John L. Austin was one of the leading philosophers of the twentieth century. The William James Lectures presented Austin's conclusions in the field to which he directed his main efforts on a wide variety of philosophical problems. These talks became the classic How to Do Things with Words.

For this second edition, the editors have returned to Austin's original lecture notes, amending the printed text where it seemed necessary. Students will find the new text clearer, and, at the same time, more faithful to the actual lectures. An appendix contains literal transcriptions of a number of marginal notes made by Austin but not included in the text. Comparison of the text with these annotations provides new dimensions to the study of Austin's work.

Austin's argument concerning the characteristics of a performative utterance are informed by a specific assumption concerning the origin and evolution of language: to wit, that language in its primitive stage was simply a collection of one-word utterances that are inherently ambiguous in terms of their individual senses. Thus, in order to refine the sense of these one-word utterances, a whole array of supplementary parts of speech evolved, and language became consequently more complex and sophisticated (71). In Austin's nomenclature, the force of a given one-word utterance was too diffuse vis-à-vis the context in which it is uttered and thus quite ambiguous from the addressee's position. In other words, a primitive one-word utterance does not provide the addressee any certainty about how she is to construe it. Therefore, the increasingly sophisticated iterations of language indicate an ongoing effort to refine the sense of an utterance, to give the force of the utterance a more specific and unambiguous valence.

However, Austin also maintains that an unintended consequence of this evolution of language is that it reaches a point where it becomes too sophisticated and thereby re-introduces the very uncertainty it was originally intended to mitigate. He claims that the various parts of speech, and the words that comprise them "lend themselves to equivocation and inadequate discrimination; and moreover, we use them for other purposes, e.g., insinuation," and thus concludes that "the trouble about all these devices has been principally their vagueness of meaning and uncertainty of sure reception" (76). In other words, there is a definite yet non-localizable threshold that an utterance must not cross if it is to remain teleologically oriented toward the clarity and accurate construal on the part of the addressee.

The speech act therefore always navigates between the Scylla and Charybdis of inadequately directed signifying force resulting from the primitiveness of the utterance on one hand, and the over-complexity of the utterance on the other. As a result, the clarity of a given utterance depends almost exclusively on the intention of the speaker; she must in some way remain cognizant of the above-mentioned threshold and therefore deploy the force of her utterance in a way that avoids being too diffuse or unmanageably polyvalent. This is not to claim, however, that the clarity of a given utterance is reducible to some Aristotelian mean; rather the clarity of an utterance depends on how well it reflects the earnestness or sincerity of the speaker. This notion of the speaker's earnestness is deduced from the circumstances surrounding the utterance, as well as the utterance's delivery, e.g., the enveloping context, the speaker's particular emphases, diction and enunciation, etc. The addressee thereby "triangulates" the speaker's specific intention through interpreting the above-mentioned features of the utterance. In short, it is absolutely essential to Austin's project that the speaker mean what she says.

It appears then that Austin's fundamental supposition is tautological: the addressee deduces/approximates the speaker's degree of sincerity through the amount of sincerity the speaker conveys in her utterance, which in turn reflects ipso facto the speaker's sincerity (as a subjective condition). In short, the speaker is found to be in earnest because she is in earnest. Only an utterance of the utmost sincerity-what Austin terms an "explicit performative"-carries with it the closest thing to a guarantee in terms of a clear and accurate construal. This further implies that clarity of utterance is ultimately an ethical consideration, rather than a linguistic or grammatical one, because the speaker's responsibility to her addressee obliges her to be earnest and therefore quite literal in her expression (see Habermas on this point). Unless of course the context in which the utterance is made is one in which it is assumed, either through mutual agreement or convention, that explicit or pure performatives are not necessarily expected nor pertinent, e.g., a comical monologue, a play, etc.

Thus, while Austin's argument in How to Do Things with Words is elegantly schematic, it nevertheless implies a somewhat simplistically idealized and unitary notion of the speaker's subjectivity. In other words, Austin's claims cannot adequately accommodate instances of insincerity that, while perhaps unanticipated, are not exactly inappropriate-such as ironical observations on an immediate situation-because such self-abrogation of the speaker's sincerity renders the utterance "infelicitous" almost to the point of being diabolically caustic with regard to the addressee's apprehension.