French Department U. C. L. A.
Los Angeles CA 90095-1550

La religion est une réaction défensive de la nature contre la représentation, par l’intelligence, de l’inévitabilité de la mort. (Henri Bergson, Les deux sources de la morale et de la religion [in Oeuvres, PUF 1970, p. 1086])

Qu’on s’imagine un nombre d’hommes dans les chaînes, et tous condamnés à la mort, dont les uns étant chaque jour égorgés à la vue des autres, ceux qui restent voient leur propre condition dans celle de leurs semblables, et, se regardant les uns et les autres avec douleur et sans espérance, attendent à leur tour. C’est l’image de la condition des hommes. (Pascal,Pensées, 199-434)

Why are there no atheists in the foxholes? Because they have all made Pascal’s bet. In times of crisis, God is present, not in some ineffable sense, but as the interlocutor of last resort. God is whoever is named by the name we call out in our panic. The customary rationale that we pray in the foxhole out of prudence, “to keep all our options open,” fails to explain why God is an option in the first place. This option exists only because humanity originates in the crisis that both God and language–language as the naming-of-God–were revealed and invented to resolve. God’s all-powerfulness is the inverse of our powerlessness in the event of human origin, the only crisis the outcome of which could be directly affected by the language of prayer.

The first word is the name of God because by the utterance of his name, the substitution of the sign for the appropriative gesture, we defer the threat of violent conflict over the object of our common desire. The source of this originary crisis is humanity’s susceptibility to mimesis, which even in circumstances of abundance creates scarcity by multiplying desirability. What we now call “humanity” was not long ago called “man,” including, yet not yet including, woman; our hypothesis explains the scandal of this exclusion. When we spoke of “man,” we named our species after its more violent gender; if we are now able to abandon this designation, it is because we now understand that the origin of the human is, not the sacralization, but the deferral of this violence, its différance through the sign of language.

We no longer contest the primacy of the vertical signifier- signified relation in the realm of signs. The dominant position of Saussurean verticality is accepted even within the Peircean world of “horizontal” semiotics. It is time to take the more radical step of hypothesizing the ultimate identity of linguistic verticality with the transcendental in general. Both the Platonic heaven of the Ideas and the more familiar paradise of the Judeo-Christian or Moslem afterlife are hypostases of the vertical relation of the sign to its referent. This is clear enough in the case of the Ideas, which are no more than reified words, but it is equally true of the eternal abode of souls. Although we inevitably situate the transcendent prior to the ethical or human realm, the ultima ratio of our gesture is itself ethical. We may call this the “paradox of transcendence”: it is the need to maintain order among human beings that generates the eternal verticality of the sacred sign.


The postulation of transcendence is the first originary hypothesis; the nascent human community discovers and invents the permanence of the sacred central being that brings peace by deferring its own appropriation. All theology is an elaboration of this originary postulate. God guarantees ethics by providing a transcendental source for ethical norms, which he promulgates as divine law. Plato’s Ideas perform the same function, but less explicitly. The Idea of the Good is not itself an ethical law, but it guarantees the possibility of such laws, indeed, suggests that explicit knowledge of the Ideas–accessible to the philosopher alone–carries with it the unique ability and responsibility to create and govern the good society.

When we pare from the myth of the cave its imaginary setting, what remains is the transcendent priority of the Ideas over the slaves who figure humankind. Both Plato and the Bible require a transcendental realm to guarantee their thematization of the ethical in the form of laws. Both metaphysics and Judaism are products of societies that have broken with what Eric Voegelin called the “compactness” of the archaic empires, Egypt and Babylon, where a supernatural guarantee of ethical norms is indivisibly linked to ritual practice.

The Mosaic rejection of concrete figuration of the divinity is not a simple triumph of enlightenment over superstition. The social payoff from ritual prolongs the deferring effect of the originary scene; worship guarantees worldly success because it averts internal conflict. But the Mosaic revelation shows that the power of communal worship is independent of the esthetic centrality of the figure. A freer, more internalized ethical system is guaranteed by a more cleanly vertical relation of transcendence, in which representation of the divinity is reduced to the minimal, originary form of the linguistic sign. Both metaphysics and Mosaic religion move in the direction of an ethic governed by moral principles rather than arbitrary ritual rules. This postulation of transcendence affirms the primordial ethical significance of the vertical relation of sign to referent, as opposed to the mystified concreteness of the “horizontal” worldly interactions characteristic of ritual.

