Emotions, Actions and Inclinations to Act

Abstract

Emotional responses to fiction are part of our experience with art and media. Some of these responses (“fictional emotions”) seem to be directed towards fictional entities—entities that we believe do not exist. Some philosophers argue that fictional emotions differ in nature from other emotional responses. (cf. Walton in J Philos 75(1):5–27, 1978, Mimesis as make-believe, Harvard University Press, Cambridge, 1990, Walton, in: Hjort, Laver (ed.) Emotion and the arts, Oxford University, New York, 1997; Currie in The nature of fiction, Cambridge University Press, Cambridge, 1990; Stecker in Br J Aesthet 51(3):295–308, 2011) The claim is supposed to be supported among others by ‘the argument from action.’ In contrast to genuine emotions, proponents of this argument claim, fictional emotions do not motivate their bearers to act. (cf. Yanal in Paradoxes of emotion and fiction, Pennsylvania State University Press, University Park, 1999; Lamarque in Br J Aesthet 21(4):291–304, 1981; Carroll in The philosophy of horror: or, paradoxes of the heart, Routledge, London, 1990; Currie 1990; Walton 1978, 1990; Suits in Pac Philos Q 87(3):369–386, 2006; Friend, in: Kind (ed.) The Routledge handbook of philosophy of imagination, Routledge, New York, 2016) This claim grounds in what may appear to be an obvious fact: that viewers and readers of are not led to act by their fictional emotions. It is certainly true that viewers and readers of fiction do not form intentions to perform actions directed towards fictional entities. In contrast to the proponents of the argument from action, I will argue that the lack of any such intentions can be explained only with reference to intending’s doxastic conditions, conditions that are unsatisfied in the fictional scenario. Decisively, this explanation does not refer to the motivational force of the agent’s emotions; indeed, it doesn’t refer to emotions at all. Thus, the lack of intentions to perform actions directed towards fictional objects provides no support for the claim that fictional emotions are no genuine emotions.

1 The ‘Paradox’ of Fiction and the Nature of Fictional Emotions

Emotional responses to fiction are part of our experience with art and media. Some of these responses, which are sometimes labelled “fictional emotions,” seem to be directed towards fictional entities—entities that we believe do not exist. Some philosophers argue that fictional emotions differ in nature from other emotional responses. (cf. Walton 1978, 1990, 1997, Currie 1990, Stecker 2011) The claim is supposed to be supported among others by the ‘argument from action’. In contrast to genuine emotions, proponents of this argument claim, fictional emotions do not motivate their bearers to act. (cf. Yanal 1999, Lamarque 1981, Carroll 1990, Currie 1990, Walton 1978, 1990, Suits 2006, Friend 2016). This claim grounds in what may appear to be an obvious fact: that viewers and readers of fiction are not led to act by their fictional emotions. In a first step, I will argue that fictional emotions raise the question as to whether they differ decisively from other emotional responses, even if a narrow cognitive theory of the emotions has to be rejected. In Sect. 2, I will modify the traditional ‘paradox of fiction’, presenting a ‘practical’ paradox, which focuses not on cognitive, but practical features of fictional emotions, and formulate on this basis the argument from action. The following sections present my discussion of the argument: in Sect. 3, I will argue that fictional emotions typically don’t lead an agent to perform intentional actions of a specific kind, namely actions that are directed at the object of the emotion. However, fictional emotions share this feature with many other emotional responses. If such responses are non-genuine, the set of non-genuine emotions will be much larger than has generally been taken to be the case. In Sects. 4 and 5, I will go on to examine the question as to whether the reason for the absence of actions on the basis of fictional emotions lies in the emotions themselves. I will briefly describe how recent theories of the emotions account for their motivational force. It turns out that among current theories of the emotions only a component theory of the emotions, together with particular assumptions about the nature of desires, could support the claim that fictional emotions lack motivational force. Nevertheless, it is certainly true that viewers and readers of fiction do not form intentions to perform actions directed towards fictional entities. In contrast to the proponents of the argument from action, I will argue that the lack of any such intention can be explained only with reference to intending’s doxastic conditions, conditions that are unsatisfied in the fictional scenario. Decisively, this explanation does not refer to the motivational force of the agent’s emotions; indeed, it doesn’t refer to emotions at all. Thus, the lack of intentions to perform actions directed towards fictional objects provides no support for the claim that fictional emotions are no genuine emotions.

