The Writing Revolution

by Amalia E Gnanadesikan

  Sounds that do not occur in English will be explained where relevant in later chapters. However, much can be learned about them from studying the IPA charts. The purpose of laying out the IPA vowel and consonant symbols in charts is that even if your language does not contain a particular sound, you can get a fairly good idea of what it sounds like from the description and its place on the chart. For instance, English does not have the [x] sound. However, the consonant chart describes it as a fricative, in the same row as [f], [v], [θ], [ð], [s], and [z], which do occur in English (as in fine, vine, thin, thine, sin, and zen). A fricative is, like any of these sounds, a sound that you can keep on making (unlike a plosive, such as [b]), but that makes a turbulent sound of rushing air (unlike, say, [l]). The column [x] is in shows that it is a “velar” sound. This tells us that it is made in the same part of the mouth as [k] and [g], at the soft palate toward the back. It is therefore the “ch” sound of Scottish loch or German ach. Most English speakers find this sound impossible to pronounce correctly. However, with the IPA they can at least talk about it, even if they can’t produce it.

  The vowel chart also has many sounds that English does not possess, such as [y]. It is described by the chart as “close” and “front.” Even if those terms do not mean anything to you, you can tell from the chart that it is similar to [i], the vowel in English see. But unlike [i], it is “rounded.” This means that it is made with pursed lips, like [u], the vowel in food. If you say [i] and then try to say it with your lips pursed, you may manage the vowel [y]. Most English speakers have a great deal of trouble with it. It is the vowel sound in French tu.

  A full understanding of the IPA is not necessary to this book. However, if you find yourself wondering what sounds the symbols in a script actually refer, you can get a rough idea by using the IPA chart.

  2

  Cuneiform: Forgotten Legacy of a Forgotten People

  Our story begins in the Middle East, in what is today southern Iraq. Nestled between the southern reaches of the Tigris and Euphrates rivers was a land whose earliest recorded name is Sumer (see map in the appendix, figure A.3). The land was rich and fertile, but dry. With irrigation, however, the land yielded enough and to spare; the extra food encouraged the growth of trade and the development of specialized professions and stratified social classes. Over time these developments led to the birth of a true civilization in the centuries between 3500 and 3000 BC.

  It was a momentous period. Facing a drying climate, people living in northern Mesopotamia moved south to avail themselves of the benefits of irrigation. The necessity of feeding a larger population was the mother of a number of inventions: the plow, the grain sled, the potter’s wheel, wheeled vehicles, and the sail. These technologies allowed people to plant more food and to store, transport, and trade it more easily. To keep pace, irrigation technology had to be improved, and larger-scale irrigation systems built. Such public works required concerted community effort, encouraging the rise of strong community leaders. With fertile land at a premium, scattered villages began to be replaced with more compact, centralized cities. As the cities grew in size and complexity of organization, so their leaders accumulated power and wealth.

  The growth of civilization required yet another advance in technology. The complex society of a city-state requires administration, and administration requires record keeping. Early attempts to meet this need included the working out of a numerical tally system and perhaps the use of tokens, which stood for farm animals, quantities of grain, trade goods, or other objects that needed to be recorded. But these advances were not enough to meet the culture’s growing bureaucratic needs. The technology that emerged to meet those needs – writing – filled a prosaic but essential purpose: accounting. The impetus behind its invention was not a desire to faithfully record language, but to record trade transactions, crop yields, and taxes – to record and preserve information, not language. It was the first – and most important – information technology revolution. It succeeded in preserving information, however, by representing language – crudely at first, but with increasing precision. As the technology matured and spread, it came to be used for languages of five separate linguistic families and inspired the development of several other scripts; it shaped forever the world that came afterwards.

  Commodities such as wood, stone, and metal had to be imported into Sumer, but thanks to the Euphrates and Tigris rivers the land was rich in mud. The right kind of mud produced clay. Almost everything in Sumer was made of clay: bricks for houses and temples, tools, and even writing surfaces. Clay left to dry of its own accord is fairly durable; baked clay is virtually indestructible. It is because of this property of clay that ancient Mesopotamian writing is preserved for us: early unbaked tablets have come down to the present rather crumbly but often still legible after five millennia, while tablets that experienced the sacking and burning of cities were merely strengthened by the process. Later tablets were sometimes intentionally baked to preserve them.

  The earliest writing known to archaeologists is found on small clay tablets unearthed from the ancient city of Uruk. Uruk (biblical Erech, modern Arabic Warka) was once a thriving city on the banks of the Euphrates. It is now a large and desolate heap 12 miles from the Euphrates, the river bed having shifted over the course of the millennia. Arguably, the name has changed less with time than the geography.

  Modern excavations began at Uruk in 1912. As is usual in archaeology, the periods of occupation have been named from the top (latest, but first to be unearthed) downward. Thus Uruk I is more recent than Uruk II, and so forth. It is during the period Uruk IV that writing is first attested, around 3400 or 3300 BC. In the Uruk III period (c.3200–3000 BC), a few other Sumerian cities also show evidence of writing. Either independently or by the inspiration of the Sumerians, writing also appears around this time east of Mesopotamia at Susa – recording the as yet undeciphered proto-Elamite language – and to the west in Egypt.

