Browse this book with Google Preview »
Man's development and the growth of civilizations have depended, in the main, on progress in a few activities—the discovery of five, domestication of animals, the division of labor; but, above all, in the evolution of means to receive, to communicate, and to record his knowledge, and especially in the development of phonetic writing. —Colin Cherry
Speech, the universal way by which humans communicate and transmit experience, fades instantly: before a word is fully pronounced it has already vanished forever. Writing, the first technology to make the spoken word permanent, changed the human condition.
It was a revolution in communication when a script allowed individuals to share information without meeting face to face. Writing also made it possible to store information, creating a pool of knowledge well beyond the ability of any single human to master yet, at the same time, available to all. Writing is regarded as the threshold of history, because it ended the reliance upon oral tradition, with all the inaccuracies this entailed. Business and administration are now inconceivable without bookkeeping to balance income and expenditures. Finally, among the innumerable benefits created by a script, writing allows us to capture our ideas when they arise and, in time, to sort and scrutinize, revise, add, subtract, and rectify them to arrive at a rigor of logic and a depth of thought that would otherwise be impossible.
How did writing come about? It is now generally agreed that writing was invented in Mesopotamia, present-day Iraq, in the late fourth millennium B.C. and spread from there to Egypt, Elam, and the Indus Valley. It is also generally agreed that other scripts developed later, independently, in China and Mesoamerica. The origin of Chinese and Mesoamerican writing is still enigmatic. In this book, I will present the archaeological evidence that the Mesopotamian script derived from an archaic counting device. This immediate precursor of the cuneiform script was a system of tokens—small clay counters of many shapes which served for counting and accounting for goods in the prehistoric cultures of the Near East. The idea that Mesopotamian writing emerged from a counting device is new. Until the eighteenth century, the origin of writing was the subject of myths crediting gods, fabulous creatures, or heroes for its invention. Then, in the Age of Enlightenment, the theory that scripts started with picture writing was put forward. This view endured until the present. In the following pages, I will show how the conception of the origin of writing evolved through time.
The oldest account of the invention of writing is perhaps that of the Sumerian epic Enmerkar and the Lord of Aratta. The poem relates how Enmerkar, the lord of Uruk-Kulaba, sent an emissary to the lord of Aratta soliciting timber, gold, silver, lapis lazuli, and precious stones to rebuild the residence of the goddess Inanna. Back and forth the messenger delivered word for word the pleas, threats, and challenges between the two lords, until the day Enmerkar's instructions were too difficult for the emissary to memorize. The lord of Kulaba promptly invented writing, tracing his message on a clay tablet:
The emissary, his mouth (being) heavy, was not able to repeat (it).
Because the emissary, his mouth (being) heavy, was not able to repeat (it),
The lord of Kulaba patted clay and wrote the message like (on a present-day) tablet—
Formerly, the writing of messages on clay was not established—
Now, with Utu's bringing forth the day, verily this was so,
The lord of Kulaba inscribed the message like (on a present-day) tablet, this, verily, was so.
One might add here that, according to the Sumerian king list, Enmerkar lived about 2700 B.C., when writing had been a common practice for five hundred years. Of course this casts doubts on the actual contribution of Enmerkar to the invention of writing!
In a second Sumerian poem, Inanna and Enki, the Transfer of the Arts of Civilization from Eridu to Erech, writing is conceived as one of a hundred basic elements of civilization held by Enki, the lord of wisdom. Inanna coveted the divine decrees for her city, Uruk, and set her mind to getting them. This was done when Enki, drunk, donated to her each and every one of the crafts. In Samuel Noah Kramer's words:
After their hearts had become happy with drinks, Enki exclaims: ...
"...O name of my power, O name of my power,
To the bright Inanna, my daughter, I shall present ...
The arts of woodworking, metalworking, writing, toolmaking, leatherworking, ...building, basketweaving."
Pure Inanna took them.
