van Oort - Performative-Constative Revisited: The Genetics of Austin's Theory of Speech Acts

Anthropoetics II, no. 2 (January 1997)

Performative-Constative Revisited: The Genetics of Austin's Theory of Speech Acts

Richard van Oort

Department of English and Comparative Literature
University of California at Irvine
Irvine, CA 92697

It may seem unlikely that so paradoxical and apparently unmaintainable a thesis as J.L. Austin's elementary distinction between performative and constative should be described as part and parcel of a general "revolution in philosophy," but this is precisely what Austin tells us at the beginning of his famous 1955 Harvard lectures, posthumously published in his How to Do Things with Words. With characteristic ironic British humility, Austin begins his lectures by politely remarking upon a "mistake" that a certain traditional philosophical attitude toward language is guilty of. This mistake arises from the view that language is preeminently the tool of constative assertion, that is, a tool primarily interested in providing statements about the world, which are characterizable as either true or false.(1) In contrast to this rather limited view, which when actually compared to the full range of utterances in the real world is forced to exclude most of them as simply nonsensical, Austin proposes a second category of utterances that are not subject to the truth/falsity conditions of propositional knowledge. Rather, these exist as acts in themselves, that is--as Austin dubs them--as performatives. The peculiarity of the performative utterance, in contrast to the constative, is that it does not describe a state of affairs independent of itself, but that it is itself the reality it describes. It is therefore a self-reflexive utterance. Austin's archetypal examples of these are the acts of naming, marrying, bequeathing and betting (see How to p. 5). Thus, for instance, when I utter, "I name this ship HMS Hermes," I do not describe a state of affairs in the real world. Rather I bring a state of affairs into existence by virtue of my utterance. The act of naming is simultaneously the reference of my statement. The performative is therefore, in the most rigorous sense, an act and not a representation of something else, at least not in the preferred constative sense of a representation.

Let us examine this situation more closely. In querying the nature of the utterance's relation to the world, Austin's conception of the performative has deeper epistemological and ontological implications. That is, it touches on the problem of linguistic reference. Are we representing a reality external to our utterance, or are we creating by the very act of the utterance the reality which we seek to define? The first type of utterance is the classic model of constative assertion, manifest in the proposition which can be verified, that is, which can be proven true or false.(2) This is the model indispensable for empirical science, for without being able to separate worldly reality from linguistic utterance, empirical science would not be able to create objective models, the validity of which is measured by the separate existence of a world conceived as ontologically prior to, and independent of, the models used to represent that world. The second type of utterance, however, does not conceive itself as merely a supplement--accurate or inaccurate, true or false--to a world against which this accuracy is measured. Rather, it presents itself as producing the very reality it names. If, given the appropriate conditions, I declare, "I name this ship HMS Hermes," then the mere fact that I have said the words produces the event to which I am referring. Thanks to my utterance, the ship is named where it was not before.


But why is this so paradoxical? Why does it cause philosophers so much consternation? One reason no doubt is the challenge the performative model presents to the ontological faith that philosophers have placed in the proposition as the fundamental unit of all our utterances. Austin's delight in taking the constative model to task is indeed one of the major themes driving his deliberations on the performative nature of language-use, and undoubtedly this undermining of constation and its concomitant concern for propositional truth is precisely the quiet "revolution" Austin regards his analyses as contributing to.

But ultimately I do not think Austin's hypothesis is radical enough. The revolution he suggests that is taking place in philosophy may appear revolutionary from inside philosophical circles, where a certain understanding of truth and language has held sway since Plato, but from the originary-anthropological perspective I wish to adopt here, Austin's questioning of the supremacy of the logical proposition is a necessary prelude to a more radical model that includes within its linguistic model a historical and anthropological content that counteracts the ideal and detemporalized models of metaphysics.

Can we then give a more radical thesis to Austin's conception of the performative? I think we can, but to do so we will have to begin from scratch and dispense with the philosophical faith--if not always explicit, certainly tacit--in the declarative sentence as the fundamental semantic unit of linguistic communication.

Why must we dispense with the declarative sentence? The simple answer is because the declarative sentence is not an originary form. In the spirit of generative anthropology, the hypothesis I wish to present is in fact based on a more economical formulation of language origin that, rather than beginning with the declarative sentence (and its younger philosophical brother, the proposition), seeks instead to explain how such a complex linguistic form originated in the first place. The assumption behind this argument is, of course, that we need a rigorous formulation of language origin at all. That such an assumption is valid is not something that can be proven indisputably. But, indeed, the very notion of such an absolute criterion of truth is itself a product of an unquestioned faith in the capacity for declarative sentences to generate ontological truth from the detemporalized scene that it is the specific capacity of the declarative to produce. The generative or originary claim is, in contrast, a more minimal one. By understanding the scene of original language production, we hope to provide not transcendental logical models, but, rather, a historical hypothesis which explains this capacity to generate abstract models of reality divorced from the scene upon which they occur. The error of traditional metaphysical thought is to grant this capacity for abstract model-building--and hence truth/falsity conditions--an ontological status. Implied in this metaphysical ontology is the belief that conceptual thought, and with it the categories of truth and falsity, is concomitant with the origin of language. But this implied model of origin, as we will see, is dependent on an anterior, more minimal model of language-use, one that does not require developed declarative sentences but is rather simply an ostensive (or indicative) use of language (as in the nominal utterance "Fire!" which indicates the presence of fire). By presenting an explicit model of the origin of language--one based on the scene of the originary performative context of the first linguistic sign--we seek to introduce precisely what is lacking in metaphysical models of language, namely, the entire scene in which language must be conceived to have evolved. It is only with the advent of the declarative, a later development of linguistic evolution, that language can divorce itself from its context, and thus appear to be wholly independent of its scene of production.(3)


