How to Do Things with Words is a collection of J. L. Austin’s William James Lectures at Harvard in 1955; it was first published posthumously in 1962. Austin’s views on the matters discussed in the lectures were first formed in 1939, and he made some use of them in his address, “Other Minds,” to the Aristotelian Society in 1946. Between 1952 and 1959, he lectured on the same topic, sometimes under the title “Words and Deeds.” Much of Austin’s philosophical reputation rests upon his incisive and acerbic criticism of the views of other philosophers—for example, his withering attacks on the views of A. J. Ayer and G. J. Warnock in Sense and Sensibilia (1962). The lectures in How to Do Things with Words, however, are constructive philosophy. Austin invented speech act theory, and his theory has been used, revised, and extended not only by philosophers but also by linguists, linguistic anthropologists and sociologists, cognitive psychologists, and speech communication theorists.

Austin begins his lectures in a remarkably modest way: “What I shall have to say here is neither difficult nor contentious; the only merit I should like to claim for it is that of being true, at least in parts.” He then recounts, with approval, attempts to recognize that some so-called statements are strictly nonsense and to determine why they are nonsense. He also lauds the discovery that some statements do not purport to state facts but aim to evince emotion or to prescribe or otherwise influence conduct. These efforts and discoveries have developed piecemeal, he thinks, but also amount to a revolution in philosophy, about which he says, “If anyone wishes to call it the greatest and most salutary in its history, this is not, if you come to think of it, a large claim.” What he proposes is a theory that describes the utterances that masquerade as statements. He calls such utterances “performatives.”


Performatives have two characteristics: First, they do not describe or “constate” anything at all and are not true or false; second, to utter the performative sentence is not merely to say something. Austin’s first examples of performatives are “I do,” uttered by a bride or groom; “I name this ship the Queen Elizabeth,” uttered by someone smashing a bottle of champagne against the bow of a vessel; “I give and bequeath my watch to my brother,” as occurring in a will; and “I bet you sixpence it will rain tomorrow.” Based upon these examples, it might be tempting to think that to say the right words is the same as to do the action at issue. However, that is not correct. In general, the words have the proper effect only if uttered in appropriate circumstances, and only if the participants are doing certain other physical or mental things—for example, breaking the bottle of champagne. Further, for some acts, words are not necessary at all. Marrying might be accomplished by cohabiting and betting accomplished by inserting a coin into a slot machine.

Austin’s examples of performatives are sufficient to prove that there is some distinction to be drawn between them and constatives: “But now how, as philosophers, are we to proceed? One thing we might go on to do, of course, is to take it all back: Another would be to bog, by logical stages, down. But all this must take time.”

Constatives are true or false; however, performatives are not—instead, because they are types of actions, they can be done well or badly. Austin, in his doctrine of infelicities, concentrates on how they can be performed badly; for one way to learn how a machine works is to see in what ways it can break. As a kind of action, performatives are subject to all the defects that any action is; as linguistic acts, they have some special problems. Without pretending that the list is...

(The entire section is 779 words.)

Explicit Performatives

These matters raise doubts about the performative/constative distinction. Is there a way to make the distinction in grammatical terms, by grammatical criteria? Many, but not all, performatives have their main verb in the first-person-singular, present tense, active, indicative mood, but “You are hereby authorized . . . ,” “Passengers are warned . . . ,” “Notice is hereby given . . . ,” and “Turn right” are exceptions. Thus, neither person, number, tense, voice, nor mood can be used as a simple criterion. The first-person, active, present tense remains, however, an attractive base upon which to build a criterion. An asymmetry exists between a performative verb in this form and the same verb in other persons, tenses, and moods. If one utters “I had bet,” “He bets,” or “They (might) have bet,” one describes a certain action; but no action is described if one utters the words “I bet.” Rather, to say “I bet” (in the right circumstances, frame of mind, and so on) is, roughly, to bet. Austin’s strategy for devising a criterion is, then, to make a list of verbs having this asymmetry and to “reduce” other performative utterances to this form, which Austin calls “explicit performative” form.

Explicit performatives should be considered a development of language that evolves out of “primary performatives,” which are vague and less explicit because they serve more than one purpose. “I will,” in contrast with “I...

