Dissertation - Rationality of emotions towards fictional characters

philosophy

This dissertation was submitted as part of my undergraduate degree in philosophy in 2024, for the University of London.

Was Radford (1975) right to claim that emotions experienced in response to fictional events or characters are irrational?

Introduction

How can we be moved by the plight of characters when engaging in fiction even though we know that the characters and / or events do not, have never, and will never exist? That is the question that Radford asks in his 1975 paper ‘How Can We Be Moved by the Fate of Anna Karenina?’, motivated by the observation that in the ‘real world context’ belief is required for emotion. Not­ finding a satisfactory answer to this dichotomy, he claims that our being moved by fiction involves us in ‘inconsistency and so incoherence’ (Radford (1975), p. 78), which is usually taken as a statement that our emotions at fiction are irrational in some way.

However, the exact nature of this claimed irrationality is unclear, leading to many attempts to clarify Radford’s argument. Whilst many such attempts consider the wider set of papers that Radford wrote on this topic, this essay will mainly focus on Radford’s 1975 paper in an attempt to understand his initial argument and to examine whether it is successful.

I shall start with an examination of Radford’s claim that belief in required for emotion in the real world context, both by outlining the cognitive theory of emotion and by examining Radford’s three ‘Belief Examples’, which I shall find to be intuitively plausible but underwhelming.

I shall then outline the usual interpretation of Radford’s paper as the so called ‘Paradox of Fiction’ that claims cognitive irrationality arises from holding three seemingly plausible but inconsistent premises. After showing why this cannot be Radford’s intended meaning, I will examine two normative claims: agent irrationality, that agents engaging in fiction fail to comply with the normative standards that apply to real world emotion, and emotional rationality, that fictional emotions fail to comply with the normative standards against which real world emotions are assessed.

I shall go on to explore, and reject, an alternative view that Radford’s paper was a causal question before putting forward my own interpretation that Radford’s paper is most simply understood as outlining puzzle of intelligibility, that ultimately results in a claim of incoherence and, therefore, irrationality. I shall also briefly look at Radford’s directly claims of irrationality in his later papers and suggest that this is a separate application of the label ‘irrational’ towards specific emotion types that include fear, leaving my interpretation of irrationality as unintelligibility to stand.

I shall ultimately conclude that, when understood as a puzzle of intelligibility, Radford’s concern has not (and cannot) been resolved and therefore he was justified in his concern.

Terms

For simplicity, I take ‘fiction’ to mean novels, poetry, plays and movies that are wholly concerned with non-factual events and people.

I take ‘emotion’ to mean only affective states that have an object. This excludes other affective states such as moods.

I use ‘emotion at fiction’ to mean ‘emotion at fictional characters and / or events’, and I use ‘emotion at fiction’ and ‘fictional emotion’ interchangeably.

Unless otherwise stated, ‘Radford’s paper’ will refer to the 1975 paper ‘How Can We Be Moved by the Fate of Anna Karenina?’.

Radford’s 1975 paper

Broadly, Radford develops his argument in four steps:

  1. The claim that belief (that an actual person has suffered) is required for emotion in the real world context. (Radford (1975), p. 68-69).
  2. The observation that we feel genuine emotion for fictional characters or situations where we do not have belief that an actual person has suffered (p.69).
  3. The examination of a number of possible solutions to this seeming dichotomy (p. 71-78).
  4. Finding none of the solutions to be satisfactory, the conclusion that being moved by works of art involves us ‘in inconsistency and so incoherence’ (p. 78).

As such, Radford’s paper is taken to be a claim of irrationality regarding emotions at fictions. Unfortunately, his paper does not make this claim clear: it neither sets out criteria for rationality nor clarifies the nature of the irrationality that he perceives.

What is clear, however, is that Radford takes the problem to arise in some way from three claims about the world that, when considered together, indicate an inconsistency. These can be summarised as follows:

  • We experience emotions towards fictional characters or events
  • We do not believe in the existence of fictional characters or events
  • To experience an emotion for something in the real world requires a belief in its existence.

The first two claims seem to be straightforward observations from life. The success of his argument, therefore, hinges on his third claim (hereafter Radford’s ‘belief claim’) that belief is required for emotion. Since this is often taken as a statement of the cognitive theory of emotion, I shall outline this in the next section, and I shall follow this with a review of Radford’s own argument for the requirement of belief.

Cognitivism

According to the cognitive theory of emotions, emotions are based on some kind of cognitive evaluation, such as a judgement or appraisal.

