I – Necessity is the mother of invention

Derrida conceived la différance as the opening-up of a paradigm. As opposed to the reflexive association of stimulus and response, as in our example of an animal’s avoidance reaction that follows the association of a behavior with an electric shock, the user of language is obliged to choose the proper word—in Derrida’s conception, from a set of similar words: colors, animals, numbers, etc. That is, the association of a signifier with a signified involves deferral of choice among the different possibilities.

But the significance of this arch-Derridean quasi-concept, which I have found so precious for GA, can be described in much more anthropologically fundamental terms. What needs to be deferred/differentiated, in a word, is violence.

The idea of a precipitating crisis, which Freud had found intuitively obvious as the source of the new behavior that would define humanity—in his case, the rivalry/inhibition connected with the Oedipus complex—has altogether disappeared from the increasingly voluminous literature on the origin of language, as though, whether or not language itself be considered a gradual development of animal communication, it cannot conceivably have been provoked into being by a newly critical situation. The passage between the Alpha-Beta hierarchy typical of the great apes and the enforced egalitarianism of hunter-gatherers cannot have escaped everyone’s attention but my own, yet the only writer I know of who evokes it, Richard Wrangham in The Goodness Paradox (see Chronicle 614, Coda) does so obliquely, portraying it as rather an effect of the birth of language than its cause.

The key point is that what is new about human language is not its function as a device but its operation in a scenic configuration. No species can survive without the means to maintain whatever collective order it requires, and the originary hypothesis proposes that the new mode of communication we call “language” reflects the need to respond to a new form of challenge to this order, one that requires the institution of a transcendental mode of interaction among ourselves and with our environment, and which manifests itself as the transmission of signs on a scene of representation.

In the terms of the originary hypothesis, the signs of language, unlike the indexical signals we share with the animals, are not only actuated by different parts of the brain, but above all originate as deviations from practical actions, directed instead at sacred/interdicted objects. The idea of an imperceptible glissement from signals to language is incompatible with the radical difference between human and animal communication. And from a heuristic standpoint, the more we focus on the specificity of language in its originary segregation from the activities of “profane” life, the more we discover in this activity that will help us to understand the specificity of the human/cultural in general.

The originary hypothesis liberates us from the gradualist paradigms of evolutionists and neo-evolutionists. Not that they are wrong; but they are seeking their solution in the wrong category of phenomena. No doubt the most interesting features of human language are the complexities of the mature languages that we know. But it is a fallacy, one so transparent to the Darwinians that they never even notice it, that human language must have emerged via some intuitively “simple” form of linguistic communication as we know it that would have evolved, after many thousands of generations, into the complex systems we now possess. Instead of thinking that what was new was a new cognitive ability, let alone a “big bang” language gene, they should have sought to model rather the critical situation which provoked a new behavior.

In a word, as are most innovations, language must have been pulled, not pushed. Necessity is the mother of invention. Not pooh-pooh or yo-heave-ho or bow-wow or ma-ma, but… fear of death.


René Girard’s most original contribution to originary anthropology is not his focus on the phenomenon of mimesis, but on that of the potential violence threatened, and often occasioned, by mimetic rivalry. Traditional Darwinians have no problem with mimesis as a concept; what is missing from their etiologies of the human is the need to control the enhanced threat of internecine violence that mimesis poses as its positive features lead to higher levels of (mimetic) intelligence.

I recently noticed in rereading La violence et le sacré that Girard uses the expression la première fois (the first time) some half-dozen times, each time to refer to the real collective murder that he believed was the basis for each killing/metamorphosis in myth. The first time there was a real scapegoat lynching, an emissary murder. Whatever you think of this hypothesis, what it shows is that, intuitively, Girard felt the need for what he calls in the conclusion an événement fondateur (founding event), even if he could not bring himself to conceive this, as GA does, in terms of the deferral of worldly into symbolic action.

Girard had no theory of language as such, but his intuition told him that the sign was born in a moment of crisis and was originarily reserved for the sacred. That he defined this crisis differently from GA is secondary to the fact that he understood that signification was something entirely new that could only be born in crisis. When was the last time this simple idea was expressed by one of the hundreds of researchers who study the evolution of language? Not that it is counter-intuitive. Even the idiotic notion that language began with the appearance of a “language gene” at least recognizes what Max Müller called the Rubicon-crossing nature of its appearance.

II – Humanism and interiority

In my recent reflections on deferral/différance, I have emphasized the continuity of this notion with the phenomenological conception of intentionality, and more particularly with Sartre’s néant, which although unconnected to language can easily be assimilated to the difference/separation between the thinking/speaking subject and the referent that he addresses, not through “instinct” or conditioned reflex, but beginning with the abortion/deferral of action, opening the point in space-time in which the “natural” reaction to an object is transformed into a sign. The recent trend in the linguistics community has gone precisely in the opposite direction. But contrary to what my readers might think, I see this as a positive development that brings a new level of clarification to the specificity of the “new” or “fourth” way of thinking that is GA.

