LANGUAGE, LINGUISTICS, AND LITERACY.

Language and writing are not the same (although people now confuse them), and the former is far older than the latter. Human spoken language developed late in the Paleolithic, probably 100,000–120,000 years ago, whereas true writing—the representation of language, element by element, in a permanent medium—was not invented until about 5,200 years ago.

Language

Our current estimate of when language itself evolved is based on skeletal evidence for a swift evolutionary push toward the modern human vocal tract during the Upper Paleolithic. These changes—away from the throat, tongue, and mouth configuration typical of other primates—made rapid spoken speech possible while making it easier to die of obstructed air passages (choking, etc.). Such difficulties for the individual must have been well offset for the species by the survival benefits of rapid, voluntary, and abstract communication, one of humankind's most powerful "ideas." Indeed, language's most important design characteristics are that it is:

1. voluntary (unlike the call systems of other primates, speech is initiated from the voluntary part of the motor cortex);
2. arbitrary (we can talk even about things that do not exist, allowing us to solve hypothetical problems and form abstractions; language is not tied to immediate reality);
3. productive (we can use already known vocabulary and syntactic patterns to generate an infinite supply of new messages, not to mention new vocabulary and patterns);
4. largely linear (messages unfold rapidly through time, a choice that places major constraints on the details of language design); and
5. two-way (individuals can both produce and understand messages).

As a result of being arbitrary, languages must be learned. Only the propensity to learn language is instinctive, not the individual languages themselves. Hence, babies must have a long period of apprenticeship to learn to communicate linguistically, a fact equally true of spoken tongues and visual languages like American Sign Language (ASL). (ASL has all the structural features and design complexities of a spoken language, only in a different—and equally evanescent—medium.) Research shows that children deprived of early exposure to linguistic communication (which happens sometimes with deaf babies) do not develop the neurological structures in the brain necessary for handling the complexities of language later.

Because languages are arbitrary, they change constantly, adapting easily to the changing world of things to talk about and social messages to send. As a result of this constant change, speakers of a single language who move away from each other will change their systems differently, a process that may continue until we perceive them as speaking different dialects or even (after many centuries) different languages. This process is called language divergence and leads to what linguists call language families.

The members of a language family are simply changed later forms of what was once one language. For example, the speakers of Latin fanned out across the Roman Empire 2,000 years ago, but after the breakup of the empire, when travel became dangerously difficult, the local versions of Latin began to diverge, producing the so-called Romance language family, consisting notably of Italian, French, Spanish, Portuguese, and Romanian (along with several other dialects and languages, less well known because they have fewer speakers). But Latin itself was a changed later form of yet older languages, going back to something we call proto-Italic (for lack of knowing what its speakers called it), and even further back to one we call proto-Indo-European.

Linguistics

Modern linguistics—the scientific study of language in all its aspects—began in the eighteenth and nineteenth centuries with the investigation of just such historical questions. The 2,500 years of continuous written records of Latin and the Romance languages provided one useful "laboratory" in which language change could be observed. The brothers Grimm (Jacob Grimm [1785–1863] and Wilhelm Grimm [1786–1859]), in recording verbatim their famous collection of German folktales, accidentally created another linguistic data set in which they kept noticing interesting relations among the Germanic dialects of the crones who recounted the tales. Elsewhere, other scholars were noting the more distant relations among Greek, Latin, Sanskrit, and other languages we now recognize as Indo-European.

Eventually linguists realized that the relationships between languages within a family were in certain ways regular (if complex), especially in how their sounds changed. This fact allowed scholars to begin to reconstruct forms of the ancestor languages, even when those forms were not attested in writing. To facilitate this linguistic reconstruction, they needed to study the structural properties of each language, since, before they could compare thoroughly, they had to know in detail what they were comparing. In this way, the systematic study of the synchronic (as opposed to historical) features of languages was added to the field of linguistics around 1900.

The most basic components of a spoken language, all needing study, are the speech sounds themselves (phonetics ), the system of using certain sounds to tell words apart in a given language (traditionally called phonemics, although other approaches to sound contrast are sometimes used), the forms of words (morphology ), the structures by which words are combined into phrases and sentences (syntax ), and the meaning components (semantics ). In the mid-twentieth century, linguists began to explore beyond the sentence level to analyze the structure of discourse, poetry, formulaic speech, and other large units.