There is much in Austin's lectures that suggests the need for a fundamental "scenic" hypothesis of language-use in general. His analysis of performatives is based on the intuition that propositional truth is a product of a more developed linguistic form. Ordinary language situations reveal this conflation of language-use and constative description to be, more a scientific ideal to be pursued, than a normal condition of the pragmatic context of language-use. The danger of the scientific ideal, Austin warns, is to hypostatize it as somehow the normal condition of all language-use:

One thing, however, that it will be most dangerous to do, and that we are very prone to do, is to take it that we somehow know that the primary or primitive use of sentences must be, because it ought to be, statemental or constative, in the philosopher's preferred sense of simply uttering something whose sole pretension is to be true or false and which is not liable to criticism in any other dimension. We certainly do not know that this is so, any more, for example, than that all utterances must have first begun as imperatives (as some argue) or as swear-words--and it seems much more likely that the "pure" statement is a goal, an ideal, towards which the gradual development of science has given the impetus, as it has likewise also towards the goal of precision. Language as such and in its primitive stages is not precise, and it is also not, in our sense, explicit: precision in language makes it clearer what is being said--its meaning: explicitness, in our sense, makes clearer the force of the utterances, or "how (in one sense ...) it is to be taken." (72-73)

Austin's well-known dissatisfaction with the traditional view that language is fundamentally propositional here receives overt expression. Rather than interpreting this passage as simply a point-blank refutation of the constative framework as a workable model (as many cultural critics have been wont to do), it is more valuable, from our point of view, to note how Austin refutes philosophy's love of the verifiable proposition. It is not that propositional knowledge is a priori an impossibility or an illusion (as an absolute skepticism would claim), but that it is an ideal, a goal, which science gives impetus to. The "mistake" (as Austin says) of philosophy is to presume that this category is originary, that it is somehow fundamental to all speech acts. For once this conclusion has been drawn, we are left with no way to explain nonconstative speech acts other than as parasitic or simply nonsensical.(4)

Austin's location of the performative is a realization from within philosophy that all language is at bottom "performative," and that constation can be better explained as a highly evolved, specialized and scientific outgrowth of more basic performative (i.e., pragmatic and communal) language-uses. Austin, however, does not seek to expand on this historical hypothesis; indeed, in the passage cited above, he deliberately refers to originary conceptions of language-use (such as imperatives) with cynicism. For Austin, the fact that there is simply no evidence for such hypotheses prohibits the question from the beginning.(5) Significantly, however, at the end of the penultimate lecture of How to Do Things with Words, he does suggest that between the performative and the constative, the illocution and the locution, "we have here not really two poles, but rather an historical development" (146). Admittedly, it is not entirely clear from this brief remark whether Austin is suggesting that the performative utterance is indeed more primitive, and that the constative is thus a specialized development from its more fundamental performative context. But given Austin's earlier remarks (and indeed the whole thrust of Austin's analysis of performative language) it seems difficult to deny that because the performative, in its most general sense, provides the more inclusive context for a consideration of speech acts than does the constative, it therefore represents an earlier stage of language development.


The problem with giving such an argument, at least in the context of Austin's work, is, of course, that Austin himself will give no historical account for such an evolutionary point of view, presumably because to do so would force him away from the rigour of empirical analysis and into the realm of unsubstantiated--that is, hypothetical--speculation and model building. But empirical analysis is not a guarantee against metaphysics; it includes its own parti pris by refusing to confront the genetic character of linguistic form. The point of the originary hypothesis is precisely to minimize our presuppositions when we theorize. The alternative, I would suggest, is to forego historical rigour for metaphysics (whether rationalist or empirical), which is clearly a position Austin himself would be unhappy with. Eric Gans, in his original anthropological analysis of speech acts, has criticized speech-act theory for this unwillingness to develop a genetic-historical account which can synthesize its taxonomic classifications of types of speech acts. Using an analogy from the history of biology, he suggests that "[p]hilosophy's dealings with speech acts involve the same kind of a posteriori classifications that we find in pre-Darwinian biology," adding that "only an anthropological hypothesis can lead us from Linnaeus to Darwin" (62, 64).(6)

How then are we to articulate our intuition that language-use cannot be explained solely on the model of constative truth functions, but rather as an expression of urgent pragmatic situations? Austin's path, as we know, is simply to point to the numerous cases of ordinary language-use that clearly cannot be defined by the truth/falsity criterion of the constative. These contexts of language-use are analysable only in terms of the total speech-act situation, or what we can here call the "scene" of the utterance. Often they do not even seem to need the complete linguistic form of the proposition, being merely single-word utterances and not well-defined sentences. But though such analyses of discrete scenes of everyday language-use allow us to question the constative model, they do not in themselves allow us to replace it with a more powerful a priori model of the general context of all language-use. Searle does talk of "illocutionary force" as his common denominator for all language-use, but this is an abstract descriptive term, not an integrated model. On the whole, both Austin and Searle remain content with a classification of illocutionary acts, making no attempt to synthesize their intuition that language operates "scenically" by proposing a general hypothesis of the scene of universal language-use. Such a hypothesis, to be rigorous, must begin with the origin of language, that is, with the creation of a universal scene of representation, which is seen as the defining aspect of the human in general. Eric Gans, whose work is dedicated to this conception of "originary thinking," proposes that what remains implicit in all individual scenes of language-use is an "originary scene"(7) which constitutes the human capacity for representation:

In the linguistic act, speaker and hearer, or writer and reader are placed in the presence of a common scene of representation. Everyday acts of language can allow us to exemplify this scene, but the only hypothesis that allows us to explain its existence is one in which the communal presence of language originates as an event. Language could not have evolved imperceptibly from prehuman forms of communication; it must have been created, or discovered, through a revelatory act. This does not imply that its structures appeared miraculously all at once, as those linguists implicitly believe who make the declarative sentence-form the unexplained basis for their grammars. But the simplest, ostensive form of utterance, which merely designates an already-present object, can only arise in the context of a collective scene, in which the scenic center is isolated at least momentarily from the human "spectators" and language-users at the periphery. ("Sacred Text and Secular Culture" 53)


We see here, I propose, an explicit articulation of Austin's notion of the performative as the context for the origin of language. The first sign is an ostensive that simply designates an already present object. For Gans, this scene in which language first occurs brings about a minimal use of language. But to avoid the circularity characteristic of previous "myths of origin," which are hence open to a deconstruction via the "always-already" argument, the originary scene must be explained by prelinguistic motivations. The designated object must have already been of interest to the group for appetitive--i.e., prelinguistic--reasons. Thus, the nascent human community surround an object attractive to appetite that all wish to appropriate but, fearing the mutual reprisal from the other members of the group, each individual is forced to abort his or her appropriative gesture, which thereby becomes instead a designative gesture, that is, precisely a representation of the object as significant to all, and therefore as forbidden to all. Language here is the solution to the potential for violent crisis brought on by excessive appetitive desire. Prehuman systems of conflict resolution depend on genetically programmed communication signals and rituals based on strict dominance hierarchies observed by ethologists in the higher animals and primates. By contrast, the aborted gesture is successful as a linguistic act only if every member participates intentionally in the act. Each member must renounce his/her appetite in order to establish the significance of the central object as being a topic worthy of a new form of perception, that is, as an object represented to all as universally significant because universally denied. The act is only a linguistic act if an intention not to appropriate the object is attributable to each designating individual. Without this establishment of what Gans calls "communal presence," there would be no intentional deferral of conflict and therefore no linguistic act.

We are now in a position to flesh out some of the performative implications of Gans's originary scene. Austin notes the "felicity" conditions necessary for performative language to occur successfully. But the ultimate "felicity condition" of all language-acts is the originary scene. Here, the "successful" performance accounts for the very origin of the human. We may assume that the conditions at the origin are wholly more urgent than those of subsequent uses of the linguistic scene of representation. The protohominid community is faced with a dangerous crisis that can only be dispersed and deferred by the mutual renouncement of instinctive appetite. The immediacy of the everyday interactional speech situation reconstitutes the original immediacy confronting the protohominid community. The plausibility of the hypothesis stems from this observation. Language must have evolved as the more effective means for controlling the dangerous immediacy of unmediated and conflicting appetitive desires for an object of universal interest. It is the immediate and threatening presence of each individual on the scene that provides the unique interactional and intercommunal conditions that are characteristic of the intersubjective basis of language-use in general. Thus, every participant is aware of the others' aborted gestures and participates equally in the creation of the moment of communal presence. This is the minimal condition necessary for language to take place. We reproduce this presence in everyday speech acts without further reflection, but at the moment of its inception, the power of the speech act was the direct response to an imminent crisis. It is worth observing that in primitive ritual situations most obviously, and aesthetic representations less overtly, this moment of crisis is explicitly--i.e., thematically--reproduced.(8)


In contrast, language performs the original resolution of the crisis simply by virtue of its form. This understanding of the human as constituted, not by (mythical and sacred) thematizations of the linguistic scene, that is, as content, but rather by form is indeed what justifies us in reconstructing a hypothesis for the origin of language. That is, rather than assuming that this origin is itself explained by a transcendental content which is generated only after language could have come into being, we seek to give a plausible account of how this linguistic scene itself could come into existence. In this originary sense, the overlap between metaphysics and religion is apparent, for both grant the content of the linguistic scene an a priori status without seeking a more minimal explanation for the development of this content via its linguistic-scenic form. Only an originary hypothesis can provide a plausible explanation for the development of linguistic content.(9)

The originary scene offers us a general anthropological model that allows us to situate more rigorously Austin's notion of the performative speech act by returning it to its original collective and ritual context. Almost all of Austin's original examples of performatives are of ritual origin and thus possess an explicitly ethical dimension (as opposed to a strictly logical one). Naming and marrying are obvious examples of social rituals that take place before the whole community. Promising and betting are not so clearly ritualistic because they do not possess the same public setting. To promise or bet, one merely needs another person with whom to undertake the promise or bet. But implied in the "local" scene upon which we engage in everyday acts of betting or promising is the equivalent understanding that our word will be evaluated according to the "public" scene explicit in our more ceremonious and externalized rituals of marrying and naming. It is indeed a measure of a society's freer ethical structure that local interactions between individuals can displace the essentially conservative domain of large public rituals, the local scene always being the site for more individual freedom.(10)

