Mental causes in the constitution of emotion
Noncognitive theories typically employ two strategies in keeping cognition out of emotional content. They either treat cognitions as contingent elicitors of emotion that do not belong to emotion proper, or they maintain that the eliciting appraisals of emotion that determine the quality of emotional content are always rather than contingently nonconceptual - subpersonal and subdoxastic - representations. Prinz (2004) employs the first argumentative strategy, whereas Robinson (2005) and affect programme theorists (e.g. Ekman, 1992, 1999; Levenson, 2003) advance the second, the latter especially for basic emotions. I shall argue against Prinz that elicitors of emotion belong to emotion proper and that they often, though not always, qualify as cognitive states. This conclusion can be turned against Robinson's view about noncognitive elicitors of all emotions. In conclusion, the elicitors of human emotions divide into cognitive and noncognitive appraisals. This means that the cognitive status of human emotions depends on whether or not cognitive processes are essentially involved in all emotions after their elicitation, a question which will be discussed in the next section.
Prinz argues that evaluative judgments contingently elicit emotions that qualify as embodied appraisals with a representational content by virtue of their evaluative function, without involving appraisals as their components. This claim is founded on a sharp distinction that Prinz makes between the initiation and response pathways of emotion. Initiation pathway contains the elicitation and calibration files of distinct emotions, whereas the response pathway is constituted of changes in the bodily state and their somatosensory perception. A noncognitive view of emotions as perceptions of bodily changes results from defining emotion exclusively in terms of the response pathway.
Two crucial distinctions for Prinz's argument are those between (1) elicitors of emotion and emotions proper on the one hand and between (2) direct and indirect causes of emotion on the other hand. The first distinction purports to establish the noncognitive view of emotions proper as perceptions of bodily changes that nevertheless, on account of the second distinction, represent their indirect causes, that is, core relational themes, rather than those bodily changes that directly cause the emotion. This idea can also be expressed by distinguishing between the real and nominal contents of emotion: "Core relational themes are the real contents of emotions, and bodily changes are their nominal contents" (ibid., p. 68). Even if emotions are reliably caused by both bodily changes and core relational themes in the organism's environment, they represent only the latter because they have the function of detecting core relational themes, which they do by registering changes in the body.
Elicitors of a particular emotion type constitute its elicitation file. These are mental states that internally trigger individual emotions of a particular type when the relevant core relational theme is instantiated in the subject's environment. Yet elicitors need not have anything else in common. For instance, the elicitation file of fear may contain perceptions of a snake or a pointing gun, as well as judgments like "That's poison!" or "I'm in danger!" Elicitation files of basic emotions have been fostered by natural selection, but response patterns of basic emotions can be retuned to respond to other properties that they were not genetically set up to detect. For instance, the response pattern of anger may come to represent infidelity (jealousy) and moral transgression (moral indignation) as well. One response type can then function as a calibrated detector of several core relational themes. Items of a particular elicitation file contain all those states that elicit an emotional response of a certain kind. However, emotions are constituted of perceptions of bodily changes alone because elicitation and calibration files are only causes, not constituents of emotion.
A major problem with Prinz's theory is its incapability to accommodate particular objects as intentional objects of emotion. Emotions as embodied appraisals have been set up to be set off by core relational themes, not particular objects that figure in the theory as mere triggers of emotion. Thus for instance "Sadness represents the loss of something valued. If I am sad about the death of a child, I have one mental representation that corresponds to the child's death and another, my sadness that corresponds to there having been a loss" (Prinz, 2004, p. 62). Since representations of particular objects are not components of emotion, individuals must scan their environment for particular objects or situations that instantiate the relevant core relational theme. The actual elicitors of emotions obviously give clues here, but the process still involves more or less interpretation because emotions do not include representations of particular objects. Worse still, psychological construction of this kind extends to the type of emotion as well since Prinz allows the same bodily response patterns to represent several core relational themes.