Nineteenth-century European scholars were not the first people, of course, to inquire about language history or structure. (For example, around the fifth or fourth century b.c.e., the Indian scholar Panini compiled a thorough grammar of Sanskrit, including what amounts to an excellent structural analysis of its sounds.) But it is their work on which modern linguistic science is principally based. They, in turn, were continually spurred on by the masses of inscriptions that archaeologists were turning up, many in unknown and unreadable scripts.

Writing

The most important scripts deciphered in the nineteenth century were Mesopotamian cuneiform and Egyptian hieroglyphics. Their importance came from the fact that they belonged to the two oldest traditions of true writing: the representation of connected linguistic structures rather than only single concepts or situations.

Jean François Champollion (1790–1832) deciphered Egyptian hieroglyphics in the 1820s. Using the Rosetta Stone discovered in 1799, he recognized that the hieroglyphic pictures represented the successive elements of a normal language (some phonological, some semantic) and not some sort of lost religious magic, since the Rosetta Stone clearly had the same message in Egyptian hieroglyphics, Demotic (everyday Egyptian script), and Greek. Champollion's progress was greatly speeded when he recognized that the language in question was the immediate ancestor of Coptic, the still-used religious language of a Christian sect in Egypt and, therefore, a version of a known language.

Cuneiform required more of a team effort, in part because "it" turned out to be more than one script, and because even the mainstream version had serviced several languages over nearly 3,000 years and had evolved considerably in form during that time. Cuneiform means "wedge-shaped" and refers to a whole family of scripts written by impressing wedge-shaped marks into clay with a stylus. The first breakthroughs occurred in the 1840s, and scholars gradually disentangled the various scripts and languages, from later to earlier (although some, like Elamite, remain undeciphered because we do not know enough about the language family represented). Most of the languages written in cuneiform belonged to the Semitic language family, although at least one (Hittite) was Indo-European and several belonged to still other families.

In fact, the very first language written in cuneiform was not Semitic, it was Sumerian, a language that we still cannot relate conclusively to any other known language or family. Sumerian cuneiform writing first appeared in the early cities of Mesopotamia (Iraq) late in the fourth millennium b.c.e.

Origin of Writing

Although many nations have claimed to have invented writing from scratch, only one culture in the Eastern Hemisphere, Mesopotamia, can show in its archaeological record the long series of steps needed to develop this significant invention. It is easy to invent a writing system once someone points out that language can be represented in a permanent form by making visible squiggles, each standing for a word or sound. But the notion that one can write at all is far more difficult to conceive the first time, truly one of the world's great ideas. Evidence shows that such an abstraction did not come easily. To trace its origin in the Near East, we must go back another 5,000 years to 8000 b.c.e., to early Neolithic sites in Syria and Iraq.

The Neolithic, or New Stone Age, differs from the Paleolithic, or Old Stone Age, in that people now had domestic plants and animals, permanent open-air settlements, and stone tools ground to a fine edge instead of merely chipped. At a number of Neolithic Near Eastern sites, along with evidence for early experiments in architecture and weaving, we find experiments with clay: hundreds of tiny balls, cones, disks, and other simple geometric shapes, which archaeologists at first took for children's game pieces. We now know they were used for keeping accounts of livestock, grain, and other possessions, each shape denoting a different commodity and each animal or unit of goods tallied by a separate piece of clay. Thus, five disks with crosses might correspond to five sheep, and three balls to three measures of grain.

The developmental history of this accounting system (which soon spread across the Near East and into southeast Europe), together with linguistic reconstructions indicating the lack of abstract number words (for example, six, seven, eight ) in any language prior to 3000 b.c.e., indicate that Stone Age peoples did not have an abstract notion of number yet. Lacking number words, they counted simply by matching the animals or jars in question to an equal number of concrete tokens. (Such systems continue to be used in a few societies.) With a simple Neolithic economy, this was sufficient for keeping track of things, although, apparently for security against swindlers, the Mesopotamians added the habit of sealing the little clay tokens of a transaction into a clay ball onto which each token had been impressed once. (This enabled everyone to see what the account was, but no one could snitch or add tokens.) But when, in fourth-millennium Mesopotamia, people began gathering in huge cities and needed to redistribute goods for labor in complex ways, they were forced to devise a more powerful and more flexible system.

The later history of writing systems shows that each new idea for restructuring writing to make it more efficient (such as moving from word writing to syllabic writing, or from syllabic to alphabetic) came from a new group learning the old way of doing things and saying, in effect, "There's got to be a better way." Just as each new generation of children restructures slightly the language data learned from its elders, so systems of writing and notation have been restructured by those with a fresh view. The same presumably happened here.

