It is both easy and hard to believe that my first book of Generative Anthropology was written over twenty years ago, before the term itself was invented: The Origin of Language (TOOL), published by UC Press in 1981, was written mostly in 1978. Now I have returned to the origin of language, which has been on the back burner since that time. Much has been written on the subject since then, virtually none of which takes TOOL’s analyses, let alone its general perspective, into account. The reason is clear: writers on the origin of language are linguists, paleontologists, primatologists, sometimes psychologists and philosophers, and I am none of these things, nor did I make an effort in TOOL to speak a language that any of these groups could recognize as relevant to their concerns. This time I intend to deal explicitly and at length with competing arguments and relevant data. The passage of years makes all the more urgent my attempt to bridge the gap between the humanities and the social sciences by demonstrating the power of the paradoxical thinking characteristic of the former in the liminal domain of language origin so important to the latter.

The point of the originary hypothesis is not to present a particular scenario for the origin of language, but to propose the necessity of a public scene of origin, of an event that originates the function of human language to memorialize events. The origin of language is also the origin of religion, art, in a word, “culture,” and gradualist hypotheses or those that begin from private, local communications have even less plausibility in explaining the public scene of human culture than that of language.

There is a constitutive difference between humanistic and social science thinking, one that in “all other” cases corresponds logically to a difference in subject matter, but that comes to a head on the particular subject of the origin of language. Social science thinking is empirical; it must formulate concrete hypotheses that facts can falsify. Humanistic thinking, on the contrary, is concerned in the first place with the production of ethically meaningful discourse; this is true whether one is analyzing a poem or defining “the human” in originary terms. Such thinking resists immediate falsifiability by data because its function is to provide a reasonably stable basis for human interaction. Such, for example, are the texts of revealed religion, the chief locus of humanistic thinking throughout the centuries. An example of an intermediate form of thought is political theory, which purports to advise us on the best system of social organization. Such thinking is not falsifiable in the Popperian sense since it seeks to impose an order, not simply to observe one. Yet political theories are falsifiable by human behavior in the long term, as most people today consider “socialism” to have been.

To the extent that the science of human origin is concerned with concrete data, its hypotheses are drastically modified from year to year as new fossils and artifacts are discovered. This is as it should be in the empirical domain. But our general understanding of the human cannot simply wait until all the facts are in; the sparse prehistoric data amount to little weighed against all we have learned from human history. That the paradoxical structure of mimesis provides the basis of human culture is revealed throughout this history, and if our contemporaries have been loath to recognize it, their reluctance is unlikely to be greatly affected by the results of further paleontological research.


But although Generative Anthropology is primarily concerned with the ethical consequences of the originary hypothesis per se rather than with any version of its “minimal” realization, if we wish GA’s scenario of the origin of language to be taken as something other than an imaginary atemporal model like the “social contract” or John Rawls’s “original position,” we must be willing to situate it, however tentatively, in the context of the current understanding of human and linguistic prehistory.

This poses to originary thinking (at least) the following empirical questions:

  1. At what stage of the evolution of the genus Homo did language arise?
  2. What kind of communication system corresponded to the tool-making of Homo erectus, neanderthalensis, and early sapiens? Is it useful to postulate lower levels of human language prior to the origin of “modern” humanity?
  3. Since prehumans, on the model of contemporary apes and (especially) monkeys, must have had a considerable vocabulary of signals, how if at all was this material integrated into the sign-system of language proper?
  4. By what mechanism did the originary sign give birth to a full-fledged phonetic system characterized by “double articulation” of morphemes and phonemes?

On the basis of my research so far, here are some preliminary answers to these questions:

    1. At what stage of the evolution of the genus Homo did language arise?

Although I have always avoided pronouncing on this question in my books, in Chronicle No. 52 (July 27, 1996), I suggested that the origin coincided with the cultural “take-off” that began around 35-50 TYA (thousand years ago). I no longer consider this a minimal hypothesis. Studies of the human vocal tract (see Philip Lieberman, Eve Spoke; Norton, 1998) would seem to indicate that the earliest modern humans of 100-150 TYA were already “designed” for speech, with the lowered larynx and extended pharynx of today. The selection of these features is difficult to explain in the absence of true human language. More fundamentally, the minimal event of the origin of language should not be presented as the flowering of a complete human culture, but as its first moment, one that would have led to the selection for the vocal traits of modern humans. Such an event must therefore have occurred at some time before 150 or 200 TYA.

