Epistemology and Emotions

Georg Brun, Ulvi Doguoglu

• Publisher: Routledge
• Year: 2016
• eBook ISBN: 9781317141969
• Edition: 1
• Topic: Philosophy
• Subtopic: Philosophy History & Theory

Undoubtedly, emotions sometimes thwart our epistemic endeavours. But do they also contribute to epistemic success? The thesis that emotions 'skew the epistemic landscape', as Peter Goldie puts it in this volume, has long been discussed in epistemology. Recently, however, philosophers have called for a systematic reassessment of the epistemic relevance of emotions. The resulting debate at the interface between epistemology, theory of emotions and cognitive science examines emotions in a wide range of functions. These include motivating inquiry, establishing relevance, as well as providing access to facts, beliefs and non-propositional aspects of knowledge. This volume is the first collection focusing on the claim that we cannot but account for emotions if we are to understand the processes and evaluations related to empirical knowledge. All essays are specifically written for this collection by leading researchers in this relatively new and developing field, bringing together work from backgrounds such as pragmatism and scepticism, cognitive theories of emotions and cognitive science, Cartesian epistemology and virtue epistemology.

Chapter One

Emotion and Understanding

Catherine Z. Elgin*

Emotions share important cognitive functions with perceptions and beliefs. Like perceptions, they afford epistemic access to a range of response-dependent properties, such as being admirable or contemptible, and provide evidence of response-independent properties that trigger them. Fear is evidence of danger; trust is evidence of reliability. Like beliefs, emotions provide orientations that render particular facets of things salient. In the grip of an emotion, we notice things we would otherwise miss. The variability and volatility of emotional deliverances might seem to undermine their claim to epistemic standing. I argue that variability and volatility can be epistemic assets, keying the subject to multiple, quickly changing features of things. Emotions, like other modes of epistemic access, are subject to refinement to increase their epistemic yield. The arts provide opportunities for such refinement.

1. The Claim to Epistemic Standing

‘Reason is and ought only to be the slave of the passions,’ Hume maintains ([1739/40], 415). Others reverse the relationship, contending that to reason well requires subduing, dominating or discounting the passions. Either way, reason and passion are antithetical. To be under the sway of emotion is to be irrational. To be rational is to be cool, calm, and deliberate; that is, to be unaffected by emotion. Let us call this the standard view. I think it is a mistake. Rather than being opposed to reason, I suggest, emotion is a facet of reason. It is an avenue of epistemic access, hence a contributor to the advancement of understanding.

Emotional deliverances are representations conveyed through emotional channels. A representation of frogs as dangerous that presents itself via fear of frogs is a deliverance of that fear. If the standard view is correct, emotional deliverances are at best epistemically inert, there being no reason to trust them. At worst, they are deleterious; there is reason to distrust them. Either way, insofar as our goals are cognitive, if we cannot subdue or silence the passions, we should ignore or discount their deliverances.

In contending that the standard view is mistaken, I do not mean that every emotional deliverance, as it stands, is epistemically acceptable. Rather, I believe that emotions provide resources that serve epistemic ends. But just as natural resources like iron ore need to be processed to yield material we value, so do emotional deliverances. We need to know how to recognize the epistemically valuable insights that emotions afford and how to use them effectively. The goal of this chapter is twofold: first, to show that emotions afford such resources; and second, to say something about how to refine the raw materials and increase their epistemic yield.

Elsewhere I have argued that an understanding is a system of cognitive commitments in reflective equilibrium. The individual commitments that comprise such a system must be reasonable in light of one another, and the system as a whole must be at least as reasonable as any available alternative in light of relevant antecedent commitments (Elgin 1996). A system’s equilibrium derives from the mutual supportiveness of the components. Its answering to antecedent commitments at least as well as available alternatives insures that the equilibrium is one that on reflection we can accept. Not all the cognitive commitments that comprise such a system are truth bearers: perspectives, vocabularies, methods and standards are as integral to an understanding as beliefs. Nor need the truth bearers in a tenable system always be true. A rudimentary science that contains only rough approximations can provide some understanding of its subject. That is why it is worth taking seriously. Finally, understanding is holistic. No commitment is fully tenable in isolation.

A systematic, interconnected network of utterly untenable commitments would not yield an understanding of the phenomena it bears on. Coherence alone does not suffice for tenability. The tie to antecedent considerations is crucial. Unless some considerations have some initial tenability prior to systematization, warrant cannot be generated. But this is not our situation.