About 3100 b.c.e., in the Mesopotamian city of Uruk, we suddenly see a new "take" on the meaning of the impressions on the clay envelopes holding tokens. Before, the impression of a small cone meant a small measure of grain and that of a big cone meant a larger measure of grain (which happened to equal sixty of the little measures, just as a gallon happens to equal sixteen cups in our system). But now the small cone was reinterpreted as meaning "1" and the big cone as "60," and another measure in-between (similar to our quart) was reinterpreted as "10" (since it was equal in volume to ten of the smallest unit). With these impressed shapes reinterpreted as designating the number intended, however, the scribe now had to specify separately the object being counted. At that moment, the system developed into two types of signs: one kind for abstract numbers, the other for names of objects. The latter corresponded to the vocabulary words of language. Thus, if we want to speak of four rings, we don't say "ring ring ring ring" (or write "oooo," which is how the Neolithic scribes did it); we use only two words, one for the number (four ) and another for the object (rings ). To write this as "4 o" would then correspond unit-for-unit to the words we speak.

The only problem with this system was that words for abstract numbers did not exist. Because we see such words hastily cobbled into being at just this time, we know that they must have been forced into existence by the reinterpretation of the impressed tokens. Now humans had two new ideas to expand upon: the notion of abstract numbers (numbers independent of any particular thing being counted), and the notion that the impressed patterns could correspond to the successive words in a spoken phrase or sentence (four rings ).

Drawing little pictures to denote words like ring or tree (pictograms) was fine, but what do you do with grammatical words and other unpicturables like the and or ? The easiest way was to use sound-alikes (homonyms or near homonyms), as in the modern game of charades. Thus, if one did this in English, orcould be denoted by a picture of an oar, and so on. (This game, called rebus writing, was easier to play in Sumerian than in English because most Sumerian words were only one syllable long.) To tell which word with that sound was meant, scribes began adding extra signs to give clues to the semantic realm. Thus the "oar" pictogram might have a schematic tree beside it when it meant "oar," warning the reader to pick the word sounding that way that denoted something wooden (eliminating or and ore ). In this way the script gradually became adequate for representing anything in the language, although it was centuries before grammatical words were consistently written along with the "content" (lexical) words. As this writing system became more adequate, it also became more unwieldy, requiring the student to learn hundreds of signs.

Spread of Writing and Literacy

Other early scripts were equally large and unwieldy, and their structures give some clues to their origins. The Egyptians appear to have gotten the idea that one could write with pictograms and rebuses from the Mesopotamians around 3100 b.c.e., almost as soon as the Mesopotamians invented true writing, and then made up their script themselves, basing some signs on traditional proprietary marks and drawing pictures for others. So far we lack evidence showing the Egyptians working up step by step to the abstract notion of writing, and we know they borrowed a number of other cultural ideas from Mesopotamia at exactly this period, so this hypothesis of borrowing is the best one we have to date. The Egyptian push to write, unlike that in Mesopotamia, seems to have been chiefly religious rather than economic, especially the desire to preserve personal names for an eternal afterlife. The very earliest Egyptian inscriptions that have come down to us are names of kings and pharaohs.

Egyptian writing maintained its predominantly religious character to the very end, even though faster cursive forms of the script were eventually devised for writing mundane secular documents. But the Egyptian script never became simpler, retaining its many hundreds of signs until it fell out of use in Roman times.

The Chinese script, which began to blossom during the first great dynasty of China, the Shang (1500–1100 b.c.e.), appears so much later than cuneiform or Egyptian hieroglyphics, in a period when Chinese already contained loan words from Iranian and possibly other Indo-European languages from the West, that probably the script was jump-started like Egyptian, by diffusion of at least the idea that one could write. The signs themselves seem to have been created indigenously, but the writing system closely resembles the Near Eastern scripts in being composed of several hundred pictograms, rebuses, and semantic determinatives. On the other hand, archaic Chinese (like Sumerian) contained mostly words of one syllable, so the Chinese were in a convenient position to (re)invent for themselves the notion of rebuses.