Let’s say I am walking home late at night and I see a dark shape in an alleyway that I take to be a man. The emotion - if any - that I feel depends on my cognitive assessment of this situation: if I consider that men lurking in alleyways in the dark are dangerous then my cognitive assessment of the situation will be that I am in danger and it is likely that I will feel fear.

Crucially, on this account, there has to be something for that assessment to be about: I have to have belief that there is a man lurking in the shadows to be frightened of that man. That is, cognitivism requires that emotions have intentionality.

Note that what is important is my belief that there is a man in the alleyway. If the dark shape happened to be a tree stump that closely resembles (at least in the shadows) a man, then my emotion is directed towards the man that I believe to be in the shadows.

On the cognitivist account, therefore, emotions requires belief in the existence of the person or situation. However, Radford does not make his belief claim by referring to this (or any other) theory of emotion, but instead argues for it through three examples (hereafter Radford’s ‘Belief Examples’). I shall therefore now review these.

Radford’s Belief Examples

All three of Radford’s Belief Examples hinge on contrasting two states of affairs or scenarios, one of which involves belief and another that does not.

  1. We read an account of the suffering of a group of people and, if we are ‘at all humane’, we are likely to feel moved. But we then discover the account is fictional and our emotion dissipates. Radford takes this to show that we must believe an account to be real – someone has to have suffered - for us to be moved.
  2. We are moved by the harrowing story a man tells us about his sister. But we then discover that both the situation and the sister have been fabricated and our emotion again evaporates. Radford takes this to show that we must believe not only the situation but also the person involved to be real or existent in order for us to be moved.
  3. We are not moved when an actor friend merely simulates extreme pain and agony. But the actor friend then enacts the death agonies of a mutual friend, or of a soldier that the actor killed and we are ‘horrified’. Radford take this to show that we can moved by past events but, again, only if we believe they actually happened to a real person.

Examples 1 and 2 move from an initially supposedly true scenario to one that is false, with a resulting loss of emotion. Example 3 moves the other way, from a knowingly fictional case to a true scenario. Since I feel it is the weakest, I shall start with Example 3.

Belief Example 3

  1. We watch an actor simulating agony

  2. We are not moved

  3. The actor now enacts the death throes of a friend or a soldier whom he killed

  4. We are moved

Conclusions: 1) belief is necessary for emotion 2) we can be moved by historical events.

However, we should note that the move between step 1 and step 3 doesn’t only involve the change from a fictional to a real situation but also a change of subject matter from ‘mere’ agony to death throes and a change of the object of the acting from an anonymous (fictional) person to a friend or to someone whom actor has personally killed. These differences mean that it is difficult to take seriously Radford’s assumption that it is solely the change from an imaginary scenario to a real one that explains the assumed difference in our emotions. I shall instead, therefore, set this aside and turn to the first two examples.

Belief Examples 1 and 2

Radford’s first two Belief Examples both follow a similar pattern:

  1. To be moved by an account of suffering is an expected human reaction.
  2. We hear of an account of suffering.
  3. We are moved.
  4. We discover the account is false.
  5. We are no longer moved.

Conclusion: belief is necessary for emotion.

In both examples, Radford takes it that it is our discovery that the account is false - that is, our loss of belief in the scenario - that causes us to change from being moved (step 3) to no longer being moved (step 5). However, I feel that Radford, by choosing two examples that are so similar, fails to sufficiently consider other variations that might have been informative. For example, later in his paper (pages 72-74) he considers real world scenarios in which belief isn’t required for emotion, and yet he fails to explore that here. I will now propose a variation of his examples that, at least tentatively, suggests that this can be the case here, too:

Variation - belief isn’t necessary for emotion

In this variation, the man starts by telling us about his stricken sister and, as before, we are moved – horrified – by the story, but also drawn in to it, Indeed, when he reveals that his sister and the scenario are made up, we are surprised by this twist, but we find we don’t lose all our emotion. We tell the man that he was naughty to deceive us but we beg him to finish the story to resolve our angst for the ‘sister’, and we find ourselves thinking it over – and remaining moved – long after we have parted company. This can be outlined as follows:

1’. To be moved by an account of suffering is an expected human reaction.

2’. We hear of an account of suffering.

3’. We are moved.

4’. We discover the account is false.

5’. We remain drawn into the story: we continue to ponder and feel moved by the plight of the protagonist.

Conclusion: belief isn’t necessary for emotion.

I suspect the plausibility of this example lies in the fabricated situation not being too harrowing - it would be ghoulish to desire that a fabricated harrowing story was actually real. It might also be objected that this example, ending where it does with us wanting to know how the ‘situation’ is resolved, is too close to the case of fictional emotion. But I disagree: the important aspect is that we started out believing the scenario was real but didn’t lose all our emotion when we discover it is fabricated, which is counter to Radford’s claims.

