2. Emotions, Value Judgments, and Existential Beliefs

While there have been recent attempts to deny that Radford was influenced by early cognitivists like Anthony Kenny and William Lyons (Friend, Reference Friend2020; Matravers, Reference Matravers2014, pp. 104–106), it is often taken for granted that the cognitive theory of emotion lies at the heart of Radford’s and Walton’s conclusions. In this section, I argue that, from a theoretical point of view, adopting a cognitive perspective does not help make sense of Radford’s and Walton’s positions.

For this purpose, let us consider the strongest form of cognitivism, often referred to as judgmentalism (Nussbaum, Reference Nussbaum2001; Solomon, Reference Solomon1976), which is frequently regarded as the most vulnerable to the paradox. According to this view, emotions are identical to evaluative judgments: fear is the judgment that something is dangerous; pity, that someone is suffering undeservedly, and so on. This account has faced several objections—for over-intellectualizing emotions (Goldie, Reference Goldie2000), being disembodied (Prinz, Reference Prinz and Solomon2004), and neglecting their phenomenal dimension (Deonna & Teroni, Reference Deonna and Teroni2012), i.e., the felt experience as a constitutive part of emotion. Nevertheless, even in this radical form, cognitivism does not imply that emotions require belief in the existence of their objects.

To see why, it suffices to recall that cognitivism’s basic equation is “emotion = evaluative judgment.” Consequently, a proponent of this view will treat emotions just as she treats other judgments, whether emotional or not. And clearly, we can—and often do—make judgments about things we know do not exist, such as possible outcomes of our actions, counterfactual scenarios, and fictional situations. For example, I might judge that if I shared my friend Diana’s secrets, she would no longer trust me; historians judge that if Napoleon had not invaded Russia in 1812, he might have preserved his empire; and in fiction, we may judge that Peter Pan is immature for refusing to grow up. These judgments are not considered irrational simply because their objects are non-existent. Nor do we treat them as “quasi-judgments,” substantially different from judgments about real-life situations. On the contrary, we typically regard them as genuine and, when appropriate to the (fictional) context, as rational. (Of course, what makes the judgment true in this case is how things are in the fiction, rather than in the real world. But, usually, this is not considered enough to claim that these are judgments of a different kind (Currie, Reference Currie, Roeser and Todd2014; Song, Reference Song2020).Footnote ²)

Since cognitivists posit that emotions are evaluative judgments, they are committed to treating them in the same way. As a result, a supporter of cognitivism may accept that we can have rational and genuine emotions about non-existent objects—such as possible or counterfactual scenarios, and fiction—just as we make rational (non-emotional) judgments about them.Footnote ³ For instance, when imagining winning the Nobel Prize, a scientist may feel content and proud, as she judges that this would recognize her hard work. Or consider Jake, who had a meeting in the North Tower just a few days before September 11. Though he was not there during the attack and the event can no longer affect him, he may still experience a genuine and rational emotion at the counterfactual thought “I would have died if my meeting had been on that day.” This response is not only intelligible and, for many, rational—it seems it is, in some sense, due. Failing to feel it might reflect Jake’s lack of appreciation for his own good fortune.

This reflection reveals a key issue in Radford’s original argument. At the end of his paper (Radford, Reference Radford1975, pp. 72–74), he addresses an objection to his claim that emotions toward fiction are irrational by comparing them with emotions about other non-existent objects—possible and counterfactual scenarios. He imagines a man weeping over two such unreal events. In the first (a possible event), the man cries at the thought of his mother’s devastation if his sister—currently on an intercontinental flight—were to die in a plane crash. In the second (a counterfactual event), he weeps at how terrible it would have been for his friend Jean—who has six children and whose greatest dream has always been to have a big family—if she had been infertile. According to Radford, both emotions are irrational, albeit to different degrees. In both cases, the thought experiment ends with the man’s wife blaming him for being ‘silly and maudlin’ (Radford, Reference Radford1975, p. 73) after finding him crying over these thoughts. However, while even in the first example the man’s reaction is deemed irrational, his response is more intelligible, for possible future events could really affect people’s lives. For Radford, the problem with this case is just that having a plane crash is very unlikely. By contrast, being sad at a counterfactual event is a blatant silliness.