When a script contains hundreds of signs and considerable ambiguity, an individual must devote enormous time and energy to becoming literate. In early societies this had two major consequences: (1) scribes/writers were specialists (fully supported by the rest of society with food, etc.), and (2) there were very few of them. In other words, literacy was not something that many could attain, and, with so few writers, only the most important things got written down. For the Mesopotamians the important topic initially was economic transactions, for the Egyptians, religious matters pertaining to eternity, and, for the Chinese, oracles about future events. Gradually other matters—both religious and secular—came to be recorded, but the cost of educating scribes was always a bottleneck, while the past investment by each of these cultures in their systems made throwing them out unthinkable.

Such was not the case for neighboring cultures. Over and over neighbors restructured these systems as they borrowed them, usually recasting them in such a way as to make the writing more efficient. Living with the Sumerians in Mesopotamia were speakers of various Semitic dialects, such as Akkadian. When Sargon of Akkad took over Mesopotamia soon after 2400 b.c.e., his scribes took the Sumerian pictograms at their phonetic value alone (no meaning whatsoever implied) in order to spell out their native Akkadian syllable by syllable (syllabic signs). Conceptually, this was a big step forward because there are so few sounds in a language compared to the huge number of words. Unfortunately, the scribes only went halfway, retaining many Sumerian pictograms as a shorthand for Akkadian words, since a Sumerian word required only one sign, in general, whereas Akkadian words, which typically contained two to five syllables, required many more. So, although the Semitic scribes had made a major conceptual breakthrough (one that the Chinese never made), it remained for others to benefit from this simplification.

The Hittites, an Indo-European group in Anatolia (modern Turkey), attempted to simplify the Sumerian system. When they borrowed cuneiform they dropped most of the duplicate ways of writing a given syllable, whittling down the syllabary to just over one hundred basic signs. But these scribes, too, found it handy to keep the Sumerian pictograms as shorthand for many of their vocabulary words, which on average were even longer than Akkadian words. (If you had to write out absolutely everything by hand, with no word processor, photocopier, or even typewriter, you, too, would look for faster ways to write, even if the system ended up more complex.)

The people of the Aegean, however, when they borrowed the idea that one could write, eliminated everything except the notion of signs for syllables, creating their own sign-shapes from scratch. They also simplified the notion of "syllable" from the four sorts used in cuneiform (V, CV, VC, CVC, where V vowel, C consonant) to two (V, CV), which meant they could manage with a mere eighty or so signs. (Well, almost: for economic accounts, they also used a roster of single signs—pictograms or ideograms—for each different commodity, but these signs were not mixed together with syllabic signs to form a sentence.) This sort of system took much less time to learn and made it possible, at least theoretically, for more people to become literate in their spare time. The people of Minoan Crete, who used this system for a short time in the mid-second millennium b.c.e., left us elegant inscriptions on jewelry and stone, economic accounts on clay, and graffiti on walls and pottery, suggesting widespread literacy. Unfortunately, they wrote most of their texts on perishable materials that have not survived.

A similar reinterpretation, leading to an even more radical simplification, occurred in the Sinai peninsula, around 1900 b.c.e., where Semites from the north encountered Egyptians from the south at the local copper and turquoise mines. The Semites apparently saw the Egyptian writing system and learned something of its structure, but concluded, "There must be a simpler way." The Egyptian script, unlike cuneiform, did not specify the vowel that a syllabic sign was to convey, only the consonant(s). So when the Semites, like the Minoans, pared away all the duplicate ways of writing syllables, they found themselves with a mere twenty-some signs, specifying that many different consonants.

This new script, the so-called Sinitic consonantal syllabary, was clearly based on Egyptian sign-shapes at first, and it had a tremendous advantage: anyone could learn this tiny number of signs in a few hours. Spelling consisted simply of sounding out the words as one wrote, so spelling did not need to be learned as such. The drawback, of course, was the ambiguity for readers. The writer knew what spoken vowels were intended, but th rdr hd t gss thm, leaving much potential ambiguity, unless writers wrote only for their own use. This may have been the dominant use, however, because variants of this simple script quickly spread up the coast all the way to Syria. Along the way it generated what became the Hebrew, Phoenician, Ugaritic, Aramaic, and Arabic scripts (and more), all of which are/were consonantal syllabaries, and generated a much wider range of uses, implying wider literacy.

We know all these scripts are related because they all use(d) the same canonical order of signs for teaching the script, an order ancestral to our modern alphabetical order. One clear and early document of this sort is a tablet from a schoolroom in the city of Ugarit, on the coast of Syria, in which the signs have been modeled for the pupil in this standard order. Because Syria was heavily influenced by Mesopotamian culture, the letters are written with wedge-shapes on clay, but those shapes have nothing to do with the shapes of signs in mainstream cuneiform. Only the cheap clay medium was borrowed.

