Imagination, Philosophy and the Arts

Matthew Kieran, Dominic Lopes, Matthew Kieran, Dominic Lopes

• Publisher: Routledge
• Year: 2003
• ISBN:
  9781134406241
• Edition: 1
• Topic: Filosofia
• Subtopic: Storia e teoria della filosofia

Imagination, Philosophy and the Arts is the first comprehensive collection of papers by philosophers examining the nature of imagination and its role in understanding and making art.Imagination is a central concept in aesthetics with close ties to issues in the philosophy of mind and the philosophy of language, yet it has not received the kind of sustained, critical attention it deserves. This collection of seventeen brand new essays critically examines just how and in what form the notion of imagination illuminates fundamental problems in the philosophy of art.

Part I

IMAGINATION, NARRATIVE, AND EMOTION

1

REASONS, EMOTIONS, AND FICTIONS

Berys Gaut

It is obvious that one can imagine having an emotion. Less obvious is whether one can have an actual emotion directed toward a merely imagined state of affairs. The latter is a quite general issue about the relation of imagination to emotion. A particular application of this issue has received extensive discussion in philosophical aesthetics in the last twenty years, in respect of emotions directed toward fictional objects. Generally the question has been put like this: is it possible to feel real emotions toward events and situations known to be fictional, and if so, how? The so-called “paradox of fiction” is then construed as the problem of how there can be real emotions directed towards known fictions. On the face of it, such emotions are puzzling: when we are afraid of something, for instance, it seems that we must believe that the feared object exists and is dangerous, and also be disposed to remove ourselves from its dangers. Yet, watching a horror film, we do not believe that the depicted monsters actually exist or are really dangerous, and far from rushing in terror from the cinema, we stay and watch with enjoyment.

Some aestheticians, including Kendall Walton, Gregory Currie, and Jerrold Levinson, have tried to dissolve the paradox by arguing that (with various nuances and qualifications) it is not, in fact, possible to feel real emotions toward known fictional events: very broadly, what is happening is that the spectator is imagining feeling these emotions (Currie 1990; Walton 1990; Levinson 1997; Walton 1997). These theorists do not deny that spectators may feel real emotions toward real-world analogues of fictional events, nor do they deny that there is something emotion-like going on in spectators when they are moved by fictions: it is simply not real fear, pity, disgust, and so forth, that they are feeling toward the fictional events. Let us dub this view emotional irrealism.

Returning a different answer to the question are philosophers such as Peter Lamarque, Noël Carroll, John Morreall, Richard Moran, Derek Matravers, and Richard Joyce, who argue that the phenomenology of spectatorship supports the claim that spectators can feel real emotions toward fictions, and that there are no good philosophical reasons for holding that the phenomenology is non-veridical in such cases (Lamarque 1981; Carroll 1990; Morreall 1993; Moran 1994; Matravers 1998; Joyce 2000). Let us term this view emotional realism.

I argue here for the truth of emotional realism. I also show that there is a distinct puzzle concerning not the possibility but the rationality of fiction-directed emotions. The latter puzzle has often not been as sharply separated from the former issue as it ought to be; for instance, Colin Radford in one of the earliest papers in the debate addressed the question of the rationality of fiction-directed emotions, but his position encouraged a conflation of this with the distinct question of their possibility (Radford 1975). Though distinct, there is nevertheless an important connection between the two issues: it will turn out that the strongest arguments for emotional irrealism stem from arguments based on rationality considerations, so that the puzzle about the rationality of fiction-directed emotions needs to be solved or dissolved in order to resolve the issue of realism.

I first examine the possibility puzzle, and show how realism can dissolve it, since it is based on a misconception of what is required for something to be an emotion. I then turn to the rationality puzzle, examine various realist attempts to solve it, and offer my own solution, based on three rationality criteria grounded on the main characteristics of emotional states. Finally, I compare the solution offered here to three recent attempts to solve the puzzles, and argue for its superiority to them.