The question as to whether fictional emotions are genuine emotions has been most influentially discussed with a focus on a cognitive feature. Take the following example: Anna is watching a movie about vampires and is mortally terrified of the bloodsucking monsters. But she knows that the film is fictional and believes that there are no such things as vampires. We seem to believe by default that the people, places and situations in such fictional stories do not exist. Nevertheless, we are often emotionally involved. This problem—the so called Paradox of Fiction—arises because each of the following three claims seem to be true, although their conjunction must be false:

• 1. Readers have genuine emotions towards fictional characters.

• 2. In order to have genuine emotions towards an object, one must believe that the object exists.

• 3. Readers do not believe that fictional characters exist.

Since Radford’s (1975) article “How can we be moved by the fate of Anna Karenina?”, philosophers have discussed different solutions to this ‘paradox.’ The core of this debate is the question concerning the nature of fictional emotions.¹ Kendall Walton’s well-known early discussion (Walton 1978) focuses on claim 1, which he rejects. A reason to reject 1 is that one accepts claims 2 and 3.

By rearranging the elements of the ‘paradox’ so that claims 2 and 3 become premises for a conclusion consisting of the negation of claim 1, we get a simple argument:

• P1: In order to have genuine emotions towards an object, one must believe that the object exists.

• P2: Bearers of fictional emotions do not believe that the objects of their emotions exist.

• Conclusion: Fictional emotions are not genuine emotions.

Within this debate, it is often said that philosophers who advance claim 2 (or something similar) have a narrow cognitive theory of emotions (Friend 2016; for a critical discussion of the role of narrow cognitivism see Cova and Teroni 2016). According to narrow cognitive theories, a genuine emotion has to involve, or be identical to a belief or judgement (Solomon 1976; Nussbaum 2001). Fear, for example, must involve a belief or judgement that the object one fears is dangerous. Because typical readers (or viewers) believe that fictional entities do not exist, it seems impossible that they simultaneously evaluate these entities as dangerous. Therefore, narrow cognitivists reject 1.

Many arguments have been advanced against narrow cognitivism, independently of the debates about fictional emotions. It has been claimed that narrow cognitivism cannot explain emotions of animals and infants who lack propositionally structured beliefs or judgements (Deigh 1994). According to a second critique, a theory that identifies emotions with judgements neglects or cannot explain emotions’ phenomenal character. The relation between a judgment and an emotion, it is further argued, is too loose to individuate the emotion (Deonna and Teroni 2012). Another worry highlights the fact that we sometimes respond to a situation with fear although we believe that there is no danger at all. A narrow cognitive theory needs to explain such scenarios on the basis of contradictory judgments on the part of the agent (Nussbaum 2001), an explanation that is rejected because it is taken to assume a wrong kind of irrationality (Döring 2008). According to a final critique, narrow cognitivism cannot explain the emotions’ intimate relation to action or behaviour (Deonna and Teroni 2012). As a result of these trenchant criticisms, narrow cognitivism is a minority position in recent philosophical and (neuro-)psychological debates about the nature of the emotions.²

If a theory of the emotions does not entail the claim that emotions involve or are to be identified with beliefs or judgements, the ‘paradox’ seems to disappear. Accordingly, many authors consider it not worth any further theoretical effort (Stecker 2011; Buckwalter and Tullmann 2014; Matravers 2014; for a critical discussion and an attempt to solve rather than dissolve the paradox see Cova and Teroni 2016). However, even if there is no paradox of fiction, there are still important open questions about our emotional responses to fiction. It seems that, independently of narrow cognitivism, there are other reasons why claim 1 is sometimes rejected (Walton 1990, 1997; Stecker 2011; Friend 2016). Fictional emotions, it is often claimed, differ in too many respects from standard emotions. The important arguments for our purposes focus on the putative fact that fictional emotions have no connection to action or action tendencies.

