In contributions published in Linguistics & Philosophy, J.R. Searle one the one hand, and K. Bach & R.M. Harnish on the other, led what was probably the most prominent 'speech-act theoretical' debate up to the present day. The issue they had was about, literally, 'how "performatives" (really) work'.
Both parties define the 'performative (utterance)' as the explicit performance of an illocutionary act; in Searle's words: "an utterance of a performative sentence token, such that the utterance constitutes the performance of the [illocutionary] act named by the performative expression in the sentence" (1989, 537). Yet their explanation of 'how' the things 'work' are very different. Bach & Harnish's explanation refers to an inference pattern which, given the performance of the illocutionary act is explicit, is available to the audience. Their example is the explicit performance of an order to leave:
(1) He is saying "I order you to leave".
(2) He is stating that he is ordering me to leave.
(3) If his statement is true, then he must be ordering me to leave.
(4) If he is ordering me to leave, it must be his utterance that constitutes the order (what else could it be?).
(5) Presumably, he is speaking the truth.
(6) Therefore, in stating that he is ordering me to leave, he is ordering me to leave. (1992, 99)
Searle's explanation of how 'performatives' work is: by means of satisfying a convention. The explanation basically assumes, among other things, the following elements:
(1) An extra-linguistic institution.
(2) A special position by the speaker, and sometimes by the hearer, within the institution.
(3) A special convention that certain literal sentences of natural languages count as the performances of certain
declarations within the institution. (1989, 548)
On Searle's account, the 'performative' works by satisfying the requirement, stated in 3, of a "certain literal sentence of natural language"; the satisfaction of this requirement entails, by virtue of the 'special convention', that the utterance counts as the performance of the illocutionary act aimed at (which Searle, who usually takes a realist stance towards the existence of 'counts-as' facts, views as sufficient for the conclusion that it actually comes into existence).
So according to Bach & HArnish, 'performatives' 'work' by virtue of the availability of an inference schema to the audience, while according to Searle 'they' basically 'work' by virtue of satisfying a convention. You will have noticed, however, that what the two parties take to be the 'working' of 'performatives' are thought of in remarkably different ways. Bach & Harnish apparently see the 'working' of 'performatives' in the achievement of understanding of what act the speaker aims at; while Searle's conception of the 'working' is the 'successful' (in the sense of 'achieved') performance of the act the speaker aims at. Do these two conceptualisations finally aim at the same thing?
They do not, and to see this leads us to the observation that in fact the two parties are talking about two completely different matters--talking perfectly at cross purposes, that means. To see the problem, let us turn back to the definition of 'performatives'. Both parties define it as the explicit performance of an 'illocutionary act'. But what definition of an 'illocutionary act' do they actually apply? As we know, Bach & Harnish have two conceptions; they one applied is that of the 'communicative illocutionary act', which amounts to (an attempt at) communication, which in turn is characterised by an intention which aims at its own recognition--on the audience's understanding, that means. Searle, in turn, at least in his classical (1969) account, targets institutional acts, acts which are constituted by a convention of the form 'x counts as y (in c)', together with the satisfaction of x in c.
So what about the two explanation of the 'working' of 'performatives'? Bach & Harnish's explanation emphasises interprets the 'working' as the achievement of communicative success (of an act aiming at communication): the fact that the success of attempts at communication is furthered by explicit devices, because these make available the the six-step inference schema to the audience. Searle, in turn, interprets 'working' as the achieved performance (of an institutional act): he emphasises that institutional acts are performed by means of satisfying the x term in a constitutive rule of the form 'x counts as y'. The two 'opposing' positions do not in the least contradict each other. The two parties have principally different conceptions of what the'working' of a 'performative' would be. Due to the different conceptions of what an 'illocutionary act' actually is, they in fact apply principally different conceptions of what a 'performative utterance' actually is.
In sum, the two parties are not giving two different answers to the question under consideration. In fact, the 'opponents' are dealing with two different questions; that means, they are perfectly talking at cross purposes. The whole debate was futile from the start, because the two parties apply different conceptions of 'illocutionary act', which leads to different conceptions of 'performative utterances', which leads to different conceptions of what the 'working' of 'these' phenomena would be, which leads to two completely different tasks. Like most other speech act theorists, Bach, Harnish and Searle turn out to in fact be unable to keep track of the diversity of the terms their use of 'illocutionary act'.
For a more detailed representation, see this draft.
Still under development