The problem is that Radford’s conclusion is not supported by a cognitive account. We have already seen that there are cases where it seems reasonable, from a cognitivist perspective, to have emotions about counterfactual events. Jake’s emotion at the thought of dying in the Twin Towers, where he was just a week before the attack, seems even rationally required. He does not come across as silly or maudlin. A cognitivist might emphasize that Radford’s scenarios are extravagant: while it is easy to imagine Jake’s emotional response at the counterfactual thought of being the victim of a terrorist attack, the man’s sorrow at the idea of Jean’s being infertile sounds decidedly odd. Yet this observation alone is not sufficient, for a cognitivist, to deem the emotion irrational.

The only source of irrationality in Radford’s examples a cognitivist might plausibly identify is that the man is overreacting. He is not simply saddened by his thoughts—in Radford’s own words (Reference Radford1975, p. 74), he feels “a lump in his throat,” and begins to “weep” and “snivel.” But on any plausible account—both cognitive and non-cognitive—the rationality of an emotional response is assessed not only in terms of the kind of emotion involved, but also of its intensity (D’Arms, Reference D’Arms, Cosker-Rowland and Howard2022; Nussbaum, Reference Nussbaum2001, p. 55; Tappolet, Reference Tappolet2016, p. 115). As Justin D’Arms (2022, p. 107) puts it, an emotion must have not only the right “shape,” but also the right “size.” This requires “a proportionality between the strength of the response and the strength or degree of value possessed by its object.” Minor achievements call for mild rather than deep pride; trivial harms justify only a modest degree of anger, and so forth. Radford’s man, by contrast, exhibits an excessive response. Moreover, it is worth noting that his reaction may also be harmful from a practical standpoint, insofar as it could prevent him from focusing on more important matters that—unlike the counterfactual scenario—he can actually influence and improve. Spending one’s days indulging in the imagination of impossible events, while neglecting what is real and significant, looks like a dangerous form of quietism. (I will return to the role of actions in the next sections.) Not by chance, the wife’s criticism points directly to this concern: “How can you sit there and weep over the dreadful thing that didn’t happen, and now cannot happen?” (Radford, Reference Radford1975, p. 74). But then, the man’s emotion is not irrational in and of itself; rather, its irrationality lies in its being too intense. In other words, irrationality here is a matter of degree, not an all-or-nothing judgment.

If my argument is sound, we can draw an important conclusion: Radford’s claim about the irrationality of emotions toward fiction cannot be grounded in a cognitive theory. The source of irrationality cannot lie in the assumption that emotions are value judgments or beliefs—since we can make rational value judgments about things that do not, and never will, exist. As long as we adopt a cognitive perspective, there seems to be no “inconsistency” or “incoherence” of the sort Radford invokes between our emotional evaluation of an object and the belief that it does not exist. This holds both for judgmentalism and for more moderate forms of cognitivism, which maintain, e.g., that evaluative judgments or beliefs are either essential components of emotions (rather than being strictly identical to them) or among their causes (Ben-Ze’ev, Reference Ben-Ze’ev2000, p. 72). On all these views, there is no reason to deny that we can have rational emotions whose objects are non-existent, just as we can have rational beliefs and judgments about non-existent objects. The only source of irrationality cognitivism seems to recognize in Radford’s examples concerns the excessive intensity of the emotional response.Footnote ⁴

The same can be said about the genuineness of fiction-directed emotions if we adopt, like Walton, a descriptive rather than prescriptive approach to the paradox of fiction: cognitivism provides no justification for postulating quasi-emotions. On a cognitive view, emotions for fiction can be genuine in the same sense as ordinary ones. This, I submit, offers an important clue to the view of emotions underlying both Radford’s and Walton’s claims.

