu A) many of the words they contained were no longer in use (Michalowski 2003).
New lists of nouns were also added, such as lists of gods, but these too were all necessary
knowledge for palace and temple administrators. ‘Word list C’ continued to be copied, and
arithmetical exercises set whose primary goal was still the correct manipulation of lengths
and areas.
Now, though, we begin to see arithmetical formats that do not simply imitate account
documents. From the cities of Shuruppag and nearby Adab there are both lists and tables of
lengths and the square areas that result from multiplying them together, that systematically
work through the length units in ascending order (Nissen et al. 1993: 135–40). From
Shuruppag come two different answers to a problem about rationing a warehouse full of
grain between a large group of workers at 7 litres each. One student has calculated a round-
number answer with a small amount of grain left over; the other has attempted an exact
solution (Melville 2002). In real administrative circumstances no worker ever received 7
litres of grain, but always some multiple of 5 or ten litres, but it is noteworthy that 7 is
the smallest integer that does not neatly divide any of the number bases used in grain
metrology. In short, the arithmetical lists, tables, and exercises together show an increasing
pedagogical interest in the properties of numbers for their own sake, beyond the immediate
needs of administrative accounting but always stemming from that context.
Out of these innovations and traditions in scribal culture also came new, non-administrative
functions of writing that we might now label as ‘literary’. For instance, amongst the earliest
literary works are the so-called Zà-mì (‘praise’) hymns from 25th-century Abu Salabikh (a
city in southern Iraq whose ancient name is unknown). They comprise a 300-line sequence
of 70 short litanies praising a city and the deity associated with it, attested in several dozen
manuscripts:
Kesh, the birthing brick,
home of the small speckled tu bird:
praise mother Nintud!
…
Girsu,
home of the princely dagger:
praise Ningirsu!
…
Isin, pole of heaven and earth:
praise Ninisina!
(Zà-mì hymns, lines 75–7, 117–19, 186–7; after Biggs 1974: 48–51)

On the face of it, the composition is little more than a list of geographical and divine
names, elaborated with epithets. How should we understand it: as list or literature? In fact
the boundaries between early Mesopotamian lists and literature are very difficult to draw.
As Rubio (2003: 205) has noted, “the same text can be regarded as a list of field names,
canals and cult places by one scholar, whereas another would interpret it as a literary
 LITERACY, NUMERACY AND THE STATE IN EARLY MESOPOTAMIA 45

text.”6 There are two basic problems at the heart of the ambiguity. First is the matter of
comprehensibility: scribes did not yet write all the grammatical elements of the Sumerian
language, making it difficult for modern readers to determine authorial intention. Second
is the matter of literary style:
“A frequent rhetorical device in Sumerian literature is what could be called ‘enumeration’.
A text may consist mainly of a listing of the terms of a lexical set. Each term is encased in
a fixed repeated formula and provided with a comment. The enumeration is inserted in a
narrative or laudatory frame.”
(Civil 1987: 37)
As Gonzalo Rubio argues, we should view Sumerian lexical lists and literary works not
as antithetical but as two complementary and mutually productive artefacts of scribal
culture:
“Nothing would seem more strictly denotative than listing. However, the fact that lists
are cognitively comprehensive but thematically selective is precisely what turns them
into connotative devices when used as templates for literary enumerations. … The link
between both kinds of textuality, lists and literary texts, lies in a common rhetorical device:
enumeration; and the compositions of some Sumerian literary works gravitates around this
very device.”
(Rubio 2003: 201)
Further, enumeration by its very definition entails counting as well as listing; it was
at the very heart of cuneiform literacy and numeracy. Thus out of recorded quantitative
bureaucracy emerged, via the humble pedagogical list, the first explorations of the aesthetic
qualities of numbers and words – namely mathematics and literature – in the middle of the
third millennium BC.