The clear vertical separation of the transcendental realm from the real world is an early stage in the dialectical process of “secularization,” the reduction of the transcendent to the immanent, the vertical to the horizontal.(1) To specify the minimal ethical principles consonant with the originary hypothesis is to confront the ultimate crux of secularization: that of constructing a model that explains without recourse to a prior sacrality the originary emergence of the vertical from the horizontal.

Whatever the usefulness of speaking of such things as the genetic “code,” the transcendent verticality of the sign does not preexist human language. In order to arrive at the Saussurian model of language as a parallel world stretched above non- linguistic reality, language has to pass through a stage at which verticality is constituted within the horizontal relations of worldly appetite. This first stage is that of ostensive, in contrast to mature or declarative language.(2) An ostensive utterance points to what is already present, as in “Fire!” or “Man overboard!” These are typical ostensives in that they are used in moments of crisis; we may even say that their use constitutes crisis. (We do not cry “fire” every time we strike a match.) The most primitive form of language reflects the chaos in which it came into being.


Derrida points out in De la grammatologie that Saussurian linguistics shares the foundational assumptions of metaphysics. But metaphysics itself is ultimately a vision of language: one that views the declarative sentence or proposition as the fundamental linguistic form, reducing the ostensive and the imperative to “defective” variants of the declarative. There is a clear parallel between Platonic metaphysics and the “declarative” name of God (ehyeh asher ehyeh) in Exodus 6.(3) But, in contrast with Plato’s denial of ostensivity, the biblical scene adds an exoteric “ostensive” name (YHVH) after the esoteric declarative one. Despite the revelation of God’s sentence-name, which affirms his existence beyond the pre-thematic appeal of the ostensive, ostensivity remains the standard of religious practice. If metaphysics understands only declarative sentences–metaphysical prayer is inconceivable–religion never forgets the originarity of ostensive presence.

In the foxhole, even those who never pray elsewhere take Pascal’s odds. Pascal’s own foxhole is a dungeon where man’s violence to man is unmediated by mortar shells. This setting reflects a profound anthropological intuition; the scene of our wager is one of purely human violence. We beseech God, the creator and creation of man, to manifest himself at the moment of crisis to stay the hand of potentially violent appropriation. Our foxhole prayer is an acknowledgment of his presence, not a reasoned proposition. Its language is not declarative, but ostensive.

We pray to God in crisis; but God lacks the power to solve most crises, or, in the believer’s terms, he forgoes the use of it. The technologically mediated human violence of war, like the deadly violence of nature, is invulnerable to the power of God, who “lets his rain fall on the just as on the unjust.” The only efficacious prayer is that of the entire human community, the common and spontaneous use of the sign as the name of God to represent the common object of desire and abort the potentially fatal appropriative movement toward it. Only in this originary configuration could God really be said to “exist,” to function objectively in the world to prevent human violence.

But moments of crisis are no time to speculate on God’s limitations. In what crisis situation can the danger of human violence be altogether excluded? Even the progress of a disease is not uninfluenced by the solicitous or hostile desires of one’s fellows. Pascal’s bet is always a good one in a pinch. Since the violence whose deferral obliged us to invent ourselves in language is not natural but human, we cannot understand violence as anything other than human. We first humanize natural violence, then call upon God to prevent it. The positive effects of prayer in bolstering our courage and reinforcing group solidarity are distant reflections of the peace established by the linguistic sign as name-of-God in the originary crisis.

When I say that “we” are incapable of understanding violence as other than human, I refer not to our “etic” or theoretical understanding but to our “emic” intuition of violence from within, at the moment of crisis. The emic-etic distinction, which translates into that between ostensive and declarative language, allows us to explain how we may pray to God in times of war without believing in him in times of peace. The ostensive is the language of crisis; the declarative, that of stability. In the foxhole, we believe “ostensively”; we speak as if in the presence of God. In contrast, when we affirm a theological proposition, the question of whether God’s existence is implicit in our enunciation is, however crucial to our argument, not essential to our meaning. But whatever the theology of our declaratives, God subsists in our ostensives, where presence in language cannot be distinguished from real presence.


