By definition, what constitutes sexual deviance and what constitutes sexual perversion relate to larger categories of social practice than those that are specifically sexual. This leads us to a useful distinction between sexual deviance and sexual perversion. Sexual deviance might be defined as the inappropriate or flawed performance of conventionally understood sexual practices. Rape, for example, is an act of sexual deviance that is rarely defined as an act of perversion. Rape becomes perverse only when performed on an individual whose inclusion in a sexual act goes beyond the limits of generally conceivable sexual practice or is performed in ways that go beyond the limits of generally conceivable sexual practice. Though we may deplore the behavior, it is possible to comprehend, and even empathically reconstruct, the experience of the rapist or his victim. In the language of the pathologists, such behaviors can be termed a disease of control.

As the character of sexual practice changes over time, so do the boundaries of the definitions of deviance. An excellent case in point is the increasing incorporation of oral sex in the scripting of conventional sexual scenarios. Although the behavior as a matter of organs and orifices remains unaltered, its collective meanings and uses for specific individuals demonstrably have been undergoing profound change (Gagnon and Simon 1987; Simon et al. 1990; Lauman et al. 1994).

Perversions, in contrast, tend to be forms of desire too mysterious and sometimes too threatening to the most elementary def initions of desire and satisfaction to be tolerated. Perversion can be thought of as a disease of desire, not only in the sense that it appears to violate the sexual practices of a time and place, but also because it constitutes a violation of common understandings that render current sexual practice plausible. The behavior of the “pervert” is disturbing because, at the level of folk psychology, we have difficulty understanding why someone “might want to do something like that”. Consistent with this approach was Krafft-Ebing’s use of the term perversion as cited by Davidson (1987). In Krafft-Ebing’s Psychopathia Sexualis (1876) there is a discussion of anomalies of appetite such as hyperorexia (increases), anorexia (lessening) and of perversions such as “a true impulse to eat spiders, toads, worms, human blood, etc.” Perversions of appetite thus involve not only the desire to eat the unthinkable, but also the desire to eat for unthinkable reasons. The same might be said of “perversions” of the sexual.

As in the case of deviance, what is considered a perversion is also subject to revision as what constitutes the thinkable changes. Many forms of what for the contemporary world constitute acts of sexual deviance tend not to occur in other social settings not because they are repressed, but merely because they are literally unthinkable. Moreover, the processes through which behavior becomes “thinkable” for a collectivity are not identical to the processes through which behavior becomes “thinkable” to the individual.

The logic that links motive and behavior is the complex, almost magical logic of representation, a logic of metaphor and metonymy that meshes personal history with social history. Concerns for the role of the symbolic in the interaction of individuals often obscure the importance of the interaction between, and even within, symbols. A homogeneity of sexual preferences inevitably masks a heterogeneity of desired emotional productions, and this perhaps is the level at which, to some extent, we may all be perverts, as even the most conventional may find their sources of sexual excitement fueled by the slightest “whiff’ of the unthinkable (Stoller 1979, 1985a).

Theories that describe the causes of behavior already presuppose the problematizing of that behavior. It is not theory that renders social practice problematic, but the emergence of the problematic in experience that gives rise to the requirement for theory, as well as the potential for its ability to find a responsive audience. In other words, theories of behavior are themselves behavior and, as such, have essentially the same basic requirements: first, they must be thinkable; second they must be plausible, that is, made either legitimate or explainable in terms of what is held to be legitimizing. In this sense I agree with Canguilhem (1989) when he notes: “[I]t is not paradoxical to say that the abnormal, while logically second, is existentially first.”