If you're happy and you don't know it, clap your hands!

A recent paper on unconscious emotional processing (Ruys & Stapel, 2008) has been getting a good amount of attention, and I thought I'd weigh in on it. You can find the abstract here.

The researchers flashed scary (e.g. a growling dog), disgusting (e.g. a dirty, unflushed toilet), or neutral (e.g. a horse) images on the screen, and they did this at two different speeds: quick and super-quick. Both of these speeds were quick enough that participants couldn't consciously perceive the content of the pictures. The participants were simply asked to judge whether the picture was flashed on the left or the right side of the screen.
The participants were then given a barrage of behavioral and cognitive measures. Here's one of them: When given the choice to complete either a "strange food test" or a "scary movie test", the people who were given scary priming images were more likely to choose the "strange food test" over the "scary movie test". The situation was reversed with the people who were given the disgusting priming images. This was, of course, the expected result. After all, who wants to think about food when there are "unflushed toilet" neurons firing in the brain?

The one message I get from this is that networks in the dorsal visual processing stream ("vision for perception"), which carry the bulk of the responsibility for processing object identification and the semantic content of visual information, may have one temporal threshold for explicit awareness and another, lower temporal threshold for implicit awareness (implicit awareness being a sort of non-thematic experience of responsiveness to specifics in the environment). The content may not be accessible to explicit awareness, but the content is still "in the brain", triggering cascades of activity that eventually result in content-based physiological, behavioral, and psychological changes.

(I wonder if the same thing would happen with spatial thresholds? If you introduce noise into the image and gradually reduce or elevate the levels of noise, would you reach a point at which people are still explicitly unaware of the content of the image, but are at some level responding to the content of the image? I might just pilot it and see . . . )

The results are interesting, but it was the discussion that really caught my eye. The authors gave the following description of unconscious emotions:

Emotions might be viewed as unconscious
when they are detected by indirect behavioral or physiological
measures, without being accompanied by conscious emotional
experience. However, what does it mean when only indirect
measures suggest the presence of an emotion?We think that the
range of emotional measures that are affected depends on an
emotion’s intensity. When emotions are full-blown, people become
aware of their emotions by perceiving their own actions
and bodily reactions. When emotions are weak, people fail to
notice their weakly related actions and bodily reactions.

This represents a turning point in the paper. Up to this point, the authors had been defining conscious and unconscious emotions on the basis of whether or not the participant was consciously aware of the content of the stimulus that induced the emotion. Here, they switch gears and define unconscious emotions as those emotions that are not accompanied by "conscious emotional experience", but which can be detected by indirect behavioral or physiological measures like the ones they used.

I assume that all emotions, conscious or unconscious, could be detected by indirect behavioral or physiological measures. If this is the case, then the only distinction being made between conscious and unconscious emotions is that the unconscious ones aren't accompanied by, well, conscious emotional experience. Quite the tautology.

This isn't the worst of it, though. It seems that defining unconscious emotions in such a way that they can be detected by others but not by oneself will lead to undesirable situations. To clarify: the problem isn't that we have behaviors and physiological responses that others are in a better position to notice than we ourselves sometimes are; rather, the problem is that we are allowing 'emotions' to slide into this category. Would it be problematic, for example, if the experimenter told the subject, "You don't know it yet, but you are afraid," and the subject happened to disagree? Or can we guarantee that these unconscious emotions are so self-evident upon reflection that the subjects would never disagree? I don't think that guarantee can be made, and I find that troubling.

The issue, I think, is the ubiquitous practice, within cognitive science, of augmenting or replacing folk-psychological definitions with the neural processes that underly the original referent(s) of the term. There is the experience of the emotion. Then there is the biological basis of that emotion. Two distinct things, and I get anxious when the boundary between them isn't respected. Even before we understood anything about the brain's role in emotion, there really was something like experiencing an emotion, and I'm guessing that this is what makes emotions interesting to cognitive scientists in the first place.

Ruys, K. I., & Stapel, D. A. (2008). The secret life of emotions. Psychological Science, 19(4), 385–391.