This chapter includes several theoretical ideas that purport to account for phenomena we have thus far examined. These ideas, unlike those of the previous chapter, are not derivations from or additions to dissonance theory, but instead, are attempted alternative accounts for "dissonance phenomena" in terms unrelated to the theory. A vast array of alternative viewpoints of dissonance-like phenomena may be found in the volume of Abelson, Aronson, McGuire, Newcomb, Rosenberg, and Tannenbaum ( 1968), but this chapter is considerably more focused. We shall discuss just those viewpoints that offer a relatively thorough- going attempt to account for several dissonance phenomena by means of alternative schemes. The most pervasive of these notions is a theory of self- perception advanced by Bem ( 1965), which is discussed first.
In one of his first statements of the self-perception thesis Bem ( 1965) proposed that cognitive dissonance experiments have their effects because of the judgmental, self-observational abilities of the subjects, rather than because of the motivational underpinnings of dissonance theory. His idea is best captured by considering an experimental situation from the standpoint of an onlooker. Suppose an observer watches someone perform an attitude-relevant behavior, such as delivering a segregationist speech before the local parent--teachers association. The observer is told nothing about the attitudes of the speaker, but he is provided with another kind of information: In one instance the speaker has been instructed by his employer and family to give the speech; in another, he issues his diatribe against integration in the absence of any apparent constraints. If asked to infer the speaker's attitude on segregation, it seems likely that a more segregationist position will be inferred when there are no pressures to deliver the