1. Imagination, Fiction and Morality

What are the limits of the imagination in morality? What role does fiction play in moral thought? Tamar Szabo Gendler (Reference Gendler2000) has argued that the imagination is limited in morality, not by logic, but by the will: we can imagine moral worlds that are vastly different from our own, including even worlds that are logically or conceptually impossible, but we are, for the most part, simply unwilling to do so. That, Gendler claims, explains what she calls ‘the puzzle of imaginative resistance’: the problem of explaining the difficulties we seem to encounter in entering imaginatively into the moral, as opposed to non-moral aspects of fiction, where those aspects differ significantly from what we take to be the case.Footnote ¹

I argue, against Gendler's view, that there is no puzzle of imaginative resistance, and that to think that there is such a puzzle is to miss almost entirely the role of fiction in moral thought and its power to change our moral minds. That, however, does not mean that there are no limits to what we can imagine in morality. The imagination, I argue, again in contrast with Gendler's view, is limited in morality as elsewhere by what is logically or conceptually possible. What that suggests is that when we enter into different moral perspectives in fiction, we are engaged in exploring what life would be like with different moral concepts. In this way, investigating the problems with the puzzle of imaginative resistance ultimately stands to show us something important about the nature of moral thought.

2. The Puzzle of Imaginative Resistance

There seems, initially at least, to be a difference in the ease with which we enter imaginatively into the moral and non-moral aspects of fiction, where they differ substantially from what we take to be the case.Footnote ² We have, it seems, for the most part, little difficulty imagining fictional worlds that depart significantly from how we take the non-moral world to be. But we do, it seems, have considerably more difficulty imagining fictional worlds that depart significantly from how we take the moral world to be.

Imagining a morally different fictional world in this sense is not simply a matter of imagining other people or societies whose values differ significantly from our own. If that were all that were involved in imagining a different moral world there would not even seem to be a problem here at all. Rather, imagining a morally different fictional world in this sense is a matter of entertaining such different values ourselves: entertaining the thought that something we take to be morally wrong is actually right (or vice versa) for ourselves, taking the thought up ‘from the inside’.Footnote ³ That, by contrast, seems to be an imaginative task of a different order of difficulty entirely.

What, then, explains the difference between the moral and non-moral aspects of fiction in this respect? Why is it apparently so easy to enter imaginatively into the one and apparently so hard to enter into the other?

Gendler calls this ‘the puzzle of imaginative resistance’ (Reference Gendler2000, p. 55), and she gives a number of examples to demonstrate its existence: Suppose, for instance, a story tells us not only that Macbeth arranged for Duncan to be murdered, but that he was right to do so, or that it was unfortunate ‘only for having interfered with Macbeth's sleep’ (Reference Gendler2000, p. 58); Or suppose that a poem tells us not only that certain white men have taken it upon themselves to educate certain non-white subjects whose lands the white men have colonised, but that in doing so the white men are merely fulfilling their moral duty of educating their moral and intellectual inferiors (Reference Gendler2000, pp. 61–62); Or suppose that a story tells us not only that Giselda killed her own child, but that she was right to do so because ‘it was a girl’ (Reference Gendler2000, p. 62); Or suppose that a story tells us not only that there are white mice, all of which are industrious, and black mice, all of which are lazy, but that the black mice are morally undeserving of help because they are lazy (Reference Gendler2000, pp. 73–74).Footnote ⁴

In each case, Gendler claims, there is a significant difference in our ability to enter imaginatively into the moral and non-moral aspects of the stories (Reference Gendler2000, pp. 62–63). Thus, for instance, Gendler writes that she has no difficulty, reading Kipling's ‘White Man's Burden’, imagining ‘that there are certain white characters who have taken it upon themselves to initiate a group of nonwhites into the ways of Western culture’, but that she does have difficulty imagining, as the poem suggests, that ‘the white characters’ behaviors are a fulfillment of their obligation to “better” those who, by virtue of their skin color, are their natural inferior’ (Reference Gendler2000, p. 62).