Inanna loaded writing and the other divine decrees onto the Boat of Heaven and started an eventful journey back to Uruk. After overcoming tempests and sea monsters, sent by Enki to recapture his possessions, she finally reached the city, where she unloaded her precious booty to the delight of her people.
According to Berossus' Babyloniaca, Oannes, a sea creature with the body of a fish and the head, feet, and voice of a man, gave to the Babylonians the knowledge of writing, language, science, and crafts of all types. In other Babylonian texts, the god Ea, the lord of wisdom, was the source of all secret magical knowledge, writing in particular." In Assyria, Nabu, son of Marduk, was revered as the instructor of mankind in all arts and crafts, including building, agriculture, and writing.
In the Bible, God revealed his will to mankind with the Tables of the Law "written by the finger of God." The source of great debates, these words were interpreted by Daniel Defoe as meaning that "the two Tables, written by the Finger of God in Mount Sinai, was the first Writing in the World; and that all other Alphabets derive from the Hebrew". Others credited Adam as the inventor of writing. In 1668, John Wilkins, one of the founders of the Royal Society and an influential and respected English scholar, commented that Adam had invented the Hebrew alphabet: "though not immediately after his creation, yet in process of time, upon his experience of their great necessity and usefulness."
The myths, from Sumer to Daniel Defoe, share one common characteristic: they present writing as emerging, on one day, as a full-fledged script. None of them conveys the notion of an evolution from a simple to a more complex system of communication. The concept of a ready-made alphabet handed down from heaven persisted until the eighteenth century.
The Pictographic Theory
In the eighteenth century, William Warburton, bishop of Gloucester, introduced the first evolutionary theory of writing. Based on his observations of Egyptian, Chinese, and Aztec manuscripts, Warburton argued that all scripts originally developed from narrative drawings. In time, he said, these pictures became more and more simplified and developed into abstract characters. The theory was presented in Warburton's book, Divine Legation of Moses, published in London in 1738. The ideas made their way into Diderot and d'Alembert's Encyclopédie, in the article entitled "Ecriture," which gave them a wide diffusion. Warburton's pictographic theory remained practically unchallenged for over two hundred years. For example, in the revised edition of A Study of Writing (1974), at present one of the bestknown modern scholarly publications on writing, I. J. Gelb still stated: "it became clear that the Mesopotamian cuneiform writing has developed from a pictographic stage."
Although the existence of cuneiform had been noticed by Western travelers as early as the fifteenth century, ancient Near Eastern scripts did not play a role in the elaboration of the pictographic theory because they were still little known in 1738. In the nineteenth century, when archaeological expeditions reaped the first great harvests of cuneiform texts and brought them back to Europe, the cuneiform script was regarded as conforming to Warburton's paradigm. In 1913, George A. Barton was of the opinion that "the investigator must proceed upon the hypothesis that Babylonian writing, like other primitive writing, originated in pictographs." The pictographic theory scheme was modified, however, to include a three-step progression from ideographic to phonetic writing. "Indeed," wrote Barton, "wherever the beginnings of writing could be traced, it took the form of picture writing, so that it seems safe to regard it as a working hypothesis, if not as a law, that all early systems of writing began in a series of pictographic ideographs, that syllabic values were developed from these and in some cases alphabetic values."
In fact, the idea that the cuneiform script started with picture writing was by no means a perfect fit. In 1928, a year before the discovery of the Uruk tablets, William A. Mason noted, "We must admit, that even in the earliest and most archaic inscriptions discovered, it is not always easy to recognize the original objects." The pictographic theory, however, was never questioned. Instead, the Babylonian scribes were blamed for the discrepancy between preconceived ideas and facts: "Owing to the limitations of primitive culture, the inexperience of the scribes and the lack of artistic ability, each scribe drew the characters in his own crude, faulty way, often incorrectly; so that it is quite impossible always definitely to distinguish the character and identify it with the object intended."
