Commentary by Alon Chasid on Explaining Imagination (with reply)

By Alon Chasid

Peter Langland-Hassan’s Explaining Imagination (hereafter: EI) presents a reductive thesis: imagining is not a sui generis mental state or attitude, but one of the basic folk- psychological attitudes such as beliefs, judgments (i.e., occurrent beliefs), desires, intentions, etc., or combinations thereof. At first sight, this novel thesis is difficult to digest. After all, imaginings do not come across as being beliefs (desires etc.): ordinarily, we seem to be able to straightforwardly distinguish between imagining a proposition and, say, believing it. Peter acknowledges this. In reducing imaginings to, e.g., beliefs, Peter does not maintain that imagining that p is identical to believing that p. Rather, he argues that it is identical to believing that p is possible or fantastical, or to believing that p is true according to a work of fiction, etc. Indeed, since imaginings can arise in a wide variety of contexts, the proposed reduction is not monolithic, but context-dependent. Yet Peter’s general principle is clear: any case that seems to involve a sui generis attitude of imagining can be fully explained in terms of more basic folk-psychological attitudes.

EI is a pleasure to read. Both lucid and sophisticated, it seems to anticipate every potential challenge Peter’s account might face. It also adduces arguments that are relevant not only to defending his account, but also to specific debates about the nature of imagining. EI is thus a significant philosophical achievement, and in my opinion, will soon be a mainstay of research into the metaphysics and psychology of imagination. In this post, I will argue against EI’s core thesis by adducing two types of cases where imagining cannot be reduced to believing, judging, etc. The first involves spontaneous imagining, the second involves imaginings that arise in response to rereading (re-watching, etc.) works of fiction.

It is widely accepted that imagining can arise spontaneously, i.e., without deliberation or intention. Spontaneous imaginative projects are, it seems to me, quite common. EI, however, does not discuss spontaneous imaginings systematically. It refers, e.g., to Walton’s famous example of Fred, a shoe salesperson, who finds himself imagining that he is rich and famous, owns fancy cars, etc. (Walton 1990, 13). Peter maintains that Fred’s imaginings are identical to Fred’s desires to be rich and famous etc., but notes that not every spontaneous imagining is a desire (EI, 25). Another reference to spontaneous imaginings can be found in Peter’s definition of daydreams as attitudinal imaginings that serve no immediate practical goal (EI, 88), and can be accompanied by mental imagery. Let us assume that ‘serving no immediate practical goal’ implies that such imaginings arise extemporaneously, without deliberation. On

Peter’s view, such imaginings can be reduced to judgments about possible but unreal scenarios. E.g., my (spontaneous) imagining that I’m driving a Mustang can be reduced, say, to the judgment that owning a Mustang will alleviate my midlife crisis. The basic idea is that non-deliberate imaginings can often be reduced to judgments; indeed, such judgments arise non-deliberately in the same sense, and to the same degree, that imaginings arise non-deliberately (as per the proposed reduction).

The problem is that judgments cannot be non-deliberate, since they must arise for a reason, in response to evidence, etc. In general, we do not simply ‘find ourselves’ believing a proposition, since beliefs are guided by evidence or reasoning. Imaginings, by contrast, need not involve reasoning or evidence. When imaginings arise non- deliberately, we ordinarily can give no reason for why we imagined what we did. But if there was no reason or evidence that induced me to imagine certain propositions, how can my imagining be considered a judgment about possible scenarios (however ‘possible’ is construed)?

Consider the following example. Suppose I find myself imagining that the cheesecake I just baked will taste like a burger. If asked whether I think it possible that this cheesecake, or cheesecakes in general, could taste like a burger, I have no immediate answer. For one thing, I have no idea why I imagined this proposition: my imagining came completely unbidden. For another, although I’ve just imagined this strange proposition, to answer the question of whether or not it is possible that my cheesecake, or any cheesecake, could taste like a burger, I must think about what would constitute a proper reply, and what evidence would warrant it. But on Peter’s reductive account, no such deliberation occurs. It follows , I contend, that my spontaneous imagining is not a judgment. (For the sake of argument, we can further assume that I have no desire for cheesecake that tastes like a burger, hence my imagining is also non-reducible to a desire etc).

That spontaneous imaginings are not judgments can also be shown by considering spontaneous daydreams involving shifts between incompatible propositions. Suppose I find myself imagining having lunch with an old friend, who is slowly transformed, before my eyes, into a monster. On Peter’s view, in imagining, I made a judgment: I judged that it is possible that I’m having lunch with my friend, and that my lunchmate isn’t, in fact, my friend, but a monster. Did I really make such a contradictory judgment? If so, what reasons or evidence did I have for that judgment? Or might I, perhaps, have made two (or more, given that the transformation was gradual) judgments about a possible scenario? If so, which of the judgments was correct and which was not? There seems to be no way to answer these questions. But if imaginings are identical to judgments, I should be able to answer them. The conclusion seems to be that spontaneous imaginings are not judgments.

The second type of case that seems to challenge the proposed reduction involves imaginings that arise in response to reading (watching, etc.) works of fiction. On Peter’s view, imagining in response to fiction is reduced, again, to making judgments. In reading a work of fiction, we judge that according to the work, certain propositions are true. In addition, we also conjure up images, have various ‘feelings,’ etc., but our primary cognitive attitude to the work’s content—i.e., the attitude that is commonly deemed to be ‘imagining’—amounts to making judgments or having beliefs about what the work takes to be true. To see why this account is problematic, consider the following two claims. First, what a work presents as true at a certain stage of its unfolding is not always what the work actually takes to be true. A work can give the initial impression that a certain person is humorless, that an event was disastrous, that a spaceship won’t lift off, whereas by the end, it has revealed that the person is amusing, the event was successful, and the spaceship lifted off perfectly. Such ‘shifts,’ often used to arouse certain emotional, conative and cognitive reactions in readers, are commonplace in fiction.

