Cuneiform Digital Library Journal
§1. The perspective of historical epistemology
§1.2. In the following, I will approach these questions from a particular point of view, that is, from the viewpoint of historical epistemology. Let me first briefly explain this perspective. It is now some 200 years ago that, in the aftermath of Napoleon’s campaign in Egypt, early writing systems and their historical origins aroused the curiosity of European scholars for the first time. They soon became a focus of never-ending attempts to decipher them and to disclose the hidden treasures of authentic information about cultures that perished long ago. Understanding these texts primarily meant retranslating them into the oral languages they represent. Thus, it does not come as a surprise that work on early writing systems was widely influenced by what can be called a philological perspective. From this point of view, a writing system is essentially conceived of as a representation of a particular language. Deciphering such a system means identifying the underlying language and reconstructing the way it is coded in the written symbols. If this language was not a familiar one that had survived, scholars felt challenged to reconstruct its grammar and its lexicon. It is an encouraging fact that this endeavor has in most cases been successful, where a sufficient number of texts has been handed down to us. We therefore know today quite well the grammar and the lexicon of the languages which have been written in systems such as Egyptian hieroglyphs or Near Eastern cuneiform.
§1.3. The enormous success of the work that has been done under the influence of such a philological perspective cannot and should not be ignored. Nevertheless, it must also be admitted that this perspective, in spite of its success, has serious shortcomings. Its deficiencies become obvious especially when one focuses on the contribution of philology to the clarification of the origins of writing. It is now a well established fact that the influence of the structures of language on a system of writing becomes weaker the further one goes back in its history. I will call in the following such incipient systems of writing with weak connections to oral language proto-writing.
§1.4. As might be expected, methods of philology are less effective if the relationship between writing and language is weak. It is possibly due to this fact that major corpora of early writing systems that do not adequately represent an ancient language have, for a long time, been widely neglected by philologists. Protowriting does not provide sufficient information about the grammar of the language of the scribe to make it a valuable source for philological research. In some cases, it is not even possible to identify the language spoken by the people who invented and used such systems of proto-writing.
§1.5. The lack of attention that proto-writing has received is, nevertheless, puzzling, because precisely these texts are exceptionally important in any attempt to solve the exciting riddles of the origins of writing. The difficulties of understanding texts written in a protowriting system result primarily from the fact that the information represented in such a system is essentially incomplete. The scribes of proto-writing apparently assumed that the readers of their texts, much like discussion partners in oral communication, knew the context of the information they wanted to transmit. Therefore, they probably took it as given that their readers were able to interpret the given information correctly, in the same way that they could understand oral statements contextualized in a natural setting of discourse.
§1.6. In modern cognitive science, such processes of decoding incomplete information have been intensively studied. A body of knowledge shared between the partners in a communication process provides cognitive frames that are triggered by the communication process, instantiated by the given incomplete information, and finally, complemented by default assumptions about the subject, which are retrieved from memory as an effect of the assimilation of this subject to the frame.
§1.7. Systems of proto-writing in this sense force us to broaden the perspective of our studies towards what I call an epistemological perspective. From the viewpoint of historical epistemology, proto-writing is not seen merely as a deficient representation of language but rather as a successful means of representing knowledge and transmitting it from one individual to another, and eventually from one generation to the next.
§1.8. From such a point of view, the philological perspective appears to be an interpretation of writing as a representation of a representation of knowledge, that is, as a written instantiation of the representation of knowledge in oral language. By and large, such a viewpoint seems to be legitimate in the case of fully developed systems of writing, but it necessarily leads to misleading consequences in the case of proto-writing. Historical epistemology poses the questions of when, where, why and how writing was invented in view of the broader perspective of studying writing as a means of representation and the historical transmission of knowledge that may or may not be intimately linked to language as a means of oral communication.
§2. Monogenesis or polygenesis of writing?