Promises and bets are instances of contractual agreement between individuals. The binding constraint they place on the individual should give us a clue as to their ritual roots. Naming, marrying, promising, betting, all operate on the same principle tacit in the originary sign. In designating the central object as collectively forbidden, the original participants provided the first "contractual" agreement between each other--the "agreement" not to appropriate the central object. The originary sign "names" the object as forbidden. The notion of an "agreement," however, is misleading. The originary sign is not a conventional contract, an agreement on already stipulated terms. It is the originary "contract," a revelatory event that produced the institution of language by deferring violence. This deferral of conflict cannot take place without the whole community participating collectively in the designation of the object. From this originary gesture stems the model for all collective acts of ritual designation. The oddity of such speech acts, in contrast to indicative and constative acts of worldly reference, is that they require the sanctioning--the commitment--of the community to be successfully performed. Hence Austin's category of appropriateness conditions. In the cases of marrying and naming, this is clear enough. But the anthropological context carries over into the cases of promising and betting. Here it is the speaker, not the entire community, who is committed to the speech-act.(11) What both types of performatives share is their emphasis on the human designators. Hence the fascination of such speech acts for perplexed philosophers who have first pinned their loyalties on the descriptive or constative model of language. The intuition behind Austin's category of linguistic performatives and felicity conditions is based on the ostensive nature of ritual. Language here can operate only in the presence of significant or, in the case of ritual, sacred objects. The ostensive designates an already present object (as in the originary scene), but it cannot abstract a model separate from the worldly scene of the linguistic utterance. This is a property of the declarative sentence, which is a later development that must be explained on the basis of the more parsimonious description of language as fundamentally an act of ostension.


At the elementary level of language-use, signification takes place as an indication of a significant object. Single-word utterances, such as the cry "Fire!", demonstrate this primitive level of ostensive utterance, which requires the context of the scene in which it is uttered to be understood. Certainly, such ostensive utterances can be seen as simply abbreviations of the declarative/propositional utterance: e.g., "(I declare) there is a fire (here)."(12) But such explanations fail to comply with the criterion of a minimal explanation of language. Thus, they assume that the scenic context must be excluded from the linguistic utterance. But this is an ex post facto hypostatization of the declarative sentence, as Austin's repeated criticism of the constative implies. The cry "Fire!" is a warning, not a proposition. The point of its utterance is for others to share in an awareness of the danger; the cry once uttered will be taken up by others. Presumably the hearers can be expected to verify for themselves whether there is indeed a fire or not, but the efficacy of the utterance lies in its ability to circulate urgently throughout the immediate community, not in any capacity to propose an abstract state of affairs that can be either true or false. Thus, at the ostensive level, we must follow Austin's argument for appropriateness conditions. The performative is not true or false, but rather felicitous or void. The cry "Fire!" when there is in fact no fire is simply a void utterance. Ringing an alarm bell has the same effect as the cry "Fire!" but we cannot attribute a truth value to either the bell or the utterance. This follows from our understanding of ostensive signification. What is signified by the elementary gesture is simply a referent deemed worthy of general communal attention. But this significance receives no measure other than that provided by the attentions of the community. There is no abstract realm to appeal to in order to apply the notion of truth to the referent. All that we can say about the ostensive object is that it is significant to all.

But is this not equivalent to a proposition? Does not the utterance assume that the speaker is asserting a fact about the world, namely, that there is a fire, the worldly reality of which will either prove or disprove the utterance? Certainly, there is the brute fact of the fire that must be present for the utterance to take proper (felicitous) effect, but what this objection fails to note is the difference between a linguistic utterance that is uttered solely in the presence of its reference and an utterance that proposes a state of affairs about the world, that is, that precisely does not need to be uttered in the presence of its referent. The confusion arises from an inadequate theorization of the distinction between the ostensive utterance (Austin's performative) and the constative utterance. The former can only be understood in context of the scene where it is uttered. Crying "Fire!" where there is no fire simply makes no sense. On the other hand, if I say "The boys have lit a fire in the park," I have depicted an abstract scene for you that does not need the immediate context of my enunciation to be understood qua statement. You can conceptualize the abstract scene regardless of my own position as interlocutor. To understand my utterance you have first to conceive a scene independent of the context in which my act is uttered; that is, you have to imagine for yourself a state of affairs where the boys have lit a fire in the park. No doubt, once you have constructed this independent linguistic scene you are free to speculate on the illocutionary force of my statement that will necessitate an examination of the context of my utterance. Such questions as the speaker's status and power, so dear to the sociopolitical interpreters of speech-act theory, necessarily come to the fore: e.g., What authority do I have? Am I merely another juvenile? an adult? a fire-inspector? Could I be lying? tattle-telling? Perhaps we are both poor out-of-work fire-fighters who have ordered the boys to start the fire.