Belief isn’t sufficient for emotion

Despite his later claim (Radford (1989a), p. 96), Radford also doesn’t explore whether belief is sufficient for emotion. He does not need this in order to make his desired point that belief is necessary for emotion, but such an investigation may well have shone additional light on the relationship between belief and emotion.

As such, then, I find Radford’s Belief Examples to be under-described and under-examined. Radford introduces too many changes into Example 3 to make its conclusion credible, whilst Examples 1 and 2 share too many similarities to provide valuable additional information, compounded by his failure to consider further variations that might have been illuminating.

Matravers also raises a concern about the ability of Radford’s examples to establish his point:

“…[Radford] argues from examples which all have the following form. Find some scenario in which an emotion causally depends on an existence claim, remove the existence claim, and note that the emotion disappears with it… What is neglected in each case is the possibility of scenarios in which the effect appears without that particular kind of cause” (Matravers (2014), p. 104).

Some have argued that a further weakness with Radford’s examples is that he fails to consider that what is required for belief in the real world context is different to that required in fiction. e.g.

“It doesn’t follow that because some condition is necessary for our ascription of a certain feeling in a particular set of circumstances that it is also necessary for the ascription of that feeling in any circumstances whatever” (Weston (1975), p. 81).

However, it is possible that this is not a claim that Radford intends to make (at least here): he clearly doesn’t take belief to be necessary for fictional emotion since he goes on to argue that we are, in fact, so moved. I take Radford here to be merely trying to establish a standard in the real world against which to compare fictional emotion and, if this is right, Weston’s objection does not stand.

Nonetheless, since he is relying on these examples to establish his claim that belief is required for emotion, I find his examples underwhelming and, perhaps, unconvincing (they may not succeed in convincing someone who didn’t already hold the view that belief is required for emotion). However, his conclusion does seem intuitively plausible and, as such, I shall proceed to a fuller examination of his paper, starting with the most common interpretation of Radford’s paper as a cognitive paradox.

Cognitive irrationality: the paradox interpretation

Radford’s paper is traditionally taken to present three prima facie plausible, but internally inconsistent, premises that form a paradox:

  • P1. We experience emotions towards fictional characters or events
  • P2: We do not believe in the existence of fictional characters or events
  • P3: To experience an emotion for something requires a belief in its existence.

On this reading, Radford’s charge of irrationality arises from the cognitive inconsistency involved in holding all three beliefs. A successful solution to the paradox by rejecting one of the premises would therefore solve the issue of irrationality.

Attempts at solving the paradox include, in brief:

  1. Rejecting the claim that we have genuine emotion (P1)

This is the claim of Walton, who holds that whilst the emotions we feel at fiction seem to us to be the same as those we feel towards real life situation, they are in fact ‘quasi-emotions’ that we develop by engaging in fiction in a similar way to how children engage in games.

  1. Rejecting the claim that we don’t believe the object of our emotion exists (P2)

This proposal claims that, at least at the point of emotion, we do not believe that the object of our emotion does not exist (or we believe that it exists), through either the suspension of disbelief or getting so caught up we forget it’s fiction, or because the object of our emotion is not, in fact, the non-existent fictional character but (for example) a real person of whom the fiction put us in mind.

  1. Rejecting the belief claim, that belief is required for fictional emotion (P3)

The most common way of attempting to resolve the paradox is to deny that belief in the plight of the character is required for emotion, either by proposing a different object of emotion (as in the case of thought theory) or by rejecting the cognitivist theory of emotions altogether, such as in the non-cognitivist claims of Jenefer Robinson.

Such attempts to solve the paradox are well rehearsed in literature and I shall not repeat them here, but shall instead now examine the claim that the reading of Radford’s as a logical paradox is erroneous.

Issues for the paradox interpretation of Radford’s paper.

One definition of a paradox is that it is a ‘set of apparent truths that are, taken together, contradictory. A paradox is, in other words, a state of affairs that appears impossible yet true’ (Yanal (1999), p. 12). One example is the Sorites Paradox, that holds that adding a single hair to the head of a bald man cannot make the difference to the fact he is bald, and so he is still bald after such an addition. This argument can be run 10,000 times, each time with the same conclusion that the man is still bald after the addition of a single hair. And yet a man with 10,000 hairs on his head would clearly not be considered bald. The general idea, then, is that with a paradox, the premises seem true but arrive at a conclusion that seems counterintuitive or even absurd.