At any given time, a person has a cluster of cognitive commitments. These include her beliefs, perceptual takings, and emotional deliverances. They also include the methods, perspectives, and sources she tends to trust, as well as the epistemic priorities and weighting factors she endorses. All such commitments are initially tenable. Their being held gives them a slight claim on epistemic allegiance. But initial tenability is not full tenability. No commitment, however firmly held, is fully warranted in isolation. Epistemic warrant accrues through systematization – the development of an integrated network of mutually supportive cognitive commitments.

Systematization requires revising as well as conjoining and augmenting, for a person’s initially tenable commitments can, and often do, clash. They are apt to be incompatible, non-cotenable or implausible in light of one another. So to arrive at an acceptable system, we need to revise and/or reject some of our initial commitments and to adopt others that we previously did not hold. A tenable system thus need not incorporate all the antecedent commitments it answers to. But if it does not, it should show why the commitments it rejects seemed reasonable when they did. It might, for example, reveal that they owed their previous plausibility to limited evidence or unsophisticated models. An individual commitment is warranted by its place in a tenable system of thought.

The crucial point is this: To say that emotional deliverances are initially tenable is to assign them only a weak and precarious epistemological status. They have that status because an agent, in the grip of an emotion, has a tendency to credit its deliverances. To experience a frisson as a fear of frogs, is to take that frisson to embed the idea that frogs are dangerous.

To credit emotional deliverances with initial tenability might seem trifling. The epistemic status of emotions turns not on whether their deliverances are initially tenable, but on whether they are fully tenable. If such deliverances are immediately and decisively overridden, then the fact that they start out with some measure of initial tenability seems insignificant. This concern has merit. For there is a clash. The standard view that emotional deliverances are unreliable is in tension with the opinions emotions embed. But the standard view is not unfounded. Emotions at least sometimes distort or derail reason. The challenge facing me is to show that the best way to alleviate the tension is to reject or revise the standard view. This does not require saying that all emotions under all circumstances are epistemically estimable, or that the deliverances of those that are must be taken at face value. It requires only showing that systems that integrate emotional deliverances are sometimes more tenable than rivals that exclude them.

Although widespread, the conviction that emotions are epistemically inert is a bit odd. Emotions are not spontaneous upwellings of arbitrary feelings. They are reactions to events. So if we can correlate emotional reactions with the events that trigger them, we can use those reactions as sources of information about the environment. Experiencing emotion a would be an indication that circumstance b obtains. The standard view considers such correlations inherently unreliable. To be sure, the existence of a causal connection between individual emotion and event pairs does not show otherwise. What we need is evidence that emotional reactions are reliably correlated with events. If an arbitrary subject were as likely to feel joy, dismay, revulsion or amusement, regardless of the trigger, then from the occurrence of a particular emotion, nothing would follow about the nature of its source. The question then is whether suitably reliable correlations can be found.

Our views about the informativeness of emotions waver. In the grip of an emotion, one normally believes its deliverances. When I am frightened, I think that the situation is dangerous. When I am infatuated, I think that my beloved is wonderful. Nor, at the time, do I consider the connection between my occurrent emotions and beliefs accidental. I am frightened, I believe, because the situation is dangerous. I adore him, I believe, because he is wonderful. In cooler moments, I may think differently. I recognize that many of my fears have proven unwarranted. I concede that I have not been drawn unerringly to wonderful men. Such failures might persuade us that suitably reliable correlations are not to be had. Then whatever we think or feel in the heat of the moment, it might be wise to defer to our cooler judgement that emotional deliverances are not trustworthy sources of information. Still, we go too quickly, I think, if we dismiss them.

One reason is biological (de Sousa 1987). The limbic system, the seat of emotion, is a product of evolution. Although evolution yields some by-products that lack survival value, the fact that a trait or complex of traits is a product of evolution is evidence that it promotes fitness. Emotions might promote fitness by being indicators of significant features of the environment. If fear is triggered by danger and is a sufficiently reliable indicator of danger, then developing and exercising the capacity for fear would be conducive to survival. If having their father present enhances the survival prospects of young humans, then a mother’s eliciting the love of someone she can live with promotes her reproductive success.