It is worth mentioning that we can find fairly similar claims in another debate, viz. the debate on empathy and simulation theory. For quite similar reasons, the affective states that occur during or as a result of simulating other mental states and processes are frequently characterised as quasi-emotions or non-genuine emotions. Max Scheler claims that the feelings that occur in the process of what he calls “Nacherleben” (“experiencing-after”) and “Nachfühlen” (“feeling-after”) are no proper feelings. Feeling what another feels, he claims, is more than just knowing what they are feeling; nevertheless what is felt in such processes is not a “real feeling” (Scheler 1923). Recently, similar claims have been made by Goldman (2006) and Kaupinnen (2010). Within the more recent debate, both cognitive states and emotions generated by simulation processes are characterized as “off-line.” One of the main arguments for this characterisation is that these states are disconnected from actions. Not only empathic states, but also imaginings or ‘make-beliefs’ are often characterized as off-line states and typically distinguished from beliefs because of their different functional role (Nichols and Stich 2000; Currie and Ravenscroft 2002; Van Leeuwen 2009, 2016; Sinhababu 2012; Kind 2016; for a critical discussion see Buckwalter and Tullmann 2017; Ichino 2019): imaginings, then, in contrast to beliefs, are also said not to motivate action. This is considered to be a decisive difference between off-line and on-line states. There are, then, important reasons for examining the argument from action.

2 A Practical “Paradox” of Fiction and the Argument from Action

The 20th-century causal theory of action, as first formulated by Davidson, included all sorts of “pro-attitudes” as possible motivators of action, but left out affective states (Davidson 1963, 4). Many contemporary philosophers and (neuro-)psychologists (LeDoux 1989; Pacherie 2002; Döring 2003; Prinz 2004; Robinson 2005, 2018; Deonna and Teroni 2012) take it for granted that emotions frequently motivate action. For Walton, motivational force seems to be a core feature of the emotions, as evidenced by his claims that “[f]ear emasculated by subtracting its distinctive motivational force is not fear at all “(Walton 1990, 202). Although the argument from action is not Walton’s main argument for his position, it plays a significant role in distinguishing fictional emotions (such as those of Charles, a film viewer who seems to be afraid of a fictional green slime in the horror film he is watching) from genuine emotions (such as those of an air traveller):

But Charles is different. The air traveller performs deliberate actions that one would expect of someone who thinks flying is dangerous, or at least he is strongly inclined to perform such actions. If he does not actually decide against traveling by air, he has a strong inclination to do so. But Charles does not have even an inclination to leave the theatre or call the police. The only signs that he might really believe he is endangered are his more or less automatic, nondeliberate reactions: his pulse rate, his sweaty palms, his knotted stomach, his spontaneous shriek. This justifies us in treating the two cases differently. (Walton 1978, 8)

Walton also seems to believe that a person who is afraid of flying has an irrational fear because in typical cases such a person will judge that flying is actually a very safe way to travel. However, as this example shows, it is not the lack of rationality that Walton treats as decisive. It seems that Walton chooses the example of the air traveller because, although fear of flying is fear of a statistically minimal danger and in this respect importantly close to Charles’ fear of the fictional green slime, he thinks that the two cases differ in one key respect: the emotions’ connection to intentional action. Although the air traveller can decide to fly, her fear will influence her actions. As Walton suggests, she has at least a strong inclination not to enter a plane. Here we find, according to Walton, the difference to Charles, who is not inclined to perform the actions we would expect if the object of his fear were not fictional.

This observation is then explained by a peculiarity of fictional emotions. Walton claims that emotional reactions toward fictional objects are generally, perhaps even necessarily dissociated from “inclinations to act” (Walton 1978, 8). According to other authors, fictional emotions lack “motivational force” (Friend 2016, 220), “motivating potential” (Currie 1990, 200) or “motivation and action guidance” (Stecker 2011, 304).

It is often assumed that proponents of the argument from action are not committed to a particular theory of the emotions, as different theories of the emotions will provide different explanations for emotions’ motivating character. Robert Stecker for instance, claims that “on both cognitive and non-cognitive theories, emotions will characteristically be connected to behavioural tendencies linked to motivation and action guidance.” (Stecker 2011, 304). On this basis he also seems to infer that the connection with action is, although not the only feature, certainly a decisive feature that distinguishes genuine from non-genuine emotions. Without committing himself to one particular theory, Stecker argues that fictional emotions either are not genuine emotions or are, at most, borderline cases because of their dissociation from action: “in being dissociated from action, our responses to fiction are in this respect further removed from mainline emotions than simulated emotions that occur during deliberation.” (Stecker 2011, 305)

We have thus arrived at a claim that can be seen as replacing claim 3 in the original tripartite schema, and to which a replacement claim for claim 2 also corresponds. The ‘paradox’ would now be generated by a practical, rather than a cognitive feature of genuine emotions that is not shared by fictional emotions and would look like this:

• 1. Readers of fiction have genuine emotions towards fictional characters.