MEASUREMENT, MATHEMATICS, AND POWER IN EARLY EMPIRES

In the last centuries of the third millennium city alliances began to coalesce into more
formal territorial empires, ruled not by nameless priests and bureaucrats in the name of
a god, but by charismatic kings who united their people through personal prowess and
force. Rulers such as Sargon of Agade (2334–2279 BC), Shulgi of Ur (2094–2047 BC), and
Hammurabi of Babylon (1792–1750 BC) united and held together huge territorial empires
the size of modern nation states through wily political manoeuvring, skilful military strategy,
and fortuitously long reigns. They each built on the core kingdoms conquered and held by
their immediate predecessors, but none was able to build stability into the system. In every
case, within a generation or two their successors, with neither longevity nor sheer force of
personality on their side, had overseen the collapse of those states back to the size of the
original core kingdoms or to total eradication.
However dominant these new-style kings, they nevertheless had to accommodate
themselves to the mighty temples, which were still the economic powerhouses of the land.
They did this through taking some of the attributes of priesthood, prioritising piety as
highly as military might. To promote and concretise this personality-orientated kingship
a new form of praise literature was created and developed to extol the ruler’s all-round
prowess. Hymns to deities and to temples, such as the Zà-mì litanies, had been recorded
since the mid-third millennium BC; hymns to kings were an entirely novel phenomenon. In
the opening lines of a praise poem of king Shulgi, for instance, enumerations of his heroism
are followed immediately by divine patronage:
I, the king, was a hero already in the womb.
I, Shulgi, was born to be a mighty man.
I am a fierce-looking lion, begotten by a dragon.
I am the king of the four regions.
I am the herdsman and shepherd of the black-headed people.
I am a respected one, the god of all the lands.
46 ELEANOR ROBSON

I am a child born of Ninsumun.

I am the choice of holy An’s heart.
I am the man whose fate was decided by Enlil.
I am Shulgi, the beloved of Ninlil.
I am he who is cherished by Nintur.
I am he who was endowed with wisdom by Enki.
I am the powerful king of Nanna.
I am the growling lion of Utu.
I am Shulgi, who has been chosen by Inana for his attractiveness.
(Black et al. 1998–2005: no. 2.4.2.01, lines 1–15)7
This particular praise poem, commissioned in commemoration of the seventh year of
Shulgi’s reign (Hallo 1970), was not only current in the royal court but also found its way
into the scribal schools, where it retained a central place in the curriculum until at least
the reign of Hammurabi’s successor Samsu-iluna (r.1749–1712 BC) (Tinney 1999). The
pedagogical role of this hymn, and others like it, was not simply to exemplify classically
beautiful Sumerian literature but to also to instil the ideals of Middle Bronze Age kingship
into the future accountants, bureaucrats and clerks whose vocation it would be to uphold
the institutions of state.
Those palace and temple managers now had to control taxation and tribute over vast
territories as well as the domestic labour and production they had been overseeing for
centuries. They responded by implementing new, more efficient management techniques.
The most significant change came in the 23rd to 21st centuries BC, parallel to the
development of courtly praise literature: the development of a single number system which
could be used to write any quantity of any commodity at all (Powell 1976; Foster & Robson
2004). We call now this the sexagesimal place value system (SPVS): sexagesimal because
the scribes chose base 60 out of all the number bases available to them – presumably
both because it was the system used to count discrete objects such as sheep and beer jugs
and as it was the lowest common multiple of the vast range of metrological units then
in use – and place value because it represented size not by the physical appearance of the
signs themselves but by the order in which they were written, as in the modern decimal
system.
Where once the scribes calculated by manipulating counters, now they used the SPVS for
multiplication and division. But they used it only for those two purposes, continuing to add
and subtract in the old way and to record numbers, weights, and measures in the old fashioned
multi-base systems. Thus sexagesimal numeration is visible in the cuneiform record only
where the accountants forgot to erase their scribbled calculations from their tablets (Powell
1976), or in the arithmetical and mathematical exercises written by trainee scribes. The
SPVS potentially set numbers free from the objects they were counting or measuring,
though in fact they were rarely liberated for long. For the scribes of four millennia ago,
numbers remained indisputably properties of real-world or imagined objects, not platonic
entities in their own right.