The possibility of fiction-directed emotions

Before turning to the paradox of fiction, we need to consider what an emotion is. In one broad, but legitimate, sense of the term, an emotion is a state that is characteristically felt, a state that characteristically has a phenomenology (unlike, say, a belief, which has none). In this sense, we may say that someone is very emotional, and simply mean that he is prone to swings of mood, or tends to feel things deeply. In a narrower sense of the term, an emotion is just one of the various kinds of state that are characteristically felt. It is hard precisely to delineate emotions in this narrower sense from other affective states—indeed, it has plausibly been suggested that emotions do not form a natural class at all (Rorty 1980:1). But it is useful at least to distinguish emotions from moods. A mood, such as happiness or sadness, need not have an intentional object: someone can simply be happy or sad, without being happy or sad about anything. An emotion in contrast has an intentional object: I am afraid of something, I pity someone. According to the dominant (and I would argue correct) cognitive-evaluative theory of the emotions, an emotion not only has an intentional object, but also essentially incorporates an evaluation of that object. So to be afraid of something essentially involves evaluating that thing as dangerous; to pity someone essentially involves construing her as suffering; to be angry with someone essentially involves thinking of that person as having wronged someone about whom one cares. (Cognitive-evaluative theories in various forms are defended by Lyons 1980; de Sousa 1987; Greenspan 1988; Roberts 1988. The particular evaluations mentioned as being essentially connected with each emotion are used for illustrative purposes only; in fact, they may be more complex than is commonly suggested, but that does not matter for our present purposes.) Finally, an emotion is a state that characteristically can motivate actions: I run away because I am afraid, I help someone because I pity her, I strike someone because I am angry with him.

There are, then, three main characteristic aspects to an emotion: the affective, the cognitive-evaluative, and the motivational. In most fully fledged occurrent emotions all three are present, but it is not necessary that all three be present for an emotion to exist. Not all of the characteristic features of an emotion are essential to it. In particular, many emotional episodes are non-occurrent: one can be angry with someone for a long time, even when he is absent and one is not thinking about him. One need not have any feeling of anger during these long periods. Nevertheless, if one’s anger does become occurrent, perhaps because the person is present, characteristically one will experience a feeling of anger. According to the cognitive-evaluative theory, an emotion is not however individuated by its particular feeling. The feelings that we experience when we are ashamed and when we feel guilty, for instance, might well be identical, and what individuates them is the content of the respective evaluative thoughts—of having done something shameful, as opposed to having done something morally wrong. Lyons, for instance, holds that an occurrent emotion involves bodily sensations of felt disturbance caused by certain evaluations; but he does not maintain that each emotion has a unique feeling associated with it by which we individuate the emotion (Lyons 1980). A feeling must be present in occurrent emotion, but what feeling it is might vary from occasion to occasion, or from person to person.

Turning back to our problem, the paradox of fiction has been discussed chiefly in terms of the emotions of pity and fear, the emotions aroused by tragic representations. But if we cast our net wider, the irrealist case becomes less compelling. Disgust, for instance, is (certainly in the broad sense) an emotion; but it is odd to say that one has to know whether something really occurred if one is to be disgusted at the thought of it. If I describe in vivid and detailed fashion the jam-and-live-mouse sandwich that I breakfasted on the other day, you might well be disgusted, without supposing that I am telling the truth. And if I tell you a joke, you don’t have to wait to find out whether the events I describe really happened before you can find the joke funny. I can also be awed by or feel wonder at a depicted landscape, even though I know that the artist invented it. Nor does there seem to be anything problematic (contra Walton 1990:203) about the possibility of admiring a fictional character, such as Superman.

Stronger, though, are the intuitions supporting the irrealist case in respect of fear and pity. How can one really fear something, or fear for someone, if one does not believe that the thing is dangerous or the person endangered? And how can one not be motivated to flee if danger really threatens one? Indeed, how could one enjoy these states if one really were afraid? Similar considerations apply to pitying someone.

If the irrealist is right about these matters, he or she owes an account of what is going on in such cases. Perhaps the most developed version of such an account is due to Kendall Walton. Walton holds that it is only fictional that Charles, the spectator of a horror movie, is afraid of the green slime that features in it, and what makes this the case is that Charles “experiences quasi fear [the sensations and physiological reactions normally associated with fear] as a result of realizing that fictionally the slime threatens him. This makes it fictional that his quasi fear is caused by a belief that the slime poses a danger, and hence that he fears the slime” (Walton 1990:245). According to this analysis, spectators actually experience the sensations and physiological reactions normally associated with fear; these reactions do not, however, constitute fear when directed toward fictions. The account thus readily provides an error theory for why spectators might falsely conclude that they really are afraid, even though they are only make-believedly afraid.

