Among the theories proposed to explain the effects of behavior on feelings, some have been directed at only one kind of response system, such as facial expression (e.g., Tomkins, 1982; Zajonc, 1985) or arousal (Schachter & Singer, 1962). Although these theories were proposed for a single kind of response system, in principle they might be extended to account for other kinds of behavior-feeling relationships—after all, any theory must begin somewhere before it expands its scope. In fact, however, most of these theories are not easily extended to a wider range of phenomena. For example, Zajonc (1985) suggested that the effects of facial expressions on feelings might be explained by a mechanism first proposed by Waynbaum more than 100 years ago. Waynbaum imagined that changing configurations of facial muscle contractions would change the flow of blood to the brain, and that the patterns of cerebral blood flows caused the different feelings.
Zajonc recognized that Waynbaum's theory was not consistent with modern knowledge about blood flows or cerebral function, so he proposed a variant in which changing patterns of facial muscle contraction would produce changes in the blood temperature in the cavernous sinus. These changes in blood temperature would then affect the temperature of the brain areas to which the blood travels next. Zajonc's studies have shown the standard facial feedback effect. They also show, as Zajonc's theory predicts, that different facial expressions produce changes in temperature of the region of the face over the cavernous sinus, and that the temperature changes are correlated with the corresponding changes in feeling (e.g., McIntosh, 1996; Zajonc et al., 1989). Thus, his theory is initially plausible.
However, three features of the research on facial expressions and feelings are difficult for Zajonc's theory to explain. First of all, the theory assumes a mechanism that varies on only a single dimension, from warmer to cooler.
His dependent variables are unidimensional as well. Consequently, his theory would have difficulty with the fact that facial expressions produce effects that are far more specific and that cannot be fit to a single dimension (e.g., Duclos et al., 1989; Ekman, Davidson, & Friesen, 1990). By itself, this problem might not be insurmountable, because the emotion system might be like visual perception of color. All light lies on a single dimension—that of the wavelength of electromagnetic radiation—yet gives rise to distinctive, separate experiences of red, blue, yellow, and all the other specific colors. However, the remaining two problems are not so easily overcome.
The second problem for Zajonc's vascular theory is that it would not readily explain the individual differences in the effects of facial feedback. One would expect that the effects of facial muscles on blood flows and temperatures would be the same for everyone, and therefore the connection from blood temperature to feeling presumably would be invariant and probably innate. There seems to be no room to explain the individual difference results. At the least, Zajonc would seem to need to add some mechanism by which different people would react differently to the same expressions.
Finally, the third and perhaps most difficult kind of observation for the vascular theory to explain is the parallel between the effects of facial expressions and the effects of postures, gaze, and so forth. Postures and gaze and hand holding are unlikely to connect in any way to mechanisms involving blood temperature. Wearing eyeglasses seems even less likely to affect cerebral blood flows, even though the effects of expressions are correlated with the effects of eyeglasses.
In sum, Zajonc's version of Waynbaum's theory will not accommodate the bulk of the results I have reviewed here. No doubt changing facial muscle activity changes blood temperatures, and no doubt cool air is pleasanter to breathe ((Mcintosh, 1996; Zajonc et al., 1989). Conceivably, the expression/blood temperature mechanism may mediate some aspects of felt pleasantness and unpleasantness. However, cerebral blood temperature does not seem likely to be the mechanism of action of most facial expressions and most feelings.
Tomkins's (1982) theory of emotional experience shares with Zajonc's an exclusive focus on facial expression. In Tomkins's early views, proprio-ceptive feedback from the facial muscles was the key element (Tomkins, 1962, 1963), but later he emphasized feedback from the skin. Tomkins's theory has fewer difficulties than Zajonc's, since it does not assume a single dimension of variation. However, Tomkins's theory has no obvious machinery to explain the individual differences in response to facial (and other) manipulations, and it does not explain why the same effects occur with manipulations of postures, tone of voice, gaze, and so on.
Some years ago Carroll Izard (1977) shared Tomkins's emphasis on the face; more recently Izard (1993) has expanded his theorizing so that it would accommodate the other effects that Tomkins could not. His theory is very elaborate and largely concerned with issues that are not relevant to self-perception, but he would seem to have become a "fellow traveler" with self-perception theory in that he identifies a role for other behaviors. Izard (1993) has argued strongly that feelings and behaviors are intrinsically connected from birth, a view that I think is dubious. I will discuss that issue in chapter 10, since I have yet to discuss some of the evidence most germane to this question. However, with the exception of this mild uncertainty, Izard's theory now seems essentially like self-perception theory.
Although their focus was on arousal rather than expression, Schachter and Singer (Schachter, 1964; Schachter & Singer, 1962) expound a theory that is similarly inadequate to deal with the range of self-perception phenomena. They proposed that arousal mediated the intensity of emotional experience, while the quality of emotional experience was determined by "cognitions" about the situation. However, a great deal of evidence has shown that the quality of emotional experience is determined by facial expressions, postures, gaze, action, and so on. The Schachter and Singer theory could easily accommodate these results by adding these factors to cognitions as determinants of the quality of experience, but then, of course, it would have become essentially identical to self-perception theory.
A second problem with the Schachter and Singer theory is their hypothesis that arousal mediated the intensity of all emotional experiences. Even their own data did not support the arousal-intensity hypothesis with respect to happy feelings, and considerable evidence indicates that arousal cannot be readily "misattributed" to all emotions. On the contrary, arousal itself seems to be relatively specific, with different patterns associated with different emotional states (Ekman, Levenson, & Friesen, 1983). In sum, only emotions that are ordinarily considered to be high-arousal emotions, such as fear, anger, and passion, seem to fit their theory.
With these restrictions in place, we can see that their theory has now become entirely consistent with self-perception theory. This is hardly a criticism, since another way of describing their theory is as one of the origins of self-perception thinking in psychology, especially as the source of misattribution research. Their initial assumptions seem to have been too narrow in regard to the determinants of the quality of feelings and too broad in the assumption that arousal was necessary for all emotions, but their general view of emotional feelings seems to have been correct.
In sum, no theory that focuses on a single component of the emotional process seems able to account adequately for the full array of data. Undoubtedly, all theories could be extended to accommodate all these results, but it seems likely that the result would be an increasing evolution toward self-perception theory.
Was this article helpful?