• 2. In order to have genuine emotions, one must be motivated to act by the emotion.

• 3. Readers of fictions are not motivated to act by their emotions towards fictional characters.

Rearranging the elements of the ‘paradox’ so that claims 2 and 3 become premises for a conclusion consisting of the negation of claim 1 and focussing on the emotions rather then their bearers, we get the following, simple.

Argument from action:

• P1*: Genuine emotions have motivational force.

• P2*: Fictional emotions do not have motivational force.

• Conclusion: Fictional emotions are not genuine emotions.

Something along the lines of P1* seems to be presupposed by those who argue that fictional emotions are not genuine emotions because they don’t lead an agent to perform intentional actions (Walton 1978, 8; 1990, 198–199; Currie 1990, 200; Stecker 2011, 305). The reason why advocates of the argument from action claim something like P2* seems to be the observation that viewers and readers of fiction are not led to act by their fictional emotions. In the following section I will discuss whether it is true that no actions are performed because of fictional emotions and whether this is a peculiarity of the fictional scenario.

3 Fictional Emotions and Actions

P2* is presumably taken to be the explanation for what might appear to be a commonplace: that viewers and readers of fiction don’t act because of their fictional emotions. However, the assumption in this perfectly general form is obviously false: my pity felt for a fictional character could motivate me to help real people, I could give up swimming in the sea because of the fear I felt of the fictional white shark or just switch off the TV because I am afraid of the fictional zombies.³ Cova and Teroni mention that admiration for a fictional character can in the same way as admiration for a real person lead to attempts to become like the admired character (Cova and Teroni 2016, 9–10).

Nevertheless, the claim that fictional emotions never motivate intentional actions seems to gain support from a comparison of two situations such as the following: in the first situation, take a person who sees a lion in her immediate environment. She is afraid of the lion and tries to escape. Even though reactions other than fleeing are certainly possible, let us, for the sake of argument and simplicity, take fleeing as the standard action of an agent who is afraid of a lion in her vicinity. In the second situation, a film viewer sees a lion on the screen. He makes no attempt to flee. He thus does not perform any action equivalent to the one performed by the agent confronted with the lion.

This comparison is problematic in several ways. The fundamental problem is, as Derek Matravers shows in his 2014 book Fiction and the Narrative, that we are comparing a situation in which someone perceives an entity with a situation in which someone perceives a representation of an entity. Clearly, not every representation is a fictional representation. If we want to find out whether emotional responses to fiction differ significantly from other emotional responses, this comparison cannot be taken as decisive. What we should be comparing is an emotional reaction towards a fictional representation and an emotional reaction towards a non-fictional representation. Although it is true that typical film viewers do not try to flee from fictional monsters, viewers of documentaries do not usually attempt to escape from, help or talk to entities they see represented on the screen. In general, viewers who know that they are confronted with a mere—fictional or non-fictional—representation typically do not try to immediately or directly act towards, or interact with the object of their emotional responses. Let’s call the kind of actions that are missing here immediate object-directed actions.⁴

Readers and viewers of fiction and non-fiction, then, don’t usually respond by performing immediate object-directed actions. This observation has apparently led some scholars to conclude that there are no decisive differences between emotional responses to fictional and non-fictional representations, at least no differences with respect to their dissociation from action (Matravers 2014; Zipfel 2012). The conclusion is, however, not entirely correct. Someone who watches a documentary about a certain existent person and responds emotionally in one way or another could try to find the person represented, talk to her and so on, whereas someone who knowingly watches a fictional movie would not try to do this. Importantly, though, actions such as searching for the place or the person we saw in the film, although directed towards the entity represented, are not immediate reactions, i.e. are not performed while watching the film. We might use the label mediated object-directed actions to pick out actions performed outside the context in which the representation is consumed. Here we have a difference: while readers and viewers of non-fiction might perform mediated object-directed actions, readers and viewers of fiction don’t.