Such evolutionary arguments are weak. The evolutionary lineage of emotions is compatible with their having no survival value at all. They could be ‘free riders’ on genes whose contribution to fitness lies elsewhere. At most, the evolutionary argument shows that there is some presumption that the capacity to experience emotions is adaptive. It does not show that the capacity to experience each emotion is adaptive. Nor does it show that if that capacity is adaptive, its adaptiveness lies in a correlation between emotions and circumstances. Perhaps the deliverances of emotions, like the contents of dreams, are independent of the circumstances in which they occur. The biological contribution of emotions might then lie in something like their expending excess neural energy or in making and breaking covalent bonds in the amygdala, not in anything that affords epistemic access to the circumstances in which they occur. Nevertheless, it is not unreasonable to think that fear promotes fitness by sensitizing to danger or that affection between parents promotes fitness by fostering living conditions that increase the likelihood that offspring will survive. The fact that emotions evolved affords some presumption that their deliverances vary with circumstances. Given that presumption, it is not unreasonable to think that they are keyed to circumstances in such a way that if we can discover the key, we can use the occurrence of particular emotions as sources of information about the circumstances in which they occur. All the evolutionary argument contends is that it is not unreasonable to suppose that emotions are capable of affording epistemic access to their objects. Weak, though it is, the argument is strong enough to do that.

2. Response Dependence

Some properties – such as being contemptible, or admirable, or amusing, or depressing – dovetail with emotions in such a way that without those emotions, the properties would not exist. These properties are response-dependent (Johnston 1989). They are genuine properties of the objects that possess them, but they owe their identities to responses they evoke. If contemptibility is the property it is because of the contempt it is apt to evoke, it would be astonishing if feelings of contempt did not afford epistemic access to contemptibility. This is not, of course, to say that my contempt for Karl demonstrates that he is contemptible. My contempt could be misplaced. But if the property of being contemptible depends for its identity on evoking feelings of contempt, then at least some evokings of contempt and perhaps other related emotions should be evidence of contemptibility. Indeed, emotional responses seem to afford our most direct access to such properties.

The foregoing considerations suggest parallels to perception. Perception is both triggered by and indicative of aspects of the environment. Perceptual systems evolved and endure because their deliverances promote fitness. Being able to see, or hear, or smell a predator, like feeling instinctively afraid of it, enhances an animal’s prospects of evading it. Perception manifestly affords epistemic access to useful information about the environment.

Secondary qualities, such as colours, tones, and odours, are response-dependent. They are known by perception, and perhaps cannot be fully known without perception (Jackson 1982). Still, mistakes are possible. Something’s looking blue to me does not establish that it is blue. An isolated colour perception is not epistemically acceptable on its own. But something’s looking blue is ordinarily evidence that it is blue. If the analogy holds, emotional deliverances are indicators, but not always accurate indicators of aspects of their objects. Just as my experiencing something as blue is evidence, but not conclusive evidence, that it is blue, my being frightened of something is evidence, but not conclusive evidence, that it is dangerous.

According to a familiar criterion for colours, something is blue just in case it would look blue to a normal observer under normal conditions. Conditions often are normal and most of us are normal colour perceivers. So in the absence of contraindications, one reasonably takes it that something’s looking blue to her is reason to believe that it is blue. But the criterion for secondary qualities is not always the response of the normal perceiver. In assessing whether a bassoon is in tune, the criterion is not how it would sound to a normal perceiver, but how it would sound to a perceiver with perfect pitch. Unless I believe myself to have perfect pitch, I assign considerably less weight to my judgement that a bassoon is in tune than I do to my judgement that the bassoonist is wearing a blue shirt. If the analogy with emotions holds, the criterion for some cases could be what a normal subject would feel in normal circumstances (with both occurrences of ‘normal’ spelt out in an informative way), while in other cases it could be what a suitably sensitive subject would feel in specified circumstances. Like perceptions, emotions might key to circumstances in any of several different ways.

The capacity to distinguish among secondary qualities is neither uniform nor fixed. One perceiver might consider two colours or tones identical, while another readily distinguishes between them. A single person’s capacity to discriminate may vary over time and with circumstances. Colours that look indistinguishable against one background are often easily distinguished against another. Moreover, perceptual abilities are subject to refinement. We can develop an ability to distinguish between properties that we previously could not tell apart. At a wine tasting, for example, one learns to taste the difference between wines that initially tasted the same.

Although secondary qualities are more tightly linked to perception than other perceptible properties, the same variability, fallibility and sensitivity to circumstances is characteristic of perception in general. Some people can identify the makes of American cars on sight; others cannot. Some can distinguish between deciduous trees in full leaf, but not when the limbs are bare. Skiers learn to differentiate qualities of snow on the basis of the way the snow looks and feels, even though originally snow seemed to them to be just one, undifferentiated sort of stuff.

Despite this variability, fallibility and sensitivity to circumstances, we readily concede that perceptual deliverances afford epistemic access to their objects. So the fact that emotional deliverances are similarly variable, fallible, and sensitive does not disqualify them from epistemic standing. Rather, the epistemic yield of emotions, like the epistemic yield of perceptions depends not on taking all deliverances at face value, but on a sophisticated understanding of when and to what extent they are trustworthy.