Prinz argues that anger, fear, happiness, sadness, and disgust are embodied appraisals that are identifiable in terms of their bodily response patterns, whereas all other emotions are blends of basic emotions or cognitive elaborations of basic emotions. The key to nonbasic emotions is calibration, a mental mechanism "that establishes a link between judgments of a particular kind and embodied appraisals" (ibid., p. 100). Thus, "for instance, guilt may be sadness brought on by the belief that one has committed a harmful transgression. [Or] pride may be happiness brought on by the belief that one has achieved a difficult task" (ibid., p. 93). But since nonbasic emotions can be elicited by both evaluative judgments and perceptual stimuli that are not part of emotion, the subject may have to rely on her emotional experience in identifying her emotion. Indeed, some theorists suggest that bodily feelings distinguish between different emotion types (e.g. Deonna & Teroni, 2012). However, Prinz does not invoke this argument. Instead, he maintains that the same bodily response patterns serve several emotion types. Consequently, the subject of emotion may have to engage in an interpretation of the environment in order to figure out which emotion - guilt or sadness; pride or happiness, jealousy or anger - she is experiencing, in addition to finding a suitable particular object for the emotion. Even if constructive processes of this kind are in some cases necessary to identify our emotions and the objects that "make us feel the way we do... there is no reason to think that this subset of cases captures the typical epistemological situation we are in when we feel emotions", as Deonna and Teroni (2012, p. 74) point out.
Another problem for a sharp distinction between initiation and response pathways emerges from the twofold nature of emotional dysfunctions. Prinz argues that dysfunctions of emotion parallel those of perception that result from accidents, diseases, lesions, substance abuse, and other organismic damages or failures. Indeed, some dysfunctions of emotion involve or emerge from failures in the emotional 'hardware', so to speak. Alexithymia, which Prinz discusses in passing, may be such a failure. Antonio Damasio presents several dysfunctions of this type in Descartes' Error (1994) and his other books. However, there are also disorders whose pathology resides in the misconnected emotional 'software', such as phobias, panic disorders, OCD, PTSD, as well as other mood and anxiety disorders. These affective disorders can be conceptualized as normal emotional responses to misplaced or inappropriate objects. Prinz mentions those phenomena as well; not as dysfunctions in the emotion system, which he identifies with the response pathway, but in another context, in discussing the initiation pathway that houses calibration files for emotional responses.
Prinz's refusal to characterize anxiety and mood disorders as dysfunctions of emotion is idiosyncratic considering that they are widely, for instance, in the DSM-V manual of psychopathology, classified as such. However, this terminological choice is in line with Prinz's definition of emotions as perceptions of bodily changes. Even so, the choice is not entirely warranted if Prinz thinks, as he does, that the function of emotions is to detect core relational themes rather than bodily changes.
This means that an inappropriate emotional response, such as fear of an innocuous object, quite literally is a dysfunctional emotion; an emotion that does not serve its function, which is to detect dangers in the subject's environment. In order to serve their function of delivering information about how we are faring, the calibration files of emotions must be "well-tuned", as Prinz (2004, p. 240) puts it. But then calibration files become part of the emotion system that cannot be limited to pathways and neural representations responsible for somatosensory perception. Prinz observes this difference between emotion and sense perception in pointing out that "vision does not require an initiation pathway. It is not mediated, under ordinary conditions, by mental states outside the visual system" (ibid., p. 239). However, he does not realize that this mediation of emotions, which is necessary for the emergence of emotions and crucial for their proper functioning, undermines the theoretical divide between initiation and response pathways. An emotional system encompasses both pathways, which allows us to see the essential role of mental causes in emotion.
The more inclusive definition of emotion in terms of both initiation and response pathways refutes Prinz's argument for the noncognitive character of all emotions. Instead, emotions appear to divide into cognitive and noncognitive, depending on the nature of their mental causes that divide into nonconceptual perceptions and conceptual judgments at the algorithmic level of analysis. Nevertheless, this conclusion can be rejected by suggesting that the immediate elicitors of emotions are always fast, automatic and noncognitive appraisals, whatever conceptual thoughts or judgments precede or follow these affective appraisals. This Robinson's strategy, who emphasizes the nature of emotions as processes but limits the role of cognition to emotion regulation.
-  In neurophysiological terms, the amygdala of a clinically anxious or phobic person, for instance, may operate normally. The dysfunction lies in the context in which the emotion is experienced rather than in the response itself. Accordingly, Davidson, Fox and Kalin (2007, p. 56) suggest that mood and anxiety disorders may be best conceptualized as disorders of the context regulation of affect, involving the expression of normal emotion in inappropriate contexts (see also Beer & Lombardo, 2007; Healy, 1990).
-  Obviously, anxiety disorders both in DSM-V and here are understood as involving several episodes of anxiety within several weeks rather than a single, prolonged anxiety episode.