But these questions of context are all contingent on the assumed conceptual ability to represent abstract scenes independent of the context of the utterance. This is the essence of all propositional knowledge, as well as, let it be noted, all fictional utterances. It is only because an independent scene can be constructed that the possibility for fiction arises. Fiction is inherent to the propositional utterance itself. The difference between an utterance that is fictional and one that is propositional is not that the former is parasitic upon the latter, but precisely that the latter forgoes the primacy of its independent construction of an imagined state of affairs to relate this conceptual scene to a worldly reality. The declarative can always in principle (if not in reality) be verified by taking it beyond the linguistic scene it creates. Thus, you can go to the park yourself, and see if the boys have indeed lit a fire. Hence the original sense behind Austin's term "constative"--to constate (from the French constater)--which means to verify or to ascertain. What distinguishes the fictional utterance from the propositional utterance is the willingness of the hearer to remain captivated by the original imaginative scene that is constructed for him/her. What is deferred here is the presence of the worldly referent, which in the case of the declarative sentence-form--the prototype of the logical proposition--is a worldly state of affairs. And this should not surprise us, for what motivated the originary sign was also the deferral of the dangerous presence of the appetitive object. The difference is, of course, that the originary sign is a more "local" form of deferral, for the referent appears in the presence of the linguistic sign.(13)

The first linguistic sign is simply a centring of the referent which by its position at the centre is deemed worthy of significance to all individuals on the periphery. This sign, in its unique situation at the origin, demonstrates the ambiguity of both directions of Searle's word/world relationship.(14) On the one hand, the sign is motivated by the worldly appetitive object that must be conceived as ontologically prior to the sign, thus demonstrating the world->word direction of fit; on the other hand, the object is sacralized as a cultural object which is forbidden appropriation by the community, thus revealing the word->world direction, where the sign creates a newly perceived reality. The act is performative precisely because it defers, via the temporal performance of the gesture, the reality of the object that threatens the stability of the community. The originary gesture takes place as a temporal deferral of this dangerous immediacy by inserting between the participants on the periphery and the object at the centre a cultural space, which is also, more specifically by our hypothesis, a linguistic space, where the object is mediated by the cultural/linguistic sign.


Now this sign is nothing other than the mutual participation of the whole community in the scene of representation. But what this collective act produces is a universal scene where the object, formerly only perceived as an appetitive object to be unproblematically appropriated, is transposed to a collective scene that forbids the individual to act out his/her appetite by forcing upon him/her the awareness that others are also designating the object. That this linguistic sign, at this primitive stage, is an act, a performance, is evident from the concrete temporalized deferral that the original sign accomplishes. The act creates the moment of collective presence, which is an ethical consolidation of the unity of the community. We may observe that the linguistic sign, at this ostensive stage, is more a musical phenomenon than a representative one, or better, that it is the musical aspect of temporalization that creates the representative aspect (in the limited--ostensive--sense used here) of the originary sign. Through temporal deferral, significance is bestowed on the central object. Northrop Frye, in a perceptive analysis of the musical aspect of spoken charms, provides us with an insightful perspective on the sacred and magical element of ostensive signification with its capacity for shutting down random, potentially destructive action: "the central idea of the magic of charm is to reduce freedom of action, either by compelling a certain course of action or by stopping action altogether" (125). Frye continues by noting the use of rhythm and sound which compels the hearer to a version of linguistic mimesis distinct from the traditional (Aristotelian/constative) conceptions of language as a mimetic doubling of worldly reality. Language in this sense is truly performative, for its mimetic efficacy does not operate through objective constation but through participation in what we have defined here as a collective performance of communal and ethical presence. Performer and performance, sign and referent, signifier and signified appear on the same scene. Hence our analogy between ostension and music, as demonstrating the "pure" form of temporal representation, is justified. The purely formal domain characteristic of secular music in the West since the Baroque era is, from our point of view, the ultimate liberation of the ostensive sign from its reference to a worldly, sacralized object. One can define music anthropologically as the secular development of the ostensive form of signification, where the sacred referent has finally been displaced into a world of pure form. Music is thus truly the most "performative" of the art-forms.

Searle's world/word distinction allows us to theorize more rigorously the dichotomous uses of the elementary ostensive sign. The cry "Fire!" demonstrates the world->word direction of fit. The utterance of the word "Fire!" does not change the nature of the reality that motivates the utterance, but spurs those within hearing into appropriate action. Very different are Austin's archetypal ritual performatives. Declarations of naming and marrying confer a cultural reality that was not present before the utterance occurred. They thus take the word->world direction of fit. By virtue of the utterance, a new (cultural) reality is created. Or, in anthropological terms of the originary scene of language-use, the newly proclaimed reality is dependent upon the voluntary consent of all the ritual participants. No doubt the temptation is to see this voluntary communal consensus as an unwarranted idealization of the sociopolitical order which is better described in terms of authority and power than in terms of communal consensus. But what such critiques fail to take into account is the essentially egalitarian context of the minimal linguistic act. There has been much discussion on the sociopolitical context necessary for authoritative acts of ordering,condemning, and such like. But such "political" explanations do not disprove the originary hypothesis; on the contrary, they lend evidence for it. It is precisely because language transcends social status and is understood across political hierarchies that we must assume the originary--i.e., fundamental--status of egalitarian linguistic exchange. Political accretions to this linguistic context are always possible, but unless we wish to hypostatize such accretions for all time, they must be understood as just that--later developments that depend upon the essentially egalitarian context of originary linguistic exchange. It is indeed precisely this originary awareness of the egalitarian context of linguistic exchange that enables the sociopolitical interpretation to gain a moral foothold. For how else could our common and basic level of linguistic understanding, let alone our moral sense that "we are created equal," be explained? It is rather those notions of language that would make the sociopolitical the epistemological condition of our linguistic heritage that unthinkingly condemn cultural reflection to a mythic interpretation of power which itself remains beyond explanation. To give the sociopolitical complete supremacy as a model of cultural reflection is to assume that hierarchy itself is originary. Rather than liberating us, this model merely leads to an ontological stalemate, for having elevated hierarchy--and more subtly the abstract concept of le pouvoir--to a transcendental category that governs the social order, we are left with no way to explain our own liberation (outside of divine salvation) from the hands of such a disenfranchised conception of the social order. (15)


