Hie question of performative utterances becomes clearer if one sees it as a particular case of the effects of symbolic domination, which occurs in all linguistic exchanges. The linguistic relation of power is never defined solely by the relation between the linguistic competences present. And the weight of different agents depends on their symbolic capital, i.e. on the recognition, institutionalized or not. that they receive from a group. Symbolic imposition - that kind of magical efficacy which not only the command and the password, but also ritual discourse or a simple injunction, or even threats or insults, purport to exercise - can function only if there is a convergence of social conditions which are altogether distinct from the strictly linguistic logic of discourse. For the philosopher’s language to be granted the importance it claims, there has to be a convergence of the social conditions which enable it to secure from others a recognition of the importance which it attributes to itself? Equally, the setting up of a ritual exchange, such as a mass, presupposes, among other things, that all the social conditions are in place to
ensure (he production of appropriate senders and receivers, who are therefore agreed among themselves. It is certainly the case that the symbolic efficacy of religious language is threatened when the set of mechanisms capable of ensuring the reproduction of the relationship of recognition, which is the basis of its authority, ceases to function. This is also true of any relation of symbolic imposition, even of the one implied by the use of the legitimate language which, as such, involves the claim to be heard, believed and obeyed, and which can exercise its specific efficacy only as long as it can count on the effectiveness of all the mechanisms, analysed above, which secure the reproduction of the dominant language and the recognition of its legitimacy. One may note, in passing, that the source of the profit of distinction, procured by any use of the legitimate language, derives from the totality of the social universe and the relations of domination that give structure to it. although one of the most important constituents of this profit lies in the fact that it appears to be based on the qualities of the person alone.
Austin's account of performative utterances cannot be restricted to the sphere of linguistics. The magical efficacy of these acts of institution is inseparable from the existence of an institution defining the conditions (regarding the agent, the time or place, etc.) which have to be fulfilled for the magic of words to operate. As is indicated in the examples analysed by Austin, these ‘conditions of felicity’ are sociaJ conditions, and the person who wishes to proceed felicitously with the christening of a ship or of a person must be entitled to do so, in the same way that, to be able to give an order, one must have a recognized authority over the recipient of that order. It is true that linguists have often rushed to find, in Austin’s inconsistent definition ol the performative, an excuse for dismissing the problem which Austin had set them, in order to return to a narrowly linguistic definition that ignores the market effect. They did this by distinguishing between explicit performatives, which are necessarily selfverifying since they represent in themselves the accomplishment of the act, and performatives conceived more broadly to mean statements that are used to accomplish an act other than the simple fact of saying something, or. to pul it more simply, the difference between a properly linguistic act (e.g. declaring the meeting open) and the extra-linguistic act (opening the meeting by the fact of declaring it open). In this way, they justified to themselves the rejection of any analysis of the social conditions in which performative utterances function. The conditions of felicity discussed by Austin concern only the extra-linguistic act; only to open the meeting effectively does one
need to be entitled to do so, as anyone can declare it open, even if his declaration remains totally ineffective.5