In the early second millennium BC, elementary scribal training underwent a revolution
(Veldhuis 1997; 2004; Robson 2001). The old lists of nouns were radically revised, and new
lists introduced that trained more abstract properties of the cuneiform writing system
and the SPVS. Thousands of arithmetical and metrological tables and several hundred
mathematical word problems attest to this new regime, in which emphasis was put more on
the ability to manipulate imaginary lines and areas in almost algebraic ways, than it did on
the ability to count cattle or calculate work rates (Høyrup 2002; Robson 2002).
A typical school problem, probably from the city of Larsa, just north of Uruk, in about
1800–1750 BC, asks for two integers whose product is 60 given that their difference is seven.
The answers – five and 12 – are demonstrated by treating the product 60 as a rectangular
area whose sides are the two unknown quantities. The rectangle is manipulated very
 LITERACY, NUMERACY AND THE STATE IN EARLY MESOPOTAMIA 47

concretely using a method we now know as completing the square (though it is perhaps
more familiar in symbolic algebraic guise).
“A reciprocal exceeds its reciprocal by 7. What are the reciprocal and its reciprocal? You:
break in two the 7 by which the reciprocal exceeds its reciprocal so that 3;30 (will come up).
Combine 3;30 and 3;30 so that 12;15 (will come up). Add 1 00, the area, to the 12;15 which
came up for you so that 1 12;15 (will come up). What squares 1 12;15? 8;30. Draw 8;30 and
8;30, its counterpart, and then take away 3;30, the holding-square, from one; add to one.
One is 12, the other is 5. The reciprocal is 12, its reciprocal is 5.”
(YBC 6967; after Neugebauer & Sachs 1945: text Ua)
Even if we allow that the invention of the sexagesimal place value system was a necessary
condition for the creation of mathematics as an intellectual game, divorced from the
mundane necessities of central administration, it alone is not sufficient to explain that
extraordinary development. Nor does it explain the particular form that mathematics
took, so clearly focused on the correct calculation of lines and areas. Jens Høyrup (1990)
has posited an oral “surveyors’ culture” in the third millennium BC in which professional
land administrators entertained and tested each other with mathematical riddles based
on their professional interests. But that hypothesis does not adequately account for the
turn to literate mathematics in the early second millennium; little remains of any third-
millennium style practical pretext in the example just discussed.
Rather, it is necessary to take a step back and look again at the ideology of the state
under which the trainee scribes were preparing to serve. Piety and militarism have already
been mentioned as key components of kingship; the third strand was justice. Royal hymns
and monumental imagery of the early second millennium BC give equal weight to all three.
The Sumerian word for justice was níğ-si-sá, literally ‘straightness, equality’. The royal
regalia of justice were the measuring rod and line, often presented to the king by a god on
royal monuments or, more usually in Sumerian literary works, a goddess:
“Nisaba, the woman radiant with joy, the true woman, the scribe, the lady who knows
everything, guides your fingers on the clay: she makes them put beautiful wedges on the
tablets and adorns them with a golden stylus. Nisaba generously bestowed upon you [king
Lipit-Eshtar, r.1934–1924 BC] the measuring rod, the surveyor’s gleaming line, the yardstick,
and the tablets which confer wisdom.”
(From A praise poem of Lipit-Eshtar: Black et al. 1998–2005: no. 2.5.5.2, lines 18–24)
Even the great goddess Inana, most especially associated with sex and war, gave up her
measuring equipment last of all when forced to strip at the gates of the underworld:
“When she entered the sixth gate, the lapis lazuli measuring rod and measuring line were
removed from her hand.