We might think that this phenomenon is just an illustration of the boundlessness of morality: the fact that morality, as Aristotle thought (Reference Aristotle and Ostwald1999, pp. 3–4 and 16–18), concerns standards we are judged by in everything we do. But that would merely postpone the problem, since then the question would be why morality is boundless in that way, and specifically why it should concern even the kind of fiction we enter into or enjoy. After all, imagining a fictional world in which the earth is flat does not compromise our belief that it is not flat, factually speaking: it is no threat to our understanding of the (factual) world beyond the fiction.Footnote ⁵ Why, then, should imagining a fictional world in which murder is right be thought to compromise our belief that it is not right, morally speaking? Why should it be a threat to our understanding of the (moral) world beyond the fiction? Again, we do not take our understanding of physics to be threatened by stories about time travel, our understanding of history to be threatened by period dramas representing events that never took place, or our understanding of zoology to be threatened by stories of talking animals. So why should we take our grasp of the moral universe to be threatened by stories in which morality is other than how we take it to be in some way? Why can't we enter into this aspect of the fiction, and safely leave it behind again afterwards?

Gendler's solution is not that we cannot enter imaginatively into (what we take to be) morally deviant fictional worlds because what we are being asked to imagine is logically or conceptually impossible: logical impossibility is, according to Gendler, no obstacle to the imagination. Her solution is rather that we will not enter imaginatively into (what we take to be) morally deviant fictional worlds, and she traces that unwillingness to a general desire she claims we have not to be manipulated into adopting views we would not otherwise endorse (Reference Gendler2000, p. 56), together with what she claims is our default assumption that the moral aspects of fiction, unlike the non-moral aspects, are there to be adopted by us and exported into our understanding of the world (Reference Gendler2000, p. 78).Footnote ⁶ Unlike the non-moral aspects of fiction, then, according to Gendler, the moral aspects are generally taken by us to be a threat to our understanding of the world beyond the fiction.

3. The Role of Fiction in Moral Thought

I am going to argue that there is no puzzle of imaginative resistance. More fully, I will argue that there is no general puzzle of imaginative resistance (wherein what we take to be morally deviant in fiction typically tends to provoke imaginative resistance in the way Gendler claims) and no special puzzle of imaginative resistance (wherein what we take to be morally deviant in fiction typically tends to provoke imaginative resistance in a way that other elements of fiction do not).Footnote ⁷ There is no general resistance to imagining morally deviant worlds in fiction, and what resistance there is to doing so is not special to deviant morality. Simply put, there just is no comparative difference between the moral and non-moral in this respect to be explained.Footnote ⁸ Moreover, I will argue that we could only think that there is a puzzle here by ignoring the role of fiction in moral thought, and the power of fiction, for good or ill, to change our moral minds.

My view, then, could be called eliminativist, since I am claiming that there is no phenomenon of imaginative resistance to be explained. As Gendler and Liao use that term, however, it is associated with a range of views that I reject, such as that there is a philosophically interesting phenomenon here, that the phenomenon is simply more restricted than Gendler claims, that the phenomenon can be explained by a lack of context, and that the phenomenon raises several distinct but familiar problems rather than a single new one (Reference Gendler, Liao, Gibson and Carroll2016, pp. 409–410; cf. Tuna Reference Tuna2020).Footnote ⁹ These views simply do not go far enough: there is no phenomenon requiring explanation as Gendler describes it. What is philosophically interesting here is a failure to look and see what moral thinking in fiction is really like that reveals itself in the idea that there is a phenomenon requiring explanation at all.

The claim that there is no puzzle of imaginative resistance might seem strange: Not only is there a basic plausibility to the idea that we have difficulty imagining to be true what we take to be morally deviant views or entering imaginatively into fiction that presents such views, but the claim itself seems problematic morally. After all, it would be worrying if you did not find it hard to enter imaginatively into the examples Gendler discusses: if, for instance, you did not find it hard to imagine that murder is right, or that it is okay to kill female children because they are female, or that racism is morally justified, that would be a sign that there is something morally very wrong with you.

But I am not claiming that we find it easy to enter into Gendler's examples. I am not even claiming that we can enter into them. Rather, what I am claiming is twofold: first, that the examples Gendler provides are bad examples, and therefore they don't show what Gendler claims them to show, since together they present an impoverished view of the ways in which morality can feature in fiction and of how fiction can draw us imaginatively into different moral perspectives; second, that when we look at other examples taken from actual works of fiction that do present us with moral perspectives different from our own, the idea that there is a puzzle of imaginative resistance at all is simply absurd. I shall take these points in turn.