In my earlier thinking on this subject, I was influenced by the line of thought exemplified by William Noble & Iain Davidson in Human Evolution, Language and Mind (Cambridge, 1996), who note the lack before about 50 TYA of creativity and planning of the sort that would suggest the presence of language. But further reflection shows that this argument, based on the cognitive function of language, should be rejected as metaphysical. For GA, language is in the first place a means not of knowing the world but of deferring intrahuman violence. What in TOOL I called the “formal” system of language and the “institutional” system of ritual were no doubt for a long time barely distinguishable; the fossil record (ritual burials date back at least 100,000 years) corroborates GA’s hypothesis by suggesting that some degree of religious culture existed among even the earliest users of language. Early language conceived as possessing an exclusively ritual or “religious” function could well have coexisted with the highly stable technologies of the Paleolithic.

This historical minimalism only increases the cognitive impact of the originary hypothesis. All the literature on the origin of language begs the question of its origin by presenting it as advantageous for survival, despite the large brain, lengthy infancy, and inefficient vocal tract it requires. Aside from gradualist non-explanations of the emergence of language in the wake of human cognitive development, no one has formulated a plausible hypothesis of the birth of language, let alone for its coincidence with religion, which (in the guise of “symbolic activity”) is nonetheless, ironically enough, the sole archaeological criterion for the presence of language. Only the hypothesis that this origin resulted from a crisis of mimetic desire explains the generation of “transcendence” from “immanence” that uniquely defines the human.

In this perspective, the cultural “take-off” of 50 TYA—a highly relative term, cultural evolution henceforth being measured in thousands of years rather than tens of thousands—would best be explained by the emergence of what I called in TOOL “linguistic universality,” the taken-for-granted characteristic of mature language that “everything” can be spoken about, a development that would put language to work–replacing the old “instinctive” signal system–in everyday practical activities. For the moment it is difficult to speculate on the specific causes for this technological and artistic acceleration.

    1. What kind of communication system corresponded to the tool-making of Homo erectus, neanderthalensis, and early sapiens? Is it useful to postulate lower levels of human language prior to the origin of “modern” humanity?

Whatever kind of language the tool-making practiced by erectus requires, what is significant about it is rather that the weapons produced (little evidence of which is available) make plausible the originary moment of “mimetic crisis.” If the early emergence hypothesis presented above is correct, the Neanderthals were descendants of the originary scene and must have had some level of human language and culture. For example, they appear to have buried their dead, although this may have been in imitation of homo sapiens.

    1. Since prehumans, on the model of contemporary apes and (especially) monkeys, must have had a considerable vocabulary of signals, how if at all was this material integrated into the sign-system of language proper?

The words of language, like signals, are part of a communication system, but words are also objects of reflection linked to conrete, event-related memories. The human ostensive “lion” is superficially analogous to an ape’s emission of a lion-call, but neither the emitter nor the receiver of the call can refer in its memory to a sign, a signifier-cum-signified.

The principle of the word, as opposed to the signal, is the deferral of appetitive activity. The one tells us to follow our instinct to devour or to flee, the other, on the contrary, tells us not to act on our desire. Since these two functions are diametrically opposed, it is difficult to assume that signals were simply taken into the language. I think rather that the early stages of language coexisted with an extensive signal system; words, at that time, would have been rare and sacred.

    1. By what mechanism does the originary sign give birth to a full-fledged phonetic system characterized by “double articulation” of morphemes and phonemes?

This is the most speculative of the four questions. The articulation of sounds could, of course, have already been present in the signal system; language differs from signal by the signifier-signified distinction, not the articulation of the signifier. But this seems far-fetched; an exchange of signals as rapid and articulated as language would, following the “looks like a duck” criterion, simply be language.

My hypothesis is that the articulation of different sounds preceded rather than followed the multiplication of meanings. The original “vocabulary” would be a set of variations on the “name-of-God” that is its first element. From this incantatory glossolalia–typical of prayer even in the “higher” religions–different meanings could come to be associated with different variants. This provides an explanation of the system of “minimal pairs” that distinguish words today. The entire phonetic system need not spring up full-blown; it can be generated through varying the constitutive elements of the originary sign.


Over the next few months I will be seeking further means of fleshing out the originary hypothesis by reference to empirical data; my goal is to produce a book about the origin of language that neither the specialists in the social sciences nor the “theoreticians” in the humanities will be able to ignore. When my research has progressed a bit further, I intend to publish a bibliography of language origin on the Anthropoetics site to which all will be invited to contribute. I would also ask readers of these Chronicles to challenge the originary hypothesis and any auxiliary hypotheses I present here or elsewhere as strongly as they can, and to point out to me any materials that they believe might falsify it.