§2.2. If the invention of writing was just a brilliant idea developed once and then copied over and over again, it may have been fortuitous circumstances that led to its invention. In this scenario, it would make no sense to compare different early writing systems in order to determine the conditions of their emergence in history. Rather, the mechanisms of transfer would have to be studied, mechanisms that could induce the development of the variety of different systems of writing that emerged one after the other in the course of history.
§2.3. Alternatively, if writing systems were invented independently on multiple occasions and at different places in the world, but in all cases leading essentially to the same result, that is, to an adequate representation of oral language, the circumstances and the internal mechanisms of the emergence of writing would have to be reconstructed, mechanisms that would in this case be responsible for the repeated occurrences of such an event.
§2.4. I shall not attempt to give here an answer to the question of which of these alternative hypotheses is more likely to be true. Let me just point out that there is probably no simple answer to this question. On the one hand, the various systems of coding oral language that have been developed at different places show a great variety. Whatever the mutual influences of writing systems of different cultures may be, this variety demonstrates, at least, that the development of writing, once it is initiated, attains a considerable degree of independence and flexibility to adapt a coding system to specific characteristics of the language to be represented. On the other hand, the historio-geographic distribution of the various occurrences of early writing systems seems to indicate, as Ignaz Gelb has pointed out in his famous Study of Writing (Gelb 1952: 212-220), that the idea spread in various directions at the beginning of the 3rd millennium BC from centers in Mesopotamia and Egypt. Proto-Elamite writing occurs only a short time after proto-cuneiform. It was used for a short period in vast areas of the Iranian plateau. In the second half of the 3rd millennium BC, writing is attested as far to the north as Ebla in Syria and to the east as the Indus culture in modern Pakistan. Minoan writing starts at Crete around the turn of the 3rd to the 2nd millennium BC. At that time, cuneiform writing is also attested further north in the regions of Anatolia. The oldest attestations of writing in China date back to the Shang period at the end of the 2nd millennium, and it is only in the 1st millennium BC that we find writing in Mesoamerica (see figure 1). This distribution of earliest attestations of writing in time and space shows the typical pattern of a spreading technology, although in some cases there appears to be no connection at all between the systems of writing that subsequently emerged.
§3. The philological perspective of the origins of
§3.2. This anachronistic projection of modern functions of writing onto its early use had the consequence that the multiple origins of writing were widely neglected. The early development of writing was interpreted as a universal process leading from a crude representation of words by pictures to the more efficient representation of words decomposed into phonemes by syllabic signs and, finally, to alphabetic writing. As is well known, Paul Sethe (1939) and later Ignaz Gelb (1952) have developed influential theories of the origins of writing based essentially on such assumptions.
§3.3. It is an irony of history that it was precisely Gelb in his Study of Writing who argued forcefully that, for systematic reasons, the Mesoamerican writing systems cannot have been based on any kind of phonetic coding, and that “even a superficial knowledge of the inscriptions of the Aztecs and Mayas is enough to convince oneself that they could never have developed into real writing without foreign influence” (Gelb 1952: 58). We know now that, in this respect, Gelb was wrong. This example should warn us not to trust seemingly plausible arguments or historical extrapolations that are not supported by direct historical evidence. We should, in particular, not rely on the plausibility of the assumption that writing was created in order to enhance or to substitute for oral communication. There is, at least, no longer any reason today to assume that the desire to represent language in an enduring medium was the only motive that triggered the invention of writing.
§4. Challenging characteristics of proto-cuneiform
§4.2. Let me first list briefly these characteristics before I go into details.
§4.3. None of these characteristics are in and of themselves sufficient to prove or disprove a particular hypothesis on the nature of proto-cuneiform writing, but taken together, they make it unlikely that proto-cuneiform was just an early step in a linear sequence of ever more successful attempts to represent human language. What is more relevant here, these characteristics raise a number of general questions concerning the emergence of protowriting that may be relevant also in contexts other than those of Mesopotamia in the 3rd millennium BC, and that may be helpful in future discussions of the origins of early writing systems.