Our practice of assessing the appropriateness of various emotions reflects such an understanding. Phobias are irrational fears. Our ability to distinguish between phobias and other fears shows that we have a lien on which fears are rational. It is hard to see what the basis for the distinction would be, if it were not that rational fears are ones that align with genuine dangers. To recognize a fear as rational is to recognize that the belief that its object is dangerous is at least prima facie rational as well. We also make more local assessments, for example, when we charge people with overreacting. We say that Harry is insanely jealous, indicating that his jealousy is far greater than it should be. This is quite different from saying that he has no basis for jealousy. We say that Fred is excessively angry, indicating that some lesser level of anger would be appropriate. We think that Sam should be proud of himself, indicating that were he to feel pride, his doing so would be reasonable. Our propensity for charging people with over- or underreacting indicates that we take ourselves to be in a position to tell when people’s emotions are apt. That is, we take ourselves to be able to reliably correlate emotions with circumstances. The robustness of our practice is evidence that we are right.

3. Is Instability a Problem?

Despite these parallels to perception, one might argue, the deliverances of emotions are untrustworthy, because emotions are far more volatile than perceptions. To investigate this charge, we need to examine two questions: (1) whether emotions are volatile and (2) whether volatility is epistemically incapacitating.

As Hume notes, not all passions are volatile (Hume [1739/40], 276). Calm passions, such as fondness for a lifelong friend, are steady, enduring, and unexciting. Nor need more violent passions be volatile. Some people carry a grudge for decades, without in the least moderating their attitude toward its object. Although some emotions may be volatile, volatility is not an invariable property of emotions. If volatility is an epistemic disqualifier then, only some emotions and their deliverances are disqualified.

But it is not clear that volatility should be a disqualifier. The question is whether emotions are, and can be known to be, so tied to circumstances that from the occurrence of an emotion we can glean information about the circumstances. There is no obvious reason why a volatile emotion should not be so tied. Suppose, for example, that fury is a volatile emotion, which normally arises only when exceedingly objectionable events occur and which ebbs quickly when objectionability wanes. In that case, if objectionability quickly waxes and wanes, fury’s volatility is simply a responsiveness to rapidly changing circumstances. This would be an epistemic asset rather than a defect. Evidently, neither the violence nor the volatility of an emotion undermines its epistemic qualifications. To do that would require showing that the emotions vary independently of circumstances – that, for example, fury arises or endures regardless of the objectionability of the object.

This might be the case. Emotional reactions seem to vary considerably from one person to another. The same object can anger one person, mildly irritate a second, sadden a third and amuse a fourth. That being so, one might think, an emotional deliverance could not possibly afford reliable information about its object. Such wide variation might suggest that emotional deliverances are entirely subjective. Pretty much anything, it seems, could trigger any emotion; so from the occurrence of an emotion, nothing can be gleaned about the nature of the object.

Again it is worth considering the analogy with perception. Similarly situated observers viewing the same scene may see different things because of differences in background knowledge and interests. A paediatrician sees a case of measles, when a parent just sees her child’s rash. A linguist hears a Texas accent, while a reporter hears a political speech, and the audience hears a call to arms. A critical grandmother sees crude, ill-mannered behaviour in her granddaughter, whereas a doting grandmother sees the same behaviour as refreshingly spontaneous (Murdoch 1970, 17–18). If the rash is a case of measles, both the parent and the paediatrician are right. If the political speech is a call to arms issued in a Texas accent, all three auditors are right. If the granddaughter’s behaviour lies at the intersection of ‘crude and ill mannered’ and ‘refreshingly spontaneous’, both grandmothers are right. The mere fact that different observers perceive the same situation differently does not show that any of them is wrong, or that perception is not reliable.

Many events are both infuriating and depressing, so the fact that some people experience anger and others sadness in response to the same event does not show that the responses do not track features of the world. But, we are apt to think, if it is infuriating, it is not amusing. So the fact that the same e...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• List of Contributors and Editors
• Guide to the Essays
• Introduction: A New Role for Emotions in Epistemology?
• 1 Emotion and Understanding
• 2 Epistemic Immediacy, Doubt and Anxiety: On a Role for Affective States in Epistemic Evaluation
• 3 Virtues, Emotions and Fallibilism
• 4 Conflict Without Contradiction
• 5 Epistemic Immediacy and Reflection
• 6 Critical Reflections on Affective Epistemology
• 7 Misleading Emotions
• 8 How Cognition Meets Emotion: Beliefs, Desires and Feelings as Neural Activity
• 9 Epistemic Feelings
• Index