Page 485 - Encyclopedia of Philosophy of Language

so the one postulated is abstract. Now, abstract verbs are figments of the linguist's imagination: by defi- nition, they have no surface manifestation and must be argued for on circumstantial evidence. Having complicated the grammar by introducing an abstract verb into deep structure, the linguist is then forced to complicate the grammar yet further, by postulating a deletion rule to remove the abstract verb from the phrase marker at some point in the progression from deep to surface structure. For the case in hand, Fraser (1974) showed that a performative deletion trans- formation would be difficult to define—and, in fact, it never was defined.
Ross (1970) presented 14 arguments based on almost unrelated bits of data in support of the hypoth- esis that the highest clause of every declarative sen- tence is a performative verb of stating. Every one of Ross's arguments has been assuredly refuted (along with additional ones in G. Lakoff 1972 and Sadock 1974) by Anderson (1971), Fraser (1974), Harnish (1975), and Allan (1986:258ff.). Despite the wealth of evidence to contradict Ross's hypothesis, it had tremendous impact when it appeared and was widely accepted. Although some of the more pernicious errors in Ross's formulation were later rectified, and his conception of deep structure was updated by, for example, G. Lakoff and Sadock, the idea that illo- cutionary force can be accounted for as one of the higher clauses in the deepest phrase-marker under- lying a sentence persisted throughout the 1970s. The question arises whether the fault lies with particular versions of the theory, or with the theory itself. Whereas the inferential theory of speech acts is a the- ory of H's interpretation of S's utterance U to H in context C, the performative analysis theory is a theory of the illocutionary potential of a given sentence. The inferential theory is a pragmatic theory that invokes H's knowledge of language L, of the use of L, of the context C, and also S's general background knowl- edge. For the performative analysis theory, all the information contained within the inference schemas of the inferential theory must somehow be located within the deepest phrase marker of a trans- formational grammar, and rules have to be specified for converting such deep structures into a surface structure equivalent to the locution. For two main reasons, this is an aim that cannot be achieved. Very briefly, such a quantity of information cannot prop- erly be accommodated within recursive structures of the kind generated by rewrite rules such as 'Sen- tence-^ NP VP,' 'VP->V (NP) Sentence' (nor the more recent versions of Chomsky-inspired grammar). Even if it could, it is very doubtful whether such deep structures could be converted into proper surface structures matching the locution, because of diffi- culties with lexical insertion and the assignment of prosody. One is forced to conclude that as a theory of speech acts, performative analysis was always a lame duck.
13. Toward a Formal Theory of Speech Acts
Linguistics has adopted two trends from language philosophy: the Fregean school of formal languages, which has developed the model-theoretic truth-con- ditional systems associated with the names of Mon- tague and Cresswell; and the notably informal school of ordinary language philosophy (see Sect. 2 above), which developed speech act theory. The formalization of speech act theory bridges the gap between the two schools. The program is, in effect, to extend the Fre- gean tradition's formal semantics of sentences by adding a formal theory of illocutionary types together with a characterization of illocutionary success (ident- ifying particular illocutions) and satisfactory cor- respondence between U and states of affairs in the world spoken of.
Katz (1977) embarked on such a program using the 'semantic markerese' of his semantic theory (Katz 1972). His theory of prepositional types sets out to describe the illocutionary potential of sentences as part of their sense. He uses the terminology of his semantic theory (Katz 1972), which purportedly models the ideal speaker-hearer's competence (cf. Chomsky 1965:4). According to Katz (1977:24):
The prepositional type of a sentence (on a sense) is the information that determines the type of speech act that a token performs in a null context.
The prepositional content of a sentence (on a sense) is the information that determines the particular speech act (within the categories specified by its prepositional type, and subtypes) a token performs in a null context.
. . . [Although for convenience we speak about a token of a sentence type performing a speech act, this is to be understood to mean that the speaker performs the act in the use of the token.
Strictly speaking, Katz does not, therefore, offer a theory of speech acts: speech acts are quintessential^ pragmatic events, and Katz eschews pragmatics, rele- gating it to an undescribed theory of performance (cf. Katz 1977:16).
Katz champions Austin's distinction between 'assertives' (Austin's constatives) and performatives, which he construes as a difference in prepositional type. He therefore distinguishes sharply between sen- tences with explicit performatives where the illocution is signaled in the semantics of the performative verb, for example, I state you will go; I request that you tell me whether or not you will go; I order you to go, and those without, for example, You will go; Will you go? Go! For Katz, the proposition is defined on the semantic content of the immediate constituents of sen- tence node in the lexically specified underlying phrase marker, and in particular the verb. In addition, Katz specifies certain conditions on the proposition to determine its illocutionary success. All types of prop- osition, 'assertive,' and performative must satisfy the denotation condition. This requires that each noun phrase must pick out the intended number of objects
Speech Act Theory: Overview
463