‘What is this?’
‘Be satisfied, Inana, a divine power of the underworld has been fulfilled. Inana, you must not
open your mouth against the rites of the underworld.’”
(From Inana’s descent to the Underworld, c.1750 BC:
Black et al. 1998–2005: no. 1.4.1, lines 154–8)
The trainee scribes, once they had finished their elementary training, memorised long
passages of beautiful Sumerian literature such as these (Tinney 1999), carefully designed
to inculcate them with the ideals of statehood and kingship, including numerate justice
in the form of equitable division of land.8 Other compositions, in the form of humorous
dialogues between trainee scribes, addressed matters of competence and incompetence
more directly:
“(Girini-isag speaks): ‘You wrote a tablet, but you cannot grasp its meaning. […] Go to
apportion a field, and you cannot even hold the tape and rod properly; you are unable to
place the field pegs; you cannot figure out its shape, so that when wronged men have a
quarrel you are not able to bring peace but you allow brother to attack brother. Among the
scribes you (alone) are unfit for the clay. What are you fit for? Can anybody tell us?’
48 ELEANOR ROBSON

(Enki-manshum replies): ‘Why should I be good for nothing? When I go to divide a plot, I can
divide it; when I go to apportion a field, I can apportion the pieces, so that when wronged
me have a quarrel I soothe their hearts and […]. Brother will be at peace with brother, their
hearts [soothed].’”
(From School dialogue 3, c.1750 BC: Vanstiphout 1997: 589)
In the light of these passages and images it becomes clear that area-based mathematics
was, at one level, simply royal ideology in another form, promoted through practice and
through literary representation. Thus, while modern scholars have chosen to portray Old
Babylonian literature and mathematics as amongst the world’s first truly creative and
non-utilitarian writings – and indeed there is much to admire in their beauty – for their
producers and consumers they represented above all idealised abstractions of the ordered
urban state, with god, king, and scribe at its centre.

CONCLUSIONS
There are many accounts of the importance of writing in development and management
of early states worldwide. But few pay more than passing lip-service to the central role of
numeracy, and even fewer consider the effects of urban bureaucracy on the development
of supposedly ideology-free intellectual pursuits such as literature and mathematics. For
early Mesopotamia, I contend, it is inadequate to consider one without the others: to
simplify greatly, numeracy drove statehood drove literacy; training in professional literacy
and numeracy for the service of the state drove mathematics and literature which in turn
supported the ideals of the state. Whether this complex nexus of relationships is peculiar
to early Mesopotamia, or whether it holds to some degree for other early state societies, I
leave for others to determine.

NOTES
1 Both stances have been thoroughly critiqued. The most important reviews of Schmandt-Besserat’s work
include Friberg (1994), Michalowski (1993), Oates (1993), Powell (1994), and Zimansky (1993). Glassner’s
book has been reviewed by Dalley (2005), George (2005), Postgate (2005), and Robson (2006) amongst
others.
2 The siglum ‘bc’ indicates uncalibrated radiocarbon dates.
3 Tokens were presumably in use at this time in northern Iraq too; the lack of evidence may be a function of
the restricted opportunities for excavation there over the last few decades since the significance of tokens
was first recognised. See Quenet (2005) for a concise survey of the origins of writing in northern Iraq. For
an excellent overview of counters and proto-literate accounting in Elam (southwest Iran) see Englund
(2004a).
4 There is no room, on this view, for Schmandt-Besserat’s “complex” or “decorated” tokens (which, with
rare exceptions, are not archaeologically attested in administrative contexts) as the precursors of proto-
cuneiform word signs (see Englund 1998: 55).
5 If this seems complicated it is useful to remember that there are 20 fluid ounces in an Imperial pint and
eight pints in a gallon; almost all pre-metric weights and measures use a bewildering variety of number
bases.