Now, there are special cases in which someone might even perform an immediate object-directed action as a result of an emotion while watching a non-fictional episode on a screen. Take, for instance, the representation on a screen connected to a security camera or a baby monitor. Most of us will try to rescue their child if they see on the screen of the baby monitor that the child is about to be abducted. Nothing of significance for our context seems to distinguish such an action from the emotionally driven attempt to rescue the child on perceiving the situation with one’s bare eyes. Typically, however, we are, as viewers of films, documentaries and so on, confronted with representations of past events or of events which, at the moment we watch the film or read the text, are not in our immediate vicinity.

Importantly, the relevant representations are frequently of entities that do often not, at least at that moment, exist. Viewers of historical documentaries or scientific films concerning possible future developments will generally neither perform immediate object-directed actions nor mediated object-directed actions. It can be concluded from this discussion that the non-performance of either immediate object-directed actions or mediated object-directed actions is no special feature of fictional emotions; rather, it is a feature of emotional reactions to representations of presently non-existent objects. These may be past, future or possible entities.

Fictional emotions, then, can motivate intentional action in general, but they do not lead to either mediated or immediate object-directed actions (“object_f-directed actions”).⁵ Given these limitations of the second premise of the argument from action, the argument seems weak. The lack of just one specific type of action is not a particularly strong reason for excluding fictional emotions from the class of genuine emotions. However, more importantly, this lack can be observed not only in cases of fictional emotions but also in cases of emotional reactions towards non-existent objects in general, like past or future entities (see Matravers 1991, 2014).

Accordingly, the role that actions play relative to emotions provide no basis for distinguishing between fictional emotions and emotional reactions to past, future or possible entities (or their representations). Hence, the observed lack of actions only justifies a distinction between emotional reactions to presently non-existent entities and their representations, on the one hand, and emotional reactions to existing entities (or their representations), on the other. If the lack of actions can indeed be explained by a peculiarity of fictional emotions, we have to assume this peculiarity for all emotional reactions to other non-existent entities. As a result, there is a bullet that would have to be bitten by someone who wanted to classify fictional emotions as not genuine on the basis of the role of actions: they would have to claim that emotional reactions to many representations and non-existent entities (like past and future entities) are in fact not genuine emotions. Nevertheless, if it turns out that this lack of actions needs to be explained by a peculiarity of the emotions involved, this peculiarity would distinguish these emotions from others. The following two sections will discuss whether the lack of actions directed towards the fictional object of the emotion is indeed best explained by the assumption that fictional emotions differ from non-fictional emotions.

4 Fictional Emotions and Their Motivational Force

As we have seen, fictional emotions can motivate actions of various kinds, like switching off the TV because one is so scared of the monster. Clearly, then, it is false that fictional emotions never involve motivational force. The claim now on the table is: fictional emotions don’t involve the motivation to perform immediate or mediated object_f-directed actions.

In order to ascertain whether this claim is true, we need to clarify the source of emotions’ motivational force. Different theories of the emotions give different explanations for the claim that emotions have motivational force. Here are three explanations that have been recently advanced, each of which relies on a different theory of the emotions:

1) According to component theories of the emotions, emotions are states consisting of several components, among which, it is frequently argued, number desires (Marks 1982; Green 1992; Gordon 1987; Sinhababu 2017). On such a view, emotions’ motivational feature is explained by, or is taken to be identical to, the desire involved.

Hence, for desire-component theories, the question as to whether fictional emotions are motivational states or have motivational force turns out to be a question concerning desires. A proponent of the argument from action who holds such a theory of the emotions must, it seems, claim (a) that fictional emotions lack a desire component (s. Tappolet 2010⁶ ), or (b) that the desire component of a fictional emotion differs from the desire component of non-fictional emotions, in particular with respect to its motivational force (Weston 1975; Currie 1990, 202; Dadlez and Haramina 2015; for arguments against this view: Düringer 2018). Thus, if a component theory of the emotions turns out to be true and if (a) or (b) also turns out to be true, we have an explanation of the second premise of the argument from object-directed action. Fictional emotions would lack their motivational force either because they lack the desire component or because their desire component lacks its usual motivational force. There would then be a decisive difference between fictional and non-fictional emotions, which would provide a reason for the claim that fictional emotions are not genuine emotions. The argument from action could then turn out to be sound.

In order to examine the plausibility of such an account, we would have to discuss whether fictional emotions entail a desire and, if so, whether this desire differs decisively from desires in non-fictional scenarios. The latter question leads into the debate about the nature of desires and cannot be discussed in length here.