The season of 1929-1930 at Uruk brought considerable new information about the beginning of writing. Hundreds of archaic tablets were unearthed that pushed writing back to the fourth millennium B.C. Signs traced or impressed with a stylus in a technique different from the cuneiform script had been termed pictographic. These archaic tablets, however, contradicted the pictographic theory. Adam Falkenstein, the German scholar who studied the texts, noted that when writing began in Mesopotamia, truly pictorial signs were rare exceptions. Those that were truly pictorial, like the signs for "plow," "chariot," "sledge," or "wild boar," were not only few but of uncommon use, represented by a single occurrence on one tablet alone. The common signs were abstract: the sign for "metal" was a crescent with five lines; the "pictograph" for "sheep" was a circle with a cross. The Uruk tablets seriously strained the pictograph theory by showing that when writing began in Mesopotamia pictographic signs were rarely used.
Edward Chiera and others tried to reconcile Falkenstein's observations with the pictographic theory. They argued that the Uruk texts represented an already evolved script and that a previous stage, consisting of true pictographs, probably had been written on a perishable material, such as wood, bark, papyrus, or parchment which had disintegrated in time and could never be recovered.
The next excavation campaign at Uruk, in 1930-1931, produced impressed tablets which continued to challenge the pictographic theory. These texts, like others found previously at Susa and later at Khafaje, Godin Tepe, Mari, Tell Brak, Habuba Kabira, and Jebel Aruda, were more ancient than the "pictographic" Uruk tablets studied by Falkenstein. They were, however, not made of wood, bark, papyrus, or parchment, as Chiera had hypothesized. This earliest form of writing consisted of wedges, circles, ovals, and triangles impressed on clay tablets and was anything but pictographic. The gulf between the evidence and the neat charts of pictographs illustrated in books became even wider.
By the second half of the twentieth century, enough data had accumulated to present a serious challenge to the pictographic theory. From Champollion in 1822 to Ventris in 1953, each great decipherment eroded the premise upon which the pictographic theory was built and determined that the early scripts all had phonetic features. Anthropologists like André LeroiGourhan entered the debate, warning against preconceptions about primitive picture writings. In his volume Le Geste et la parole, he argued that "the linguists who have studied the origin of writing have often conferred to pictographic systems a value which derives from literacy." Leroi-Gourhan noted that the only true pictographic scripts were recent phenomena; that most had emerged in groups which did not have writing prior to contacts with travelers or colonists from literate countries. He concluded: "therefore it seems impossible to use Eskimo or Indian pictography in order to understand the ideography of preliterate societies."
Scholars began to question the idea that writing emerged as the rational decision of a group of enlightened individuals, such as that put forward by V. Gordon Childe in What Happened in History: "the priests ... have agreed upon a conventional method of recording receipts and expenditures in written signs that shall be intelligible to all their colleagues and successors; they have invented writing." For, like other human inventions, writing did not come ex nihilo and, in Chiera's words, "there never was a first man who could sit down and say, 'Now I am going to write.' That supreme achievement of mankind, the one which makes possible the very existence of civilization by transferring to later generations the acquisitions of the earlier ones, was the result of a slow and natural development. "
Above all, the pictographic theory was not consistent with modern archaeological research. In recent years, Near Eastern excavations have focused on the beginning of agriculture and cities and have tried to determine how these events affected society. Viewed in the perspective of the urban phenomenon, the first "pictographic" tablets of Uruk (and for that matter the earlier impressed tablets) are out of step with other socioeconomic developments. These first documents occur in level IVa of Uruk, lagging far behind the rise of cities and the emergence of the temple institution, which was already well under way some two hundred years earlier (in levels X-IV). If writing emerged so late, it could not have played a role in state formation. How, then, did the Mesopotamian city-states function without record keeping?
Although popular books have continued to present the traditional pictographic theory, as early as the 1950s scholars had begun to anticipate the discovery of an antecedent of the Mesopotamian script. Some, like V. Gordon Childe, searched for it in seals while others looked into potter's marks, but to no avail. Most, like Seton Lloyd, foresaw an even earlier script: "The degree of competence... [attained by script of the Uruk IV tablets] ... suggests that earlier stages in its development may eventually be recognized elsewhere, perhaps in levels corresponding to Uruk V and VI." David Diringer simply referred to "another more primitive writing" or "an at present unknown, early script, which may have been the common ancestor of [the Indus Valley Script and] also of the cuneiform and early Elamite writings." I propose that the antecedent of writing was not an earlier script, but a counting device. What had been missed—or dismissed—were the humble tokens that had been used for centuries and that were, I argue, the immediate precursor to writing.
