In early September I participated in a colloquium on esthetics in Porto Alegre, the southernmost large city in Brazil. On this occasion a young scholar from Rio named João Cezar de Castro Rocha introduced himself as one of the two collaborators on René Girard’s latest book, Les origines de la culture (Descleé de Brouwer, 2004), and pointed out to me a passage in which, in response to a question from his interlocutors, Girard briefly discusses my work–the only time, I believe, that he has ever done so. Perhaps predictably, his judgment is dismissive:

« Comment un simple geste, tout « ostensible » qu’il soit, peut-il empêcher les doubles symétriques de s’entretuer ? Comme si, à ce moment-là, la violence n’existait pas ! C’est encore une autre façon de nier la violence. Je crois qu’il s’agit d’une manœuvre de plus pour nier la primauté du religieux dans la culture humaine. Pourquoi pas un « contrat social » au sens traditionnel ? Pour avoir le langage, il faut d’abord une forme embryonnaire de culture, une sorte de protection culturelle contre la violence. À mon avis la solution fondamentale du problème de la violence est forcément religieuse, et, selon moi, ne peut résulter que du mécanisme victimaire, du rassemblement mimétique spontané contre une victime arbitraire. Les rites sacrificiels et l’immolation de victimes viennent en premier. C’est là qu’on doit placer l’origine de tout le reste, à commencer par le langage. . .« . . . en rétrécissant le rôle du mécanisme victimaire, [Gans] appauvrit la théorie mimétique et il laisse tomber des pans entiers de l’anthropologie archaïque, ce qui nous prive de toute une masse de correspondances significatives et savoureuses, parfaitement éclairées par cette théorie. Rien ne me paraît justifier son attitude sinon cette allergie moderne au religieux dont parle Cesareo Bandera. » (179-81)

“How can a simple gesture, as “ostensible” [sic] as it may be, prevent the symmetrical doubles from killing each other? As if, at that moment, violence no longer existed! This is yet another way of denying violence. I believe it is just one more maneuver to deny the primacy of the religious in human culture. Why not a “social contract” in the traditional sense of the term? To have language, there must first be an embryonic form of culture, a sort of cultural protection against violence. In my opinion, the fundamental solution to the problem of violence is necessarily religious, and, to my mind, can only proceed from the victimary mechanism, the spontaneous mimetic union against an arbitrary victim. Sacrificial rites and the immolation of victims come first. It is there that we should situate the origin of all the rest, beginning with language. . .

“. . . in limiting the role of the emissary victim, [Gans] impoverishes mimetic theory and lets fall away whole areas of archaic anthropology, depriving us of a whole mass of significant and revealing [lit: tasty] correspondences that are fully illuminated by that theory. Nothing seems to me to justify his attitude other than that modern allergy to the religious described by Cesareo Bandera.”

I have, of course, responded to these criticisms before; but now that this text is in print, I think a more extended response is in order. Nothing I say here should be taken as detracting from my lifelong respect and affection for René Girard.


Clearly Girard is not interested in following the order of events as I describe them; nor does he seem to be aware that I have always made clear that, at the origin, language cannot be separated from “the religious.” One could in fact situate the moment of representation easily enough at the center of the Girardian scenario. The victim has to be designated, and this designation would afford an alternative way of describing the originary scene of language; it would suffice that a moment of designation/ representation be interjected into the “mechanism” for it to be humanized, without sacrificing any of the violence that Girard fears I am neglecting. I might find the description of the originary scene in terms of emissary victimage less than fully parsimonious, but I would recognize the homology.

But Girard would not accept this homology. He wants the “mechanism” to be purely mechanical, uncorrupted by the Sartreannéant that separates the human mind from its objects–not that Sartre ever recognized the interdependence of this “free” space of consciousness with language. Thus the religious in Girard’s view can only succeed in generating the world of culture if it operates unconsciously. Girard never tells us exactly at what point human beings began using signs to represent what they were doing so that it could be reproduced as ritual rather than simply repeated–if not at the beginning, when? One wonders why the idea that language and religion are coeval, acceptable even to a respectable social scientist such as the late Roy Rappaport, is incompatible with mimetic theory, or with any plausible theory of hominization. For Girard, I’m just another social contract theorist–as though, incidentally, the “scenic imagination” of the contract scene, which is the direct ancestor of Freud’s father-murder, and thence, of Girard’s own scene of emissary victimage, were without interest. I reply that Girard is just another gradualist, never wanting to specify exactly when or how language and symbolic thought emerge from animal behavior. His humanizing “mechanism” is not just violent, it is scenic; but woe unto him who attempts to describe its operation as a scene, even a minimally ordered one, with a memorable climactic moment of hesitation/consciousness/ representation that the sign, the “word” that is the name-of-God, commemorates. In order to accept Girard’s dismissal of my version of generative anthropology, it is necessary to reject the central role in hominization played by the “radical shift in communication strategy” (Deacon) that language represents.

I will only mention in passing the fact, rather embarrassing to the maximalist theory of human sacrifice, that the earliest known representations of what are presumably sacred objects are almost exclusively of animals and sometimes of women, with men playing at best an auxiliary (always hunting) role. This suggests that hominization can be precipitated by mimetic rivalry over meat that does not reach the paroxystic intensity at which all appetitive interest is abandoned in a purely aggressive rage.