The ostensive operation of the sign is independent of our hypothetical reconstruction of the originary scene and our derivation from it of the idea of God. It is immaterial to our ostensive belief whether or not the originary hypothesis is more powerful than those of metaphysics or theology; to use language is itself to express this belief. The believer differs from the atheist only in affirming the applicability of their common ostensive experience to the domain of declarative language. No ultimate demonstration is available for either position. The nature of the existence of the originary signified of the ostensive sign can never be determined separately from the scene on which it manifests itself.

The very fact of my appeal to God implies my solidarity with the rest of the human race. A God conceived as a wholly personal daemon would have no power over the other parties to the scene of human conflict that is the originary model of crisis. It is inconceivable, for example, that the God I pray to for a cure to my illness be without the power to cure another’s. Our ostensive belief in God supplies the core of our ethical intuition: the sacred being is only present to me because it is present to all, because all have transformed their gesture of appropriation into a sign.

But the foxhole is no place to work out the social contract. Our first explicit ethical notion can only emerge when, the crisis having passed, we relax our fixation on the center and become aware of the other members of the community. At this point we become susceptible to the scandalized reaction of resentment that reveals the moral exigency of reciprocity among the users of human language.

What we call our “sense of justice” is first experienced through the scandal of injustice. We need no reflection to feel resentment when we see ourselves refused a privilege granted to another. The model we apply to such situations is that of the symmetrical exchange of signs in the originary scene of language. The originary crisis is averted by the enunciation of the sign as name-of-God by the entire human community. At this moment there is no hierarchy, no alpha individual; the exceptional being that resolves the crisis is God, not man. Resentment is our scandalized reaction to the existence of situations where this symmetrical configuration is not maintained. Unequal treatment of anyone constitutes a disequilibrium that is scandalous because it seems to threaten the community with return to originary chaos. I am not merely upset at my own ill-treatment; I am in terror of the potential disintegration of the entire social order.

Our resentful reaction to inequality reveals our belief in the moral model–an ostensive belief like the foxhole belief in God. Resentment points to the act of injustice, makes it known. God remains the implicit audience of our resentment as he was of our plea for help, but now we expect the rest of the human community to share our reaction. Where the foxhole renews the terror of the originary crisis that compels the use of the linguistic sign, the scene of resentment reproduces the moment in which language has already brought peace by deferring appropriation of the central object. In the first case, there is no preexisting model of resolution; we put ourselves in the hands of God. In the second, the community is expected to close ranks against a threat to an already established stability.

The equalitarian moral model is the minimal basis of ethics, just as ostensive belief in God is the minimal basis of religion. The traditional claim that this model, like the idea of God, is implicit in humanity itself is sharpened by its identification as that of the originary exchange of signs. But the resentment that reveals this model to us is no more a demonstration of its truth


than foxhole belief is a demonstration of God’s existence. The conviction that accompanies our inner sense of morality is ostensive rather than declarative. And just as we ask God to solve problems caused by phenomena outside the anthropological domain where alone the idea of divine control finds justification, so our resentment is aroused even when the non- reciprocity that offends us is defensible on broader ethical grounds.

How does the originary deferral of violence through representation lead to the formulation of ethical codes? The prescription of concrete acts, however elaborate, has for its basis the differentiating deferral or différance of the appropriation of the central object. At the core of the acts of ritual observance prescribed in ancient codes is the reenactment of the originary interdiction of the center.

The advantage of interdiction mediated by language over the constraints of animal behavior is the superior flexibility of software over hardware. Because animal constraints cannot be thematized in language, they cannot be transgressed. The originary core of interdiction is deferral through representation–not absolute denial, but postponement. The achievement of the originary scene is on the side not of constraint, but of liberation.