2) Sabine Döring advocates a theory according to which emotions are affective perceptions, arguing on the way against the view that emotions are or entail desires. According to Döring, emotions have their motivational force because of their affective properties. Beside their motivational force, emotions have a representational content. This representational content can, but does not always provide reasons to act. Thus, an action can be explained as being motivated by a certain emotion without presupposing a desire:

“It is the emotion’s affect which gives it motivational force, rather than any desire being ‘part’ of it. Unlike a desire, an emotion’s affect can still move its subject to act even if it is not necessary or actually impossible to change the world in such a way that it fits the emotion. In this case, which I claim to be the paradigm case of expressive action, the emotion’s representational content fails to provide an end for action, which in turn means that the action cannot adequately be explained by means–end reasoning.” (Döring 2003, 224).

Standing in a different tradition, basic emotion theorist Silvan Tomkins also explains emotions’ motivational force with reference to their affective component (Tomkins 2008).

In recent analytic debates, a consensus has emerged, according to which fictional emotions don’t differ from non-fictional emotions with respect to their affective properties (Radford 1975; Walton 1978, 1990; Currie 1990; Neill 1991; Hartz 1999; Stecker 2011). But if it is true that emotions have motivational force because of their affective component and if it is also true that fictional emotions don’t differ from non-fictional emotions with respect to their affectivity, then fictional emotions have the same motivational force as non-fictional emotions. P2* would turn out to be false and the argument would be unsound.

3) Finally, according to what are frequently called “motivational theories” of the emotions, the primary task of an emotion theory is to explain the relation between emotions and actions. Recent such theories identify emotions with action tendencies (Scarantino 2014) or the feeling of these action tendencies (Deonna and Teroni 2012), where “action tendency” means a bodily state of action readiness. Accordingly, the relation between emotions and actions is explained by the bodily state of action readiness.

The key question for our problem here is: might fictional emotions generally, or even necessarily, fail to generate a readiness to act that accompanies, or can be identified with certain bodily changes? As we have seen, fictional emotions are not dissociated from actions in general, but at least typically do not motivate mediated or immediate object_f-directed actions. Because fictional emotions can motivate actions that are not directed toward a fictional object, fictional emotions can cause or comprise states of bodily readiness to perform these actions. Therefore, the general claim, that fictional emotions do not generate states of bodily readiness to act is wrong. But can fictional emotions cause or comprise states of bodily readiness to perform object_F-directed actions?

I think the answer is positive. Walton’s famous film viewer Charles is subject to bodily changes typical of those we experience in fear: Charles “says that he is afraid, and he is in a state which is undeniably similar, in some respects, to that of a person who is frightened of a pending real world disaster. His muscles are tensed, he clutches his chair, his pulse quickens, his adrenalin flows” (Walton 1978, 6).

Tensed muscles, a quickened pulse and high adrenalin production are processes or states that are typically identified with a readiness to action (for empirical research on viewers’ bodily responses to fictional films, see Kreibig et al. 2007; Child et al. 2014; Fukumoto and Tsukino 2015). These seem to be the same processes Bradley and Lang refer to when they speak of “preparatory metabolic changes that occur in muscles and glands” (Bradley and Lang 2000, 244). There is no plausible reason why Charles’ bodily state should not count as an action tendency, understood as bodily action readiness.

As a readiness to act does not have to result in the performance of an action, Charles’ not fleeing or hiding is irrelevant for the question of whether he is in a state of bodily readiness. If this state of action readiness is identified with the motivational force of emotions, fictional emotions, then, are not or at least not in general or necessarily dissociated from their motivational force.

Furthermore, Charles is in a state of bodily readiness to flee or avoid something, presumably the green slime. If this is true, Charles is in a state of bodily readiness to perform an object_f-directed action. It is worth emphasising that fictional emotions going along with or generating such bodily states is anything but unusual. Hence, if we identify the emotions’ motivational force with a bodily readiness to act, viewers and readers of fiction are in this sense motivated to act, even motivated to perform object_f-directed actions. Therefore, P2* is, according to this interpretation of the claim, once again false and can hence provide no support for the conclusion that fictional emotions are not genuine emotions.