The word->world direction of fit, then, explains the voluntary consent of the participants to bestow a cultural reality, such as a name or a marriage, upon a worldly reality. Such consent mirrors the irenic path of the originary sign which allowed the community to pass from crisis to consensus. Minimal speech acts such as greetings and apologies (Austin's "behabitives," Searle's "expressives") reproduce this irenic collective context. "Hello" clearly cannot be explained either as a constative act, nor indeed as an explicit performative, as in Austin's original examples of marrying and naming. However, its usage becomes clear as a minimal ritual inherited from the originary scene. Thus, uttering "Hello" to a stranger is an attempt to reestablish the original equilibrium that must exist in ritual performatives for consensus to occur. The greeting is a prelude to more practical speech acts, but its fundamental nature is evident in the fact that it has itself no content, being merely a ritual to be performed before communal presence can allow the interlocutors to engage in more functional speech acts. No doubt greetings have become trivial affairs where a grunt, or simply a nod, is adequate for a civil threshold of presence to be established. But this is a consequence of the ever-increasing awareness that communal presence is not a unique contingent phenomenon but a condition of human existence. Acts of apology function in much the same way. The utterance "Sorry" serves to avert conflict in the event of an action that disturbs the balance of communal equilibrium.

In conclusion, we can refer back to our initial suggestion that performatives are not simply the polar opposite of constatives. As we have seen, Austin himself took care not to reduce the two types of speech acts to a simple case of binary opposition. But Austin's hypothesis, I have argued, is not radical enough. His performative-constative distinction is better explained along genetic lines, where the performative is understood as the condition of possibility for more specialized constative speech acts. Our own investigation of Austin has led us to Gans's conception of the originary scene, which is a radical reformulation of the concept of the performative from the perspective of generative anthropology. Such an originary hypothesis allows us to explain the genetic unity behind such seemingly trivial acts as greeting, as well as more overtly ritual acts, such as naming, and finally also fictional and logical speech acts which are dependent upon a fully developed declarative sentence-form. At the basis of our analysis is the model of the originary sign which deferred conflict via the creation of a communal scene of representation. All linguistic-cultural acts have this deferral as their basis, and all such acts are--as was indeed already the originary sign--reflections on this moment of origin. The current analysis, by seeking to perform its own act of reflection on our collective origin, makes the same wager that created humanity in the first place. There is no logical proof for this conception of origin, but that is why it is a performative wager. We bet our existence on the ethical efficacy of theorizing our own origin, just as the original humans wagered their existence on designating consciously, for the first time, something other than the human--the inaccessible sacralized object which could generate the cultural and ethical space necessary for a temporary solution to the crisis. To be sure, our current theorizing does not blindly wager its faith by designating for the first time a unique object as sacred, as the original protohumans must have done. The wager we place is the originary hypothesis, which gives a plausible account of the entire scene in which this original act of faith must have occurred. This scene is both a consequence of, and a heuristic for, interpreting the cultural acts that we, as the descendants of the first humans, have inherited. Speech acts, because they exemplify the minimal institution of language itself, are only the most fundamental of these cultural acts. But they are also therefore the best indicators of our common heritage as creatures who have wagered their existence on the communal efficacy of language-use.



1. See p. 3: "Not all true or false statements are descriptions, and for this reason I prefer to use the word 'Constative.' Along these lines it has by now been shown piecemeal, or at least made to look likely, that many traditional philosophical perplexities have arisen through a mistake--the mistake of taking as straightforward statements of fact utterances which are either (in interesting non-grammatical ways) nonsensical or else intended as something quite different."(back)

2.This is meant in the broadest sense. Naturally, the interpretation of the evidence will influence one's judgment, but the fact remains that the proposition by offering an independent model--i.e., in grammatical terms, a predicate of a topic--always leaves itself open to the criterion of correspondence--true or false--to a world from which it is definitively severed. This is both why there indeed exists a notion of "truth" at all and why such truth can never be "absolute" but enters into the intermediary space--the "hermeneutic circle"--between word and world. The performative, however, in obvious ways does not offer such an independent model. The power of Austin's analyses lies precisely in his realization that the performative is tied to its scene of utterance in a way the constative is not.(back)

3. As will become evident in this paper, the distinction between performative and constative is best explained as the difference between an utterance that includes the scene of its own production and one which must be understood first as a predicating a state of affairs about the world on a plane separate from the scene upon which it is conceived. This latter, more complex, utterance requires the characteristic subject-predicate construction of declarative sentences. The error of linguistic and philosophical models of language is to presume that the declarative sentence is the elementary unit of language. It might be objected that though perhaps linguistics and analytic philosophy presume the declarative sentence to be fundamental, speech-act theory makes no such equivalent assumption, since it deals, not specifically with elements of grammar, locution, or propositional meaning, but with the illocutionary force of utterances, i.e., what gets done by the total speech-act. But this remark fails to note that speech-act philosophy's concern with elements of context is not specifically an undermining of the general faith in the declarative as an elementary unit, but indeed precisely an assumption of its validity as a semantic unit embedded within the context of ordinary language usage. Thus does, for instance, John Searle describe the illocutionary acts he seeks to classify as "the illocutionary force of an utterance and its propositional content," which he symbolizes as: F(p), where F = force, and p = proposition (Expression and Meaning 1).(back)