The first problem with Gendler's examples is simply that they are bad as literature, and part of the reason that they can seem persuasive is simply that we do not compare them with examples that are similar in that respect.Footnote ¹⁰ Whereas Gendler invites us to call to mind, as examples of (what might be called) ‘factually deviant’ fictions, works like Macbeth, The Adventures of Sherlock Holmes, and The Lord of the Rings (Reference Gendler2000, pp. 56–57) that are developed at length and (to many) highly engaging, Gendler's examples of morally-deviant fictions are neither of those things: they consist of a sentence or two or a few lines, and (with the possible exception of Kipling's poem) they display little in the way of literary merit. It is not that you cannot produce something engaging in so few words.Footnote ¹¹ It is that it takes a great deal of skill to do so, and Gendler's examples just do not display that kind of skill. If we compare cases that are equally good or equally bad the differences disappear.Footnote ¹²

The second problem with Gendler's examples is that they all involve cases where the moral of the story is simply stated, and stated, moreover, in the authorial or narratorial voice. We are not shown what a different moral perspective looks like in practice. We are not shown characters with moral views or attitudes apparently different from our own in the stream of their lives. We are not brought to care about, sympathise with, like, admire or trust those characters in any way. We are not shown what pressures lead them to think or act in the ways they do, or how their behaviour impacts on them and those around them. Instead, we are told to imagine that Macbeth was right to kill Duncan, we are told that white men have a burden to ‘better’ their inferior non-white subjects, we are told that Giselda was right to kill her female child, and we are told that the lazy black mice are undeserving of the industrious white mice's help.

But this is not how good fiction typically gets us to imagine anything, let alone how it gets us to imagine our way into fantastical worlds or moral perspectives that differ significantly from our own. Stories about time-travel, for instance, to take one of Gendler's examples, rarely simply state that time-travel is possible, and if they do, it is typically a character within the fiction rather than an authorial or narratorial voice that does so, and the statement is often greeted by that character or others with a resistance that mirrors our own, only for their resistance (and, hopefully, ours with it) to be broken down as the narrative proceeds.Footnote ¹³

The third problem with Gendler's examples is that Gendler treats them as if they all involve a sharp distinction between the moral and non-moral. It is that distinction, according to which the moral can be neatly compartmentalised, rather than the fact/value distinction per se, that is crucial to the idea that there is a puzzle of imaginative resistance to be solved.Footnote ¹⁴ But arguably that distinction is problematic even in relation to some of Gendler's own examples. The facts we are asked to imagine in Kipling's poem, for instance, that certain white men have taken it upon themselves to educate their non-white subjects who have responded with ingratitude to these attempts at their betterment, are in that context not separable from the basic moral premise of the poem, that the white men are morally and intellectually superior to those subjects, who stand in need of being civilised against their will. Likewise, the basic facts of the story of the mice, that there are white mice, all of whom are industrious, and black mice, all of whom are lazy, are obviously bound up with the racist moral of the story from the start; hence, if we resist entering imaginatively into the moral of the story, we should resist entering imaginatively into the basic facts of the story as well, even if some of those facts (e.g. that there are white mice and black mice) are not only possible, but actual.Footnote ¹⁵

More importantly, the idea that morality tends to feature in literature in such a neatly compartmentalised way, let alone that it tends to feature that way in good literature, or that it always features in literature in that way, simply does not stand up to scrutiny. To give just one example, the Iliad's moral reflection on war, with very few exceptions,Footnote ¹⁶ is not contained in neatly compartmentalised statements about or comments upon the facts of the war. It is contained in the descriptions of the facts themselves: in the graphic descriptions of death in battle (e.g. Homer Reference Homer and Fagles1998, Bk. 4, ll. 600–616, Bk. 5, ll. 72–83, 321–29, and 666–78, Bk. 13, ll. 749–51, Bk. 16, ll. 389–413 and 854–65) and the contrast between those deaths and the injuries to Aphrodite (Bk. 5, ll. 380–494) and Ares (Bk. 5, ll. 973–1050), in the glimpses of lives now lost and of those (parents, wives, children) left behind (e.g. Bk. 5, ll. 76–83 and 168–77, Bk. 6, ll. 14–33), in the metaphors used to depict the killing (e.g. Bk. 5, ll. 637–40, Bk. 8, ll. 349–53, Bk. 13, ll. 452–58) or the battle more generally (e.g. Bk. 4, ll. 517–28, Bk. 11, ll. 76–86, Bk. 15, ll. 722–40), in Hector's estrangement from various aspects of civilised life on his return to the city (Bk. 6, ll. 280 ff.), in the breakdown of the heroes’ values as the battle progresses (e.g. Bk. 6, ll. 53–77 and 77–83, Bk. 16, ll. 866–73), in the use of the second-person to address a character immediately before their death (Bk. 16, ll. 914–16 and 943–45), in the depiction of the cosmos and the place of war within human life on Achilles’ shield (Bk. 18, ll. 558–709), in the self-destructive image of Achilles killing Hector when Hector is dressed head-to-toe in Achilles’ armour (Bk. 22, ll. 374–86) or of the Greeks laying siege to what is depicted as a recognisably Greek civilization (Bk 6, ll. 283 ff.), in the fact that Achilles, having killed Hector, does not then sail home to a long life in Phthia, preferring instead an early death at Troy, though there is nothing except his self-evaluation to stop him leaving (Bk. 24, ll. 559–790), and so on. In none of these cases is the moral aspect separate from the facts of the story. It is a matter of what the facts are: of which facts are represented and how they are represented. Thus, to enter imaginatively into the facts of the story here is to enter imaginatively into the moral world of the story: there is no doing one without the other.Footnote ¹⁷

