Speech Act is a technical term in linguistics and the philosophy of language. The contemporary use of the term goes back to John L. Austin's doctrine of locutionary, illocutionary, and perlocutionary acts. ''Speech acts'' are commonly taken to include such acts as promising, ordering, greeting, warning, inviting and congratulating.
Contents |
Speech acts can be analysed on three levels: A locutionary act, the performance of an utterance: the actual utterance and its ostensible meaning, comprising phonetic, phatic and rhetic acts corresponding to the verbal, syntactic and semantic aspects of any meaningful utterance; an illocutionary act: the semantic 'illocutionary force' of the utterance, thus its real, intended meaning (see below); and in certain cases a further perlocutionary act: its actual effect, such as persuading, convincing, scaring, enlightening, inspiring, or otherwise getting someone to do or realize something, whether intended or not (Austin 1962).
The concept of an illocutionary act is central to the concept of a speech act. Although there are numerous opinions regarding how to define 'illocutionary acts', there are some kinds of acts which are widely accepted as illocutionary, as for example promising, ordering someone, and bequeathing.
Following the usage of, for example, John R. Searle, "speech act" is often meant to refer just to the same thing as the term illocutionary act, which John L. Austin had originally introduced in How to Do Things with Words (published posthumously in 1962).
According to Austin's preliminary informal description, the idea of an "illocutionary act" can be captured by emphasising that "by saying something, we do something", as when someone issues an order to someone to go by saying "Go!", or when a minister joins two people in marriage saying, "I now pronounce you husband and wife." (Austin would eventually define the "illocutionary act" in a more exact manner.)
An interesting type of illocutionary speech act is that performed in the utterance of what Austin calls performatives, typical instances of which are "I nominate John to be President", "I sentence you to ten years' imprisonment", or "I promise to pay you back." In these typical, rather explicit cases of performative sentences, the action that the sentence describes (nominating, sentencing, promising) is performed by the utterance of the sentence itself.
Searle (1975)[1] has set up the following classification of illocutionary speech acts:
In the course of performing speech acts we ordinarily communicate with each other. The content of communication may be identical, or almost identical, with the content intended to be communicated, as when a stranger asks, "What is your name?"
However, the meaning of the linguistic means used (if ever there are linguistic means, for at least some so-called "speech acts" can be performed non-verbally) may also be different from the content intended to be communicated. One may, in appropriate circumstances, request Peter to do the dishes by just saying, "Peter ...!", or one can promise to do the dishes by saying, "Me!" One common way of performing speech acts is to use an expression which indicates one speech act, and indeed performs this act, but also performs a further speech act, which is indirect. One may, for instance, say, "Peter, can you open the window?", thereby asking Peter whether he will be able to open the window, but also requesting that he do so. Since the request is performed indirectly, by means of (directly) performing a question, it counts as an indirect speech act.
Indirect speech acts are commonly used to reject proposals and to make requests. For example, a speaker asks, "Would you like to meet me for coffee?" and another replies, "I have class." The second speaker used an indirect speech act to reject the proposal. This is indirect because the literal meaning of "I have class" does not entail any sort of rejection.
This poses a problem for linguists because it is confusing (on a rather simple approach) to see how the person who made the proposal can understand that his proposal was rejected. Following substantially an account of H. P. Grice, Searle suggests that we are able to derive meaning out of indirect speech acts by means of a cooperative process out of which we are able to derive multiple illocutions; however, the process he proposes does not seem to accurately solve the problem. Sociolinguistics has studied the social dimensions of conversations. This discipline considers the various contexts in which speech acts occur.
Searle has introduced the notion of an 'indirect speech act', which in his account is meant to be, more particularly, an indirect 'illocutionary' act. Applying a conception of such illocutionary acts according to which they are (roughly) acts of saying something with the intention of communicating with an audience, he describes indirect speech acts as follows: "In indirect speech acts the speaker communicates to the hearer more than he actually says by way of relying on their mutually shared background information, both linguistic and nonlinguistic, together with the general powers of rationality and inference on the part of the hearer." An account of such act, it follows, will require such things as an analysis of mutually shared background information about the conversation, as well as of rationality and linguistic conventions.
In connection with indirect speech acts, Searle introduces the notions of 'primary' and 'secondary' illocutionary acts. The primary illocutionary act is the indirect one, which is not literally performed. The secondary illocutionary act is the direct one, performed in the literal utterance of the sentence (Searle 178). In the example:
Here the primary illocutionary act is Y's rejection of X's suggestion, and the secondary illocutionary act is Y's statement that she is not ready to leave. By dividing the illocutionary act into two subparts, Searle is able to explain that we can understand two meanings from the same utterance all the while knowing which is the correct meaning to respond to.
With his doctrine of indirect speech acts Searle attempts to explain how it is possible that a speaker can say something and mean it, but additionally mean something else. This would be impossible, or at least it would be an improbable case, if in such a case the hearer had no chance of figuring out what the speaker means (over and above what she says and means). Searle's solution is that the hearer can figure out what the indirect speech act is meant to be, and he gives several hints as to how this might happen. For the previous example a condensed process might look like this:
Searle argues that a similar process can be applied to any indirect speech act as a model to find the primary illocutionary act (178). His proof for this argument is made by means of a series of supposed "observations" (ibid., 180-182).