If Radford’s paper were setting out a cognitive paradox, then, he would be arguing for the truth of his three premises but applying Yanal’s definition we would expect the outcome to be a state of affairs that seems impossible but is true (for example, ‘despite our best intuitions, there is, in fact, a world where all three premises are true’). However, it seems more accurate to say the Paradox of Fiction is a ’set of apparent truths that are, taken together, impossible yet impossible’. As Yanal points out, Radford presents his belief claim as a fact of the world and therefore it is true whether we accept it or not. This would mean that it would not be ‘we who would be involved in ’incoherence’ but the universe’ (Yanal (1994), p. 55), referenced in (Friend (2020), p. 7).

A stronger problem for the paradox interpretation is that a paradox is a cognitive inconsistency and, as such, is on the part of the theorist of emotion. Yet Radford makes clear here that his puzzle is about what we do: it is our being moved by works of art that involves us in inconsistency and as such the claim of irrationality is on the part of the experiencer. (Friend (2020), p. 7).

I agree with Friend’s point here. Radford also says ‘How can the seeming incongruity of our doing this be explained and explained away’ (Radford (1975), p. 71), which illustrates both the fact that he does not consider emotion at fiction an impossibility and that his puzzle is about our ‘doing’.

Additionally, whilst Radford does use the word ‘paradox’, he doesn’t do so in such a way as to commit himself to a cognitive paradox (Matravers (2014), p. 102) and indeed seems to use it interchangeably with ‘puzzle’.

If Radford’s paper is not a cognitive paradox then the irrationality cannot be located, as many have taken it to be, in the commitment to incompatible premises. How, then, can it be understood?

I shall now outline two ways that emotions at fiction might be measured against the norms that apply to real world emotion and found to be irrational - agent irrationality, and emotional irrationality - which I shall argue both share a similar issue as interpretations of Radford’s 1975 irrationality claim.

The normative claim #1: agent irrationality

Friend argues that Radford’s belief claim should be understood as a normative statement (NS): ‘To experience a rational emotion for something requires a belief in its existence’. She rejects that NS should simply replace the belief claim as P3 of the paradox, on the grounds that this would not substantially move us forward. Instead, she considers his three statements to be part of a valid deductive argument:

  1. We experience emotions towards fictional characters or events
  2. We do not believe in the existence of fictional characters or events
  3. (NS) To experience rational emotion for something requires a belief in its existence.

Conclusion: our emotions towards fictional characters or events are not rational.

That is, if NS is true, then fictional emotion is irrational. The project then, on this account, becomes one of identifying the truth or otherwise of NS (Friend (2020), p. 12).

How does Friend make the move from Radford’s belief claim to NS?

First, she rejects the traditional claim that Radford is a cognitivist since whilst his belief claim is consistent with the cognitive theory, he also is committed to P1, the premise that we do, in fact, have genuine emotions at fiction. Since, as we have seen, cognitivism holds that we can’t have emotions at things or situations that we do not believe exist, she claims that this commitment indicates that Radford can’t be a cognitivist (Friend (2020), p. 8).

This leaves her with the task of locating his irrationality elsewhere, which she does by claiming normative irrationality. What is the norm or standard that she takes as the basis of this claim? She notes that Radford’s examples of the problems of emotions in the real world are ones of knowingly violating the norms of fittingness. In the real world, emotions that are misplaced - e.g. fear about something that is not dangerous, are recalcitrant emotions. Since emotions at fiction are a step even beyond this - they are emotions at something we know does not exist - this explains Radford’s charge of irrationality.

Friend bases her interpretation on a broader survey of Radford’s papers. However, I don’t find the normative interpretation of Radford’s 1975 paper to be compelling. In this paper, he doesn’t actually use the term irrational (or indeed rational), which might be expected if his claim was a normative one.

Furthermore, Friend’s claimed normative assessment is of our having emotions towards fiction - that is, it accepts that the emotion towards fiction has arisen in us and then locates the difficulty, if there is one, between having that emotion and the standard or norm by which we measure the correctness of having emotions in the real world. However, in his 1975 paper, Radford does not seem to be asking ‘what makes our having fictional emotions fiction irrational’ but instead the question ’how can we have such emotions?’.

This problem is also seen in the interpretation of Radford’s claim of irrationality as that of the emotion itself, which I shall look at now.

The normative claim #2: emotional irrationality

A different way of measuring fictional emotions against the standards of those in the real world context is to consider the criteria that we use to judge a real world emotion as rational or otherwise.