All thought realizes the liberating movement of self-understanding through the sign. The truest anthropology is the one that best exemplifies the power of representation to reduce the verticality of sacred interdiction to horizontal, worldly relations, a process sometimes called deconstruction. The principle of the maximization of deferral defines an ethic of both thought and action. What must be deferred through representation is the conflict engendered by de-differentiation, the convergence of the two sides of the mimetic triangle on a single object. An act of positive ethical value is one that creates a new significant difference, that promotes openness rather than closure.

Yet the creation of difference cannot be dictated by ethical law. Ethical norms must share the accessibility of signs, not the scarcity of things, the originary model for which is the unique central object. The interdictions and corollary precepts that compose an ethical code are such that all are presumed capable of carrying them out. The only way to claim inability to obey an ethical law is to plead insanity.(4)

But no law can prescribe the significant differences that the “open society” seeks to multiply. The market system that promotes the creation and circulation of such differences operates on the inherently fallible basis of economic judgment. In contrast with the preestablished differential roles of ritual societies, the values of the marketplace are determined a posteriori through exchange. Laws regulate the fringes of the exchange operation; they do not touch its central core. In principle, market transactions are non-coercive: with some well-defined exceptions, I can offer whatever I please at the price I choose.

The market system accustoms us to the divergence between the social imperative of increasingly freer deferral and ethics in the narrow sense of a set of obeyable rules. We are so used to this fundamental modern tension that we fail to be struck by the anomaly of a situation in which our ethical needs cannot be solved by an ethic. Our only enforceable ethical code is the minimal morality of criminal justice. The law assimilates gross violations of reciprocity (theft, rape) to the forbidden appropriation of the center; it is indifferent to the quality of the interpersonal exchanges on the periphery.


Ethical laws are never merely instrumental rules; even where there is no claim of divine promulgation, their imperative force is transcendentally guaranteed. This guarantee not merely to laws, but to the ethical propositions of metaphysical systems as well, even those that claim to be “beyond good and evil.” The use of morally charged concepts like “the good” or “justice” makes it impossible to avoid the slippage from “is” to “ought.” This invisible displacement of the ethical question from the anthropological to the conceptual plane was effective throughout the pre-Nietzschean history of metaphysics; it is still very much alive, if a little less well, today. Once one goes beyond the genealogical discussion, or what I call the originary analysis, of concepts such as “the good” to give one’s own definition of them, that is, once one accepts their validity as concepts–as is more or less obligatory in Anglo-Saxon moral philosophy–one cannot avoid going beyond description to prescription. A “theory of justice” is not a theory in the usual sense of the term. We cannot create a theoretical model of justice without conveying the imperative implication that we should act in the manner theorized as just.

Theories of the just society are as old as conceptual thought. The birth of Greek metaphysics accompanies the loosening of ritual constraints in an exchange-based economy, the first step toward a modern market-system. In such a world, where behavior can no longer be regulated by ritual prescription, metaphysical ethics, the ethic of the concept, finds the truth of the social order by projecting the moral model implicit in the use of language into the conceptual content of the Idea of the Good. The reasoning that goes into the establishment of such a concept is circular. One first postulates the existence of a universal Good from which mimetic rivalry is somehow excluded– whereas the common-sense notion of the good, as expounded by Callicles in the Gorgias, is indexical, my good and yours not being necessarily identical or even compatible. Then one claims that the “good society” is one whose members simply act, as all enlightened people must, for “the Good.”

The construction of this good society nevertheless requires political reflection; people need to be instructed as to the Good that is their own best interest. In contrast, the Gospel utopia of the Kingdom of Heaven obeys only the fundamental moral principle of reciprocity; the Platonic parallel to this transcendent realm is not the Republic but the heaven of the Ideas. But in either case, the transcendent is a hypostasis of the world of representation; the common referent of both the heaven of the Ideas and that of the Word is the sign.

If the moral model is based on the reciprocal exchange of words, the notion of justice originally concerns the distribution of things.(5) The Greek dike refers to just or equitable distribution before it acquires the meaning of “justice” in the narrow or judicial sense. The notion of “social justice” still implies distributive equality. John Rawls’ A Theory of Justice(6) is concerned not with crime and punishment but with the distribution of goods. The fiction that the author calls the “original position” is a metaphysical version of the originary scene, where equality is imposed by an artificially constructed “veil of ignorance” rather than being a necessary condition of the hypothetical emission of the sign in the originary event. The passage from ignorance to knowledge in Rawls’ scheme may be understood as an allegory of the historical evolution from primitive equalitarianism to modern market society.

