It doesn’t seem like Prinz has much familiarity with the history of emotion before Descartes, but this is really interesting in light of it. [for what he means by “modular”]
“Examples are easy to multiply. They strongly suggest that emotions are modular in the Fodorian sense. Once an emotion is triggered, judgments and other higher cognitive states are relatively ineffective in making it go away. Of course, judgments can exert some influence. Judgments can cause emotions, and judgment can change emotions. When a judgment causes an emotion, it is not directly interacting with the emotional response system. It is only triggering bodily changes that set the response in motion. Judgments may change emotions in the same way. They may trigger a change from one pattern of bodily responses to an- other. In both cases, the effects are indirect. The efficacy of our judgments is further delimited by a second fact. Calibration files can include perceptual triggers that operate at a fairly low level of processing. A person with a fear of flying may have a link between visual representations of airplanes and the amygdala. Be- cause this link operates from within the visual system, judgments about the safety of air travel cannot preempt the fear response. In other words, initiation pathways may contain informationally encapsulated links between perceptual representations of emotion elicitors and the mechanisms that orchestrate bodily response. This makes emotions doubly impervious to direct higher cognitive control.
“As Griffiths (1997) points out, the modularity of emotions offers a good ex- planation of their passivity. We do not seem to be in control of our emotions. They just happen to us. This common observation is etymologically enshrined in the term “passion.” A comparable kind of passivity can be found in paradigm cases of perception. When we look at a ripe tomato, we cannot help but see red. No judgment or desire can block the path from retina to redness, just as no judgment or desire can block the path from body to affect. In both cases, passivity owes to modularity.
“Some authors have argued that emotions are less passive than they may appear. Solomon (1976), for example, argues that emotions can be chosen. A similar theme appears in Aristotle (1985), who stresses emotional training. There is no conflict between the claim that emotions are passive and the claim that they can be controlled. Emotions are voluntary in a double sense. Thinking about some- thing in the right way can certainly influence our emotions, and calibration files can be modified through education and experience. We exert control over emotions by choosing what to think about, and by cultivating calibration files. But emotions are also involuntary in a double sense. First, the thoughts and images contained in an established calibration file may set off emotions automatically. If one happens, by choice or accident, to activate a representation in a calibration file, an emotion will ensue. Second, once an emotion has been initiated, we can- not alter it by direct intervention. Initiation pathways and response pathways both operate without the luxury of control.”
Jesse Prinz, Gut Reactions, 236