§5. Proto-cuneiform does not match the syntax of oral
§5.1. There may be exceptions. I am thinking in particular of the so-called Tribute List, attested by nearly sixty textual witnesses, which exhibits structures resembling epic iterations and has, therefore, been interpreted by Bob Englund and Hans Nissen as the earliest example of real literature (Englund & Nissen 1993: 25-29). Unfortunately, this text is still so badly understood that no conclusion can be derived about the technique of coding language that might have been used to create it.
§5.2. A number of questions that concern not only proto-cuneiform but proto-writing in general can be derived from the observation that, with this exception, the texts written in proto-cuneiform do not show any direct relation to oral language. Such questions are:
§6. Semantic categories
§6.2. Proto-cuneiform thus shows that in proto-writing systems the meaning of signs and sign combinations may be restricted to a few dimensions determined by their specific functions. The question has to be asked for any early writing system:
§7. Lack of phonetic coding
§7.2. The first explanation of this phenomenon is, of course, that the meaning of such sign combinations is coded phonetically. This, however, seems not to be the case. Phonetic coding is, in fact, not the only possible explanation that can be given for such an independence of depiction and meaning. In a restricted semantic field such as administration in an archaic society, any other convention of coding could easily serve the same purposes as phonetic coding, provided that the context of its application makes it possible to memorize the represented meanings (see figure 4).
§7.3. If a writing system does not have to represent the full range of semantics of oral language, to depict an object or to use some kind of phonetic code are obviously not the only ways in which meaning can be conveyed. The possibility has to be taken into account for every early system of writing that a variety of different techniques may have been used for coding information. The question must therefore be raised:
§8. Range of conventionalization
Table 1: Statistics of the use of non-numerical protocuneiform signs based on ca. 40,000 occurrences of 1,617 non-numerical signs in ca. 6,000 texts and text fragments.
§8.2. Proto-cuneiform consists of more than 1500 non-numerical signs attested by more than 40,000 occurrences of these signs in the corpus of approximately 6000 preserved texts and text fragments. Among these signs, there is a comparatively small group of frequently used signs; about 100 signs are attested more than 100 times, of which the two most frequent ones, the signs ENa and GALa, are found more than 1000 times each. There is, however, on the other side a much larger group of signs that were used very rarely; more than 500 signs are attested only once, a further 600 signs less than ten times each.
§8.3. This remarkable irregularity of sign usage suggests that proto-cuneiform writing was based on a core of standardized signs. These could, however, be flexibly complemented by modifications of existing signs or by the creation of new signs that were used only in specific contexts, and that never developed into standardized signs of cuneiform writing (see figure 5).
§8.4. Further evidence for this assumption is provided by another statistical irregularity (see table 2). There are signs and sign variants that are well attested, but that were nevertheless used by scribes in specific groups of texts, while not in others. Such a text group is defined by the texts of the former Erlenmeyer collection together with some other texts in private collections (MSVO 3, 1-90). We can infer from persons and institutions represented in these texts that the collection must have been written at the same place and at the same time. Sign statistics demonstrate that some of the signs are frequently used in this small group of 90 texts, but are rarely used in the approximately 6,000 other proto-cuneiform documents.
Table 2. Statistics of proto-cuneiform signs used in the text group MSVO 3 (90 texts) in comparison to their use in the total text corpus (ca. 6,000 texts).
§8.5. The general question which is suggested by this example is the following:
§8.6. This question introduces a new dimension into my arguments: The development of writing cannot be adequately described on the technical level of coding information only. The social environment has to be taken into account because it obviously had a great influence on what was written down and how. Writing is not just a technique developed to serve a universal human need, but rather it is a social process of knowledge representation based on human interaction and historical continuity.
§8.7. From this point of view, it cannot be considered only an incidental condition of proto-cuneiform writing that it was initially used predominantly or even exclusively to document administrative activities. This restricted context of application which influenced its formal structure and its semantics may even be considered constitutive of its origin. For any early writing system the questions have to be posed:
§9. Precursors of writing
§9.2. Clay tokens were used to record objects and quantities. Uniform containers were used to standardize quantities. Seals were used to represent legitimacy and property based on social power in a form that makes them independent of personal conflict. The invention of proto-cuneiform writing turned out to have been only the last step in a long tradition in the development of a prehistoric means of administration.