In order to generalize this sketch of an indirect request, Searle proposes a program for the analysis of indirect speech act performances, whatever they are. He makes the following suggestion:
With this process, Searle concludes that he has found a method that will satisfactorily reconstruct what happens when an indirect speech act is performed.
For much of the history of linguistics and the philosophy of language, language was viewed primarily as a way of making factual assertions, and the other uses of language tended to be ignored. The work of J. L. Austin, particularly his How to Do Things with Words, led philosophers to pay more attention to the non-declarative uses of language. The terminology he introduced, especially the notions "locutionary act", "illocutionary act", and "perlocutionary act", occupied an important role in what was then to become the "study of speech acts". All of these three acts, but especially the "illocutionary act", are nowadays commonly classified as "speech acts".
Austin was by no means the first one to deal with what one could call "speech acts" in a wider sense. Earlier treatments may be found in the works of some church fathers,[2] and scholastic philosophers,[3] in the context of sacramental theology,[4] as well as Thomas Reid,[5] and Charles Sanders Peirce.[6]
Adolf Reinach (1883–1917) has been credited with a fairly comprehensive account of social acts as performative utterances dating to 1913, long before Austin and Searle. His work had little influence, however, perhaps due to his death at 33 in the German Army at the onset of war in 1914.
The term "Speech Act" had also been already used by Karl Bühler in his "Die Axiomatik der Sprachwissenschaften”, Kant-Studien 38 (1933), 43, where he discusses a Theorie der Sprechhandlungen and in his book Sprachtheorie (Jena: Fischer, 1934) where he uses "Sprechhandlung" and "Theorie der Sprechakte".
Dore (1975) proposed that children's utterances were realizations of one of nine primitive speech acts:
Computational speech act models of human-computer conversation have been developed.[7]
Speech act theory has been used to model conversations for automated classification and retrieval.[8]
Another highly-influential view of Speech Acts has been in the 'Conversation for Action' developed by Terry Winograd and Fernando Flores in their 1987 text "Understanding Computers and Cognition: A New Foundation for Design". Arguably the most important part of their analysis lies in a state-transition diagram (in Chapter 5) that Winograd and Flores claim underlies the significant illocutionary (speech act) claims of two parties attempting to coordinate action with one another (no matter whether the agents involved might be human-human, human-computer, or computer-computer).
A key part of this analysis is the contention that one dimension of the social domain- tracking the illocutionary status of the transaction (whether individual participants claim that their interests have been met, or not) is very readily conferred to a computer process- independent of whether the computer has the means to adequately represent the real world issues underlying that claim. Thus a computer instantiating the 'conversation for action' has the useful ability to model the status of the current social reality independent of any external reality on which social claims may be based.
This transactional view of speech acts has significant applications in many areas in which (human) individuals have had different roles- for instance- a patient and a physician might meet in an encounter in which the patient makes a request for treatment, the physician responds with a counter-offer involving a treatment she feels is appropriate, and the patient might respond, etc. Such a "Conversation for Action" can describe a situation in which an external observer (such as a computer or health information system) may be able to track the ILLOCUTIONARY (or Speech Act) STATUS of negotiations between the patient and physician participants even in the absence of any adequate model of the illness or proposed treatments. The key insight provided by Winograd and Flores is that the state-transition diagram representing the SOCIAL (Illocutionary) negotiation of the two parties involved is generally much, much simpler than any model representing the world in which those parties are making claims- in short- the system tracking the status of the 'conversation for action' need not be concerned with modeling all of the realities of the external world- a conversation for action is critically dependent upon certain stereotypical CLAIMS about the status of the world made by the two parties. Thus a "Conversation for Action" can be readily tracked and facilitated by a device with little or no ability to model circumstances in the real world other than the ability to register claims by specific agents about a domain.
In making useful applications of technology to domains such as healthcare, it is helpful to discriminate between problems which are very, very hard (such as deep understanding of pathophysiology as it relates to genetic and various environmental influences) and problem which are relatively easier, such as following the status of negotiations between a patient and a health care provider. Speech Act (Illocutionary) Analysis allows for a useful understanding of the status of a negotiation between (for instance) a health care provider and a patient INDEPENDENT of any well-accepted credible and comprehensive understanding of a disease process as it might apply to that patient. For this reason, systems which track the status of PROMISES and REJECTED-PROPOSALS and ACCEPTED-PROMISES can help us to understand the situations in which (human or computer) AGENTS find themselves as they attempt to fulfill ROLES involving other agents, and such systems can facilitate both human and human-computer systems in achieving role-associated goals.
Multi-agent systems sometimes use speech act labels to express the intent of an agent when it sends a message to another agent. For example the intent "inform" in the message "inform(content)" may be interpreted as a request that the receiving agent adds the item "content" to its knowledge-base; this is in contrast to the message "query(content)" which may be interpreted (depending on the semantics employed) as a request to see if the item content is currently in the receiving agents knowledge base. There are at least two standardisations of speech act labelled messaging KQML and FIPA.