Two Views of Emotional Perception: Some Empirical Suggestions

Two Views of Emotional Perception: Some Empirical Suggestions

I. Perception and Modularity

One stream in contemporary philosophical and psychological study of the emotions argues that they are perceptual capacities (e.g., Prinz 2004).² For instance, Jesse Prinz has recently defended the view that emotions are perceptions of bodily changes and, via these, of "core relational themes" (2004, 224–25). Core relational themes are, roughly, relations an individual has to his/her environment that pertain to that individual's welfare (2004, 15–16). The present project is to compare and contrast two possible models of emotional perception. The central difference between these models is the notion of modularity that they use, and the corresponding overall view of the nature of the mind. I will suggest some empirical tests that might adjudicate between these different kinds of emotional modularity, and hence between these two models of emotional perception. I will conclude with some remarks about the extent of the relevance of this issue.

So, suppose that emotions are perceptual capacities. How would one go about conceiving of and studying such capacities? A natural way to proceed is to use other perceptual capacities as a model: identify important characteristics of uncontroversial perceptual capacities, then see whether emotions have identical or similar features. This is Prinz's method (2004, 221–22, and throughout the same chapter). One of the characteristic features of ordinary perceptual capacities, and the only one which will concern us here, is their modularity. Prinz argues that emotions share the characteristics of classical, Fodorian modularity (more or less; both he and Fodor 1983 acknowledge that modularity can be a matter of degree; Prinz 2004, 232–36). This means that they share a very important feature with uncontroversially perceptual capacities, such as vision. Thus an important part of the overall case for emotions as perceptual capacities is in place.

This view of perceptual modularity is connected to a particular overall view of the structure of the mind. Prinz gives us an explicit statement of this view:

The mind is divided into different kinds of information-processing systems. There are perceptual systems that provide inputs, action and motor systems that provide output, and, perhaps, higher cognitive systems that engage in reasoning, planning, problem solving, and other mental operations that mediate between inputs and outputs when we move above the level of reflex response.

(2004, 221)

If emotions are perceptual capacities, then they fall into the input part of this view of the mind. They provide information to, in all likelihood, both the action-production systems and higher cognition.

Susan Hurley has called this view of the mind into question. She calls it the "classical sandwich" view of the mind (1998, 20–21): higher cognition is the filling between the input and output layers that have more direct contact with the environment.³ Hurley describes this view of the mind as being structured with vertical modules. The metaphor of verticality is used to capture the constitutional independence of modules in this view of the mind. On this view, higher cognitive processes and action-production systems are both constitutively distinct from all perceptual capacities. Classical, Fodorian modules have their natural home in this vertically modular, classical sandwich view of the mind. If emotion is a classically modular perceptual capacity, then it too is constitutively distinct from the modules that realize higher cognition and output.

Hurley argues that this view of the mind, and specifically its vertical modularity, has been called into question by neuroscience. Instead of a mind composed of constitutively distinct vertical modules, Hurley argues that neuroscience reveals a mind structured by horizontal modules. Horizontal modules are content-specific and task-specific systems that " [loop] dynamically through internal sensory and motor processes as well as through the environment." (Hurley 1998, 21; see 408 for discussion) These systems are "modular" in virtue of their content- and task-specific functionality. However, they are importantly different from vertical modules. First...