One evident disadvantage that such an account has over realism is its far greater complexity: whereas the realist holds simply that Charles is afraid of the Green Slime, the irrealist holds that he is in the complex state described above, which would involve his possession of concepts such as quasi-fear. We can indeed ascribe such concepts to spectators if they are necessary to explain the phenomena, even if spectators are not explicitly aware of possessing the concepts (think of how Chomskyans ascribe knowledge of deep grammatical structures to competent speakers, though such speakers have no explicit knowledge of them). But if there is a simpler account available, which adequately explains the phenomena, we ought on general heuristic grounds to adopt that. And this means that, unless the irrealist has good grounds for showing that realism is inadequate, then realism ought to be the preferred account.

We have already adverted to the kinds of considerations to which irrealists can appeal; they can be organized around the three aspects of an emotion: the affective, the cognitive-evaluative, and the motivational.

First, if a fiction-directed emotion is genuine, how could one enjoy feeling it? After all, real fear is intrinsically unpleasant, and people go out of their way to avoid it. But spectators enjoy horror films, and also a wide variety of other fictions in which they fear for characters. That strongly suggests that such spectators are only make-believedly afraid, not really so. However, this argument fails. For one of the many merits of Walton’s account is that it allows for make-believe fear to have exactly the same phenomenology as real fear—that is what quasi-fear is. And any adequate account of make-believe fear must keep its phenomenology the same as that of fear, because it is a truth of introspection that one can be in a state that feels like fear when watching a horror film. But if one insists that one cannot enjoy fear, because the feel of fear is unpleasant, then one must equally insist that one cannot enjoy make-believe fear, since, sharing the same phenomenology, that too must be unpleasant. Moreover, as I have argued elsewhere (Gaut 1993), it is a mistake to hold that fear is intrinsically unpleasant. According to the cognitive-evaluative theory it is not the feeling that individuates the emotion, but the cognitive content; and this allows for at least some cases where fear can be felt as pleasant (perhaps because of its associated “arousal jag”). Bunjee-jumpers, mountaineers, and many others who are confronted with real danger, can seek out and enjoy fear; the cognitive theory shows how this is possible.

The second, cognitive-evaluative aspect of an emotion seems to provide a stronger ground for the irrealist case. Real emotions require their subjects to have appropriate beliefs: real fear requires one to believe that someone about whom one cares is threatened. If a man in a pub tells you a harrowing story about his sister, then you will cease to be harrowed if the man then admits that he made the story up, and does not have a sister (Radford 1975). However, the example does not prove that an emotion requires a belief: if one’s emotion does evaporate on being told that the sister is an invention, that merely shows that if an emotion is grounded on a belief, removal of that belief tends to remove the emotion grounded on it. It does not support the more general claim that having an appropriate belief is a necessary condition for the existence of an emotion.

The claim that to have an emotion, a person must have the appropriate evaluative belief, might seem to follow from the cognitive-evaluative theory of the emotions. But in fact cognition is a broader category than belief, including imaginings in a broad sense—unasserted thoughts, construals, seeings-as, and so on. The more restricted version of cognitivism, which holds that emotions require beliefs about their objects has been dubbed judgmentalism, and there are good reasons to reject it. Consider Fido, an old and harmless dog. Due to a traumatic childhood incident, Fred has a neurotic fear of dogs, and Fred fears Fido. Fred does not believe that Fido is dangerous: Fred has, for instance, no tendency to warn others about him. Fred simply cannot help seeing Fido as dangerous, and he responds to Fido with fear. So it is possible to feel fear while not possessing a belief that the object of one’s fear is dangerous (Greenspan 1988: chapter 2). Or consider cases where I vividly imagine my hand being mangled by a machine, and I feel fear, though I firmly believe that my hand is not caught in the machine (Carroll 1990: 78–9); or when I vividly imagine plunging to my death on the rocks below, though I believe that I am safely behind the parapet on the cliff-top. It is also possible to fear merely counterfactual cases (Moran 1994). For instance, it makes sense to be afraid of what would have happened if Hitler had won the Second World War. And a visitor to the CN Tower in Toronto can have the experience of walking on a glass floor several hundred feet up on the observation deck of the tower, seeing the ground far below one’s feet. As one edges gingerly across the floor, one certainly believes that it is safe to do so (otherwise it would be madness to walk across), but the experience of fear is perfectly genuine: the appearance of extreme danger is a very vivid one, and grounds one’s fear of falling.

