1
Speech Acts and Intentions: The Things We Do with Words
People do things with their words. They order and promise and criticize and apologize and so on. In other words, to use language is to perform an action. As obvious as this is, an action view of language has had a relatively short history, and many of the implications of viewing language use as action have yet to be pursued in any detail. Moreover, viewing language as action, especially as a social action, makes clear many of the social psychological underpinnings of language use.
There is a strain of thought in philosophy, developed primarily at Oxford University beginning in the 1930s, that is concerned with the analysis of the ordinary use of language. Two related lines of research within this tradition are of particular importance, and I will use them as the starting points for this chapter. The first, termed speech act theory, was developed by John Austin and later elaborated on by John Searle. Speech act theory explicitly conceptualizes linguistic meaning as use; what we mean with an utterance is the use to which the utterance is put. In the first half of the chapter I describe this approach and the issues that it raises for understanding language use. The second strand of thought revolves around the concept of conversational implicature, an approach to nonliteral meaning that was developed by Paul Grice. Griceâs views have been extremely influential and a central concern within the field of pragmatics. In the second half of this chapter I examine his ideas, placing particular emphasis on the recognition of speaker intention in general and illocutionary force in particular. Much of this discussion will be centered on the compre- hension of indirect speech acts and the various processing models that have been proposed and tested.
JOHN AUSTIN AND THE DEVELOPMENT OF SPEECH ACT THEORY
In the early part of the last century there existed a major theoretical approach to languageâtermed logical positivismâthat claimed all utterances were to be evaluated exclusively on the basis of their verifiability. In this view, if the truth of an utterance could not be determined, the utterance was viewed as meaningless. The intent of logical positivism, of course, was to eliminate the imprecision inherent in human languages. But the outcome of the approach was extremeâmany everyday conversational utterances were viewed as simply meaningless. A truth-conditional approach to language (without the requirement of verifiability) exists today in the various logical approaches to language, particularly those attempting to develop a formal semantics (e.g., Kaplan, 1979).
It is against the backdrop of logical positivism that the development of speech act theory should be viewed. Both Ludwig Wittgenstein (1953) and John Austin (1962)1 independently, and almost simultaneously, proposed action- oriented approaches to language that clearly articulated the weaknesses inherent in the truth-conditional view. I give priority to Austinâs work here because, as the developer of speech act theory, his work continues to have a specific relevance for language theorizing.
Austinâs (1962) speech act theory arose from his observation that it simply is not possible to determine the truth value of many utterances. For example, the truth of the utterance âI promise to do it tonightâ cannot be determined. The utterance has no relationship with the external world, and so truth conditions cannot be established. As we will see in moment, this utterance can be evaluated on many different dimensionsâit may or may not be sincere, for exampleâbut attempts to ascertain its veracity are useless; it is neither true nor false.
This led Austin to propose a (preliminary) distinction between performative utterances and constative utterances. Constatives are utterances for which a truth value conceivably could be determined. Thus, one could ascertain the truth of the utterance âItâs raining outâ by looking out the window. Performatives (e.g., âI apologizeâ), on the other hand, are used in order to perform some act (their occurrence changes the world in some way), and hence they are not amenable to a truth-conditional analysis. Although one cannot determine the truth value of performatives, there are various ways in which they can go wrong, or to use Austinâs term, be infelicitous. For example, if I utter the performative âI declare war on Canada,â I will fail to substantially alter the world. My remark will have no effect (it misfires according to Austin) because I have no authority to declare war. My utterance, although neither true nor false, is clearly infelicitous.
Austin proposed three sets of conditions required for the felicitous perfor- mance of performatives. First, there must be a conventional procedure perfor- med by an appropriate person in an appropriate context that has a conventional effect. For example, a minister can perform a marriage by uttering, in the appropriate context, âI now pronounce you man and wife.â Second, the proce- dure must be executed correctly and completely. Third, the person performing the act must have the requisite thoughts, feelings, or intentions (e.g., to perform a felicitous promise the speaker must intend to perform the promised act).
