In addition to its obvious debts to two of my teachers, J. L. Austin and P. F. Strawson, this book owes much to helpful advice and criticism from the many people who read and commented on portions of the manuscript: I am especially grateful to Julian Boyd, Noam Chomsky, R. M. Harnish, Benson Mates and Hans Sluga.

The nucleus of this work was my D.Phil, thesis on Sense and Reference submitted in Oxford in 1959. Several of the ideas presented here have appeared in articles by me, and I wish to thank the editors and publishers of Mind, The Philosophical Review, The Encyclopedia of Philosophy and Messrs Routledge & Kegan Paul and Allen & Unwin for permission to use some of this material again.

Thanks are also due to the American Council of Learned Societies for a grant which enabled me to work on these and related matters in 1963–4, Miss Ruth Anderson for supervising the typing, R. M. Harnish and M. Shapira for work on the index, D. Parfit for help with the proofs and R. B. Kitaj for designing the cover. Most of all I wish to thank my wife for continuing help and advice.

In this chapter I wish to expose three related fallacies in contemporary philosophy, and then, using the concepts and methods of the first part of this book, to offer a diagnosis of them and an alternative explanation of the relevant linguistic data. The three fallacies, as I shall attempt to show, are interrelated and all stem from a common weakness, the failure to base particular linguistic analyses on any coherent general approach to or theory of language. Linguistic philosophers of what might now be called the classical period of linguistic analysis, the period roughly from the end of the Second World War until the early sixties, showed a nice ear for linguistic nuances and distinctions but little or no theoretical machinery for handling the facts of linguistic distinctions once discovered. One of the aims of this work is to provide us with the beginnings of a theory of speech acts. Such a theory if adequate ought to be able to deal with certain kinds of linguistic distinctions in a more adequate way than the ad hoc methods of the classical period. This chapter, therefore—in addition to being an exposure of the fallacies—will be both an application of the theory to current philosophical problems and, to the extent that the theory is capable of dealing adequately with these problems, a further confirmation of the theory.

As I am about to make some criticisms of contemporary linguistic philosophy, perhaps this is a good place to remark that I regard the contribution made by this kind of philosophy as truly remarkable.