(The entire section is 433 words.)

The Locutionary Act

Up to this point Austin has been contrasting saying and doing. A new approach is required, one that focuses on the senses in which saying can be doing. Austin notices that every case of saying something, in the full sense, what he calls the “locutionary act,” is a case of doing something. Every locutionary act consists of a phonetic act, a phatic act, and a rhetic act. The phonetic act is the act of merely uttering noises; a parrot is capable of performing a phonetic act. The phatic act is the act of uttering certain words in a grammatical sequence, that is, noises that belong to a language, and of uttering those words as belonging to a language. The as requirement is important; a parrot utters words but because it is not aware of them as words or as having a meaning, it does not perform a phatic act. The rhetic act is the act of uttering the words with a more or less definite sense and a reference. The terms “sense” and “reference” are those of German philosopher Gottlob Frege, but the doctrine is Austin’s. For Frege, all meaningful words have both a sense and reference; for Austin, reference belongs to words that are correlated to objects by “demonstrative conventions”; sense belongs to those words that are correlated to general things by “descriptive conventions.”

The difference between the phatic act and the rhetic act is brought out by the different ways of reporting them. A phatic act is reported by direct quotation: He said, “The cat is on the mat.” A rhetic act is reported by indirect quotation: He said that the cat is on the mat. The difference is critical. One who reports a phatic act is claiming, in effect, to be offering a verbatim report of the speaker’s words and is not committed to the proposition that its speaker had achieved any reference; there might have been no cat to which to refer. A person who reports a rhetic act is not claiming that its speaker used the very words in which the report is cast; the speaker might have said, “The feline pet is lying upon the fabric used for protecting the floor.” The person is committed to the proposition that the speaker’s words had a definite sense and reference.

Illocutionary and Perlocutionary Acts

To report a rhetic act is not to report a speech act fully, for such a report leaves out the force of the utterance. Was his saying that I was to go to the store an order or merely advice or a suggestion? The force of a speech act is its “illocutionary force”; the act is an illocutionary act. The illocutionary act is governed by and conforms to conventions, and it should not be confused with something else that is done in a speech act, a perlocutionary act. A perlocutionary act is an act that produces certain effects on the feelings, thoughts, or actions of the audience, or even the speaker, as a consequence of the illocutionary act. These effects are natural consequences and not conventional ones, such as follow illocutionary acts. Although it is only a rough linguistic guide, people commonly report illocutionary acts as things done in saying something and perlocutionary acts as things done by saying something. In saying it, one warns another person (illocutionary act); by saying it, one persuades another person (perlocutionary act). These linguistic formulas do not, however, yield a criterion. By saying something, a person might have been joking or insinuating, but joking and insinuating are not perlocutionary acts. And in saying something, a person might have made a mistake, but making a mistake is not an illocutionary act.


The original contrast between performatives and constatives was a false dichotomy. Illocutionary acts are performative, in Austin’s original sense of that term, and some of them have truth values. “I state that . . .” is on a par with “I argue that . . .” and “I promise that . . .” Like performatives, statements have felicity conditions. A statement often presupposes the existence of a referent, so if no referent exists, the attempted statement fails. Also, like performatives, statements require that the speaker be in a certain position; without evidence, the speaker cannot state when the world will end, although he or she may guess or prophesy it. Stating, it appears, is not unique; it is just one of many kinds of...

(The entire section is 304 words.)


(Student Guide to World Philosophy)

Additional Reading

Berlin, Isaiah, L. W. Forguson, D. F. Pears, G. Pitcher, J. R. Searle, P. F. Strawson, and G. W. Warnock. Essays on J. L. Austin. Oxford: Clarendon Press, 1973. This book of memoirs of J. L. Austin written by former students offers criticism and commentary on his work. These essays show what Austin tried to do as a philosopher and why.

Burr, John R., ed. Handbook of World Philosophy: Contemporary Developments Since 1945. Westport, Conn.: Greenwood Press, 1980. This survey of world philosophy covers Austin among many other philosophers. The book reveals that no agreed-upon doctrines, method of...

(The entire section is 455 words.)