The pictographic theory will remain a landmark in the history of ideas because it was the first evolutionary explanation of writing, departing from the former belief that a full-fledged script had been communicated to humans by divine revelation. The theory was based, however, on Egyptian, Chinese, and New World models that were irrelevant. Over the course of the twentieth century, archaeology has generated new evidence that contradicted the paradigm. At the same time, excavations steadily produced small tokens that, as I will show, were the antecedent of writing.
The immediate precursor of cuneiform writing was a system of tokens. These small clay objects of many shapes—cones, spheres, disks, cylinders, etc.—served as counters in the prehistoric Near East and can be traced to the Neolithic period, starting about 8000 B.C. They evolved to meet the needs of the economy, at first keeping track of the products of farming, then expanding in the urban age to keep track of goods manufactured in workshops. The development of tokens was tied to the rise of social structures, emerging with rank leadership and coming to a climax with state formation.
Also, corresponding to the increase in bureaucracy, methods of storing tokens in archives were devised. One of these storage methods employed clay envelopes, simple hollow clay balls in which the tokens were placed and sealed. A drawback of the envelopes was that they hid the enclosed tokens. Accountants eventually resolved the problem by imprinting the shapes of the tokens on the surface of the envelopes prior to enclosing them. The number of units of goods was still expressed by a corresponding number of markings. An envelope containing seven ovoids, for example, bore seven oval markings.
The substitution of signs for tokens was a first step toward writing. Fourth-millennium accountants soon realized that the tokens within the envelopes were made unnecessary by the presence of markings on the outer surface. As a result, tablets—solid clay balls bearing markings—replaced the hollow envelopes filled with tokens. These markings became a system of their own which developed to include not only impressed markings but more legible signs traced with a pointed stylus. Both of these types of symbols, which derived from tokens, were picture signs or "pictographs." They were not, however, pictographs of the kind anticipated by Warburton. The signs were not pictures of the items they represented but, rather, pictures of the tokens used as counters in the previous accounting system.What fascinated me most in this study was the realization that the token system reflected an archaic mode of "concrete" counting prior to the invention of abstract numbers. There were no tokens for "1" or "10." Instead, a particular counter was needed to account for each type of goods: jars of oil were counted with ovoids, small measures of grain with cones, large measures of grain with spheres. Tokens were used in one-to-one correspondence: one jar of oil was shown by one ovoid, two jars of oil by two ovoids, etc. The consequences of this discovery are significant. Namely, writing resulted not only from new bureaucratic demands but from the invention of abstract counting. The most important evidence uncovered is that counting was not, as formerly assumed, subservient to writing; on the contrary, writing emerged from counting.
Studies on Tokens
Tokens came my way by chance. It all started in 1969-1971, when I was awarded a fellowship from the Radcliffe Institute, Cambridge, Massachusetts (now the Bunting Institute) to study the use of clay before pottery in the Near East. This led me to systematically examine Near Eastern archaeological clay collections, dating from 8000 to 6000 B.C., stored in museums of the Near East, North Africa, Europe, and North America. I was looking for bits of Neolithic clay floors, hearth lining, and granaries, for bricks, beads, and figurines, and I found plenty of these. I also came across a category of artifacts that I did not expect: miniature cones, spheres, disks, tetrahedrons, cylinders, and other geometric shapes made of clay. I noted their shape, color, manufacture, and all possible characteristics. I counted them, measured them, drew sketches of them, and entered them into my files under the heading "geometric objects." Later, when it became obvious that not all the artifacts were in geometric form, but some were in the shape of animals, vessels, tools, and other commodities, the word was changed to token.
