What's Wrong with the Psychoanalysis of Literature?

Though it uses clinical psychoanalysis as its model and bases itself upon psychoanalytic theories of pathology, cure, and so on, literary psychoanalysis is quite different from therapeutic psychoanalysis. Indeed, the relationship between these two fields is highly obscure. In the following pages, then, I should like to look at object and method in the psychoanalysis of literature, seeking to determine the degree to which literary psychoanalysis is justified by the clinical and theoretical work it invokes.

In clinical analysis the question of object is easily answered. The primary object of analysis, for both parties, is the analysand. (There is also the analyst's self-analysis, which is absolutely crucial in clinical work and should no doubt play more of a role in literary psycho-analysis; I leave this aside, however, as it does not appear to have any special properties in literary study.) In the case of literary psychoanalysis, things are much less clear. As Norman Holland has pointed out, a literary psychoanalysis may focus on one of three objects: the author of a work, the reader of a work, or the work itself (8). The author and the reader are relatively unproblematic as objects of psychoanalysis. Though methodological difficulties arise in each case, authors and readers are subjects with unconscious beliefs and desires, oedipal experiences, repressed fantasies, and so on, just like real analysands in clinical psychoanalysis.

The case of texts, however, is quite different. The usual way of proceeding with the psychoanalysis of a text is to focus on character. But although literary characters are like people in the way we imagine them, they just are not people. They are imaginative constructs only. Thus there can be no question of a character having the sorts of unconscious beliefs and desires that analysands, authors, readers, and all other real people share (Norman Holland 296-309).

What, then, can sensibly be said about characters in a psychoanalytic vein? There are two obvious possibilities. The first possibility is to return to authors or readers and the ways in which they imagine characters. In this case any claims (about, say, a character's unconscious motivation) would pertain not to the character but to the author's or a reader's fantasy of the character. The second possibility is to treat characters as if they were real—while fully and explicitly recognizing that they are not. In this case one's interpretive claims would concern hypothetical cases of real individuals modeled upon fictional characters. I should emphasize that both cases are quite different from the supposed analysis of characters per se. The analysis of character per se is purely speculative in method and leads to questions that have no correct or even incorrect answers. By contrast, the alternatives just proposed are purely empirical and do have right and wrong answers—about the fantasies of readers and authors in the first case and, in the second, about typical etiologies for certain types of symptoms.

In all cases, clarity about the object of one's literary psychoanalysis is absolutely crucial. Too many psychoanalytic studies jump from character to author to reader without any sense on the part of the interpreter that no single object is being consistently examined and thus no cognitive aim is being fulfilled by the criticism.

Before going on to method, I should like to look very briefly at the psychoanalysis of texts per se, principally as practiced by certain psychoanalytically oriented post-structuralist critics. This form of criticism applies psychoanalytic notions to literary texts without reference to human subjectivity (real or fictional) so that the language or the text itself is invested with the relevant powers of agency. For example, condensation and displacement "in the text" are seen not as a function of the author's or even the reader's primary process thinking, but as processes of, say, language.

This view is based upon an implausible notion of linguistic autonomy, a notion that is rarely if ever questioned, even though it is widely denied within linguistic science (see, for example, Chomsky 19-46). Furthermore, it spuriously extends to semantics a structuralist theory of phonology that does...