§9.3. The discovery that such means developed into proto-cuneiform with virtually no discontinuity is in good accordance with the limited function proto-cuneiform seems to have had in the period immediately after its emergence towards the end of the 4th millennium BC. The general questions implied by this observation concerns any introduction of writing into a culture:
§10. Adaptation to oral language
§10.2. Thus, cuneiform writing apparently developed in two stages. In the first stage, writing was fairly independent of phonetic coding, but its application was restricted to narrowly defined contexts and its signs and sign combinations did not yet represent universally applicable words, but rather specific entities and activities in the context of administration. In the second stage, phonetic coding made a new type of application possible, the written representation of information in the same way as it was previously transmitted orally. While at the beginning this new application played only a minor role, in the course of later history, fostered further by its application to the Akkadian language, the relative importance of the different elements of the writing system changed and, in its turn, the nature of cuneiform writing.
§10.3. This model of the development of early writing is surely specific to the development of proto-cuneiform into the cuneiform writing system. It shows, nevertheless, a further aspect of the development of writing that has to be taken into consideration. This development is not necessarily a linear process, but may have different tracks combined with internal reorganizations of the total writing system in certain periods. In any case, the model challenges the historical reconstruction of the origins of writing by suggesting another general question:
§11.2. With regard to the structures of early writing systems, these questions are focused on two areas. The weakness of philological methods applied to protowriting demonstrates, first, that it is necessary to pay special attention to non-linguistic structures of syntax. The specific use of proto-writing demonstrates, second, that the nature of the semantic fields of early writing systems, and the various techniques used to represent meanings, have to be carefully analyzed.
§11.3. With regard to the social contexts of early writing systems, the questions focus on the social mechanisms that establish shared meanings and on the explanation of the structures of proto-writing that can be derived from the social context of its emergence.
§11.4. With regard to the historical dynamics of the development of writing, these questions are focused on the continuities and discontinuities in the development from the pre-literate means for representing knowledge to proto-writing and, finally, to the representation of oral language. This development has to be analyzed in view of the important fact that in pre-literate societies many techniques of representing knowledge other than writing exist that finally merge, by substitution or integration, into real writing.
§11.5. Given the great variety of non-linguistic structures, functions, and techniques of knowledge representation discussed here, it no longer makes sense to speak about some presumed unified origins, or linear development of writing. The term “origins of writing” is related to historical developments in huge geographical areas over a time-span of some 2000 years. Already the comparatively small region of the Near East from which I took my examples shows an enormous richness of different developments, which all contributed to the emergence of writing. If one takes into account developments in such different settings as those of Egypt, China or Mesoamerica, the complexity of different developments is increased even more.
§11.6. This brings me back to the problem of monogenesis or polygenesis of writing that I raised at the beginning of the present paper. Whatever the eventual solution might be to this problem, it has to take into account the quite different ways writing in one culture may influence the emergence of writing in others. Even if we accept the monogenesis hypothesis, the complexity of the historical emergence and development of writing will not be significantly reduced. Proto-Elamite did not developed in the same way under the influence of proto-cuneiform as did Hittite or Minoan writing under the influence of cuneiform, or Japanese writing under the influence of Chinese, to say nothing of developments such as the emergence of the Indus script, of Chinese itself, or of Maya writing, which cannot easily be related to models of monogenesis and diffusion. Even the final outcome of the development, that is, the developed system of real writing, can be quite different in different cultures, depending on the different structures of the languages that are finally represented.
§11.7. This historical variety should be kept in mind when we compare semiotic developments across many cultures. In fact, due to its historical variety, questions relating to the development of early writing systems will probably only be resolved through the cooperation of specialists from different fields.
Recent Publications on Proto-Cuneiform and Its Precursors
Version: 28 January 2006