Performatives, then, are a class of utterances that are conventionally used as a means of performing certain actions. They can be either successful or not but are neither true or false. Constatives, on the other hand, report or describe the world and hence should be amenable to a truth-conditional analysis. But it is Austinâs fundamental insight that these acts, too, are subject to felicity conditions. To assert something felicitously requires, among other things, that the speaker sincerely believe that what is being asserted is true. To perform an assertion is to take a particular stance regarding the nature of the world. If I say to someone, âItâs raining out,â I commit myself to that proposition; I have taken a specific stand and performed a specific action. So even assertions have a performative feature. On the basis of these and other problems, Austin abandoned the performative-constative distinction in favor of a theory of illocutionary forces or speech acts.
On this view, all speech acts have a dimension of meaning (or propositional content) and a particular force. In other words, one is doing something with oneâs words. But what exactly is one doing? In Austinâs speech act theory, any utterance involves the simultaneous performance of a number of different acts. First, one is performing a locutionary act. That is, one is making certain sounds (a phonetic act) that comprise identifiable words that are arranged on the basis of a particular grammar (a phatic act), having a certain sense and reference (a rhetic act). When John says to Mark, âI will do it tonight,â the sound produced by John will be recognized by Mark as words with a particular meaning and a particular reference (e.g., âIâ refers to John, the speaker). In a sense, the locutionary act involves the dimensions of language (phonetics, syntax, and semantics) with which linguists have traditionally been concerned.
In addition to the locutionary act, the speaker is also performing a particular âact in saying,â or what Austin termed an illocutionary act. The illocutionary act is the conventional force associated with the uttering of the words in a particular context. Thus, Johnâs utteranceââI promise to do it tonightââwill have the force of a promise (if performed felicitously).
Finally, a speaker is simultaneously performing what Austin termed a perlo- cutionary act. The perlocutionary act refers to the effects the utterance has on the hearer. For example, the effects of Johnâs verbal promise may include Markâs recognition that John is making a promise, his relief that he has done so, his hope that he will keep it, and so on. Perlocutionary acts are hearer based. Although obviously prompted by the speakerâs utterance, the effects may extend well beyond anything intended by the speaker. Thus, perlocutionary acts are indeterminate and not necessarily intentional.
It should be clear from this brief review that Austinâs speech act theory, together with the latter Wittgenstein (1953), marked the beginning of a new way to view language. The emphasis on language as action rather than as an abstract system for describing reality marked a fundamental shift. It raised new issues and posed new questions, and although Austin dealt with some of these issues, they have been developed further by other theorists. It is to those issues and theorists that we now turn.
JOHN SEARLE: SPEECH ACT TAXONOMY AND FELICITY CONDITIONS
Beginning with his dissertation, Searle (1969) systematized and extended speech act theory in several directions. For the present discussion, his most important contributions include his specification of illocutionary force via the felicity conditions, his taxonomy of speech acts, and the notion of indirect speech acts. Each of these issues are considered in turn.
Maintaining a view of language use as action, Searle (1969) argued that the conditions for the felicitous performance (in Austinâs sense) of a speech act can be viewed as constituting the performance of that speech act. That is, the felicity conditions for a particular speech act must be met for that act to be performed successfully, and the meeting of these conditions constitutes the performance of that speech act. Moreover, speech acts differ in terms of the specifics of the conditions underlying their performance, and hence these conditions (in conjunction with some others to be discussed below) serve as a framework for comparing different speech acts.
The original felicity conditions proposed by Searle (1969) have been modified only slightly over the years, and other important components of illocutionary force have been proposed. These original four conditions, however, remain central to understanding the nature of what it means to perform a particular speech act. Note that there are other conditions that are not unique to speech act performance that must be met also (e.g., both speaker and hearer must speak the same language). A summary of these conditions, with illustra- tions for the act of requesting, is presented in Table 1.1.
Condition | Example for Request |
Prepositional Content | Predicates future act by the hearer |
Preparatory | Act has not yet been performed Speaker believes hearer is able to perform act |
Sincerity | Speaker sincerely wants hearer to perform act |
Essential | Utterance counts as an attempt to get the hearer to perform act |
- 1. Prepositional Content: A fundamental feature of speech act theory is the conceptualization of an utterance as having both a propositional meaning and an illocutionary force (Austinâs locutionary and illocutionary acts). Logically, speech acts are often represented as having the form F(p), where F is the illocutionary force and p the propositional content. Although it is possible for two utterances to have the same force yet different propositional meaning, or the same meaning but different forces, force and propositional content are not completely independent. The propositional content condition states that certain illocutionary forces specify what is acceptable in terms of propositional content. For example, promises and requests require the specification of future courses of action, and hence âI promise to study hard last nightâ and âPlease, study hard last nightâ make little sense.