4. Hence the bemusement that so-called fictional "constative" speech acts cause. Austin, sensibly perhaps, simply ignored the question of such acts, referring to them, infamously, as merely parasitic on ordinary speech acts. Subsequent attempts to resurrect a speech-act model of fiction usually begin with the constative assertion and then add rules that allow us to account for the non-verifiability of such assertions. Thus, Searle talks of the suspension of the "horizontal conventions" relating the statement's words to the world (Expression and Meaning, Chapter 3), and Samuel Levin of a "higher sentence" that implicitly prefaces all literary statements with a warning that we are entering a fictional world ("Concerning What Kind of Speech-Act a Poem Is"). But such models cannot avoid the tacit assumption that fiction is indeed an afterthought parasitic upon the real constative business of language. Consequently, they revert back to the model Austin was intent on undermining. See Thomas Pavel, "Ontological Issues in Poetics," for a criticism of this "breakdown model" of fictional speech acts.(back)

5. But how could there ever be empirically observable evidence for the origin of language?(back)

6. See Originary Thinking, Chapter 4. For a full exposition of language evolution see his The Origin of Language (1981). Much of what I say will be a development of the generative model in the area of speech-act theory, which strikes me, as I hope will become clear, as a particularly fruitful place for such analysis.(back)


7. Cf. also, The End of Culture, pp. 16-17: "As soon as the phenomenon of representation is no longer accepted as a given, either 'natural' like the genetic and other 'codes,' or miraculous and incapable of explanation, then it is obvious that any explanation of categories of representation, let alone individual acts of representation, must refer at least implicitly to an explanation, that is, a theory, a hypothesis concerning the phenomenon of representation-in-general. And the only hypothesis that can in any sense explain representation as a historically given activity peculiar to our species must be a generative one that proposes a model of its emergence from an earlier state in which it was absent."(back)

8. See René Girard, Violence and the Sacred, for a comprehensive analysis of the sacrificial crisis in primitive ritual and aesthetic-cultural representations.(back)

9. It should be noted that the originary hypothesis is not another version of the social contract model popular in the Enlightenment period. The originary hypothesis is not a myth or fiction that seeks to describe on the model of already existing categories a narrative of the origin of the social order. The paradox of all such aetiological/mythical derivations is, as Derrida, de Man, and others have delighted in pointing out, precisely that the social categories for which the explanation is sought must already have been present at the origin. At its most radical, the deconstructive argument gives this paradox its own hyper-substantive (metaphysical) status. Thus, Derrida's différance becomes the epistemological condition of language and meaning itself. But rather than seeking recourse to metaphysics (even as anti-metaphysics), the originary hypothesis, by postulating that the originary sign is the deferral of appetite via the simultaneous creation of a uniquely significant centre, historicizes the philosophical (non)category of Derridean différance. For generative anthropology, the first linguistic difference is precisely the temporal deferral of the appetitive object.(back)

10. The great insight of speech-act theory is to return language to the collective scene of the human community. It is only in context of such a collective scene that we can understand the radically communal origin of representation. Many speech-act interpreters have focused on this collective scene in order to undermine, on the one hand, the objectivist notion of truth and value, which ignores the anthropological context for a metaphysical conception of ideal form estranged from its human context, and, on the other, romantic myths of the autonomous self whose individual creativity is understood as independent of the communal context. But the parallel ontological danger also applies to pragmatist interpretations of truth; for once we assume that because everything is at root collective, it is all too easy to simply invert the objectivist and romantic paradigms by excluding tout court questions of objective truth and individual subjectivity. Hence, for instance, Stanley Fish's interest in speech-act theory is reflected in his affection for the notion of "interpretive communities." That Fish can offer no explanation for such good-natured communal valuation should alert us to the ultimate anthropological nullity of his hypothesis as an explanatory category. The same logic lies behind Fish's use of speech-act theory to analyse Coriolanus (Is There a Text in this Class?, Chapter 9). For Fish, Shakespeare's hero is an example of the error of believing that the individual exists independently of the community. But what Fish neglects to comment on is precisely why such an error should occur in the first place. For us, Coriolanus is not a metaphysician guilty of constructing a false ontological model of the human, but an illustration of the real human resentment that is produced by the communal scene, which, we note, is not only a site for collective representation via the linguistic sign, but also the moment of rejection, where each individual is denied, by the collective designations of the community, the supremacy of occupying the centre of the scene of representation. The source of Coriolanus's rage--like his archaic Greek precursor, Achilles--is the resentment of being denied the communal centrality that each man assumes is his by right as "first" among warriors.(back)

11. This is also the "liberation" of the individual from the more conservative scene of public ritual, brilliantly thematized by Shakespeare in Coriolanus. Coriolanus's preference for making promises and his scorn for participating in public rituals reflects the tension between the romantic/modern belief in the uniqueness of the individual and the premodern subordination of this self to the communal context of public ritual. It would be an error to believe, however, that Coriolanus's extreme individuality is a mere ethical aberration. As our derivation of promising shows, the romantic scene requires an ethically advanced conception of the self prohibited in the collective context of public ritual, as Shakespeare's play makes clear.(back)