Rawls postulates that things should by right be equally distributed; any other distribution must be justified by the greater good of the least-favored members of the community. This


attempt to articulate moral equality with ethical differentiation is a step beyond Benthamite utilitarianism, which treats society as an aggregate of human atoms.(7) But the “original position” goes beyond the familiar fiction of the social contract, the parties to which have full knowledge of their situation: Rawls’s fictional scene is only conceivable as directed from without. Once the fiction is converted to a hypothesis, the authoritarianism that is the only plausible context of this scene becomes implicit in the scene itself. For there to be the kind of uncertainty that an original position, as opposed to an originary hypothesis, requires, it would have to be imposed by the equivalent of a Platonic guardian. This anthropological critique of the authoritarian basis of liberal/socialist redistributionism is not coincidentally similar to that of such modern conservatives as Hayek.(8)

Justice in the narrow sense is the restoration of an imbalance, the punishment of a crime. The Oresteia recounts in mythical terms the emergence of the civic judicial system that breaks the endless chain of reciprocal retribution by offended parties and their families. Justice in this sense is not concerned with the original distribution of goods; it intervenes only to counteract acts of redistribution that threaten to subvert the deferral of conflict. Justice is a supplement to the originary deferral of appropriation; it punishes the violation of a norm that was enforced in the originary scene by the sacred power that manifests itself in the presence of the community as a whole.

Throughout all previous history, the historical diminution of this originary power has been recuperated by means of religious or secular eschatologies that assign as the end of history the making-eternal of the deferral of violence with which it began. The egalitarian moral model, which presides over the origin of man as an ethical animal, becomes humanity’s immanent goal. But to set the symmetry of originary deferral as a final goal imposes closure upon the open paradoxical structure of human mimesis. The claim that history returns in the end to the model that deferred conflict at its beginning–thereby justifying present violence on its behalf (“the end justifies the means”)– reduces history as a whole to the eternel retour of the originary scene. But the doctrine that deals with the total social order as though it were perpetually in the throes of the originary crisis is precisely what we call “totalitarianism.” What purports to be the self-closure of history is really the imposition of a gnostic tyranny.

From the standpoint of the originary hypothesis, the rough material equality that follows the division of the central object is not the beginning but the end of the originary process of distribution. True economic activity depends upon the deferral of ritual centrality; only after this activity has been performed away from the communal center is its product returned to the center for evaluation. Long before it becomes a dominant force, and however strongly a given social order attempts to resist its influence, there is always a “market”: a locus where value is determined through exchange. At the very least, the competition of rival societies in war provides a locus for determining value–not only that of weapons, but of social development in general. The creation of the free “capitalist” market that makes all goods fungible thematizes and liberates this process of exchange; it does not create it.

Market society realizes the Christian intention of making the moral model into a ethic, but in a minimal rather than a maximal mode. Between the Kingdom and the market lies all the difference between the mutual love of intimates and the free


exchange of goods among strangers. But before goods can be exchanged and distributed, they must be produced. Ever- expanding, ever-diversifying production is the primary imperative of the market system because it continually expands the ability of the members of society to acquire the differentiated personalities that oblige their fellows to recognize them as individual persons. The deproletarization of the working class has gone hand in hand with the expanding productive capacities of the industrial market-system.

The free market, as opposed to ritually controlled systems of exchange, obliges the individual to evaluate the product of his own labor before testing that evaluation in the marketplace. This social reliance on both subjective evaluation and its subsequent correction by the a posteriori objectivity of the market generates a free-floating resentment in virtually all members of market society. Although in crisis situations such as the Great Depression this resentment can come to focus on the market system itself, it normally attaches to specific “injustices” that can themselves become items of exchange. The elective legislative bodies characteristic of the mature market system constitute markets for ethical ideas and for the resentments, real or potential, that underlie them. The democratic political order permits the negotiation of these individual resentments on the basis of an expanded version of the moral model: one person, one “voice” or vote. The effectiveness of the “moral” domain of politics as a corrective to the “ethical” domain of economics derives from the fact that, like the economic market, the ethical market is wiser than any a priori set of norms.