6 Thus some scholars have argued that ‘Word list C’ is not an elementary scribal exercise but an esoteric
literary work (e.g. Westenholz 1998).
7 Ninsumun: the divine mother of the legendary hero Gilgamesh, from whom Shulgi’s dynasty claimed
descent. An: the supreme deity, very distant from human affairs. Enlil: The father of the gods, and spouse
of the goddess Ninlil; particularly involved in the doings of mankind. Nintur: an important birth goddess.
Enki: the god of wisdom and cunning, who was thought to reside in the Abzu, the fresh-water sea under
the world. Nanna: the moon god, worshipped especially in Shulgi’s home city of Ur. Father of the sun god
Utu and the goddess Inana. Inana: The most important goddess of Sumer, worshipped for her sexual
relations with her lover Dumuzid and others, and for her warlike aspect.
8 Enumeration inspired by lexical lists remained a central feature of curricular Sumerian literature in the
early second millennium BC: see Veldhuis (2004) for a comprehensive and convincing discussion.
 LITERACY, NUMERACY AND THE STATE IN EARLY MESOPOTAMIA 49

BIBLIOGRAPHY
Akkermans, P.M.M.G. (ed.) 1996. Tell Sabi Abyad: George, A.R. 2005. Review of Glassner 2003. Bulletin
the Late Neolithic settlement. Nederlands Historisch- of the School of Oriental and African Studies, 68: 107–9.
Archaeologisch Instituut te Istanbul, Leiden & Glassner, J.J. 2003. The invention of cuneiform: writing
Istanbul. in Sumer. [Trans. Z. Bahrani & M. Van De Mieroop.]
Akkermans, P.M.M.G & Verhoeven, M. 1995. An The Johns Hopkins University Press, Baltimore.
image of complexity: the Burnt Village at late Hallo, W.W. 1970. The cultic setting of Sumerian
Neolithic Sabi Abyad, Syria. American Journal of poetry. In Finet, A. (ed.), Actes de la XVIIe Rencontre
Archaeology, 99: 5–32. Assyriologique Internationale: 117–34. Comité belge de
Algaze, G. 2004 [2nd edn]. The Uruk world system: the recherches en Mésopotamie, Ham-sur-Heure.
dynamics of expansion of early Mesopotamian civilization. Høyrup, J. 1990. Sub-scientific mathematics:
University of Chicago Press, Chicago. observations on a pre-modern phenomenon. History
Biggs, R.D. 1974. Inscriptions from Tell Abu Salakabikh. of Science, 28: 63–86.
University of Chicago Press, Chicago. Høyrup, J. 2002. Lengths, widths, surfaces: a portrait of
Black, J.A., Cunningham, G.G., Ebeling, J., Flückiger- Old Babylonian algebra and its kin. Springer, New York
Hawker, E., Robson, E., Taylor, J.J. & Zólyomi, G.G. & Berlin.
1998–2005. The electronic text corpus of Sumerian Jasim, S.A. 1985. The Ubaid period in Iraq.
literature. The Oriental Institute, Oxford (http:// Archaeopress, Oxford.
etcsl.orinst.ox.ac.uk). Jasim, S.A. & Oates, J. 1986. Early tokens and tablets
Blackman, M.J. 1999. Chemical characterization of in Mesopotamia: new information from Tell Abada
local Anatolian and Uruk style sealing clays from and Tell Brak. World Archaeology, 17: 348–62.
Hacınebi. Paléorient, 25: 51–6.
Matthews, R. 2000. The early prehistory of Mesopotamia:
Civil, M. 1987. Feeding Dumuzi’s sheep: the lexicon 500,000 to 4,500 BC. Brepols, Turhnout.
as a source of literary inspiration. In Rochberg-
Melville, D. 2002. Ration computations at Fara:
Halton, F. (ed.), Language, literature, and history:
Multiplication or repeated addition? In Steele, J.M.
philological and historical studies presented to Erica Reiner:
& Imhausen, A. (eds), Under one sky: mathematics and
37–55. American Oriental Society, New Haven.
astronomy in the ancient Near East: 237–52. Ugarit-
Collins, P. 2000. The Uruk phenomenon: the role of social Verlag, Münster.
ideology in the expansion of the Uruk culture during the
Michalowski, P. 1993. Tokenism. American
fourth millennium BC. Archaeopress, Oxford.