3. Action Tendencies in Emotions for Fiction

In the last section, I have shown that a cognitive view does not justify Radford’s and Walton’s conclusions. In the remainder of the article, I will argue that their claims can be better understood by assuming a motivational theory of emotion. I am not suggesting that they were aware of this theory, whose first full-fledged examples came out only in the 1980s (Frijda, Reference Frijda1986). What I contend is that we can make better sense of Radford’s and Walton’s arguments by adopting a motivational perspective.

According to motivationalism, emotions are forms of action tendencies (D’Arms & Jacobson, Reference D’Arms and Jacobson2023; Frijda, Reference Frijda1986; Scarantino, Reference Scarantino, D’Arms and Jacobson2014). This means that they are dispositions to (re)act in certain ways in response to stimuli. The evolutionary function of emotions—motivationalists argue—is precisely to help us deal with their objects, by preparing our body to respond to their challenges (in this sense, they often speak of “action readiness”). Emotions incline us to act in certain ways in order to pursue specific goals: fear is the tendency to avoid a potential danger; anger is the tendency to respond to unjust harms, and so forth (Tappolet, Reference Tappolet2023, ch. 5). As Frijda (Reference Frijda1986, p. 71, original emphasis) puts it, emotions “can be defined as modes of relational action readiness, either in the form of tendencies to establish, maintain, or disrupt a relationship with the environment or the form of modes of relational readiness as such.” All major motivational accounts stress that a pivotal feature of the action tendencies emotions amount to is that they have “control precedence” (D’Arms & Jacobson, Reference D’Arms and Jacobson2023, pp. 105–135; Frijda, Reference Frijda1986, p. 78; Scarantino, Reference Scarantino, D’Arms and Jacobson2014, pp. 168–171). In other words, they capture the subject’s attention so as to prioritize the pursuit of their characteristic goals over competing ones. As D’Arms and Jacobson (Reference D’Arms and Jacobson2023, p. 112) put it, “emotions ready the agent for direct and urgent ways to pursue the emotion’s goal, which constitute its action tendency.” They “prioritize a specific goal and prejudice the means taken to meet it” (D’Arms & Jacobson, Reference D’Arms and Jacobson2023, p. 105, original emphasis). Or, in Scarantino’s (Reference Scarantino, D’Arms and Jacobson2014, p. 170) own words, “an action tendency takes precedence over other actions and states of action readiness.”

Control precedence is crucial for the motivational theory to account for the urgent character of emotions, which enjoy a kind of priority over other goals precisely because they prepare the body to deal with the environment. This feature is also central to distinguishing emotions from other action tendencies. For instance, D’Arms and Jacobson (Reference D’Arms and Jacobson2023, p. 112) define anger as the tendency to retaliate. Yet not all tendencies to retaliate count as instances of anger: “Although someone can desire to retaliate on strategic grounds, such as for deterrence, that does not constitute having the emotional motivation of anger.” Similarly, imagine that, because of his fear of flying, Vince is motivated to take a pill to calm down.Footnote ⁵ Although his behavior is motivated by the desire to extinguish his fear, on most motivational accounts this would not count as a genuine emotional action tendency, since its goal is not the one characteristic of fear—namely, avoiding danger by fleeing from it—but rather the elimination of the emotion itself.

What I am going to argue is that, by assuming a motivational framework, it is possible to make better sense of Radford’s and Walton’s conclusions. For if the essential element in defining emotions is action tendency, it becomes easier to see what distinguishes ordinary emotions from fiction-directed emotions. In fact, it seems reasonable to suppose that the latter lack (or at least involves a different kind of) motivational force. Many would be willing to accept that we do not get ready to help Mercutio in Romeo and Juliet or save Anna Karenina. Or, as Walton (Reference Walton1990) suggests, when Charles is scared of the Green Slime on the screen, he has no tendency to flee the cinema or call the police.