All in all, then, Gendler's examples are not fit to make the kind of point she wants them to make. Together, they present too crude a picture of both fiction and morality to establish the existence of a phenomenon that requires explanation here. If we want to know if there is a general difference in the ease with which we enter imaginatively into ‘factually deviant’ and morally deviant fictions, then, we need to look elsewhere, at actual cases of fiction that present us with moral perspectives different from our own. But when we do that, the idea that we do typically or generally struggle to enter imaginatively into such perspectives is simply absurd.Footnote ¹⁸

Take (the idea of) revenge, for instance. Most of us probably do not think that revenge is good or right or just. Most of us, I would guess, would agree with Socrates, in principle at least, when he says that we should never do wrong or, when wronged, inflict wrong in return, and that we should never mistreat anyone or, when mistreated by someone, mistreat them in return (Plato Reference Plato and Grube2002, ll. 49b–c). Nevertheless, many of us enjoy revenge fiction, and it is often central to our appreciation of it that we enter imaginatively into the possibility that revenge is required, morally, in doing so. If, in reading the Iliad, you do not share the sense that Achilles has to kill Hector in revenge for Patroclus, then you will not be able to appreciate how the poem unfolds after Patroclus’ death. If you cannot even imagine that Achilles might be required, morally, to take revenge upon Hector, then you will not be able to understand why Achilles does what he does there at all. Likewise, if, in reading Aeschylus’ Oresteia (Reference Aeschylus, Shapiro and Burian2003), you cannot imagine how it could seem right to Clytemnestra to kill Agamemnon in revenge for Iphigenia, or how it can seem right to Orestes to kill Clytemnestra in revenge for Agamemnon, you will miss entirely the moral force of the trilogy. That is not because the Oresteia is unambiguously in favour of violent revenge as means of justice, but because it is not unambiguously for or against it. The trilogy as a whole is an argument without a clear conclusion: to appreciate that argument, we have to be able to appreciate both sides of it.Footnote ¹⁹ In both examples, we cannot explain what goes on when we enter imaginatively into these works simply in terms of any pleasure we may experience in imagining revenge: What we have to be able to imagine, if we are to understand them at all, is that revenge is required, morally, whether we like it or not.

Similarly, take police corruption, for instance. Most of us would, again, I think, hold that police corruption is morally bad. But corruption is a staple of police dramas, and in many cases our engagement as viewers is premised on our wanting the characters involved to get away with it without being appropriately punished. In the drama Spiral (MHZ 2013), for instance, the character Gilou repeatedly engages in corrupt behavior, often dragging other members of his team into further corruption as a consequence: he accidently shoots someone, tampers with evidence, steals from a crime scene, and so on. But if, as a viewer, you do not want him and his colleagues to succeed in covering up his actions, if you are hoping that he is caught and appropriately punished, then you are not entering properly into the events of the drama. It is crucial to our engagement with those events that we are invested sufficiently in his character to want him to get away with it, even if, outside the fiction, we would not, for moral reasons. The tension that we experience as viewers that makes the drama successful as a drama comes not from the hope that Gilou and his colleagues will be caught and appropriately punished for their corruption, but from the hope that they will not, combined with the fear that they might. For us to feel that tension, we have to view as desirable or good a course of events that we would not otherwise view as such.

