A behaviorist (and radical empiricist) theory of emotion

Behavior and Philosophy is an odd, but important journal. It has, for four decades now, provided a venue for papers about the philosophy of behaviorism and the relationship between behaviorism and related disciplines. The table of contents for the recent volumes, and more information about the journal, can be found here. Alas, being run by a small group of generous academics, they aren't always on top of the little things you would like a journal to do... like notifying you when your article comes out. Thus, it wasn't until fairly recently that I found out about last year's printing of:

A Behaviorist Account of Emotions and Feelings: Making Sense of James D. Laird's Feelings: The Perception of the Self, by Charles, Bybee, & Thompson

As the title implies, we present a system for explaining the phenomenon of emotions and feelings within a behaviorist tradition. The discussion is set off by the consideration of ideas presented in Jim's book, which build's upon William James's approach to emotion. The short version goes something like this:



Laird's book presents tremendous amounts of evidence in favor of the textbook version of the James-Lange theory of emotion. This simplified version is merely about the order of behavior and the feelings of emotions; for example that one feels happy after covertly being made to smile (William James's most famous argument regarded feeling fear after running from a bear). If you want to know the state of research on the effect of behavioral manipulations on the feeling of emotions, there is no better source. Oh, and the work is fascinating, so you want to know about it.

While I don't really know what Lange was up to, I am sure that is a hollow caricature of what James was up to. Laird's agenda, growing out of a more direct connection with James's work, is much more ambitious: Laird want's to argue that certain types of behaviors-in-circumstances constitute emotions, which - once present - can then be felt. This is an externalization of feelings that would be a strong compliment to the externalization proffered in recent work on embodied cognition. As Laird puts it:

As patterns or patterns of patterns of action in context, emotions and other psychological states are all observable. The necessary observations are not easy, but the difficulty arises because the patterns are complex and abstract, not because the events observed are occurring in some inner, intrinsically private space. (Laird, 2007, p. 219)

Nice and behaviorist, in the descriptive mentalist vein. Alas, Laird never quite goes as far as William James to assert that behaviors cause feelings, and so he does not work through the implications of that claim. Laird has some weird misgivings about causality, and he tries to write in a very genial way, which sometimes masks the implications of this way of thinking, or at least leads to them not being developed as explicitly as one might like. His resistance to this is linked to a fear of being misunderstood as a "vulgar reductionist", which he is not. There are also a series of concerns over whether behaviors and feelings are "ontologically distinct", which they would not be if one were a constituent component of the other. That is:

It is possible that William James’s bear-fear could neither cause nor be caused by his fearful behavior. Neither causal link is possible if, when we talk of being “afraid of the bear,” we are describing the relation between the-appearance-of-the-bear and the-start-of-fearful-behavior. (Charles, Bybee, & Thompson, 2011, p. 8)

These concerns are each addressed in turn.

We propose (among other things) that much confusion can be caused by better distinguishing between emotions and the feelings of those emotions. As readers of this blog know, I have a great respect for metaphor, which I got in part through my post-doctoral work with Thompson. The metaphor of feeling entails a thing being present, which is then felt, and this should be taken seriously. This requires a world with (at least) three "levels":

Level 1. The world is full of objects and events, including behavioral events. (Bears exist and are encountered, people run, etc.)

Level 2. The circumstances conventionally associated with feelings cause the behaviors conventionally associated with those feelings, and that causal relationship constitutes an emotion. (The sight of a bear can cause me to flee; that the bear caused me to flee is my bear-fear.)

Level 3. Feelings are responses to that causality in our behavior; when I respond to that causal relation in some way, I can be said to have “felt” it. (When stopping to catch my breath, my backward glances—to see if the bear is following me—are in response to the causal relationship between the bear and my fleeing, which is itself my fear.)

Level 1 (repeat). Feelings, as a type of event, are part of the environmental context that can cause subsequent behaviors. (Because I am avoiding the woods next time I jog through the field.) (p. 8-9)

 This can be represented graphically:

It is worth noting that:

Any attempt to interpret Laird’s argument must be consistent with Laird’s clearly and repeatedly stated belief that feelings are perceptions. The book is, after all, titled Feelings: The Perception of Self. But is responding to causal relations an example of perception? We think yes. In the world of experimental psychology that Laird inhabits, perception is synonymous with discriminative response; when something is responded to in any way—diverts the flow of the animal’s behavior in any way—it has been perceived. A researcher within this tradition seeks to discover the antecedent conditions necessary and/or sufficient for subjects to respond differently to two presentations. For example, early studies of color perception in animals simply trained animals to respond to a specifically colored light, then measured the animals’ responses to lights of different wavelengths but identical intensities. “Perceiving red” is thus nothing more than responding as a function of a particular wavelength; being red–green color blind is nothing more than responding to a variety of wavelengths in the same manner, etc. On this understanding, “the self” has been perceived when a person has responded to some feature of his own behavior; that is, whenever a person’s past behavior is a part of the circumstances causally connected to current behavior. (p. 10-11)

 Also, while I won't put the whole thing here, it is worth mentioning that we have an improved version of the running-from-a-bear scenario (complete with college room mates and a bear suit), which we work through zealously. The conclusion goes further to connect this way of thinking with the other works of William James, after which we feel obligated to ask:

But why should we care that Laird’s efforts flesh out aspects of William James’s approach to psychology? James died over 100 years ago, and though his theory of emotion still draws some attention, few in contemporary psychology or philosophy would self-identify as “radical empiricists.” Why should anyone outside the narrow field of emotion research care about this contribution? (p. 15)

People should care because, as you might have guessed based on my anticipation above:

The major challenge faced by supporters of embodied cognition is the same as that faced by radical behaviorists for over 100 years: How can they deal with the psychological phenomena of emotion and feelings? .... for many who are outside these specialties, the lack of ability to deal with emotion-related phenomena has often been a deal breaker.... [and] this has led many to discount the importance of behavior analysis, molar behaviorism, ecological psychology, and embodied cognitive work.

....Thus, there is great value in a non-reductive behaviorist interpretation of emotions and feelings, such as the one offered here, supported by the systematic evidence that Laird marshals. Any of a number of classic books in the behaviorist or ecological traditions—as well as any of the recent books about radical embodied cognition—could be complemented by Laird’s book to provide an empirically grounded interpretation of the full range of traditional psychological phenomena. (p. 15-16)