The foregoing discussion suggests that we may derive from the originary hypothesis, not a prescriptive ethic, but two meta- ethical principles:

  1. The reciprocal exchange of signs is the fundamental (“moral”) model of human interaction. This model justifies our intuition that “we are created equal.”
  2. The fundamental operation of the social order is the deferral of conflict through the generation of significant differences (différance).

The first principle is the basis of all ethics. But it is insufficient to govern society. Conversely, the increasingly minimal application of the moral model to human interaction that accompanies the historical process of secularization allows ever greater scope for the operation of the second principle, which, as we have seen, is not susceptible to ethical regulation. The free market minimizes constraint on the circulation of goods and services that regulates the society’s production of differences.

But to claim that the wisdom of the market is superior to that of the theoretician is to affirm, as a derivative principle, the necessity that the market continue to operate, since its loss could not be made up by any theory. This principle counsels us to reject attempts to replace market society with one or another kind of socialist utopia, and to support political forces that expand the circulation of both economic goods and ethical ideas. At the same time, in accordance with the first principle above, we must defend the capacity of the members of the social order to engage in the reciprocal linguistic exchange that is the essence of human behavior. This proviso is analogous to the requirement in Rawls’s distributional model that social inequality contribute to the welfare of the least favored, but the minimal requirement of human welfare cannot be expressed in material terms. What is essential is to maximize participation in the social dialogue.


But the prerequisites of this participation at a given historical moment cannot be specified in advance. Nor can the freedom of a society be measured within a single generation; the parents’ sufferings may be redeemed by their children’s or grandchildren’s success. Because political, like economic activity takes place in a market that is wiser than any theory, no overall theoretical position can serve as an infallible guide for action within it. Hence the theory that puts its faith in the market explains why this faith cannot be made to justify political fanaticism, including the fanaticism that would unconditionally subordinate political exchange to the economic values of the marketplace.



1. Generative anthropology articulates our postmodern dissatisfaction with the Enlightenment version of secularization, which either denies the transcendental altogether or reduces it to the most abstract version of the metaphysical “first mover” (Deism), without ever explaining the transcendentality of the language it uses in the process. Revolutionary atheism is an inverted religious fundamentalism that makes use of verticality to tell us that the vertical does not exist.(back)

2. My The Origin of Language (Berkeley: University of California Press, 1981) contains a fuller discussion of this distinction.(back)

3. I discuss the Exodus passage at length in Science and Faith (Savage, Md.: Rowman & Littlefield, 1991); the parallel with Plato is taken up in a forthcoming work entitled Paradoxes of Mimesis.(back)

4. Saint Paul claims that the very point of the Hebraic Law is to create guilt for our inability to fulfill it. But we only feel guilty because the source of this inability lies in the rebelliousness of our will rather than our incapacity to follow the Law’s instructions.(back)

5. In the originary circle of participants surrounding the sacred object, there is no “justice” because there is no material distribution. In the ensuing sparagmos, in which the object is torn to pieces and distributed among the participants, the force of originary resentment against the center diminishes the rivalry among the participants, resulting in a roughly equal division. Justice among the participants is a consequence of the community’s vengeance against the center for having withheld itself from its members.(back)

6. Cambridge, Mass.: Harvard University Press, 1971.(back)

7. As a consequence, the apparently benign utilitarian calculus can be made to justify any degree of sacrificial violence for the sake of making the “greatest number” of atoms cohere. The economy of Girard’s scapegoat model, where a single individual is sacrificed for the greater good of the whole, is both the nec plus ultra and the reductio ad absurdum of utilitarianism.(back)

8. Rawls’s model lacks the very notion of production, let alone that of entrepreneurial innovation. His list of “primary goods” the distribution of which is hidden by the “veil of ignorance” includes not merely the basics of survival but “wealth,” as though the latter were found in nature and apportioned by a central authority rather than created through risk-bearing economic activity.(back)