- 2. Preparatory Condition: For any particular speech act, there are one or more conditions, usually concerning the beliefs and desires of the interlocutors, that are presupposed in the (felicitous) performance of the act. To felicitously perform a request, a speaker must believe that the hearer has the ability to perform the requested act and that the hearer would not perform the act unless requested to do so. The felicitous performance of an assertion requires the speaker to believe the truth of the proposition and that the hearer is not aware of the proposition. An apology requires the belief that the act for which one is apologizing is bad. And so on. In some instances the preparatory conditions specify the nature of the social relations existing between interlocutors. For example, to order someone to do something requires the speaker to have power over the hearer.
- 3. Sincerity Condition: In performing a particular speech act, the speaker expresses a certain psychological attitude regarding the propositional content of the utterance. For example, in performing a promise a speaker expresses an intention to do the act that is being promised. A person can make a promise with no intention of keeping it, of course, but such an utterance would not be felicitous. The sincerity condition, then, is a requisite psychological state the speaker must have in order to perform a particular speech act; it is the specification of a particular relationship between the speakerâs mental state and the prepositional content of his words.
- 4. Essential Condition: The uttering of a particular expression in a specific context âcounts asâ the performance of a specifiable act; the essential condition specifies the particular illocutionary point of an utterance. In most contexts, the uttering of âPlease close the doorâ counts as a request for the hearer to shut the door.
Now, the essential condition states that an utterance in context will have a conventionally recognized illocutionary point, and according to Searle (1979), there are five basic, primitive illocutionary points. These five points are exhaustive, and mutually exclusive, and they are derived from a consideration of the possible relations between oneâs words and the world (as it is or could be). This represents an important attempt to classify, in a systematic manner, actions that speakers can perform with their utterances. Although other taxonomies have been proposed (e.g., Bach & Harnish, 1979), Searleâs is probably the most well known. His taxonomy is in the following list and summarized in Table 1.2.
- 1. Directives: A directive counts as an attempt to get the hearer to perform some future action. Prototypes include requesting, ordering, and questioning. With these speech acts, a speaker is attempting to alter the world in some way with words. Hence, directives represent a world-to-words fit; the speaker is attemp- ting to bring the world in line with words.
- 2. Assertives: An assertive counts as an attempt to represent an actual state of affairs, to commit the speaker to something being the case. Prototypes include asserting, concluding, informing, predicting, and reporting. With assertives a speaker is attempting to depict the nature of the world. Hence, rather than attempting to get the world to match oneâs words (the world-to- words fit of directives), one is attempting to get oneâs words to match the world; assertives represent a words-to-world fit.
- 3. Commissives: A commissive counts as an attempt to com- mit the speaker to a future course of action. Prototypes include warning, promising, threatening, and guaranteeing. As with
Illocutionary Point | Direction of Fit | Examples |
Directive | World-to-words (hearer) | Request, order |
Assertive | Words-to-world | Conclude, predict |
Commissive | World-to-words (speaker) | Promise, warn |
Declarative | World-to-words and Words-to-world | Declare war Perform marriage |
Expressive | Null | Thank, apologize |
- directives, the speaker of a commissive is attempting to alter the world in some way; commissives thus reflect a world-to-words fit. Unlike directives, however, it is the speakerâs (rather than the hearerâs) subsequent actions that will alter the world.
- 4. Declaratives: A declarative counts as an attempt to bring about a change in some institutional state of affairs. Prototypes include declaring war, performing a marriage, and calling a base runner âout.â For a declarative, the relationship between the world and a speakerâs words is bidirectional; declaratives have a double direction of fit (both words-to-world and world-to- words). The point of a declarative (e.g., declaring war) is to alter the state of the world (world-to-words) by stating that the prepositional content matches the state of th...