This is most clear in the cognitive account that suggests that emotions are rational if they are fitting, warranted, and coherent.

Let’s say I am angry with my boss. My anger is fitting if it is the case that she has (or I take her to have) done things that would naturally lead someone to be angry - for example if she has singled me out unfairly or has given feedback in a way that she does not normally do to others.

An emotion might, however, be warranted even if it not fitting, if it is in line with the evidence available to the agent. For example, my fear of the man walking alongside me wearing a balaclava and carrying a gun may well be warranted even though he is a harmless student on his way to a fancy-dress party.

Finally, an emotion is coherent if it is consistent with my other beliefs. For example, if the snake in front of me is a harmless grass snake, but I believe that all snakes are deadly, my fear is coherent (Scarantino and de Sousa (2021)).

Crucially, such appraisal requires that emotion has intentionality - that there is something that the emotion is about. Since in fiction there is not an external (mind-independent) object at which my thought is directed, this can lead to the claim that emotion at fiction cannot be rational.

This has led to attempts to find other objects of fictional emotion - such as belief in the ‘truths in the fiction’, or of thoughts - to restore rationality. However, it is notable how Radford continually rejects these suggestions. The issue, once again, seems to be that this interpretation is assessing the emotion that we have at fiction and fails to address Radford’s question as to how we can have emotion at fiction.

This latter question is examined in the causal interpretation, which I shall now look at.

A causal puzzle: the causal interpretation

Kim (Kim (2024)) argues that Radford was not necessarily claiming irrationality but asking how we can have fictional emotion in the first place. Kim calls this the causal question (Kim (2024), p. 30).

This interpretation holds that belief (in the existence of the object of our emotion) is amongst the causes of our emotion, thus Radford’s belief claim can now be characterised as a causal statement: for us to feel emotion, it is necessary that we have belief. (Matravers (2014), p. 103)

Kim arrives at this conclusion as follows:

  1. Attempts have been made to address the issue of conflicting premises by explaining how emotion at fiction is compatible with the lack of belief. One such attempt is Walton’s claim that our emotion is not genuine.
  2. These solutions address questions of the kind ‘How can we pity Anna Karenina, knowing that she is a fictional character?’ - in Walton’s case, the answer is that our pity isn’t real emotion, but quasi-emotion.
  3. But Radford also asks more general questions about emotion, such as ‘How can we weep while knowing it is fiction’.

Thus she claims:

“… the question is not about what we are moved to by fiction, but how we can be moved by fiction in the first place, how we can have emotional responses to fiction at all” (Kim (2024), p. 36).

Since this interpretation might suggest that Radford was not, in fact, making a claim of irrationality it would, if plausible, dissolve the question that we are answering today. I shall therefore examine it in more detail in the next section.

Examination of the causal interpretation

As we have seen, Radford concludes that having fictional emotion involves us in ‘inconsistency and so incoherence’. In philosophy, these terms are usually associated with irrationality, however, this need not be the case in more common parlance. Definitions for inconsistency in the Oxford English Dictionary include ‘lack of accordance or harmony’, ‘want of agreement or harmony between two things’, and those for incoherence include ‘want of cohesion’ and ‘want of connection; incompatibility’ (https://www.oed.com).

These terms, then, would be as comfortable with a causal question, ‘how can this even be the case, given the inconsistency’ as they would be with a question of rationality, ‘how can this be rational, given the incoherence’.

Other statements from Radford’s paper seem compatible with this interpretation (although I shall argue for a better interpretation later). Emphasis mine:

“What seems unintelligible is how we could have a similar reaction to the fate of Anna Karenina, the plight of Madame Bovary or the death of Mercutio. Yet we do” (Radford (1975), p. 69)

“So if we can be and if some of us are indeed moved to tears at Mercutio’s untimely death, feel pity for Anna Karenina and so on, how can this be explained ? How can the seeming incongruity of our doing this be explained and explained away?” (ibid p. 71)

“But all over again, how can we do this knowing that neither she nor Mercutio ever existed, that all their sufferings do not add one bit to the sufferings of the world?” (ibid, p. 75)

“But this is my difficulty. For we are saddened, but how can we be? What are we sad about?” (ibid, p. 77)

I agree that this seems at least ambiguous. It is true that ‘how can we’ questions can be an informal statement of irrationality: ‘how can you claim a man is bald when he has 10,000 hairs (that’s just irrational!)’. But they can also be causal questions: ‘what could cause you to claim that the man is bald, given that he has 10,000 hairs?’.