The strength of these examples, of course, depends not only on what moral views, and attitudes we have to start with, but also on whether we do respond, individually, to the examples in the ways described. But part of what I am suggesting is not just that it is common for us to enter imaginatively into different moral worlds in response to works of fiction,Footnote ²⁰ but also that, in cases like these, it would be a sign of our failure as readers or as viewers were we not to do so: we would be bad as readers or viewers were we not to enter into such perspectives in responding to them.

That might sound like a version of what Wimsatt and Beardsley (Reference Wimsatt and Beardsley1949) called ‘the affective fallacy’: the fallacy of confusing the meaning of a work with its effect upon the reader. But first, what we are talking about in connection with the puzzle of imaginative resistance is precisely the effect of the work upon the reader (or audience), and second, the affective fallacy just is not a fallacy anyway, since the effect upon the reader is often crucial to the meaning of a work and its significance for us: trying to make sense of what is so great about the Iliad without mentioning how Achilles’ story is apt to move us is hopeless; trying to explain what makes Oedipus the King peerless as a tragedy without mentioning the combination of elation and despair it is apt to produce in us as Oedipus vainly, hopelessly, seeks to reassert control over his fate by plucking out his own eyes, would be absurd. If you leave out the effects these texts are apt to have upon us, you leave out an essential part of what makes them so great.

To think that there is a general problem entering imaginatively into different moral views and perspectives in fiction, then, is to miss almost entirely the role that fiction plays in moral thinking, and its power to change our moral minds: to draw us into new moral perspectives, for better or worse, and to show us what they are like from the inside. Entering imaginatively into a different moral perspective in this way may take very different forms. We may be brought to feel the necessity of a certain course of action. We may be lead to sympathise with a character or understand more fully their situation and choices. We may come to appreciate similarities or differences that we were not previously aware of, or analogies or distinctions we did not see could be made. We may be brought to share an emotional response or set of expectations that we would otherwise not share. We may simply become accustomed to certain conduct, through its repeated representation to us. In these kinds of ways, we may, reading the Iliad or Oresteia, be brought to imagine that revenge is good; We may, watching Spiral, find ourselves on the side of the corrupt; We may, reading Lolita, begin to see things from the perspective of a pædophile; We may, reading Paradise Lost, come to feel sympathy for the devil. The point is that it is not even rare to find ourselves, in responding to fiction, taking up a position morally that we would not otherwise want to occupy, or that outside the fiction we would disown, whether we realise it or not. The fact that these things are possible, if the fiction is good enough, is what makes fiction so important, morally, and potentially so dangerous too.Footnote ²¹

None of this means that we experience no difficulty at all in imagining different moral worlds in fiction, no matter what the views or attitudes on display or how good or bad the fiction. But then we may also experience difficulty in entering imaginatively into all kinds of other things in fiction too. We may be alienated by the politics of Ayn Rand, the sexual fantasies of E. L. James, the social world of Marcel Proust, the ghoulish horror of H. P. Lovecraft, the tedious interruptions of George Eliot, or the overbearing stage directions of George Bernard Shaw. We may not find the Marvel universe so marvelous. We may struggle with Karl Ove Knausgård's My Struggle. We may even experience difficulty imagining what we fear is right, morally, too, perhaps even more so than what we think is wrong: we may resist acknowledging that we are wrong even when we know it, or simply that we might be wrong when we know that too.Footnote ²² The idea that we encounter a quite special resistance to imagining what we take to be morally wrong in fiction is just that: fiction.

4. The Limits of the Imagination

I have argued that to think there is a puzzle of imaginative resistance is to miss almost entirely the role of fiction in moral thought and its power to change our moral minds. The appearance of a phenomenon of imaginative resistance, of a difference in the ease with which we enter into morally and non-morally deviant fictions, ultimately depends upon a failure to look and see what role fiction actually plays in moral life.

But that does not mean that there are no limits at all to what we can imagine in morality, and it would be a mistake to conclude that there are none. What I shall argue now, again in opposition to Gendler, is that the imagination is limited in morality by logic: in morality as elsewhere, we cannot imagine what is logically or conceptually impossible. Therefore, if ‘murder is right’, say, is logically or conceptually impossible, then it is unimaginable too.