It seems that Walton fails to see a connection between Charles’ bodily states (tensed muscles, quickened pulse) and actions. Walton characterises them as mere unconscious reflexes, thus ignoring the fact that he has provided evidence against his own claim, at least on one plausible reading (Walton 1978, 8–9). But obviously, it cannot be the case that it is this sense of “inclination to act” that Walton has in mind. Advocates of the argument from action must therefore have something else in mind. Because the actions viewers and readers typically avoid are intentional actions, we should turn to the question of whether readers and viewers intend to perform the relevant object directed actions. I think the answer is, in typical cases, straightforwardly negative. We can assume that those viewers and readers who do not even try to help the heroine, to hide with a victim or have a conversation with their favourite characters make no such attempts because they don’t intend to do so. It is not as if they are prevented from doing such things, frustrating their intentions. If we—somewhat unusually—understand claims about the lack of inclinations to act (Walton 1978) or dispositions to perform object-directed actions as claims about a lack of intentions, then, but only then, the claims are true.

We normally distinguish between the claim that emotional responses to fiction generate no motivation or inclination to perform object-related actions and the claim that they don’t generate intentions to act. The fact that no relevant intentional actions may be performed allows no particular conclusion as to whether the agent may have been motivated to perform some such action. Why, then, do emotionally responding consumers of fiction typically not form intentions to perform object_f-directed actions?

5 Intentions and Doxastic Conditions

In order to answer this question, we have to qualify the claim that object_f-directed actions are not performed. Actually, it is not quite right, that readers of fiction don’t perform object_f-directed actions, as we can see if we consider certain mental actions that readers or viewers sometimes do perform, such as thinking of the protagonist of a novel or comparing her with another character. The latter may even be in part a physical action, if it, for instance, involves writing down attributes. The type of action fictional emotions do not lead to must be specified as object-directed actions that involve physical interaction with the object. The peculiarity of actions belonging to this restricted class is that viewers and readers of fiction don’t have the ability to perform them. Or as Lamarque puts it by agreeing with Walton in this respect “[w]ithin their world, Othello can kill Desdemona and within ours I can kill you, but there is a logical barrier that prevents them from killing us and us from killing them.” (Lamarque 1981, 292)

The reason why readers cannot perform these actions is metaphysical. According to all standard accounts of the metaphysical status of fictional characters, it is not possible to physically interact with them. On the one hand, according to fictional realists, there are fictional entities, but these entities are non-physical objects. Thomasson, for instance, claims that fictional entities are created but non-physical artifacts (Thomasson 1999; Voltolini 2003). On the other hand, anti-realists claim that fictional characters simply do not exist (Yagisawa 2001; Everett 2005). Although these accounts are very different, they agree that the nature of fictional characters make interaction with fictional characters impossible. We can of course shake the hand of an actor who is playing the role of Sherlock Holmes, but we can never shake Holmes’ hand.⁷

What is decisive for our purposes is the fact that typical consumers of fiction believe that they cannot physically interact with fictional characters.⁸ In the following, I will explain why readers’ beliefs about their inabilities prevent them from forming intentions to perform the actions in question.

There is a widespread consensus among otherwise strongly differing theories of intention that intending has some doxastic condition. According to some philosophers, an agent’s intending to φ requires she believe that she will φ (Grice 1971; Velleman 2007), or that it is at least possible for her to φ (Audi 1991; Adams 1986). Setiya (2008) argues that an agent must know how to φ and that she must know she is able to φ. Finally, some authors have argued for merely negative conditions, such as the lack of a belief that she won’t φ or that it is impossible for her to φ (Bratman 1987, 1999; Mele 1992; Roughley 2016).

I will not argue here for any of these different accounts. Rather, I want to highlight the fact that, whatever our chosen doxastic condition, a typical consumer of fiction will not satisfy it: typical readers and viewers of fiction lack the belief that they can perform object_f-directed actions. Indeed, they characteristically believe that they cannot do so. And readers and viewers certainly cannot know how the relevant actions could be performed, as they are impossible.

The belief that one cannot perform an object_f-directed action might not be conscious the whole time during which one is watching a film or reading a book. Particularly in moments of high absorption, this belief will not be consciously in the viewers or readers mind. However, most of the object_f-directed actions we are focusing on here would be actions naturally performed after deliberation. If a viewer or reader is so absorbed that the negative belief that she cannot perform an object_f-directed action is not conscious to her, we still can assume that this belief would manifest itself consciously as soon as she starts a process of deliberating. It has been argued that practical deliberation is itself constitutively dependent on the lack of a belief in the impossibility for the agent of the action being considered (Roughley 2016, 165). If this is correct, then we can even assume that the advent of any such conscious belief would immediately put a stop to any deliberation on whether or how to perform any object_f-directed action which involves physical interaction.