The irrealist will deny the possibility of these cases. First, it will be objected that all of these examples are ones in which one does have a belief, albeit an irrational and probably subconscious one. Several phobic fears may be like this: someone who fears flying may well have an irrational belief that flying is dangerous. However, while this may be true of some phobias, it fails as a plausible explanation for the full range of cases considered above: I do not, for instance, have a subconscious and irrational belief that Hitler really did win the Second World War, or that I will fall to my death when I walk across the glass floor—if I did have the latter belief, I would not walk across. The better explanation is that the vivid imagining or appearance of danger is sufficient to ground real fear.

Second, it has been objected that anti-judgmentalist accounts mistake the intentional object of emotions. Carroll dubs his version of the anti-judgmentalist view the Thought Theory, and holds that Charles fears the thought of the Green Slime, and this thought need not be asserted (Carroll 1990). A related position is defended by Lamarque (Lamarque 1981). Walton has objected to the latter that one does not fear the thought, but the object of the thought—the Green Slime (Walton 1990:203). The anti-judgmentalist should simply agree: the theory should hold that one fears not the act of thinking of the Green Slime, but the thought-content: that is, the Green Slime. Judgmentalism holds that asserted thoughts, that is, beliefs, are necessary for emotions; and that the intentional object of the fear is the belief-content, that is, what the belief is about; anti-judgmentalism holds that unasserted thoughts are all that may be required for an emotion, and should relatedly hold that the object of fear is the thought-content, that is, what the thought is about.

Finally, an irrealist can object that the Fido case and its ilk show only that it is not necessary to believe that an object is dangerous to fear it; but it is compatible with this that one must believe that the object exists in order to fear it. So fear does not require characterizing beliefs that something is dangerous, but it does require existential beliefs, and hence the paradox of fiction is not solved by the anti-judgmentalist stratagem. Something like this suggestion is made by Levinson (Levinson 1997). However, such a half-way position lacks motivation: fear has as its formal object the dangerous, so the fact that one believes that the object of fear exists, even though it lacks the property of being dangerous, would be no less puzzling than the case where one does not believe that object of fear exists at all. And far from it being the case that fear must be directed to what one believes exists, fear is often felt toward what may happen in the future, not what one believes is presently so.

The third, motivational, aspect of emotions grounds what some irrealists have seen as their most powerful argument. Walton remains neutral on Greenspan’s anti-judgmentalism, but holds that there are nevertheless good reasons for thinking that Charles, unlike Fred who fears Fido, is not really afraid, for Charles is not motivated to flee, and “fear emasculated by subtracting its distinctive motivational force is not fear at all” (Walton 1990:102). Similarly, the irrealist can hold that I am not motivated to help Anna Karenina, and that this undermines the claim that I pity her.

In reply, the realist needs to get clearer about what the motivational aspect of an emotion involves. Consider a parallel argument to the above in respect of historical victims: it might be claimed that I do not really pity Anne Boleyn, since I am not motivated to help her in any way. But I really can pity Anne Boleyn; and this is so by virtue of the fact that I have a genuine desire or wish that she had not suffered as she did. It is the presence of such a desire or wish that constitutes the motivational aspect of the emotion of pity. If, for instance, I said that I pitied Anne, but had no desire at all that she not suffer as she did, there would be defeasible grounds for questioning whether I really did pity her. But I do not try to help Anne, since I lack the belief that there is something I can do to help her. It is absence of this belief, not of the desire (or therefore of the emotion), that explains why I do not act. So the absence of motivated action does not prove that I do not pity Anne; rather, I lack the appropriate beliefs about being able to help her. In the same way, I may genuinely pity Anna Karenina, since I wish that she had not suffered as she did, but I lack the belief about being able to help her in any way; so I do not try to do so. The point applies not just to historical cases, but also to anyone whom I believe I cannot help. If I watch a live television broadcast of a hurricane sweeping through a small Florida town, I can genuinely pity the inhabitants, though I do not try to get them out of the way of the hurricane, since I know that there is nothing I can do to prevent it from engulfing...