The causal interpretation seems to be supportable by the examples (hereafter, Radford’s ‘Solution Examples’) Radford gives in his fourth potential solution (Radford (1975), p. 72 - 74), in which he explores putative cases of emotion in the real world context where there is no belief. I shall examine these now.

Radford’s real world examples without belief

In his first Solution Example, Radford imagines a mother’s emotional reaction upon hearing of the fatal accident of her friend’s child. Her own children don’t understand her emotion (‘what’s wrong with you?’ they ask), but, Radford says, the ‘explanation is obvious’. Whilst her children’s question seems to be an accusation of irrationality, what seems to satisfy Radford is the explanation of why she is emotional.

His second Solution Example sees a man imagining the impact on his mother of the death of his sister. His wife, Radford claims, might think him silly - again, seemingly a claim of irrationality, but Radford says that although his emotion at the mere thought of his sister’s death is silly and maudlin, it is ‘intelligible and non-problematic’. Here again, we seem to see a statement of irrationality being glossed over and that what Radford seems to be looking for is the thing that makes it intelligible - the explanation.

Radford now looks at a case where the sister was in very little danger at all. He says that we might find the man’s behaviour ‘worrying and puzzling’. However, ‘we can explain his divergent behaviour, and in various ways’. Once again, he seems to be comfortable with the statement that could be more aligned with irrationality - the ‘worrying and puzzling’ - but moves to his resolution which is the explanation of the behaviour.

Finally, he tests his view by considering his man displaying emotion over something that could have happened but didn’t: his sister’s failure to have children. Radford claims, ‘It is by making the man… a man whose imagination is so powerful and vivid that, for a moment anyway, what he imagines seems real, that his tears are made intelligible, though of course not excusable’. Again, ‘not excusable’ seems to be a statement of irrationality, but Radford is satisfied with what makes the tears ‘intelligible’.

If we take ‘intelligible’ to be aligned with a question of causality (though we will look at a different interpretation), it does seem that much of Radford’s 1975 paper is consistent with a causal reading. However, Kim overlooks that his questions that she takes to be causal are not the main focus of the paper: he uses them in order to make his claim of inconsistency and incoherence. If this is right, then contra Kim’s reading, Radford’s causal questions are not the main point of the paper but a vehicle through which to arrive at his actual main point.

Furthermore, Kim overlooks that Radford explicitly denies that a causal explanation will solve the puzzle ((Radford (1982b), p. 530), (Radford (1990), p. 353))

I do not, therefore, accept Kim’s interpretation of Radford’s paper, but I think there is something in her claim that Radford was asking a different question to how is often interpreted. I shall now go on to outline my own tentative interpretation of his paper as a statement of unintelligibility.

My interpretation of Radford’s paper as agent inconsistency: a puzzle of intelligibility

Unlike Friend, I do not read Radford as a non-cognitivist: he remains committed to the necessity of belief for emotions in the real world context (at least at an intuitive level: his objections to Boruah’s characterisation of cognitivism (Radford (1990), p. 351) possibly suggest a lack of interest in a formal cognitivist theory). It is, I take it, this very commitment that makes him so troubled by our ability to be moved by fiction. I take him to be making the straightforward observation that since, in his view, belief is required for emotion in the real world context, it is a puzzle or paradox - something to be explained - as to how we can have emotion towards fiction.

Whereas Kim takes his ‘how can this occur’ questions to be ones of causality, I take them to be ones of intelligibility: given that belief is required for emotion in the real world context, how can we, as rational agents, make sense of the fact that we have emotion in fiction where no belief is involved. Radford is asking ‘how can our emotion be intelligible given that there is no real person who is suffering?’.

This is an even more straightforward reading than Kim’s causal interpretation, since it takes Radford’s questions about intelligibility and the need for explanation to be exactly that.

This interpretation also allows us to understand Radford’s curious final paragraph of his 1975 paper which notably does not ask a causal question but applies Radford’s notion of inconsistency and incoherence to two real world scenarios - that is, to already existing scenarios. Again, this is consistent with a straightforward question of unintelligibility.

If I am right, this interpretation also illuminates his Solution Examples. These can seem puzzling at first glance since they seem to contradict his previous point that belief is required for emotion in the real world. But an examination of them shows that in each case what satisfies Radford is that the situation described could happen to a real person: it is this, on his account, that makes it intelligible. The mother becomes aware that the accident that happened to her friend’s child could happen to her own (real) children. The case of the man thinking of the impact that his sister being involved in an accident would have on his mother again involves a real person: his mother (I shall ignore his curious decision to make this example involve two real people: his sister and his mother, with the man’s emotion being caused by the latter). These examples seemingly do not trouble Radford because the existence of a real person as the object of emotion means they are intelligible, even if he finds the emotion itself to be irrational or, at the least, silly.