Anthropologist, 95: 996–9.
Cooper, J.S. 2004. Babylonian beginnings: the
Michalowski, P. 2003. An Early Dynastic tablet
origin of cuneiform writing system in comparative
of ED Lu A from Tell Brak (Nagar). Cuneiform
perspective. In Houston, S. (ed.), The first writing:
Digital Library Journal (http://cdli.ucla.edu/pubs/
script invention as history and process: 71–99. Cambridge
CDLJ/2003/CDLJ2003_003.html).
University Press, Cambridge.
Neugebauer, O. & Sachs, A.J. 1945. Mathematical
Dalley, S.M. 2005. Review of Glassner 2003. Technology
cuneiform texts. American Oriental Society, New
and Culture, 46: 408–9.
Haven.
Englund, R.K. 1998. Texts from the Late Uruk
Nissen, H. 1986. The archaic texts from Uruk. World
period. In Bauer, J., Englund, R. & Krebernik,
Archaeology, 17: 317–34.
M., Mesopotamien: Späturuk-Zeit und Frühdynastische
Zeit: 15–233. Universitätsverlag, Vandenhoeck & Nissen, H., Damerow, P. & Englund, R.K. 1993. Archaic
Ruprecht, Freiburg & Göttingen. bookkeeping: writing and techniques of administration in
Englund, R.K. 2004a. The state of decipherment the ancient Near East. University of Chicago Press,
of proto-Elamite. In Houston, S. (ed.), The first Chicago.
writing: script invention as history and process: 100–149. Oates, J. 1993. Review of Schmandt-Besserat 1992.
Cambridge University Press, Cambridge. Cambridge Archaeological Journal, 3: 149–53.
Englund, R.K. 2004b. Proto-cuneiform account-books Pittman, H. 1999. Administrative evidence from
and journals. In Hudson, M. & Wunsch, C. (eds), Hacınebi: an essay on the local and the Colonial.
Creating economic order: record-keeping, standardization, Paléorient, 25: 43–50.
and the development of accounting in the ancient Near East: Postgate, J.N. (ed.) 2002. Artefacts of complexity:
23–46. CDL Press, Bethesda, MD. tracking the Uruk in the Near East. British School of
Foster, B.R. & Robson, E. 2004. A new look at Archaeology in Iraq, London.
the Sargonic mathematical corpus. Zeitschrift für Postgate, J.N. 2005. New angles on early writing.
Assyriologie, 94: 1–15. Cambridge Archaeological Journal, 15: 275–80.
Friberg, J. 1994. Preliterate counting and accounting Powell, M.A. 1976. The antecedents of Old
in the Middle East: a constructively critical review Babylonian place notation and the early history of
of Schmandt-Besserat’s Before Writing. Orientalistische Babylonian mathematics. Historia Mathematica, 3:
Literaturzeitung, 89: 477–502. 417–39.
Friberg, J. 1997–98. Round and almost round Powell, M.A. 1994. Review of Schmandt-Besserat
numbers in proto-literate metro-mathematical field 1992. Journal of the American Oriental Society, 114:
texts. Archiv für Orientforschung, 44–45: 1–58. 96–7.
50 ELEANOR ROBSON

Powell, M.A. 1995. Metrology and mathematics colonial contact. In Rothman, M.S. (ed.), Uruk
in ancient Mesopotamia. In Sasson, J.M. (ed.), Mesopotamia and its neighbors: cross cultural interactions in
Civilizations of the ancient Near East: 1941–58. the era of state formation: 265–305. School of American
Scribners, New York. Research Press, Santa Fe.