This has not always been the case. In Chomsky’s famous 1959 review of Skinner’s 1957 Verbal Behavior, the entire linguistics community stood behind Chomsky’s rejection of a behaviorist stimulus-response model of dialogue. Intuitively, in the everyday “humanistic” world, it is obvious that, to use one of Skinner’s examples, the command (or mand) “Tell me what this is” said when pointing to a violin, as a means of provoking the response “violin,” suggests a stimulus-response relation implicitly short-circuiting the conscious decision behind the reply. If I think that someone considers “violin” the response to a “stimulus,” I may say something else just to show that I can, whereas an animal having learned a conditioned reflex cannot make such a decision. Such arguments seemed at the time to settle the question. Which is to say that, abstracting from the idiosyncratic manner in which Chomsky considered the mind to operate in the case of language, in his response to Skinner he was defending in his fashion the classical humanist conception of interiority, or what we call the internal scene of representation.

But contemporary constructivist linguistics considers Chomsky’s cure worse than the disease. For one thing, Chomsky’s notion of the linguistic pour-soi is far from the néant within Derrida’s différance. For Chomsky, language is the affair of a computer-like “module” independent from the rest of the nervous system, and only lightly touched by ontogeny. The Language Acquisition Device (LAD) with which the child is presumed to be born is a genetically determined machine that awaits only the setting of a few parameters (“switches”) to become fully operational. The “poverty of the stimulus” to which the language-learning child has presumably been exposed is offered as proof of the language module’s nearly complete independence of epigenesis. In other words, in Chomsky’s hands, the interiority whose history goes from Descartes to Husserl to Sartre to Derrida is reduced to something almost as mechanical as the stimulus-response link he rejects in Skinner.

I first encountered a thorough exposition of the revised attitude toward Skinner in Chapter 8 (entitled “Revisit: Skinner, Chomsky, and construction grammar”) of Julie Tetel Andresen’s Language and Evolution (Cambridge, 2014: 229-259), one of the more interesting and original books on language origin/evolution I have encountered. This work has the considerable ambition of being, in the words of its introductory blurb, “the first rethinking of an introductory approach to linguistics since Leonard Bloomfield’s 1933 Language.” Although, unlike Bloomfield’s classic, it is more the exposition of a program than a possible textbook for Linguistics 101, it develops at length a constructivist perspective based on “developmental systems theory” that defines its subject, the essence of human communication, not as the Saussurian langue or even as language, but as the activity of languaging. Thus its fundamental components are not words or even sentences but speech-acts, in the tradition of Austin and Searle, to the study of which Andresen sees Skinner’s explorations of “verbal behavior” as, surprisingly, closely related.

In this perspective, rather than attempting to improve on Chomsky’s idea of what our internal language machine must look like, we put the whole “mental” world in brackets and study “languaging” as an activity with no need to posit “hidden parameters” at all. Andresen does not claim—and she assures us in her own terms that this is true of Skinner as well—that the space-time we refer to as différance does not exist, merely that it is unnecessary to the explanation of an interactive system. If to the question “What is the capital of France?” instead of “Paris” someone answers “I don’t know,” or “London?” or “I’m not going to tell you,” these are all speech-acts that can each be understood as producing its own specific illocutionary and perlocutionary effects, without requiring us to concern ourselves with what went through the speaker’s (and hearer’s) brains before (and after) it was selected.

Not being myself a professional scholar of either language or languaging, I cannot vouch for the superiority of this or of any other elements of constructivist methodology. But by all indications, Andresen’s work is an eloquent defense of the path that linguistics is taking today. Politics aside, I would say that anything that can liberate linguistics from decades of Chomskian dominance is a good thing. And I might add that orthodox Chomskians’ dogmatic reduction in practice of all “language” to more-or-less-English is certainly an ironic commentary on Chomsky’s radically “anti-imperialist” political stance. Just ask Anna Wierzbicka (see in particular Chronicle 597).


By making far clearer than before the separation between humanistic and empirical science, this return to behaviorist-like positivism is of great benefit to GA. The element of différance is precisely the point at which they diverge, which is to say, the point at which the uniquely human begins. The emergence of this separation in the case of language coincides, by no means coincidentally, with the historical moment in which the evolution of computers has begun to show us that anything humans can do with language can be simulated by machine, and in simple cases at least, improved upon. This in no way dissolves the difference/différance between the two; on the contrary. But it demonstrates the incommensurability of these two “ways of thinking”—and as a corollary, the paradoxical impossibility that the mechanical universe of stimulus-response could have produced by itself machines that simulate the human without having first produced the transcendence embodied in the human itself. We may if we like consider this as a demonstration of the “incompleteness” of the mechanical universe, analogous to Gödel’s proof of the incompleteness of arithmetic.

In short, the existence of machines that can truly simulate thinking makes it all the more urgent for us to preserve the intuition of transcendence outside the sphere of empirical science, which would seem to have increasingly less use for it. Just as philosophy has allowed for reflection on the human condition by “bracketing” the anthropological source of the transcendental intuition that permits this reflection in the first place, so empirical science, in studying our specifically human mode of communication, now finds itself obliged to function in the same manner, the henceforth unsatisfactory alternative being the interposition of a black box of “universal grammar” between an abstract “meaning” and a physical output. I am happy to see the end of this procedure, which results in a much more parsimonious model of how “languaging” operates in the brain—one that can hopefully be filled out by future research into the neuronal processes involved.

Neuroscience is not our domain. Our task is rather to flesh out the newly reinforced humanism/empiricism split with the products of our reflection. For although this new way of thinking is one made possible by the originary hypothesis, it is by no means exhausted by the mere fact of its enunciation.