Such a disanalogy between ordinary emotions and fiction-directed emotions is not significant from a cognitive perspective—for while judgments often motivate action when combined with a desire, they do not always do so. Cognitivism is not committed to the claim that emotions must always motivate action, even though they often do. By contrast, the basic assumption of motivational theories is that emotions are action tendencies, i.e., states of bodily action readiness. A necessary consequence of this premise is that nothing that counts as a proper emotion can be devoid of its motivational component. As Christine Tappolet (Reference Tappolet2023, p. 78) underlines, motivational theorists are compelled to describe even emotions such as joy or sadness in terms of readiness to action or inaction—respectively, as an “action readiness with no specific goal” and a state “of action readiness, this time with the goal to not relate” with the environment. Thus, we can say that, for the motivational theory,

Moreover, it is plausible to expect that, on this view, the rational assessment of emotions should be connected to some extent to their motivational element. It is worth noting that, while the normative debate on the paradox initiated by Radford commonly focuses on the “rationality” or “irrationality” of fiction-directed emotions, several contemporary philosophers of emotion—most notably proponents of motivational theories—tend to distinguish two notions by means of which emotions can be rationally assessed: fittingness, or appropriateness, and warrant, or justification (D’Arms & Jacobson, Reference D’Arms and Jacobson2000a, Reference D’Arms and Jacobson2000b; Deonna & Teroni, Reference Deonna and Teroni2012). An emotion is said to be fitting when its object merits the emotion, or—on an alternative formulation—when it possesses the relevant evaluative property (e.g., the bear we fear merits our fear because it is dangerous). By contrast, an emotion is warranted when it is rationally justified as fitting from the subject’s perspective, given the evidence available to her. While the two notions are closely connected, they are not identical: an emotion can be fitting but unwarranted (if I fear something that is genuinely dangerous but, given the evidence I have, I am not justified in doing so) and, vice versa, warranted but unfitting (e.g., I may be justified in feeling sad upon seeing my partner walking hand in hand with another person, even if—unbeknownst to me—that person turns out to be a doppelgänger.). Nonetheless, for the sake of simplicity, I will assume that in the cases discussed here the two dimensions overlap: whenever I say that an emotion is fitting, I will also assume that the subject is justified in having it, given the evidence available to her. This will allow us to reflect more clearly on the commonsense notions of emotional rationality and irrationality to which Radford appeals.

The notions of fittingness and warrant are necessary for any account that aims to explain why we ordinarily assess emotions in rational terms—for instance, by saying that it is irrational to get angry at a friendly comment on one’s book manuscript, or that it is rational to be resentful of government corruption—without adopting the cognitivist strategy (i.e., without making the rationality of emotions depend on the rationality of the judgments or beliefs that are either constitutive of emotions or identical to them). This point is clear in D’Arms and Jacobson’s (Reference D’Arms and Jacobson2023, p. 137) account. They claim that “an emotion is fitting when its object really is as the emotion appraises it.” However, since they reject cognitivism, they cannot understand emotional appraisals as “constitutive thoughts by another name.” Rather, appraisals must be of a different nature from beliefs, judgments, or thoughts:

Emotions can be said to appraise their objects in virtue of being “discrete motivational systems” (D’Arms & Jacobson, Reference D’Arms and Jacobson2023, p. 108). Hence, stripping away the metaphor, we may substitute any reference to emotional appraisals with reference to action tendencies. As a result, we may adopt the following premise within a motivational view:

This principle follows from the idea that an emotion is fitting when its appraisal is appropriate to its object (that is, when the object possesses the relevant evaluative property), together with the claim that emotional appraisals are not evaluative thoughts but rather action tendencies. As a result, from this perspective, Wendy’s fear of a rabid dog is fitting, since avoidance behavior is appropriate in the presence of a rabid dog; John’s anger at a benign comment is unfitting, since there is no hostility to react to. We can derive such a normative standard for the assessment of emotions both from what Tappolet (Reference Tappolet2023) calls the “simple motivational view” as well as from other theories that assign significant importance to action tendencies, such as Julien Deonna and Fabrice Teroni’s attitudinal theory.Footnote ⁶ Their approach challenges the traditional “representational” view, according to which emotions represent their object as having distinct evaluative features. (According to judgmentalism, they represent their objects by being evaluative judgments; other views describe emotions as different kinds of representations, such as concern-based construals (Roberts, Reference Roberts2003) or perceptions of values (Tappolet, Reference Tappolet2016).) Deonna and Teroni argue that emotions are not evaluative representations but rather evaluative attitudes. For them, the evaluative component of emotions consists in the fact that they are experiences of one’s body as being ready to act toward an object:

An example makes clear that, like D’Arms and Jacobson, they too take emotions to be forms of evaluation precisely insofar as they prepare the body for certain actions:

Now, how should we conceive of our emotions for fiction within this paradigm? Clearly, by the law of the excluded middle, either

In the first case (P3), we must say that emotions for fiction do not have motivational force. However, since a basic assumption of the motivational theory is (P1) that action tendencies cannot be absent in a proper instance of emotion, we are led to conclude that

This is the conclusion famously suggested by Walton (Reference Walton1990), who labeled these non-genuine emotions quasi-emotions. Not by chance, two major proponents of the motivational account such as D’Arms and Jacobson (Reference D’Arms and Jacobson2023, p. 114) explicitly endorse Walton’s view that Charles is not really afraid of the Green Slime, arguing that “the spectator’s lack of fear-specific motivation makes it implausible to attribute literal fear to him.” This is also compatible with a suggestion proposed by Cova and Friend (Reference Cova, Friend and Scarantino2024, pp. 437–438; see also Friend, Reference Friend and Kind2016), according to which the reason we apply different normative standards to ordinary emotions and fiction-directed emotions is that in the real world “the demand for action is pressing or immediate.”Footnote ⁷

Others, of course, may want to deny (P3). Matravers (Reference Matravers1991), for instance, has argued that our emotions for fiction have motivational force too: when we pity a fictional character, we do have a tendency to relieve their pain. We do not concretely act simply because we know we cannot causally affect the fictional world. It is worth noting that this position is not as easy to argue for as it might seem. In fact, I mentioned earlier that the kind of action tendencies which, according to motivationalism, constitute emotions are characterized by control precedence. Thus, in order to claim that emotions for fiction are action tendencies in the sense assumed by motivationalism, it is not enough to say that they are accompanied by an inclination to act. Rather, one must show that this inclination to act (or action readiness) takes priority over the subject’s other goals. Recall the example of anger as the tendency to retaliate: if someone desires “to retaliate on strategic grounds, such as for deterrence, that does not constitute having the emotional motivation of anger” (D’Arms & Jacobson, Reference D’Arms and Jacobson2023, p. 112), because such a tendency lacks control precedence. Recall also D’Arms and Jacobson’s (Reference D’Arms and Jacobson2023, p. 112) definition, according to which “emotions ready the agent for direct […] ways to pursue the emotion’s goal, which constitute its action tendency.” This definition rules out as genuine action tendencies—those constituting emotions—many of the indirect behaviors fiction may motivate, such as the decision to volunteer for an NGO after watching a film about fictional refugees. Unless one shows that the action tendencies of fiction-directed emotions have control precedence—i.e., that they are direct tendencies to act toward the very object of the emotion—Walton’s idea that those for fiction are quasi-emotions is not refuted, since they lack an essential part of ordinary emotions.

