Some prior authors have noted that people have no unique access to why they believe what they believe (e.g., Nisbett & Wilson Reference Nisbett and Wilson1977). Others have gone a step further and postulated that people do not know their own attitudes (e.g., I like ice-cream) but must construct them when needed from other available information that they either retrieve from memory (e.g., ice cream tastes good) or extract from the immediate context (e.g., Schwarz & Bohner Reference Schwarz, Bohner, Tesser and Schwarz2000). Carruthers takes this “constructivist” position to the ultimate extreme by arguing that people have no direct access to any attitudes or relevant beliefs. According to this view, introspection does not exist, and is merely an illusion. Furthermore, he provides many examples where people either clearly or plausibly are confabulating when they express what they believe. In his view, at best individuals only know what they feel and perceive, not what they think. In our view, it is not clear why an intelligent organism would have evolved to have direct access to its feelings and perceptions but not its cognitions.

Nevertheless, Carruthers has an important point. Whenever someone expresses a belief or has a thought, some degree of interpretation likely is involved, if only to understand the meaning of the “inner speech” in which the thought is expressed. Thus, if a person has a positive reaction to some stimulus (ice-cream), this can be translated into “it's good” (Fazio 1985). Or even if the word “good” immediately springs to mind, the meaning of the word must be understood by the self if an internal thought, or by others if expressed. However, this very minimal form of “interpretation” is very different from the kind of interpretation involved in most of the examples of confabulation provided by Carruthers. Indeed, we argue that it may not be wise to think of introspection and interpretation as dichotomous categories in which to place any given judgment. Rather, there are various degrees of interpretation. At the low end of the interpretation continuum, judgments are introspection-like in that they involve at most some trivial interpretation. At the other end of the continuum, the judgment is totally confabulated from external sources also available to outside observers.

Although dichotomous categories can be useful in understanding some phenomena, as illustrated by the distinction between primary and secondary (meta-) cognition, we believe that it is not conducive to understanding human information processing to simply lump all judgments into the same overarching “interpretation” category and stop there. This is because putting all judgments into the same category might suggest that there are no meaningful differences within the catetory. In contrast to lumping all judgments together into one interpretation category, we espouse a continuum view in which people express beliefs based on very little interpretation in some cases but based on extensive confabulation in others. We further argue that differences in the degree of interpretation are meaningful.

Previous research on psychological elaboration provides one instance of the usefulness of the continuum view. The term “elaboration” is used in social psychology to describe that people add something of their own to the specific information provided, for example, in a persuasive communication. In the domain of social judgment, variations in elaboration are consequential. For example, when people are relatively unmotivated or unable to think, they are more likely to rely on immediately accessible information that originates either internally (one's attitude) or externally (e.g., the attractiveness of the message source). When people are more motivated and able to think, then these initial reactions and the judgments that follow from them can be overridden by more complete interpretative analyses. Furthermore, judgments based on high levels of elaboration are more consequential than those based on low levels (Petty & Cacioppo Reference Petty and Cacioppo1986).

Viewing interpretation as a continuum has a number of implications. Most obviously, it means that interpretation can go from zero (i.e., introspection) to extensive. More interestingly, the continuum view suggests that the point on the continuum that corresponds to minimal or trivial interpretation has more in common with zero interpretation than it does with extensive interpretation. One can draw an analogy to a distinction that attitude theorists used to favor between attitude formation versus attitude change. Different mechanisms of influence were thought to be operative depending on whether a person had an existing attitude or did not (a categorical view). Today, it is more common to think of attitudes as falling along a continuum such that they can vary in how accessible they are or upon how much knowledge they are based. An attitude formation situation would be present when a person has no prior attitude. But, a continuum approach to attitudes suggests that a person who has an attitude that is difficult to bring to mind and based on little information (Person B) has more in common with an individual who has no attitude (Person A) than a person who has an attitude that comes to mind spontaneously and is based on much knowledge (Person C). That is, the first two individuals – A and B – are more similar to each other in psychologically relevant ways than they are to C, despite the fact that a dichotomous approach places Person B in a different category from A and in the same category as C. So too is it the case that a judgment based on minimal interpretation (B) is closer to a judgment based on no interpretation (A) than it is to a judgment based on extensive interpretation (C; see Fig. 1).

Figure 1. Continuum of extent of interpretation underlying judgment

In sum, we conclude that an all-or-none frame regarding the existence of introspection may not be the best way to make the most progress in understanding social judgment. Instead, drawing from the literature on elaboration and attitude strength, we suggest that it might be more fruitful to approach interpretation as a continuum where the low end is anchored at introspection. That is, sometimes interpretation can be quite minimal as when people recall their birth-date or liking of a favorite book. At other times, interpretation can be quite extensive, such as when there is either nothing relevant to recall or one's interpretation totally overwhelms any mental content introspected.