That point is important not simply as a corrective to Gendler's views, but for what it suggests about the role of fiction and the imagination in moral thought and the nature of moral thought itself. What it suggests, I shall argue in the final section below, is that when we enter imaginatively into different moral perspectives in fiction, we are involved in revising our moral concepts to bring new possibilities into view. Ultimately, I shall argue, it is that role for fiction and the imagination in moral thought that we overlook in taking there to be a puzzle of imaginative resistance at all. In this section, I begin with the question of whether we can imagine what is logically or conceptually impossible in general, before returning specifically to morality.

Why, then, think that we cannot imagine what is logically or conceptually impossible? Take Wittgenstein's example: ‘Every rod has a length’ (Reference Wittgenstein, Anscombe, Hacker and Schulte2009, §251). The problem with imagining a rod that has no length is not that it is so very hard to do. The problem is that there is nothing to do here at all. To imagine a rod is to imagine something with a length. To imagine something with no length is not to imagine a rod. There is no room in the middle for the imagination to occupy. We cannot imagine what is conceptually impossible because there is no such thing.Footnote ²³

But if you cannot imagine a rod with no length, then you cannot even say that you cannot imagine a rod with no length: that is, you cannot say it and mean it, because there is nothing there to mean. The words are simply senseless. That, ultimately, is the problem with the idea that the imagination is not limited by what is logically or conceptually possible: there is nothing beyond that limit for the imagination to get hold of. The words we use to try to capture what we are supposed to be imagining are empty.Footnote ²⁴

To say that the imagination is limited by logic, then, is not to imply that there are things that we can specify or point to beyond that limit that we cannot imagine: the limit is not a boundary between two spaces as the term might suggest. One could, therefore, just as well say that there is no limit to the imagination at all insofar as there is nothing that we can specify or point to that we cannot imagine, so long as that is not taken to imply that we can imagine what is not logically possible after all. The point is simply that whatever is imaginable is possible too (cf. Wittgenstein Reference Wittgenstein, Pears and McGuinness1974, §3.02–3.031 and §5.4731).

It is of course possible to imagine someone who sincerely believes that they can imagine something logically or conceptually impossible as, for instance, Graham Priest claims that it is possible to imagine a box that is empty and contains a statue (Reference Priest1997, pp. 575–76 and 579–81).Footnote ²⁵ But that does not make it plausible that we can imagine such things. We may think that we can imagine, or think, or believe, or even just say something, and still be wrong about that. Merely insisting that we can imagine something conceptually impossible is not an adequate means of defending the idea that we can.

To defend that idea, to make it remotely plausible that we can imagine something conceptually impossible, we would need to be able to say something substantive about what the impossibility is that we are supposed to be able to imagine.Footnote ²⁶ Without that, we have no reason to think that there is anything here to be imagined at all. But we cannot do that without thereby articulating a genuine possibility that could make sense of our words where currently they seem to have none. Thus, anything we could say in defence of the idea that we are imagining something would itself count against the idea that what we are imagining is conceptually impossible: it would, if successful, show that what we are imagining is not impossible after all (cf. Conant and Dain Reference Conant, Dain, Read and Lavery2011, pp. 69–72). Hence, the idea that we can imagine what is conceptually impossible is indefensible: it cannot coherently be defended except by mere insistence, and mere insistence is no defence.

Priest tries to defend the claim that we can imagine an empty box that contains a statue by telling a story in which the characters discover a box that they take to be empty despite knowing it contains a statue. But the story adds nothing to the basic claim that we can imagine this. In particular, it does nothing to make intelligible the claim that the box is empty despite containing something or to explain why the characters in the story believe it is empty despite believing it to contain something. The characters in the story seem either not to understand the words they use, or to mean something quite different by them.

We might think nevertheless that things are different when it comes to more complex cases (in mathematics, for instance) where we are not sure whether what we are trying to imagine is conceptually impossible, or where it takes a great deal of thought to appreciate that it is. If, for instance, it is impossible for there to be an even number that is not the sum of two primes, it is not obvious that it is. In that kind of case, it is quite natural to think that we can imagine what is conceptually impossible. In fact, that might even seem to follow from the fact that we can consider whether these things are possible at all.

But again, we may think that we imagine something, and we may think that our words make sense, but that does not show that we are imagining something or that our words do make sense. What we discover, when we discover that something we took to be thinkable is not possible at all, is precisely that there is nothing to be imagined where we thought there was. We discover that we only thought we could imagine something because we did not understand what we thought we could imagine, or that whatever we were able to imagine is not what we thought it was. We discover that we were no different from someone who thinks they can imagine a rod with no length, or a shade of red that is not a colour.

