Emotional Gestalten

Cognitive–behavioral therapy (CBT) is committed, as its name suggests, to the modification of behavior and of cognition. Maxwell and Tappolet’s paper raises two important theoretical questions: whether the desired change in behavior is supposed to come as a result of the change in cognition, and whether what counts as cognition amounts primarily to propositional belief. On the first question, the authors cite discouraging empirical evidence, and they suggest that the explanation for this lies at least in part in the fact that the answer to the second is that the relevant sort of cognition is a mode of perception rather than a mode of belief. What therapeutic change requires, then, is not belief, but something more like a Gestalt shift.

Several considerations make this plausible. Unlike changes in unemotional belief, which are fairly predictable on the basis of the strength and validity of evidence or arguments, Gestalt switches are not only involuntary but also somewhat unpredictable. This is also the case with changes in emotional responses. Furthermore, emotions often seem to be informationally encapsulated in a way that resembles the modularity of perception. Thus, if one is in the grip of an emotion that seems to incorporate a belief, evidence against that belief often fails to shake it. Although common sense suggests that the belief forms the core of the emotion, it is as if the causal order were precisely the opposite, with the emotion forming a protective shield against any change in the belief it appears to rest on. Indeed, it’s worse than they say, for there is evidence that if a belief is false and motivated in the first place, it is likely to be reinforced rather than weakened by contrary evidence (Nyhan and Reifler 2010).

Despite its plausibility (and my own bias in favor of the perceptual model), there are a number of points that remain obscure. I raise them without any confidence in my own ability to shed light on them.

To begin with, it is a common assumption of CBT that wherever therapy is indicated in the first place, there are erroneous beliefs that need to be corrected. Thus, Maxwell and Tappolet (2012, 2) remind us, “the therapist may have recourse to such cognitive interventions as ‘Socratic dialogue,’ ‘guided discovery,’ and ‘collaborative empiricism.’” Yes, but these interventions presuppose that what will be uncovered and corrected are mistakes. What if one is faced with a situation that is objectively such that being demoralized is the only reasonable response? Or, to put the same question differently, how can we be so sure that no such situations exist? Are there not situations to which sadness or even depression is an appropriate response? Does CBT then amount simply to attempting to spin things to produce an encouraging illusion?

A second question concerns the assimilation, implicit in Maxwell and Tappolet’s discussion, of the perceptual with the “nonconceptual.” Some have, to be sure, argued that perception involves a nonconceptual component; but others have denied this (Sedivy 1996). Besides, there are many ways in which a state can be “nonconceptual,” notably if the state in question is entirely constituted by subpersonal components. If these subpersonal components or aspects consist in conditioned responses or learned habits, it is indeed unlikely that they will be modifiable by means of conceptual change, but by the same token it is not obvious what should earn them the title of ‘perceptions.’ Perhaps we need to compare emotions to visual perceptions as interpreted by the dual-pathway hypothesis (Goodale and Humphrey 1998). One pathway involves the conscious estimation of a size and shape; the other disposes the perceiver to a specific haptic behavior. The former, but not the latter, is linked to consciousness, and the latter can plausibly be thought to be procedural rather than representational.

Many people have suggested that emotions are a way of construing or seeing the world or some specific situation. I have done so myself. But that view raises the question of the mechanism by which a Gestalt can be changed. Many well-known examples, such as the Necker cube or the classic two...