Given the textual evidence for the intelligibility question and Radford’s seeming unconcern of statements of emotional irrationality in the Solution Examples, my suggestion, then, is that Radford’s project is to raise the issue of the unintelligibility of us, as rational agents, having fictional emotions when belief is required for emotion in the real world context.

I take this to be a plausible reading of Radford’s 1975 paper, which is the focus of this essay. However, it may be objected that since (as we see in the above reference) in later papers he does directly refer to the irrationality of both agents and (separately) emotions, I should take this into account in my interpretation of this paper. I shall therefore now examine his later claims about irrationality.

Irrationality of certain emotions (e.g. fear)

Radford’s 1975 paper is almost entirely focused on the general emotion of ‘being moved’, with a particular focus on the titular emotion of pity. It ends with the claim of inconsistency and incoherence which I have claimed should be understood as a kind of irrationality arising from unintelligibility.

The first time that Radford talks directly about a different fictional emotion, that of fear, is also the first time that he uses the term ‘rational’: ‘…such a belief is a necessary condition of our being unpuzzlingly, rationally, or coherently frightened’ (Radford (1977), p. 210) (my emphasis). In this example, Radford has switched the discussion from pity to fear to illustrate a point, seemingly considering both types of emotion to be interchangeable.

Yet it seems that at least at an intuitive level, Radford may be aware of a difference between these emotion types. In later papers, when the debate is almost wholly about fear, Radford has no issues with using the term ‘irrational’, and whilst he does do so in relation to all kinds of emotions, he seems most comfortable in claiming irrationality when talking about fear at fiction.

Indeed, this can be seen at times within the same paper. For example, here Radford is seemingly merely puzzled by fictional emotion:

“And, rather as I have been puzzled by how we can be moved by characters or events that we know to be fictional” (Radford (1989b), p. 69)

but finds fear at fiction to be irrational:

“That they can be frightened by something which they know cannot harm them… is perhaps paradoxical (it is certainly irrational) (ibid, p. 73).

In a later paper, Radford again finds fear at fiction irrational:

“I once argued that it is irrational to be frightened … of anything which we believe cannot harm us” (Radford (1990), p. 349).

but for other fictional emotion claims that we cannot feel them:

“I also argued (ibid.), at greater length, that we cannot feel certain emotions, e.g. sadness, concern, etc., for creatures we know and believe to be unreal, for their unhappiness is unreal.” (ibid, p. 349).

Tentatively, then, there does seem to be a difference in how Radford thinks of different types of emotion. Given this, it seems possible that in his later papers, Radford is arguing for two different kinds of irrationality: his original claim of irrationality arising from the unintelligibility of having fictional emotion, which applies to all fictional emotion, and a second, more specific kind of irrationality that applies only to certain emotions such as fear.

That a distinction can be made between the rationality of different emotion types is easily seen in the real world context by extending Radford’s Belief Examples. Let’s say the suffering of a group of people that we encountered in Example 1 was caused by an escaped prisoner having gone on a violent rampage, and that it took place on the other side of the world. Radford – we have seen - takes it to be a normal human response to feel pity for his victims. But applying his own thinking from his Solution Examples suggests that he would consider us, or the emotion, silly, if we were frightened of the prisoner, given that there is very little chance that he can harm us.

Solomon warned of the danger of assessing emotions as if they are a generic kind, pointing to the difference between ‘engagement emotions’ such as fear and ‘spectator emotions’ such as sympathy (and I would add pity (Solomon (1990)). A review of this distinction is beyond the scope of this essay; however, it is enough for us to note that there is a difference of type between fear and pity, and I am suggesting that Radford’s writing suggests he was aware of this, at least at an intuitive level.

If this is right, then, Radford argues for two different kinds of irrationality: the normative irrationality of fear that we see in his later papers and the irrationality arising from the unintelligibility of our having fictional emotions at all, and this goes some way to explain why he later directly uses the term ‘irrational’.

Plausibility of Radford’s claim of irrationality

Returning, then, to the question of this essay, Radford could be said to be ‘right’ to make a claim for rationality if a) his argument for such is successful and b) it has no defeaters. I have already shown that his argument for the requirement of belief for emotion is underwhelming, and given the difficulties in interpreting his paper, it is hard to consider it wholly successful.