The confusion behind Priest's view is that terms like ‘possible’, ‘impossible’ and ‘necessary’ are themselves terms that apply to possibilities, as if some possibilities are merely possible, others are necessary, and others are impossible.Footnote ²⁷ They are taken to function as terms for categorising possibilities, not for distinguishing between what is possible and what is not. In that way, we seem to be able to ask whether certain impossible possibilities are necessarily impossible possibilities, or whether there could be worlds in which they are possible possibilities. But to say that something is impossible is not to say that it is a possibility that is impossible, and so that it is both possible and impossible at the same time. It is to say that it is not a possibility at all. Thus, when we find ourselves asking whether something (an even number not the sum of two primes, say) is possible or not, we are not asking whether the possibility we are imagining is merely possible or also (as well as being possible) impossible too: we are asking whether there really is a possibility here to be imagined or not.Footnote ²⁸

That confusion itself rests on the idea that we understand sentences compositionally, by understanding what the individual words mean (words such as ‘box’, ‘empty’, or ‘statue’), together with the way that they are combined to form a sentence, so that we seem then to be able to understand sentences purporting to express impossible possibilities. But the fact that we cannot elaborate on, or articulate, or specify what the words, in that combination, mean in this context is itself evidence for thinking that this is not all there is to meaning and understanding: that the fact that we are familiar with these words, and these kinds of combinations, in other contexts gives rise to an illusion of understanding something here where there is, as yet, nothing to be understood.Footnote ²⁹

Gendler, by contrast with Priest, argues that we can imagine something conceptually impossible by focusing our attention on certain features of the relevant concepts not others: we can, for instance, Gendler claims, imagine that twelve is not the sum of five and seven by imagining something which shares some of the features of twelve, while ignoring others, such as that it is the sum of five and seven (Reference Gendler2000, pp. 67–70). But it really matters which features what we are imagining has. It cannot merely be that what we are imagining shares with twelve the feature of being a number (like fourteen) or of being signified by the sign ‘twelve’ (which we might use to signify another number, like fourteen): that would just be to imagine a number that is not the sum of five and seven, and there is nothing impossible about that. For what we imagine to be conceptually impossible, for it to be that twelve (not some other number) is not the sum of five and seven, we need something other than that. The problem is that the features that whatever we are imagining needs to share with twelve, if what we are doing is to count as imagining twelve at all, are those features that are essential to that concept. But it is precisely those features that what we are imagining must also not have if we are to imagine something conceptually impossible. Thus, we would have to imagine something that has and does not have those features. Nothing Gendler says makes it plausible that we can do that, and so her account does not help to show how we could imagine something logically or conceptually impossible.Footnote ³⁰

Like Priest, Gendler defends her claim by telling a story: in this case, one in which God decrees (among other things) that five and seven do not make twelve (Reference Gendler2000, pp. 67–68). But again like Priest, it is not clear how the story is supposed to make it plausible that this claim is intelligible or imaginable. To say that this claim is true in the context of the fiction is not yet to tell us what, if anything, it means there. So we can grant, as Gendler says, that she has told a story in which this claim is true (Reference Gendler2000, p. 68). The problem is that we don't know what it means, and hence we don't know what it would be to imagine it is true.Footnote ³¹ For it to be intelligible, we would need much more detail about what exactly it involves. But again, providing such details would be a way of giving a sense to the words by articulating a possibility, such as that God decrees a simple change in the use of the word ‘twelve’.Footnote ³²

Before turning back to morality, I want to discuss an objection to my argument here, which is this: If, as I have argued, we cannot even say what is conceptually impossible, then some of what I have written in arguing for that view is itself empty nonsense. I have, for instance, given various examples: of a rod that has no length, an empty box that contains a statue, twelve not being the sum of five and seven. But if we cannot say what we cannot imagine, then when I have given these examples I have said nothing at all.Footnote ³³ This is true. The words we use to try to capture what we want to say in such cases fail to make sense.Footnote ³⁴ But what that shows is not that we can say or think such things after all. It shows how easily we are taken in by the mere appearance of sense where there is none, and so just how big a challenge faces those, like Priest and Gendler, who want to insist on the basis of such appearances that we can imagine something where there is nothing to be imagined.