  Though the earliest known writing comes from Uruk, we cannot be entirely certain that this is where writing was actually invented. However, later Sumerian legend also places the first writing there. The invention is ascribed to Enmerkar, said to have been king of Uruk after the Great Flood of Sumerian legend. Enmerkar was engaged in a contest of wills with the lord of far-off Aratta. He had sent three messages demanding tribute and had been denied three times. His final move was to send a written clay tablet. The written message reduced the lord of Aratta to submission, perhaps in recognition of the significance of the invention.

  The first written tablets are in a script we call proto-cuneiform. About 85 percent of them are of an administrative or accounting nature, while the remaining 15 percent are lists of words. The latter were spelling lists, used by scribes practicing the signs for the various professions, agricultural produce, and commodities. The same word lists, written in the same order, were used for hundreds of years; conveniently, this fact allows modern Sumerologists to use the later lists to identify early proto-cuneiform signs.

  Proto-cuneiform was scratched, or drawn, into damp clay tablets (see figure 2.1). Numerals figure prominently in the early business records: of the roughly 800 different signs that have been identified, 60 or so are numerals. This is a lot, compared to the 10 digits we use today. But the early Sumerians did not use numerals abstractly, without reference to what was being counted. Instead, different systems of numerals were used for counting different things: for discrete entities, for areas of land, for periods of time, for quantities of grain, and so forth. This was probably a holdover from the tally systems of the pre-literate period, when numerals that told you something about what was being counted were an advantage rather than a cumbersome inconvenience.

  Of the remaining signs, some were straightforwardly pictographic – stylized pictures of identifiable objects – like the sign for “head,” sag, or the sign for “fish,” ku, shown in the left-hand column of figure 2.2. In such cases the shape of the sign indicated fairly clearly what the mea
ning of the word was. In other cases the relationship between the appearance of the sign and its meaning was looser, and would only be apparent after one already knew what the meaning was. Still other signs were from the beginning entirely arbitrary or highly abstract, like the sign for “sheep,” udu.

  Figure 2.1 A proto-cuneiform tablet. Administrative tablet with cylinder seal impression of a male figure, hunting dogs, and boars. 3100–2900 BC. Jamdat Nasr, Uruk III style, southern region, Mesopotamia. Clay, H. 2 in. (5.3 cm). Purchase, Raymond and Beverly Sackler Gift, 1988 (1988.433.1). The Metropolitan Museum of Art, New York, NY, U.S.A. Image copyright © The Metropolitan Museum of Art/Art Resource, NY.

  Yet to have a separate sign for each word (even just those consisting of a single morpheme, like sheep) requires a great many signs, and if one is relying on pictography there is the problem that not all words are easily drawn. Modern players of Pictionary can attest to this fact. The early Sumerians came up with a number of workarounds for this obstacle. Less easily pictured words often used the same sign as one that was more easily pictured but had a similar meaning. Thus the sign for “mouth,” ka, which indicated the position of the mouth on a picture of the head, could also mean “tooth” (zu), “word” (inim), “voice” (gu), or “speak” (du), depending on context. Another way to represent words was to combine or modify other, simpler signs. Thus a sign that showed a jar (dug) became, with the addition of stippling or cross-hatching, kaS, the beer that was kept inside such a jar.

  Figure 2.2 The development from proto-cuneiform, through early Sumerian cuneiform, to later Akkadian cuneiform. Proto-cuneiform signs were often pictographic, though not always, as the sign for “sheep” shows. By late cuneiform the pictographic origins are hard to spot. At some point between the second and third columns the orientation of signs (and tablets) was rotated by 90 degrees.

  Some signs were not meant to be read aloud, but functioned as determinatives – unpronounced signs that told the reader what class of thing was being referred to. Early determinatives marked divine names, wooden objects, and male and female names. The use of determinatives is common in logographic writing systems, but they are not unknown elsewhere. In English, for example, we vary letters between upper and lower case. The words frank and Frank are pronounced the same and spelled with the same letters, but the capitalization of Frank informs us that it is a name, while frank is an adjective. The use of upper case thus serves as a determinative in English.

  In contrast with determinatives, some signs were used precisely because of their pronunciation. For example, gi meant “reed.” But the word that meant “to render” in Sumerian was pronounced precisely the same, so the reed symbol was also used to mean “render.” Similarly, ti meant “arrow,” but it could also mean “life,” so a sign depicting an arrow could mean either one. This use of an easily pictured object to stand for its more abstract homonym is known as rebus writing. Protocuneiform used it sparingly, but enough to allow us to identify protocuneiform as the writing of language rather than merely concepts, and to convince some scholars that the language represented was indeed Sumerian. Without these homonym pairs we would know very little about the pronunciation of proto-cuneiform.

