Hereby explained: an event-based account of performative utterances

Regine Eckardt 0 0 R. Eckardt (&) University of Gottingen , Gottingen, Germany Several authors propose that performative speech acts are self-guaranteeing due to their self-referential nature (Searle 1989; Jary 2007). The present paper offers an analysis of self-referentiality in terms of truth conditional semantics, making use of Davidsonian events. I propose that hereby can denote the ongoing act of information transfer (more mundanely, the utterance) which thereby enters the meaning of the sentence. The analysis will be extended to cover self-referential sentences without the adverb hereby. While self-referentiality can be integrated in ordinary truth conditional semantic analysis without being a mystery, the resulting account shows that self-referentiality in this sense is a necessary but not a sufficient condition for performative utterances. I propose that the second ingredient of performative utterances consists in an act of the speaker defining her utterance to be an act of the respective kind. The final theory can successfully predict the performativity, or lack thereof, of a wide range of performative sentences, and leads to an explicated interface between compositional sentence meaning and speech act. - (1) Peter promised Susan to come and see her. (2) I promise you to come and see you. The descriptive sense of (2) shows up, for instance, when the sentence is used to describe habitual promises, like in Whenever I make a phone call to you, I promise you to come and see you (a variant of Searle 1989, p. 538; also Jary 2007). A simple check of verbs, tense, signal words or other linguistic parts of an utterance alone can not predict whether or not an assertion is performative. Still, we would say, there is something about the meaning of (2) that turns it into an eligible candidate for performative utterances that (1) lacks. I will adopt Searles terminology (1989) and call verbs like promise performative verbs. I use performative sentence for sentences like (2) which can potentially be used to issue a speech act. If an utterance is indeed a speech act of the respective kind (e.g. a promise), it is called a performative utterance. Examples like (1) and (2) give rise to a number of questions concerning the interaction between semantics and speech act. Do performative sentences denote propositions when they are used in a performative utterance? Specifically, truth conditional semantics offers an explicit way to compose such propositions. Is any of these the denotation of the sentence, if used in a performative utterance? Are performative acts primarily statements of a proposition, which then bring about the performative act (as claimed e.g. by Bach and Harnish 1992)? Or is there a primary performative act? How does it relate to the proposition? How can the very same sentence, with the same meaning, figure sometimes as a statement, and sometimes as a performative utterance? Several authors propose that self-referentiality might be the key to understanding performativity (Jary 2007). Searle (1989) assumes that performative utterances are statements about themselves, and their logical structure is made explicit by the use of adverbial hereby. He proposes that self-referentiality, together with a declarative act which establishes new linguistic facts, suffices to derive performativity. At the outset of the paper, Searle promises that his attempt to understand how performatives work is not just a fussy exercise in linguistic analysis but the key to understand human communication. In this paper, I will add more fussy exercise in linguistic analysis to Searles picture and spell out more precisely how the truth conditional denotation of sentences helps to derive their performative use, a project that has been formulated as early as Szabolcsi (1982, p. 531). I propose a Davidsonian analysis of performative verbs which assumes that they take an event argument. Utterances are events, and hence we can explicate in what sense a performative utterance e talks about itself. The second ingredient to performativity consists in what I will call definition. In making a performative utterance, the speaker in (2) not only describes her utterance e as a PROMISE but expresses the intention that e be an instance of this kind: I define my utterance to be a promise. This second ingredient obviously corresponds to Searles declaration. However, his term caused substantial confusion and I believe that a new label will benefit the account.1 Even though my truth conditional analysis takes up proposals of Searle (1989), it substantially extends range and quality of his theory. First, his prose rendering of the analysis was still open for interpretation. For instance, the version of his theory that Bach and Harnish (1992) reconstruct and criticize can perhaps be read into the formulations in Searle (1989). However, I do not think that their interpretation does justice to the original proposal. I specify a fully explicit syntaxsemantics interface for performative utterances which should exclude many misunderstandings that hindered reception of Searles account. Second, Searle in 1989 could not afford himself of semantic operations at LF which are standardly used in contemporary semantic analysis. For instance, we know that arguments of the verb can be saturated by several processes like existential binding, indexical instantiation in context, or saturation by an explicit syntactic constituent. We will see how such choices can explain different uses of performative sentences. The choice between various possible LFs of performative sentences does not depend on lexical ambiguity of the performative verb, unlike what Bach and Harnish assume in their critique. Third, it can be shown that the analyses by Searle, and by Bach and Harnish are mutually consistent parts of one overarching analysis. There is independent linguistic evidence in favour of two types of logical structure that achieve selfreferentiality. One of these structures shows all properties that seem to underlie Searles work in Searle (1989). The other structure, however, fits Bach and Harnishs characterisation of how performatives come about. The two logical structures come about by two different ways to instantiate the event argument of the performative verb: existential closure in one case, and instantiation by utterance e in the other. Neither Searle nor Bach and Harnish took such operations into account. Consequently, they perceived their analyses as mutually exclusive, and most likely one correct and the other incorrect. As an exciting consequence of the present fussy exercise in linguistic analysis, it turns out that both analyses are possible, and both warranted by linguistic data. For some philosophers, it seems dangerous to attempt to derive self-referentiality from the truth conditional meaning of an utterance. Bach and Harnish take a very reserved stand with respect to such an enterprise. To suppose that the self-refe (...truncated)