When I wrote about Inside Out I mentioned that I wasn’t a fan of basic emotions theories, and felt a bit sniffy about Riley’s limited repertoire. Now an approach I’m drawn to, constructivism, is in the news. Whenever I’m asked to recommend a paper in my field that would be interesting to general audiences I always pick Barrett’s “Are Emotions Natural Kinds?”
Alva Noë both appreciates and critiques Barrett’s Times op-ed. I critique the critique.
It’s worth noticing that there are two ways to interpret this idea that emotions are contextual.
One might simply take it to say that emotions look and feel different in different settings. Or that without information about context, it can be hard to know what you’re feeling, as I noticed as a little boy.
But there is a more radical interpretation as well. Emotions, one might say, are contextual all the way down. What we call fear in one setting might be totally different from what we call fear in another setting.
The first interpretation pushes us to revise a simple-minded theory of emotions as entities inside us. The second urges us to deny that there are emotions.
First of all, emotions could both be “contextual all the way down” and still exist. Noë understands a category, in the sense in which emotions like anger can be categories, as “having a family resemblance.” In the case of fear, for example, the easiest way to describe it is with reference to the motivational situation — fear is a bad feeling that you feel when there is a future bad outcome you want to avoid (anger a feeling that your goals are frustrated by something you ascribe agency to, sadness a feeling when you’ve lost something, etc). There may be some instances of fear that involve a racing heart, and some that don’t, but a racing heart is very nonspecific to fear. There may be instances of fear that evoke paralysis, and some that evoke aggression, so the most frequent characteristics occurring in this family are the qualities “bad” and “outcome you wish to avoid” — the latter, a context, being so important to understand what knits the family together that it’s fair to say fear is “contextual all the way down.” But it wouldn’t mean fear doesn’t exist; it would just mean that “fear” is a human, social label — a lens we’ve chosen. Social categories are real things.*
As she argues in “Are Emotions Natural Kinds?”
Emotional responding exists, can be functional, and is very likely given to us by evolution. But that does not necessarily mean that anger, sadness, and fear are useful categories for conducting science.
I think even if she’s right about the fact that they’re not natural kinds, they might be useful as long as they’re not too reified or misunderstood, but in any case, I understand her position to be that this way of carving up experience and expression was invented by people.
I also think he may be misreading Barrett when he complains about her complaint about airport security being trained to identify emotions from facial recognitions. He quotes her:
“When airport security officers are trained on the assumption that facial and body movements are reliable indicators of innermost feelings, taxpayers’ money is wasted.”
Then goes on:
Now, it may very well be that it is a waste of taxpayer money to try to train airport security officers to recognize emotional states on the basis of what passengers say and do. But nothing Feldman Barrett has said supports this claim at all. The fact that you can’t define emotions in terms of a strict set of a behavioral rules does not entail that you can’t learn to be more sensitive to a person’s emotional state by careful observation, just as the fact that variation is the norm with biological species does not entail that you can’t sort animals into species by examining them.
I think he misunderstands Barrett both in that he doesn’t see that she really is arguing what he calls the “radical” view, and also perhaps misses the work that “reliable” is doing in that sentence. I doubt she’d argue that it’s impossible to to become more in tune with implicit emotional communication, just that it’s error prone without a variety of other kinds of information from context. To argue otherwise is almost by definition absurd: of course emotional responding is communicative; that’s one of its functions; the question is whether there are discrete categories of responses that map onto our linguistic labels.
Anyway, I’m drawn to Barrett’s approach and find it appealing, but I don’t want to take the strongest stance on its correctness. Ekman and Keltner say she misrepresents the literature. I also think the paragraph of her op-ed dealing dealing with the lack of evidence for neural signatures that correspond to basic emotion categories is frustratingly argued. It doesn’t prove much that there’s no voxel (that’s the term, analogous to pixel, for a unit of volume in MRI imaging) sensitively and specifically associated with fear, other than that the grossest and most naive localization approach is not the way to understand how emotions are instatiated in the brain. Our imaging tools are so crude that not being able to find a unique voxel doesn’t say much. A friend of mine recently linked to this video and joked: “Just looks like a uniform voxel with a linear response profile to me.”
*I was about to write, “they are just different kinds of things than natural kinds,” but I realized that was a sentence I don’t feel competent to defend. I suppose it’s possible to argue that all the organization of matter into things can be sensical only because there’s some perceiving agent interacting with them. IANAPhilosopher.