When we speak of the classification of illocutionary acts we consider illocutionary acts as types (as opposed to tokens): we are looking for kinds, or groups, of types. In principle, a classification of illocutionary acts (act types) does not necessarily correspond to a classification of sentences. However, the kinds of sentences used for performing the various kinds of illocutionary acts, as well as the kinds of verbs used for performing illocutionary acts explicitly, were often taken into consideration as relevant to the classification of the latter.
According to Austin, there are at least as many illocutionary acts (act types) as there are verbs that can be used performatively. Thus, Austin based his classification of illocutionary acts on a list of verbs that he took to be performatives. His classes are fuzzy sets allowing for overlaps, and are characterized by the intuitive descriptions of some salient features of the procedures in which their prototypical members consist (1962:151 -163).
Searle aimed at a neater subdivision of illocutionary acts into classes. As criteria of classification, he selected three dimensions of the illocutionary act:
· the point or purpose of the act, expressed in its essential condition;
· the direction of fit, i.e., whether the words (or more precisely, the details of their truth-conditional meaning or expressed propositional content) have to match the world, or the world has to match the words;
· the expressed psychological states, i.e the speaker’s psychological attitudes with respect to the propositional content, which satisfy the sincerity condition of the illocutionary act (1979: 2-5).
He also linked each of his classes to a standard deep structure of the sentences used (1979:20-27).
Other attempts at classification can be considered as reformulations or refinements either of Austin’s classification or of Searle's (Vendler 1972; Wunderlich 1976: Bach & Harnish 1979; Sbisa 1984). Some proposals are attentive to linguistic facts such as speech act verbs (Ballmer & Brennenstuhl 1981), sentence types (Croft 1994), modal verbs (Zaefferer 2001). Weigand has elaborated a classification in terms of dialogic action games (see Weigand 1994). Searle’s classification of illocutionary acts has been by far the most influential one and has often been taken as a basis for the further investigation of particular areas. Recently, some attention has been paid to Austinian categories, such as that of exercitives, in connection with socially relevant issues (see e g., McGowan 2003, Sbisa 2006). (For a critique of classification attempts in general, see Verschueren 1983 and 1985.)
Modes of understanding
Are illocutionary forces understood by virtue of the semantics of their linguistic indicators or by means of pragmatically invited inferences?
Illocutionary force occupies an ambiguous position between semantics and pragmatics, it could be considered as a purely semantic phenomenon, wholly dependent on the codified meaning of words, only if it were possible to assign illocutionary forces to speech acts on the sole basis of the linguistic indicating device (or set of indicating devices). But this is not the case. The presence of clear cut indicators in the uttered sentence does not by itself determine the actual, serious and felicitous performance of the speech act (cf. Davidson 1979). Is then illocutionary force wholly pragmatic? This solution would involve a minimization of the contribution of linguistic illocutionary indicators to the understanding of illocutionary force.
However, the proposal has also been made to admit of different modes of understanding for the so-called direct and respectively 'indirect' speech acts. While direct speech acts display appropriate illocutionary indicators, indirect speech acts are performed in uttering sentences which do not contain indicators of their intended force, so that the hearer has to understand such force by inference (Searle 1 975). Strategies for performing and understanding indirect speech acts have been related to politeness phenomena (Brown & Levinson 1987) and to different socio-cultural environments (Blum-Kulka, House & Kasper Eds. 1989).
It should be noted that the notion of an indirect speech act relies on Grice's theory of implicature and is therefore liable to be rejected by those who do not accept that theory. Indirect speech acts may then be traced back again to some kind of convention, script or schema.
Speech acts and truth
There is a tendency in philosophy to draw a distinction between assertive or descriptive language on the one hand, and all the uses of language that are not true or false on the other. There is another tendency, in philosophy and in particular in logic, to consider sentences as having truth values quite apart from their actually being uttered in a context.
Speech act theory proposes a different perspective, according to which assertions are speech acts just as well as orders. promises, apologies, appointments, and no sentence as such can be said to be either true or false. The issue of truth or falsity can arise only when a sentence is used in performing an assertive speech act. However, this perspective is not without problems.
First of all, is speech act felicity a precondition for the truth/falsity assessment (as Austin 1962 and Strawson 1950 put it) or is it a mere matter of appropriateness, while the truth/falsity assessment independently relies on truth conditions (as Grice 1975 would say)? Secondly, what exactly is it that we are calling true/false: the whole assertive speech act, or a locutionary or propositional component of it? Although the debate about these topics in philosophy cannot be considered as settled once and for all, one widely shared view is that what is deemed true or false is the propositional content of an assertive speech act (Searle 1968; Strawson 1973). Contextualis
t developments of this view have stressed that the proposition to be evaluated is not determined by the sentence uttered alone, but depends on many types of information provided by the situational or the cognitive context (see e.g .Travis 2000; for criticism, Cappelen & Lepore 2005).
A further related problem is whether there are assessments of non-assertive speech acts related to the correspondence to facts and thus parallel to the truth/falsity assessment. According to Austin, there are ways in which we relate non-assertive speech acts to facts in an "objective assessment of the accomplished utterance''; e.g., a piece of advice can be good or bad (Austin 1962: 141-42). Searle (1976) tackled the issue in a different way by distinguishing the two main ''directions of fit', from world to words and from words to world: in the case (for example) of an order, what corresponds to the truth of an assertive is obedience.