12. Benveniste's observation that deictic demonstratives such as "here," "there," "now," are explicit indicators of performativity is an a posteriori insight based on the declarative sentence's ability to thematize the instance of utterance which is the sole condition of the ostensive utterance. That there is a fire "there/here/now" is already implicit in the utterance "Fire!" See Benveniste, Problems in General Linguistics, Chapter 21.(back)


13. This generative model resolves most elegantly the problem--apparently intractable to speech-act theorists--of why fiction can use propositional utterances without thereby being restricted to the conditions of verifiability that constrain such utterances in ordinary language situations. From our point of view, fictionality is precisely a function of the originary deferral which renounces worldly appetite for the moment of peaceful linguistic presence. Fiction capitalizes on the aesthetic pleasure of this renunciation, for in prolonging our separation from worldly contact we gain an imagined satisfaction that is more powerful, because more total, more ideal, than its worldly counterpart in the culmination of the sign with the subsequent division of the appetitive object. Austin intuits the fundamental status of originary deferral when he comments on the impossibility of stating with absolute conviction "There are fifty people in the next room" (138). But rather than seeing this example as the inherent temporal condition of all declarative forms which must begin with independent scenes (such as the room next door) before they are (or, in the case of fiction, are not) verified, Austin stresses the immediate context of the utterance, believing that the utterance depends on the same kind of felicity conditions that characterize his prototypical performatives. What is lacking in this explanation is the realization that the felicity conditions of the declarative or constative utterance are dependent upon a notion of a sincere speaker in the way that the felicity conditions of such ostensive performatives as marrying and naming are not. To be sure, Austin is well aware of the collective felicity--as opposed to individual sincerity--conditions of his ritual performatives, but his inability to prov ide an overall model for this collective context--an originary scene--leads to the kind of paradoxes that Austin encounters with his performative understanding of statemental utterance.(back)

14. Gans makes this observation. See Originary Thinking, p.77.(back)

15. For a one-sided account that transforms Austin's analyses of speech acts into a sociopolitical theory of linguistic usage, see Sandy Petrey, Speech Acts and Literary Theory. Petrey accepts with enthusiasm Austin's emphasis on the collective context of speech acts, but, sadly, he parodies Austin's account of performative language by turning the latter's concern for appropriateness conditions into the ontological condition for all language-use as unproblematically "sociopolitical." But this is simply to beg the question of linguistic usage that interested Austin in the first place. By declaring that political context alone explains the speech act, we have not solved the problem of why language functions as it does; we have only moved from a narrower context--the specific occasion for linguistic utterance itself--to a greater one--the illimitable context of human social and political action. This interpretation simply forgoes tout court the notion of the linguistic. Thus becomes clear Petrey's interest in speech-act theory, namely, to evacuate Austin's empirical analyses of speech-act situations and turn the notion of language-context into a moral platform for general criticisms of authority and power. Such a theory does not help us explain language, but merely contributes to the veil of sacrality that continues to shroud it under a blanket of moralizing. The kernel of truth behind Petrey's (and others') sociopolitical position is that the cultural order is ubiquitous. But this only proves the strength of the originary scene as the only rigorous model that can provide substance to the inchoate criticisms of the culturalists, who assert the universality of the cultural but can provide no model of their own to account for this blanket pronouncement.(back)


Works Cited

Austin. J.L. How to Do Things with Words. 1962. 2nd ed. Cambridge: Harvard UP, 1975.

Benveniste, Emile. Problems in General Linguistics. Trans. Mary Elizabeth Meek. Coral Gables FL: U of Miami P, 1971.

Fish, Stanley. Is There a Text in this Class?: The Authority of Interpretive Communities. Cambridge: Harvard UP, 1980.

Frye, Northrop. Spiritus Mundi: Essays on Literature, Myth, and Society. Richmond Hill, Ont: Fitzhenry, 1991.

Gans, Eric. The Origin of Language: A Formal Theory of Representation. Berkeley: U of California P, 1981.

----. Originary Thinking: Elements of Generative Anthropology. Stanford: Stanford UP, 1993.

----. "Sacred Text and Secular Culture." To Honor René Girard. Saratoga: Anma Libra, 1986. 51-46.

Girard, René. Violence and the Sacred. Trans. Patrick Gregory. Baltimore: Johns Hopkins UP, 1977.

Levin, Samuel P. "Concerning What Kind of Speech Act a Poem Is." Pragmatics of Language and Literature. Ed. Teun A. van Dijk. North-Holland Series in Theoretical Linguistics 2. Amsterdam: North-Holland, 1976. 141-60.

Pavel, Thomas G. "Ontological Issues in Poetics: Speech Acts and Fictional Worlds." Journal of Aesthetics and Art Criticism. 40 (1981): 167-78.

Petrey, Sandy. Speech Acts and Literary Theory. New York: Routledge, 1990.

Searle, John R. Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge UP, 1979.

Anthropoetics subscribers may copy or download this text from the network, but its distribution or publication shall constitute an infringement of the Author's copyright.

Anthropoetics - The Electronic Journal of Generative Anthropology
Volume II, number 2 (January 1997)
ISSN 1083-7264