I have also made clear that the moment of representation is not the end of the scene, but is followed by the sparagmos of the central victim whose remains, rather than being simply destroyed, are shared in a proto-sacrificial meal. This evacuates Girard’s objection that I deny the scapegoat and the violence done to him. The scapegoat mechanism as Girard describes it is a specific mode–not necessarily the most primitive–of the more general phenomenon of the scene, in which the violence of the group is concentrated on a central figure. Nor is Girard’s postulation of a mimetic crisis leading to a total breakdown of order in the protohuman community a parsimonious one. A full-fledged scapegoating operation in which the energies of a “war of all against all” come to be concentrated on one or a few emissary victims presupposes not a proto-human society that can no longer control its mimetic rivalries with the pecking-order hierarchy in place and must consequently evolve a more powerful means of maintaining order to survive, but a society governed by a human order that has now broken down. The degree of lability of mimetic aggression in Girard’s model is conceivable only in a group that has already crossed the “Rubicon” into humanity; this lability is, to be sure, the result of greater mimetic capacity, but this capacity can be actualized only when embodied in a system of symbolic signs, the “arbitrariness” of which is the cultural counterpart of the arbitrary choice of the victim.

The scene is the minimal component of the distinctively human mode of mediated interaction that we call representation or metaphorically, “culture.” The human imagination is not a tabula rasa but an internalization of the communal scene whose content, even when it emerges from the secret recesses of one’s memory, is mediated by communally understandable signs. All significant human experience is mediated by this communal scene, which has its minimal form in language, its maximal form in ritual.

Girard sees violence as itself productive of culture: la violence est le sacré. But blind repetition of blind aggression cannot lead to insight; we can only speak of true violence, that is, of human violence, within a preexistent framework of human culture. The scapegoat mechanism can only be said to explain hominization to the extent that the victim is or becomes an object of representation. Mere repetition of the “mechanism” will not lead to symbolic language. Girard’s idea is that language takes its origin from behavior around the already-immolated victim. But sign-creation is not an act of violence but an act of representation that preserves its object from appropriative and aggressive violence until it can become the referent of a sign. Nor can the moment following the “lynching” be that of the first human scene; after the participants tear the victim to pieces, they have no longer any reason to stay together. It was mimetic desire that created their unanimity; now that the desire is satisfied, they have been relieved of what bound them together. The only way in which the human scene of representation can be born of collective mimetic desire is if the desire remains unabated until and during the emission of the sign. Once this sign of individual renunciation has been exchanged, the violence can be unleashed, because it can no longer threaten an order that includes it. The violence of the sparagmos subsequently exercised on the central object must take place within the confines of nascent human culture rather than outside it. If the “emissary mechanism” functions prior to the representation of the victim by a sign, it can be recalled by the members of the group only in prehuman fashion, as a mechanism rather than as an event, and the victim remembered only in a generic sense as an element of the mechanism. In contrast, a divinity is a specific being with a name. The climax of my hypothetical scene is the naming of the central object/victim by the emission of a sign; this quintessentially human act has no counterpart in Girard’s system.


It might appear that a debate over originary scenarios that cannot possibly be empirically tested is a nugatory at best. But the differences I have raised here have consequences for the future of what has come to be called mimetic theory.

The most serious of my differences with René is in our approach to modernity. René sees Christianity as the ultimate anthropological revelation, whereas in my view the evolution of the social order and of the technologies that emerge within it are accompanied by progress in our understanding of mimetic desire and in the political and social forms that both reflect and seek to control it. However powerful the Christian revelation and its influence on history, it is not “the end of history” and should not be treated as such. Nor can Girard’s apocalyptic view of modernity, which, having thrown off the sacrificial defenses of the pagan era, no longer has the means to protect itself against the anarchic force of mimetic desire, be articulated into a genuine theory of the modern. It is surely true that the modern world must find a way of assimilating the Gospels’ anthropological insight that the inherent form of human relations is that of moral reciprocity, but acceptance of this touchstone does not suffice to provide a model of social organization. It follows from the scenic nature of the human that moral reciprocity is not an immediate relation but is necessarily mediated through forms of representation: language, ritual, art, economic exchange. It is the institutions that have replaced sacrifice, most notably those of liberal democracy: the market and representative government, that are the legitimate heirs of the Christian revelation. The historic success of this “ultimate” political system is predicated on its capacity for deferring the resentment it inevitably generates. Resentment is part of the human condition, and neither individuals nor nations can renounce it once and for all, even in the face of the threat of annihilation. But the human as defined by the scene of representation, with its tension between desire and renunciation, fear and security, has a fighting chance of surviving through the deferral of violence without the necessity of anex machina universal conversion from war to peace.


Originary thinking as I conceive it is a new way of thinking that offers a concrete scenic framework for human institutions that have heretofore been studied by means of an undisciplined mix of empirical data and a priori models abstracted of their interactivity by metaphysics. René Girard’s mimetic anthropology is an indispensable source of this new mode of thinking, but not its full realization; Girard’s representation of the scene of human origin is missing the most crucial element of all: the origin of representation itself. This, and not our respective attitudes toward violence, is the essential difference between our anthropologies.