There are two important limitations to such a system. Firstly, every different type of good for which you want to make a record must have its own distinctive sign. We saw how the increasing complexity of economic life led to a great proliferation of styles of tokens. Each of these tokens now had to be rendered by its own sign, and, of course, all the signs had to be learned. The second limitation concerns not the range of goods available, but their quantity. Recording a delivery or disbursement of three jars of oil by writing the oil-jar symbol three times is simple and convenient. Recording a delivery or disbursement of several hundred jars of oil the same way is no longer so convenient and is also a system to prone to error. The large temple complexes that developed in the late fourth millennium, such as the temple of Inanna at Uruk, were considerable economic enterprises, dealing in large quantities of goods and labor. Gradually, a new system had to be developed.

The first great innovation after the act of writing was the separation of the quantity of the good from the symbol for the good. That is, to represent three units of grain by a symbol for 'three' followed by a symbol for 'grain-unit' in the same way that we would write 3 sheep or 3 cows or, more generally, 3 liters or 3 kilometers. A system of this sort is a metrological numeration system, a system of weights and measures. The 'three' symbol is not completely abstract, but is given value by its context, by having the units appended. The development of this concept over the third millennium is a fascinating and extremely complex story that is as yet only partially understood.

Whereas we use the same number signs, regardless of their metrological meaning (the '3' for sheep is the same sign as the '3' for kilometers or jars of oil), the Sumerians used a wide variety of different symbols. Nissen, Damerow and Englund have identified around 60 different number signs, which they group into a dozen or so metrological systems (see Chapter 6 of 'Archaic Bookkeeping'). For example, the Sumerians used one system for counting discrete objects, such as animals, and other systems for measuring areas or volumes. Each system had a collection of signs denoting various quantities.

Any metrological system contains a number of different-sized units with fixed conversion factors between them, so that, for example, there are 12 inches in a foot and three feet in a yard, and so on. Just as in our old weight and measure systems, Sumerian metrology featured all sorts of conversion factors, although it is notable that they were all simple fractions of 60.

In the basic sexagesimal system used for counting most discrete objects,
a single object, a sheep or cow or fish, is denoted by a small cone. Ten
cones equaled one small circle, six small circles equaled one big cone,
ten big cones equaled was a big cone with a circle inside it, six of those
was a large circle and ten large circles was given by a large circle with
a small circle inside. This last unit was then worth 10x6x10x6x10 = 36000
base units. Note that the circle and "cone-shape" could be easily made
by a stylus pressing on the clay, either vertically for the circle or at
an angle for the cone.

For discrete ration goods a 'bisexagesimal' system was used with conversion
factors 10, 6, 2, 10 and 6, so that the symbol for the largest quantity,
this time a large circle containing two small circles, denoted 6x10x2x6x10=7200
base units.

Yet another system was used for measuring grain capacity. Here the conversion factors were 5, 10, 3, and 10, so that the largest unit, a large cone containing a small circle, was worth 10x3x10x5=1500 of the small units.

Adding to the confusion for modern scholars attempting to unravel these complex metrological systems was the fact that a single sign might be used in several systems, where it could mean different multiples of the base unit. In particular, the small circle could mean 6, 10 or 18 small cones, depending on context (as well as other multiples of base units denoted by other symbols). For an example of an archaic ration tablet see here.

Gradually, over the course of the third millennium, these signs were replaced by cuneiform equivalents so that numbers could be written with the same stylus that was being used for the words in the text.

The final step in this story, occurring probably some time in the Ur
III period, right at the end of the third millennium, was the introduction
of a sexagesimal place value system. The number of signs was reduced to
just two: a vertical wedge derived from the small cone often used for the
base unit, and a corner wedge, derived from the small circle. The corner
wedge had a value of ten vertical wedges. In the sexagesimal counting system
described above, the next size unit was the large cone, worth six circles.
In the place value system, this unit was denoted by the same-sized vertical
wedge as the base unit, and it was worth six corner wedges. Now the pair
of symbols could be repeated in an indefinitely larger alternating series
of corner and vertical wedges, always keeping the same conversion factors
of 10 and 6. The price paid was that a vertical wedge could now mean 1,
or 60 (6x10), or 3600 (60x60), and so on. Its actual value was determined
by its *place*.

The sexagesimal place-value system greatly facilitated calculations,
but, of course, at the end of the day, the final answer had to be translated
back into the underlying metrological system of units. So a problem would
be stated in proper units and the solution would be given in proper units,
but the intermediate calculations were carried out in the new sexagesimal
place value system.

Further reading: Far and away the best discussion of this topic is in
Archaic
Bookkeeping.

Go back to Tokens.

Go on to Old Babylonian.

Go up to Mesoptamian
Mathematics Home Page.

Last modified: 3 October 2002 Duncan J. Melville

dmelville@stlawu.edu