I became increasingly puzzled by the tokens because, wherever I would go, whether in Iraq, Iran, Syria, Turkey, or Israel, they were always present among the early clay assemblages. If they were so widely used, they must have had a useful function. I noted that the tokens were often manufactured with care and that they were the first clay objects to have been hardened by fire. The fact that people went to such trouble for their preparation further suggested that they were of importance. The tokens appeared to be part of a system, for I repeatedly found small and large cones, thin and thick disks, small and large spheres - even fractions of spheres, such as half and threequarter spheres. But what were the tokens used for?
I asked archaeologists about the tokens and learned that everyone who had excavated early sites had encountered them in their trenches. No one, however, knew what they were. I looked in site reports and noted that tokens were usually omitted or relegated to such headings as "enigmatic objects" or "objects of uncertain purpose." The authors who did risk an interpretation identified the tokens as amulets or game pieces. Carleton Coon is among those who simply wondered. He jovially reported about the five cones he found at Belt Cave, Iran, as follows: "From levels 11 and 12 come five mysterious conical clay objects, looking like nothing in the world but suppositories. What they were used for is anyone's guess." The data I collected on tokens seemed at first to be of little significance, but ultimately made it possible to recognize the importance of these artifacts. The information on the Neolithic counters turned out to be the piece of a puzzle that, finally, gave a clue to the entire picture.
Many archaeologists, starting with Jacques de Morgan at Susa (1905) and Julius Jordan at Uruk (1929), should be recognized for excavating, preserving, and publishing tokens, even though they seemed insignificant at the time. Vivian L. Broman is to be credited for including in her work the study of hundreds of tokens from Jarmo. When Broman completed her thesis in 1958, she too had no alternative but to guess from the shape of the objects what they might have been. Consequently, she attributed a different function to each particular type. She viewed the cones as being perhaps schematic figurines and the spheres as sling stones or marbles. She also earmarked cones, spheres, and hemispheres as possible counters, noting that some Iraqi shepherds today keep track of their flocks with pebbles. At the time, her insight could not be supported by archaeological evidence. Only one year later, however, the use of counters in the ancient Near East was documented.
A. Leo Oppenheim
In 1959, A. Leo Oppenheim of the University of Chicago wrote an article on counters of the second millennium B.C. which proved to be the key to understanding what the tokens were . His paper concerned a peculiar hollow tablet recovered in the late 1920s at the site of Nuzi, in northern Iraq (fig. 2). This egg-shaped tablet belonged, together with a normal tablet bearing an account of the same transaction, in the family archive of a sheep owner named Puhisenni. The cuneiform inscription on the hollow tablet read as follows:
Counters representing small cattle:
- 21 ewes that lamb
- 6 female lambs
- 8 full grown male sheep
- 4 male lambs
- 6 she-goats that kid
- 1 he-goat
- 3 female kids
The seal of Ziqarru, the shepherd.
When opening the hollow tablet, the excavators found it to hold forty-nine counters which, as stipulated in the text, corresponded to the number of animals listed.
That hollow tablet proved to be the Rosetta stone of the token system. The counters (Akkadian abnu, pl. abnati, translated "stone" by Oppenheim), the list of animals, and the explanatory cuneiform text leave no possible doubt that at Nuzi counters were used for accounting. Although no other example of a cuneiform tablet holding counters has ever been encountered in Nuzi or, for that matter, in Mesopotamia or the Near East, Oppenheim made a case that abnati were commonly used in the bureaucracy. He suggested that each animal of a flock was represented by a stone held in a container. The tokens were transferred to various receptacles to keep track of changes of shepherds or pasture, when animals were shorn, and so on. He based his argument on short cuneiform notes found in archives, referring to abnati "deposited," "transferred," and "removed" as follows:
- These sheep are with PN; the (pertinent) stones have not been yet deposited.
- Three lambs, two young he-goats, the share of PN, they are charged to his account (but) not deposited among the stones.
- One ewe belonging to PN, its stone has not been removed.