  The written messages of proto-cuneiform tablets are all rather telegraphic. All grammatical information, such as verb tense or noun case, was omitted, and only the core morphemes of words were shown, without any prefixes or suffixes. The signs were arranged in boxes outlined on the tablets, one statement per box. The order of the signs within a box did not follow the order of spoken language, and some have even described it as random. Yet there do seem to be patterns: they tended to first record numerals, then the objects counted, and then other relevant information about them, such as “3 sheep temple,” meaning, presumably, that three sheep had been given to the temple. Different types of transactions appear to have been organized differently. But there is much about the arrangement of proto-cuneiform signs that is not obvious to the modern reader; the ancient scribes would have been able to use a significant amount of contextual understanding that is lost to us.

  The proto-cuneiform system was limited and full of ambiguities. However, for the purposes to which it was put it was quite adequate, and context provided the necessary disambiguation. As a technological and intellectual development, it was like nothing the world had seen before.

  The dawn of the third millennium BC saw the emergence of true cuneiform. The name cuneiform refers specifically to the wedge-shaped impressed lines that make up the individual signs of the script. Drawing curved lines on clay is relatively hard; stamping marks into clay is much easier. And so over time the curvilinear signs of protocuneiform gave way to the angular signs of true cuneiform, pressed into the clay with a stylus made from the end of a reed. In the process the pictographic origins of the signs (where there were any) became obscured (as in the second column of figure 2.2).

  The writing now ran consistently from left to right within the boxes. Over time the boxes widened, so that the writing eventually ran in lines across the full face of the tablet. The number of signs grew substantially to about 1,200, then shrank again as the writing system became systematized and the phonological aspect of the script grew. By the middle of the third millennium BC the number of different signs in use was about 800, and by the end of the millennium it had stabilized at about 600.

  It is early in the third millennium BC that we begin to truly recognize the Sumerian language in its writing. The proto-cuneiform tablets contain relatively little phonological information, and no grammatical information, so they tell us relatively little about their language. From later tablets we learn that Sumerian was an agglutinative language, which means that it expressed grammatical information and the relationships between words by adding prefixes and suffixes to its words. As in English, grammatical features such as verb tense, possession, and plurality would be expressed with affixes (the collective term for prefixes and suffixes). Thus there were Sumerian analogs for the English past tense -ed, the possessive -’s, and the plural -(e)s. But unlike in English, other grammatical features were also expressed with affixes. We use prepositions like from, to, and with and modal verbs like will, would, and could as free-standing words, while in Sumerian these concepts would have been expressed with affixes. Nouns and verbs carried up to three suffixes, while verbs could have up to six prefixes and nouns one or none.

  None of these many affixes was written in early Sumerian texts. The recording of affixes began about 2900 BC with a greater use of rebus writing, and increased gradually over the next millennium, spurred by increasing numbers of readers for whom Sumerian was a foreign language. These readers could not so easily fill in missing affixes from context and a personal knowledge of the language. By the time the affixes were being fully represented, Sumerian had probably ceased to be a living, spoken language. It had become a classical language of literature and science, much as Latin did in the centuries following its development into the separate Romance languages. With the death of Sumerian as a spoken, everyday language, the written language was called upon to preserve knowledge of the grammatical details. The death of spoken Sumerian is impossible to date exactly, but probably occurred early in the second millennium BC.

  The recording of the first affixes, and of personal names, forced an expansion of the writing system. The early signs inherited from proto-cuneiform were logograms representing the core morphemes of words – the core nouns and verbs apart from their affixes. Having a sign for simple, unaffixed words like “ox” and “barley” and even abstract words like “life” and “give” is one thing. Being able to represent the arbitrary sounds of a foreigner’s name, or to indicate grammatical details such as verb conjugations and a system of 10 noun cases, was a significant step forward.

  The way the Sumerians went about expanding their writing system was to listen to the sounds of their words. Already in the protocuneiform period they had occasionally used rebus writing, using one logogram to represent another word
that sounded the same. Now they began to use signs to represent only a part of a word, to represent just a syllable, regardless of whether that syllable meant anything in itself. This syllable could be used in spelling out an affix, a personal name, or a foreign word.

  This use of syllabograms was a significant advance, not just in writing, but in linguistics. The logical individual unit of language is the core morpheme, like sheep, house, or go. Pronouncing incomplete pieces of words by themselves is not natural. Given a word like feeling, we might find ourselves with reasons to mention the core morpheme, feel, but we rarely find ourselves called upon to pronounce the suffix -ing by itself, and we are even less likely to think about the fact that when we speak, the final syllable of feeling is in fact -ling, including part of the core as well as the suffix. Yet the Sumerians realized that their words were made up of smaller, pronounced units – syllables. And so a sign like gi came to mean not only “reed,” and not only “render,” but also merely the syllable [gi], independent of what word it appeared in.

  Thus cuneiform became what we call a logosyllabary, a mixed system in which some signs were logograms and some were syllabograms. Many of the signs could be either one depending on context. The syllabograms were also pressed into service as phonetic complements: a word spelled with a logogram could take an additional syllabic sign to indicate which of various possible pronunciations of the logogram should be used. Thus ka, “mouth,” plus the phonetic complement me was the word that had to do with the mouth but ended in the syllable [me], in other words eme, “tongue.”