Despite these problems, suppose that someone convincingly claims that (P4) fiction-directed emotions share their motivational component with ordinary ones—being action tendencies with control precedence. In that case, Walton’s view should be rejected, since emotions for fiction would have all the essential components required to count as proper emotions. However, this second path entails the conclusion

In fact, we have assumed in (P2) that an emotion is fitting (or rational), within a motivational perspective, if the action tendency it amounts to is appropriate to its object. Yet it clearly makes no sense to try to help the fictional characters we feel compassion for, or to try to flee from a danger that is not real but merely fictional, precisely because we do not believe that the object of the emotion exists. This conclusion sounds familiar because it’s a close kin of Radford’s—even though it is argued for in a different way. We can now identify where the rational conflict to which Radford appealed lies. The inconsistency he points to is between the tendency to act toward an object and the belief that the object does not exist. In this way, his conclusion that emotions for fiction are irrational (or silly) acquires an intuitive force that is lacking in the original formulation (Cova & Teroni, Reference Cova and Teroni2016). For many may be willing to accept that it is irrational to try to act toward something non-existent. But while a pure cognitive view of emotion does not support this conclusion, a motivational one does.

4. Radford and Walton Revisited

We are now in a position to revisit some of Radford’s examples, having a clearer grasp of the role played by belief in the (probable) existence of the object of our emotions. We saw that Radford criticizes the man’s emotions toward unlikely and counterfactual events, while allowing that an emotion directed at a probable future event may be rational. If we assume that emotions are action tendencies—i.e., reactions that prepare the body to deal with their objects—it becomes easier to justify such a conclusion. Being afraid of a probable stock market crash is rational because, if emotions are states of action readiness, they prepare us to deal with a danger we are likely to face. By contrast, it is not rational to prepare for action in response to events that are highly improbable or that can no longer occur, as in the counterfactual scenario. It would be like writing a highly specific paper—unsuitable for submission elsewhere—for an essay prize or a special issue whose deadline has already passed. In this sense, there is something puzzling about emotions toward fiction, insofar as they prepare us to act toward something we believe does not exist.

One may object that there is at least a sense in which emotions (as action tendencies) toward non-existent objects can be seen as rational: if a state of affairs is true in the fiction, it is fitting to desire to act accordingly. It is fitting to desire to help Anna Karenina because it is fictionally true that she is suffering undeservedly. And if she really existed, it would be appropriate for our body to be ready to help her. This move, which Friend (Reference Friend2022) calls the substitution approach, parallels the arguments I have offered regarding the cognitive view of emotions toward fiction: just as a cold judgment about a fictional scenario is rational if it is appropriate to what is described in the fiction, so too is an emotional response, for a cognitivist. Fictional truth is substituted for real-world truth as the standard for evaluating the appropriateness of fiction-directed emotions (Song, Reference Song2020).

However, it seems to me that while the substitution approach may be defensible from a cognitive perspective, it is much less easily applied within a motivational framework. For, as noted above, motivational theories explicitly challenge the classical representational approach, according to which emotions represent their objects as having evaluative properties (Tappolet, Reference Tappolet2023). From a motivational view, emotions are not representational states: they are states of action readiness. As Deonna and Teroni (Reference Deonna and Teroni2012) explicitly note, we may say that emotions evaluate their objects in the sense that they prepare the body to act in certain ways. Fear evaluates something as dangerous not by representing it as such, but by readying the body to flee or fight. Consequently, this view is committed to the claim that one is not motivated to act because they have first represented an object as having an evaluative feature; rather, “the evaluative nature of the attitudes is derived from the underlying states of action readiness” (Tappolet, Reference Tappolet2023, p. 90). Saying that it is appropriate to desire to help Anna Karenina because it is fictionally true that she suffers undeservedly appears to presuppose a representational account of how emotions motivate action. It amounts to claiming that, since the representation of Anna Karenina as undeservedly suffering in fiction is accurate, we are rationally motivated to help her. But this is not the kind of connection between emotion and action tendency to which motivationalism is committed. As a result, while the substitution approach may be available to representational theories of emotion, it is not a viable option for motivational accounts.