However, if I am right that we should understand his charge of irrationality as the unintelligibility of having emotions at fiction given that there is no real person for it to be directed to, it seems this cannot be resolved: “But, of course, anything but (what we believe to be) real persons or situations will not render our responses rational or coherent anyway” (Radford (1990), p. 354).

That is, the only thing that would satisfy Radford’s concern would be a real person as an object of the emotion but this would, of course, immediately preclude it from our scope of interest about fictional emotion. It therefore seems that the puzzle relating to fictional emotions is unsolvable and thus Radford’s concern remains.

Therefore, based on the interpretation of irrationality that I have outlined, it seems that Radford was right to raise it as a concern in his 1975 paper, although his arguments are at time both unclear and underwhelming.

Conclusion

I have looked at the claim of irrationality in Radford’s 1975 paper and have examined different ways of interpreting this. I have rejected the cognitive paradox interpretation based on Radford’s claim that the issue is located in the having of emotion and not in holding contradictory beliefs. I then looked at two normative interpretations: Friend’s claim of agent irrationality that takes at the normative standard the existence of the object of emotion, and the emotional irrationality interpretation that uses the cognitive criteria of rationality to assess the rationality of emotion. I argued that both these interpretations are flawed in that they assess the rationality of existent emotion whereas Radford’s questions angle more towards the why do we feel the emotions in the first place. I found Kim’s causal interpretation to be more aligned with this aim of Radford, but I didn’t find the causal question to be the sole focus of the paper.

This led me to re-examine Radford’s paper and take his question to be much more straightforward than many interpretations have taken it to be: that it is a straightforward question as to how emotion at fiction can be intelligible given that there is no real person as the object of the emotion and that, since no attempt to solve Radford’s puzzle will plug this gap, Radford’s concern will go unsolved. After briefly examining Radford’s more direct claims of irrationality in later papers, I tentatively find them to relate to a different kind of emotion, leaving my interpretation of his 1975 paper untouched. I conclude, then, he was right to claim (on this interpretation) that there is irrationality in having emotion at fiction.

7319 words

Bibliography

Friend, Stacie. 2020. “Fiction and Emotion: The Puzzle of Divergent Norms.” British Journal of Aesthetics 60 (4): 403–18. https://doi.org/10.1093/aesthj/ayaa010.

Kim, Seahwa. 2024. “The Real Puzzle from Radford.”

Matravers, Derek. 2014. “The (so-Called) ‘Paradox of Fiction’.” In Fiction and Narrative, First edition. Oxford, United Kingdom: Oxford University Press.

Radford, Colin. 1975. “How Can We Be Moved by the Fate of Anna Karenina?” Proceedings of the Aristotelian Society, Supplementary Volumes 49: 67–80. https://www.jstor.org/stable/4106870.

———. 1977. “Tears and Fiction.” Philosophy 52 (200): 208–13. https://doi.org/10.1017/s0031819100023159.

———. 1982a. “Philosophers and Their Monstrous Thoughts.” Brit J Aesthetics 22 (3): 261–63. https://doi.org/10.1093/bjaesthetics/22.3.261.

———. 1982b. “Stuffed Tigers: A Reply to H. O. Mounce”. Philosophy, 57(222), pp. 529–532. Available at: https://doi.org/10.1017/S003181910005614X.

———. 1989a. “Replies to Three Critics.” Philosophy 64 (247): 93–97. https://doi.org/10.1017/S0031819100044065.

———. 1989b. “Emotions and Music: A Reply to the Cognitivists.” The Journal of Aesthetics and Art Criticism 47 (1): 69. https://doi.org/10.2307/431994.

———. 1990. “The Incoherence and Irrationality of Philosophers.” Philosophy 65 (253): 349–54. https://doi.org/10.1017/S003181910005765X.

Solomon, Robert C. 1990. “Review of Fiction and Emotion.” The Review of Metaphysics 43 (3): 620–21. https://www.jstor.org/stable/20128911.

Walton, Kendall L. 1978. “Fearing Fictions.” The Journal of Philosophy 75 (1): 5–27. https://doi.org/10.2307/2025831.

Weston, Michael. 1975. “How Can We Be Moved by the Fate of Anna Karenina?” Proceedings of the Aristotelian Society, Supplementary Volumes 49: 81–93. https://www.jstor.org/stable/4106870.

Yanal, Robert J. 1994. “THE PARADOX OF EMOTION AND FICTION.” Pacific Philosophical Qtr 75 (1): 54–75. https://doi.org/10.1111/j.1468-0114.1994.tb00120.x.

———. 1999. Paradoxes of Emotion and Fiction. Kindle edition. University Park, Pa: Pennsylvania State University Press.