The final step in developing the most efficient possible writing system was to add separate signs for vowels, eliminating the guesswork without adding many more signs. Remarkably, this step was first taken around 1400 b.c.e. in Ugarit, six centuries before the Greeks produced their alphabet. The structure of Hamito-Semitic languages is such that vowels are not so important for determining the basic semantic contents of a message as they are in virtually all other languages. (It is surely no accident, therefore, that it was Semites who made the big step in simplification to the consonantal syllabary.) But Ugarit had two main linguistic populations: Western Semites and Hurrians. The Hurrian language belonged to yet another family, centered in the Caucasus to the north. The Hurrian merchants, searching for a better way, reused a sign here and added a sign there to obtain the vowel signs they needed to write their language readably.

As they did so, however, the system cracked into two entirely new types of signs—signs for single vowels and signs for single consonants—such that the latter were no longer signs for entire syllables (of a consonant plus an implied vowel). Each sound in the language now had a sign of its own, and each sign represented a single sound (the definition of a true alphabet). Now readers could read unambiguously what the writer wrote, and for only the cost of learning a couple of dozen symbols.

Unfortunately, this brilliant idea was almost immediately lost in the massive destructions curtailing the Bronze Age (1200 b.c.e.), and it remained for the Greeks to reinvent the same step around 800 b.c.e. as they pondered the ambiguous consonantal syllabary still used by Phoenician traders. The canonical order of the Sinitic script thus passed to the Greeks and is still visible despite the reuse of some unneeded consonantal signs as vowels, the addition of a few letters, and the retention of some unnecessary signs. The Greeks passed one version of their alphabet to the Etruscans, who passed it to the Romans—each time with minor modifications. With a few more changes in medieval times, it came down to us today as the "Roman" alphabet. (The Cyrillic alphabet used in Eastern Europe branched off in medieval times directly from the Greek alphabet.)

Alphabetic Literacy

The simplicity of the alphabet should have meant that now everyone would learn to read and write. Yet the evidence tells a somewhat different story: everyone could learn to write, but for centuries few did. Early Greek inscriptions are few, mostly labels and short dedications. The first person we know who used writing systematically to record everything he wanted to "make note of" was the great Athenian statesman Solon, living around 600 b.c.e., who apparently kept personal notes of what he learned abroad and what he thought about political matters. This is a use of writing not seen before, but soon to become commonplace. (The fragmentary nature of our evidence may, of course, have hidden some other earlier examples in other scripts—one might suspect the Third Dynasty Egyptian genius Imhotep (fl. c. 2650 b.c.e.) of similar practices—but clearly such a use was not common elsewhere.) When Solon had something he thought important for his countrymen to remember, he habitually put it into verse so they could memorize it easily, a fact indicating that most of the population was not literate.

By 500 b.c.e., however, citizens of Athens (that is, those inhabitants who were free and male) were expected to know how to read the public notices and write their own ballots, although anecdotes show that not all could. And throughout the fifth century—the Golden Age of Athens—literature was still primarily oral, still experienced communally in the theater, as the great dramas of Aeschylus (c. 525–456 b.c.e.), Sophocles (495–406 b.c.e.), Euripides (484?–406 b.c.e.), and Aristophanes (c. 448–385 b.c.e.) attest. Written copies of the plays were, however, increasingly available. Our first reference to a new form of literacy, reading to oneself for one's own idle pleasure, occurs in Aristophanes's The Frogs, written in 405 b.c.e., in which a not very bright Athenian citizen says he was sitting around on shipboard between practice maneuvers reading one of Euripides's tragedies to himself. In fact, at this time both Aristophanes and Socrates (469–399 b.c.e.) voiced concern over the rise of mass literacy, worrying that a flood of written information would cause people to forget the important things of life.

Then the flood hit: from then on great drama on the stage was replaced in Greek society by great books to be read—the philosophical works of Plato (427–347 b.c.e.) and his successors (which start, interestingly, as "oral" dialogues in form, although too deep to be fully appreciated in one hearing). And there were new types of lesser works, such as novels to be read purely for fun. A major watershed had been crossed.

The new literacy entailed new problems. How could so many literate people obtain the works they wanted to read? Copying by hand was very slow, so the idea of libraries, where numerous readers could share copies, was conceived. The famous library at Alexandria (destroyed in the mid-seventh century c.e.), although the largest, was only one of many. But with increasing amounts to read, people wanted to be able to read faster, so in Roman times letter shapes became simpler and more regular, allowing the eye to take in and distinguish the letters more rapidly.

