Language and the Origins of Psychoanalysis: By John Forrester. New York: Columbia University Press, 1980. 285 pp.

If asked to consider the influences which most profoundly operated upon Freud in his creation of psychoanalysis, most students of the field would mention the biological and physical sciences as Freud knew them. In this always interesting and frequently exciting book, John Forrester, a Cambridge University philosopher of science, offers the alternative thesis that Freud and many of his early followers were even more influenced by their understanding of language generally and, more specifically, by that nineteenth century branch of linguistic study known as philology. As the title suggests, to Forrester "language is the central concern of psychoanalysis." He attempts, as did Freud, to explain the way in which the curative process is related to the power of words. To assist in this rather large task, Forrester offers the reader a friendly critique of certain Freudian texts and letters, an extensive discussion of the history of aphasia and its paramount conceptual significance in the development of psychoanalysis, and reflections upon nineteenth century linguistics (especially philology), symbolism, grammar, and the concept of the unconscious. As the author correctly notes, it is not easy to find a disciplinary niche for this work. It is full of surprises for the traditionally educated psychoanalyst.

If psychoanalysis as conceived by Freud can be said to be a study of the unconscious and of the rules which govern it, it might not be unfair to characterize Forrester's work as an exploration of the theory of consciousness implicit in early psychoanalytic thought. The exploration begins with the observation that Freud had two different conceptualizations of the unconscious. One was the familiar metapsychological thesis that there is a system of unconscious ideas and a separate system of word presentations with which the unconscious ideas articulate. This emphasis upon the primacy of the unconscious contrasts with a second conceptualization in which the inferential