- Altogether 23 sheep of Silwatesup, PN brought ... their stones have not been transferred.
- x ewes that have lambed, without (pertaining) stones, belonging to PN.
Marcel Sigrist has pointed out further texts which probably also allude to counters in the Third Dynasty of Ur, ca. 2000 B.C. For instance, a tablet dealing with oxen reads: "The remaining part of the account is held in the leather pouch."
When Oppenheim wrote his article, no one knew what the counters looked like. Of course, the abnati mentioned in the texts were not described and those held in the Nuzi hollow tablet were lost. They were simply referred to as "pebbles" in the site report, with no information as to their shapes or the material of which they were made. The next important piece of the puzzle was provided by Pierre Amiet in Paris.
Pierre Amiet studied seals used in the administration of ancient Near Eastern cultures to validate documents. In particular, he came to study the impressions of such seals on globular clay objects from Susa. These artifacts were hollow and contained small clay objects. Following Oppenheim's lead, Amiet interpreted the small clay objects enclosed in the clay envelopes as "calculi" representing commodities. The proposition was daring, since the Susa envelopes came two thousand years earlier than the Nuzi egg-shaped tablet, with no known example in the interval. It was a leap of great importance for three reasons. First, the counters were revealed: they were miniature clay artifacts modeled in various, mostly geometric, shapes. Second, the Susa envelopes showed that counters held in envelopes were not restricted to the literate period but extended into the time before writing was invented. Third, Amiet foresaw the possibility that the calculi were an antecedent of writing. In his words: "One might ask whether [the scribe] had in mind the little objects that were enclosed in the envelopes, and that very conventionally would symbolize certain goods."
Amiet's contribution was a tremendous step, but it revealed only one time frame of the token system at a particular location: namely, Susa about 3300 B.C. It should be kept in mind that in 1966, prehistoric tokens were not known and the only published parallels to the Susa envelopes were those recently excavated in Uruk. Six years later, in 1972, when Amiet published the Susa envelopes in his Glyptique susienne, he still described the markings on envelopes as follows: "a series of round or long notches, similar to the ciphers featured on tablets and corresponding to the number of calculi enclosed inside. The shape of the notches, however, is not as varied as that of the calculi." More recently, Amiet summarized his position: "I was thus wondering whether writing was inspired by certain of these calculi enclosed in the envelopes."
Maurice Lambert, conservator of Western Asiatic antiquities at the Louvre, took Amiet's insight two steps farther. He clearly recognized that the first impressed signs of writing were reproducing the shape of the former calculi: "Writing, here as elsewhere, imitated true things." Consequently, he assigned the values 1, 10, 60, 600, and 3,600 to, respectively, the tetrahedron, sphere, large tetrahedron, punched tetrahedron, and large sphere, a route that turned out to have been, partly, a false one.
The recognition that the tokens constituted an accounting system which existed for five thousand years in prehistory and which was widely used in the entire Near East was to be my own contribution. I was also able to draw parallels between the shapes of the tokens and those of the first incised signs of writing and to establish the continuity between the two recording systems. Finally, much later, I realized the mathematical importance of the tokens as an archaic reckoning device, preceding the invention of abstract counting. I recall vividly when, in 1970, two pieces of the puzzle snapped together for me. In order to prepare a class lecture, I pulled from my files Amiet's 1966 article, which I had not seen since I began collecting tokens. I could not believe my eyes when I saw the small clay cones, spheres, and tetrahedrons illustrating the paper. Until then, I had instinctively dismissed the idea that the Susa artifacts could have anything to do with the tokens which had been found in Neolithic villages. After all, the calculi from Susa were held in envelopes and the Neolithic tokens were loose; moreover, the objects were separated by thousands of years. The next day, however, I was intrigued enough to check several excavation reports of fourth-, fifth-, and sixth-millennium sites and saw the possibility that tokens might have been used, with no discontinuity, between 8000 and 3000 B.c. The rest was hard work. My first publications on tokens and their relation to writing date from 1974 to 1978, those on tokens and concrete counting from 1983 to 1986.