We cannot, then, I have argued, imagine what is logically or conceptually impossible, even if we want to: the imagination is limited by what is logically possible. That goes for morality as for anything else. If it is not even possible for, say, murder to be right, then there is simply no imagining or saying that it is. For something to count as murder in that case just would be for it to be wrong.Footnote ³⁵ (I am not here saying that ‘murder is right’ is conceptually impossible: maybe it is; maybe it isn't. What I am concerned with is what would be the case if it were conceptually impossible.) Likewise, if, for instance, it is part of our concept of cruelty that cruelty is morally wrong, then we cannot simply imagine that cruelty itself, just as such, is kind. To imagine that cruelty itself is kind we would have to imagine that what we call ‘cruelty’ takes the place in our conceptual lives that kindness occupies: for instance, that everything that we previously would have called ‘cruel’ we now call ‘kind’, and that we now respond in the ways we respond to kindness to what we previously considered cruel, and so on (see Dain Reference Dain2014, pp. 5–6). But if cruelty is by definition wrong, then there is no doing that: anything that we think of in those ways will not be cruel. We could only imagine this if cruelty referred, not to something by definition wrong, but to some set of actions or occurrences, picked out in entirely naturalistic terms, which we could think of in different ways: now as good, now as bad.

Although fiction has the power, then, for better or for worse, to draw us imaginatively into new and different moral perspectives, it cannot help us to enter imaginatively into moral perspectives that are, or would be conceptually impossible. There simply is no doing that, because there is nothing there to be imagined at all. We are limited in what we can imagine in morality by logic, as everywhere else.

5. Imagination, Fiction and Morality again

I have argued that there is no puzzle of imaginative resistance, and to think there is, is to miss almost entirely the role of fiction and imagination in moral thought. But that, I have also argued, does not mean there are no limits to the imagination in morality: we cannot imagine what is logically impossible, because there is no such thing.

Together, these two claims have significant consequences for our understanding of the nature of moral thought itself, and the role of fiction and imagination within it, and I want in conclusion to bring those consequences out by pointing to what might at least seem to be a tension between the two claims as I have argued for them. In arguing for the first claim, for instance, I have suggested that we can imagine things like revenge being good, but in arguing for the second I have suggested that we might not be able to imagine that murder is right. What is the difference supposed to be between these two cases? Why should one be imaginable and the other not?

I want to suggest, surprisingly perhaps, that there is no difference between them. When we imagine that revenge is good, we imagine what life would be like with a different concept of revenge, one that shares enough in common with our concept of revenge for it to be recognisable as an alternative to that concept, rather than something wholly unconnected to it, but one that nevertheless occupies a very different role in our thinking, with a range of different connections to other concepts, judgments, and actions.Footnote ³⁶ We imagine a different moral practice, woven from different moral concepts: an alternative form of life. Thus, in Aeschylus’ Oresteia, for instance, we do not simply see a clash between two contrasting attitudes towards a single concept of revenge: we see two different forms of life, centering on two different concepts of revenge, vying with one another. What we are asked to do as the audience is not simply to consider whether revenge is good or bad, but to weigh two different forms of life against one another, centering on two different conceptions of revenge: one in which it is something required of us; one in which it is not. What we imagine on this account, then, when we imagine revenge being good, is not something logically impossible, but it is all the same something that is not available for thought with the concepts we start out with. We have to rethink the concept to make sense of what we are being asked to imagine. There is nothing to stop us rethinking or revising any of our moral concepts in this way too, including even the concept of murder.Footnote ³⁷

Reflecting on, rethinking or revising our moral concepts in this way is the work of the imagination, and there is no right or wrong, valid or invalid method of doing that. No matter how good our reasoning in morality may be, then, we are always at the mercy of the imagination and its contribution to how we conceptualise the moral world. Thus, the only certainty there can be in morality is the certainty of commitment, not of knowledge: our moral knowledge is always relative to the concepts we use in making sense of the world.

The imagination, then, I want to suggest, plays a much more fundamental role in moral thought than we might think, shaping our understanding of the moral landscape itself through the concepts we bring to bear upon it, and thinking through the problems with the puzzle of imaginative resistance can lead us to see that. The power of fiction to change our moral minds, the power that Gendler and others miss in thinking that there is a puzzle of imaginative resistance at all, is the power to reshape our entire moral world in this way. Fiction is not the only thing that can do that,Footnote ³⁸ but it is perhaps uniquely well placed to do it.Footnote ³⁹