A second possible objection concerns a different kind of action fiction may elicit. Imagine that, while fearing the Green Slime, Charles is prompted to hide his face behind a pillow or to close his eyes. One might argue, against Walton, that this is a genuine action tendency with control precedence, and, against Radford, that it is not even irrational, since such a reaction might be considered appropriate to a fictional monster. Yet it seems to me that the motivational theory can provide Walton and Radford with some resources to reply. First, it can be questioned whether this example really counts as a genuine action tendency constitutive of fear. To see why, it suffices to recall the example of Vince discussed in Section 3. There, I argued that Vince’s tendency to take a pill to calm his fear while flying does not qualify as a genuine emotional action tendency on most motivational accounts. The reason is that the goal of his action is not the one characteristic of fear (avoiding danger) but rather the elimination of the emotion itself. The most plausible interpretation of Charles’s behavior—it seems to me—is similar. By hiding his face in a pillow, he is not attempting to escape a potential danger, but rather to put an end to what has become an unpleasant or overwhelming experience.

It is worth noting that reactions of this sort are more common with visual media, whereas other forms of fiction, such as novels, are much less likely to elicit them. This comparison is instructive, since it helps to clarify that the relevant action tendencies are not those typical of emotions like fear: if one closes a book because the story has become too frightening, one is not trying to flee from a danger, but rather to terminate an unpleasant experience. If this is correct, we are again in a position to question either the genuineness or the rationality of this fiction-directed emotion. For if the action tendency is not an attempt to escape the Green Slime, then the emotion cannot be said to be genuine, as Walton would argue.Footnote ⁸ Conversely, insofar as we insist on interpreting this tendency as an attempt to flee from the danger posed by the Green Slime—the object of the fiction-directed emotion—rather than as an attempt to end an unpleasant experience, the emotion should be regarded as irrational, since it prepares the body to escape from something Charles knows to be a non-existent threat.

For these reasons, it seems to me that, by adopting a motivational perspective, it is easier to understand why Radford claimed that emotions for fiction are irrational, and Walton argued that they are not genuine.Footnote ⁹ To dispel any remaining doubts, I will conclude by highlighting some textual evidence from the original works of Radford and Walton that supports this view.

Even though—as I have repeatedly pointed out—Radford did not argue for his view by referring to the motivational theory, as I have done here, a final example from his well-known paper seems to suggest that the line of argument I have proposed is compatible with his own. In his final remarks, he compares our being (irrationally) moved by fiction to the idiosyncratic behavior of some tennis players:

Radford’s reference to our behavior is revealing, and it suggests that his conclusion can be more convincingly supported by assuming a motivational view of emotion. The tennis player’s idiosyncratic behavior seems irrational because it cannot affect the trajectory of the shot, and the player knows this. It is, in Radford’s own words, silly. Now, if this example is supposed to resemble our emotions for fiction at all, then the reason for considering fiction-directed emotions as irrational should be that they prime our body to act toward something we know does not exist.

The reference to action tendencies also helps us explain Walton’s distinction between ordinary emotions and quasi-emotions. Indeed, on a closer look at his Mimesis as Make-Believe, it is easy to see that the essential difference between them is not that quasi-emotions lack a belief in the existence of their object, but that they lack motivational force. Walton (Reference Walton1990, p. 202, my emphasis) is explicit, for instance, when he writes: “Fear emasculated by subtracting its distinctive motivational force is not fear at all.” Or consider the passage where Walton rejects Lamarque’s (Reference Lamarque1981) thought theory, according to which what we fear in fiction is the thought of something dangerous:

Once again, what seems essential for an emotion to be genuine is not the belief in the existence of the object, but the readiness to act. The most revealing instance of such a view is the comparison Walton makes between our emotions for fiction and recalcitrant emotions. Borrowing Greenspan’s (Reference Greenspan1988) example, he imagines that Frances, due to an irrational phobia of dogs, fears the harmless Fido—the old, toothless dog of one of her friends—despite judging that he is not dangerous. Walton comments on that case as follows:

If—according to Walton himself—in both cases there is no belief of being in danger, the only difference lies in the action tendencies. For Frances—Walton argues—does flee from Fido despite judging him to be non-dangerous. By contrast, Charles does not escape from the cinema where he sees the Green Slime, nor does he call the police.