We can, then, explain the non-performance of object_f-directed actions with reference to the non-fulfilment of intention’s doxastic condition. We don’t need to refer to the agent’s emotional responses. Accordingly, this explanation does not require that fictional emotions differ in any significant way from non-fictional emotions. No additional class of emotions is necessary. The explanation is therefore more parsimonious than an explanation that works with the argument from action.

Now, the explanation for the lack of action in the fictional scenario cannot rule out that additionally fictional emotions have no motivational force. As we saw in the previous section, different current emotion theories provide different explanations for the relation between emotions and action and for the motivational force of the emotions. The explanations offered by Döring’s perception theory of the emotions and by the leave no room for the idea that fictional emotions lack the motivational force of non-fictional emotions. According to both theories, motivation is the essence of the emotion and state without motivational force is no emotion, not even a non-genuine emotion. My explanation is perfectly compatible with them.

According to some desire component theories, a desire is a necessary component or part of an emotion.⁹ These theories, however, explain the motivational force of the emotion by reference to the motivational force of the desire component of the emotion. On the basis of such a component account, Derek Matravers has argued both that fictional emotions are genuine emotions and that inaction in fictional emotion scenarios can be explained by the lack of a relevant instrumental belief. Thus, Matravers also sees the non-fulfilment of a doxastic condition—either of intentional action in general or of specific types of intentions—as the key explanatory factor. Remember, though (see Sect. 4) that, according to some philosophers, there are no desires concerning fictional entities or only quasi-desires, that is, desire-like states that lack motivational force. The truth of either of these claims would also be sufficient to explain the lack of relevant actions in fictional scenarios and would support the claim that fictional emotions differ decisively from other emotions. Hence, a defender of a component theory needs to show that fictional emotions have a proper, motivation-encompassing desire-component, if they also want to claim that fictional emotions are genuine emotions.¹⁰

6 Conclusion

Fictional emotions can certainly lead us to perform various kinds of actions. The kinds of actions to which fictional emotions typically don’t give rise are object-directed actions where the relevant objects are fictional, i.e. object_f-related actions, where these would involve physical interaction with the object. As we have seen, the disconnection from action directed at representational objects is by no means a unique feature of fictional emotions.

At the core of the argument from action is the assumption that we need a special explanation for the fact that consumers of fiction neither perform nor intend to perform object_f-directed actions. The explanans posited is a motivationally attenuated category of fictional emotions. Their peculiar lack of motivational force is what is supposed to prevent fictional emotions from counting as genuine emotions. If this explanation were true and the related categorisation proposal were correct, we would require a special category of non-genuine emotions that also subsumes other emotional responses to non-existent entities (like past and future objects), where the emoter is aware of the non-existence of her emotion’s object.

Where the proponents of the argument from action explain the lack of intentions by fictional emotions’ putative lack of motivational force, I have argued that the absence of object_f-directed intentions is best explained with reference to the doxastic conditions of intending. Typical readers and viewers of fiction don’t believe that they can physically interact with fictional characters; moreover, they characteristically believe that they cannot interact with them. The non-satisfaction of the doxastic condition on intending is typical for, but obviously not exclusive to consumers of fiction. It is equally true of bearers of emotional reactions towards other non-existent entities who are aware that the objects of their emotion do not exist. The non-satisfaction of either the positive or the negative doxastic condition can explain why no object_f-directed actions are performed. For this explanation we don’t need to assume a peculiarity of fictional emotions.

It is often claimed that the argument from action is independent of a theory of the emotions. My comparison of three different types of recent emotion theories shows that this is not the case: among the theories discussed, only a component theory is compatible with the existence of motivationally inert emotions. Although component theories are consistent with the explanation from the doxastic conditions on intending, proponents of such theories need to show that the motivational force of fictional emotions does not differ from that of emotions towards non-fictional objects. In contrast, both Döring’s perception theory and motivational theories of the emotions are incompatible with the claim that emotions can be motivationally inert. According to both kinds of theory, a motivational component is the essence of emotions. Subtract that feature—be that a particular form of affect or of bodily action readiness—and you are left not with a category of motivationally attenuated emotions, but with no emotions at all.