In the West, most of this progress was lost with the destruction of the Roman Empire around 400 b.c.e. by illiterate hordes. Literacy retreated to the monasteries, where the storing, copying, and reading of books became sacred tasks. Letter shapes ceased to be clean and simple, since now the criteria were different. The desire to shorten the task of copying (rather than speed the reading) led to complex ligatures and other methods of abbreviation, while seriphs, curlicues, and other scriptal flourishes embellished the page even as they impeded easy recognition.

Books had been printed in China since at least 868 b.c.e. (and probably earlier), using carved wooden blocks, but Johannes Gutenberg's (c. 1397–1468) idea of printing with movable type, in the mid-1400s, changed all that. His first printed books are very difficult to read, but simplicity of letter design returned quickly as easier access to books fostered greater literacy, and greater literacy, together with the possibility of multiple copies, fostered more new texts to read. The Renaissance of learning, already started among the rich merchants of southern Europe, could now blossom fully.

But there was a new bottleneck: the high cost of a printing medium. Vellum and parchment were very expensive and paper a little-explored craft (although invented long ago by the ancient Egyptians and Chinese). With the invention of pulp paper late in the eighteenth century, however, books could be printed so cheaply that even lower-class workers could afford them, carrying in their pockets not just novels but printed manuals for the increasingly complex machines they had to operate. In a sense, it was manuals printed on pulp paper that fueled the Industrial Revolution. In the twentieth century, people were so inundated with printed materials that computers were invented to manage the glut, and those computers are inevitably changing patterns of literacy once again.

See also Communication of Ideas: Orality and the Advent of Writing ; Diffusion, Cultural ; Language and Linguistics ; Prehistory, Rise of ; Reading .

bibliography

Barber, E. J. W. Archaeological Decipherment: A Handbook. Princeton, N.J.: Princeton University Press, 1974. Discusses history, theory, and methods of archaeological decipherment.

——. The Mummies of Ürümchi. New York: Norton, 1999. Discusses literature on early Chinese loans from Iranians and other evidence for early trans-Eurasian contact.

Daniels, Peter, and William Bright. The World's Writing Systems. Oxford: Oxford University Press, 1996. An encyclopedia of all known writing systems.

Deacon, Terrence W. The Symbolic Species: The Co-Evolution of Language and the Brain. New York: Norton, 1997. Physical evidence for human development toward language, including evidence from brain studies for change to voluntary speech centers.

de Saussure, Ferdinand. Cours de linguistique générale. Paris: Payot (1964 reprint of 1915 original). First major treatise on descriptive linguistics; early analysis of language as arbitrary and linear.

Friedrich, Johannes. Extinct Languages. Translated by F. Gaynor. New York: Philosophical Library, 1957. Story of all the major decipherments (including Egyptian and cuneiform), written by one of the decipherers of Hittite.

Gelb, Ignace J. A Story of Writing. 2nd ed. Chicago: University of Chicago Press, 1963. Analytical history of the development of scripts.

Hockett, Charles. A Course in Modern Linguistics. New York: Macmillan, 1958. Chapter 64 contains an important discussion of arbitrary and productive properties of language.

Justus, Carol. "Indo-European Numerals Since Szemerenyi." In The Emergence of the Modern Language Sciences, Vol. 2: Methodological Perspectives and Applications, edited by S. Embleton, J.E. Joseph, H.-J. Niederehe, 131–152. Philadelphia and Amsterdam: John Benjamins, 1999. Discussion of lack of number words in early languages and the process of their introduction.

Lieberman, Philip. Uniquely Human: The Evolution of Speech, Thought, and Selfless Behavior. Cambridge, Mass.: Harvard University Press, 1991. Discusses evidence for the rapid push toward a vocal tract usable for speech.

Pedersen, Holger. The Discovery of Language: Linguistic Science in the Nineteenth Century. Translated by J. W. Spargo. Bloomington: Indiana University Press, 1959. Development of historical linguistics, with discussion of key discoveries of scripts and texts, and the basic theories generated about the language families of much of the world.

Schmandt-Besserat, Denise. Before Writing, Vol. I: From Counting to Cuneiform. Austin: University of Texas Press, 1992. Development of the Near Eastern token system into writing and numbers.

E. J. W. Barber