Quenet, P. 2005. The diffusion of the cuneiform Stein, G.J. 2002. The Uruk expansion in Anatolia:
writing system in northern Mesopotamia: the
a Mesopotamian colony and its indigenous host
earliest archaeological evidence. Iraq, 67/2: 31–40.
community at Hacınebi, Turkey. In Postgate, J.N.
Robson, E. 2001. The tablet house: a scribal school
(ed.), Artefacts of complexity: tracking the Uruk in the Near
in Old Babylonian Nippur. Revue d’Assyriologie, 95:
39–67. East: 149–71. British School of Archaeology in Iraq,
Robson, E. 2002. More than metrology: mathematics London.
education in an Old Babylonian scribal Strommenger, E. 1980. Habuba Kabira: eine Stadt
school. In Steele, J.M. & Imhausen, A. (eds), Under one vor 5000 Jahren. Ausgrabungen der Deutschen Orient-
sky: mathematics and astronomy in the ancient Near East: Gesellschaft am Euphrat in Habuba Kabira, Syrien.
325–65. Ugarit-Verlag, Münster. Philipp von Zabern, Mainz am Rhein.
Robson, E. 2006. Review of Glassner 2003. American Tinney, S.J. 1999. On the curricular setting of
Journal of Archaeology, 110: 171–2. Sumerian literature. Iraq, 61: 159–72.
Rothman, M.S. (ed.). 2001. Uruk Mesopotamia and Vanstiphout, H.L.J. 1997. Sumerian canonical
its neighbors: cross-cultural interactions in the era of state compositions. C. Individual focus. 6. School
formation. School of American Research Press, Santa
dialogues. In Hallo, W.W. (ed.), The Context of scripture,
Fe & James Currey, Oxford.
I: Canonical compositions from the Biblical world: 588–93.
Rubio, G. 2003. Early Sumerian literature:
Brill, London, New York & Cologne.
enumerating the whole. In González Blanco, A.,
Vita, J.P. & Zamora, J.Á. (eds), De la tablilla a la Veldhuis, N. 1997. Elementary education at Nippur:
inteligencia artificial: Homenaje al Prof. J.L. Cunchillos the list of trees and wooden objects. Unpublished Ph.D.
en su 65 aniversario: 197–208. Instituto de Estudios thesis, Rijksuniversiteit Groningen, Groningen.
Islámicos y del Oriente Próximo, Zaragoza. Veldhuis, N. 2004. Religion, literature, and scholarship:
Schmandt-Besserat, D. 1992. Before writing: from the Sumerian composition ‘Nanše and the Birds’ . Brill
counting to cuneiform. 2 vols. University of Texas Press, Styx, Leiden.
Austin. Veldhuis, N. 2005. The archaic lexical corpus. Digital
Schmandt-Besserat, D. 1996. How writing came about. corpus of cuneiform lexical texts. University of California
University of Texas Press, Austin.
Berkeley, Berkeley. http://cuneiform.ucla.edu/dcclt.
Stein, G.J. (ed.) 1999. The Uruk expansion: northern
Westenholz, J.G. 1998. Thoughts on esoteric
perspectives from Hacınebi, Hassek Höyük and
Gawra. Paléorient, 25: 7–171. knowledge and secret lore. In Prosecký, J. (ed.),
Stein, G.J. 1999. Rethinking world systems: diasporas, Intellectual life of the ancient Near East: 451–62.
colonies, and interaction in Uruk Mesopotamia. University Academy of Sciences of the Czech Republic.
of Arizona Press, Tucson. Oriental Institute, Prague.
Stein, G.J. 2001. Indigenous social complexity at Zimansky, P. 1993. Review of Schmandt-Besserat
Hacınebi (Turkey) and the organization of Uruk 1992. Journal of Field Archaeology, 20: 513–17.