Secondly, knowing in advance what a work of fiction posits to be true needn’t thwart our imaginative experience. Many works of fiction invite rereading, and each rereading generates the same overall imaginative experience as the initial reading did. We may be well aware that, say, a certain fictional character turns out to be nasty, but in rereading the early chapters, which describe him as pleasant and considerate, we have no difficulty imagining him as such, despite our knowledge of his ‘true’ character. Indeed, just as knowledge of truths simpliciter does not usually thwart our engagement with a work of fiction, knowledge of what that work takes to be true does not usually thwart our engagement with the work. Combining these claims, we can conclude that imagining in response to fiction is not judgment or belief about what the work takes to be true. In rereading a work’s early chapters, we can easily imagine that which we know to be false according to the work. It follows that holding beliefs about what the work takes to be true is one thing, imagining the work’s content is another.

At this point, the reductionist could perhaps refine his account, and claim that the core attitude we have in response to reading fiction is judging, not what the work sets down as true, but what the work presents as true at each stage of its unfolding. This move, however, would be problematic for Peter’s explanation of emotional responses to fiction (EI, chapters 10 and 11). Peter rejects explanations of emotional responses that invoke sui generis imaginings and desires, and argues that emotional responses can be fully explained by beliefs and desires vis-à-vis the work in question. For instance, we are ‘afraid’ since we believe that something dangerous is happening in the work, and we want that, according to the work, the dangerous event will end. However, given the suggested revision, this explanation is untenable, especially in cases where, knowing the story, we know that nothing dangerous ultimately transpires. It seems that the revised reductive explanation of emotional responses must explain our ‘fear’ in terms of a belief that the work is presenting something as dangerous (though it will later be revealed to have been harmless), and a desire that it not be presented as such by the work. But this account of our ‘fear’ seems odd: why would our preferences and beliefs about what the work, at a specific stage in the narrative, presents as dangerous, generate the said fear?

There is much more to say about these, and similar, challenges facing the reductionist stance. Elegant and refreshingly unconvoluted, Peter’s account of imagining deserves a comprehensive discussion, which mandates going into detail, and exploring the wide variety of contexts in which imaginings arise. I anticipate that EI will generate both new debates about the nature of imagination, and exploration of aspects of imagining that have escaped philosophical scrutiny.


Langland-Hassan, Peter (2020). Explaining Imagination. Oxford: Oxford UP.

Walton, Kendall L. (1990). Mimesis as Make-Believe. Cambridge MA: Harvard UP.

Reply from Peter Langland-Hassan

I am very grateful to Alon Chasid for his thoughtful commentary and many well-framed criticisms.  They touch on issues that are difficult to satisfactorily address in a short blog post. But I will sketch some replies and am happy to continue discussion in the comments.

Alon’s first set of questions concerns what he calls “spontaneous imaginative projects.”  Intuitively, these are imaginings that seem to just pop into our heads from out of nowhere and to bear little if any bearing on what we are doing at the moment.  It may be hard to see how any of these imaginings could be reduced to states like judgments, desires, and decisions in the way I propose to do for imaginings across the board.  Alon makes this case first by arguing that certain imaginings that I might wish to reduce to judgments cannot be judgments because “judgments cannot be non-deliberate.”  Unlike imaginings that may seem to arise spontaneously, judgments “must arise for a reason, in response to evidence.”

This is a good point.  If a would-be imagining is really to be a judgment in disguise, then it must be the conclusion of a bit of reasoning in disguise.  Judgments cannot—or, at least, should not—arise spontaneously, if by “spontaneously” we mean without being based on justified beliefs or good evidence.  On the other hand, if by “spontaneously” we mean only that the judgment arises without our being consciously aware of the reasoning that led to it, then we certainly can have spontaneous imaginings that are judgments.  Because I think that much reasoning occurs below the level of conscious awareness, I am quite open to imaginings being “spontaneous” judgments in this latter sense.  We may indeed “find ourselves” making a judgment without knowing the basis on which it was formed.  This may account for some of the cases Alon has in mind. 

However, Alon gives an example of a spontaneous imagining with an unusual content, where it hard to see how this could be a judgment of any kind. “Suppose I find myself imagining that the cheesecake I just baked will taste like a burger,” he writes.  “If asked whether I think it possible that this cheesecake, or cheesecakes in general, could taste like a burger, I have no immediate answer.”  Because he does not believe the cheeseburger will taste like a burger, nor even have any settled beliefs about whether it is possible for it to do so, it is hard to see what judgment this imagining could possibly be. 

But now the form of objection seems to have little to do with the requirements of judgments.  What is doing the work is the oddity of the content.  For any brief imagining with a very odd content—one that has no clear relation to any of one’s beliefs or desires—it will be difficult to say what sort of more basic mental state it could be.  I do not attempt, in the book, to establish that such imaginings simply cannot occur.  Instead, I focus on contexts where we need to appeal to a person’s imaginings to explain some ability—such as pretense, hypothetical reasoning, or enjoying fiction—and argue that, in each of those instances, we can give a reductive account of the imaginings involved.  In this way, I aim to put pressure on the idea that there are other “spontaneous” imaginings that play no such role but that are nevertheless the appropriate target for a philosophical and scientific theory of imagination. 

I think that most of our spontaneous imaginings and daydreams stem from our desires, our concerns, our plans, our memories, and so on, even if they elaborate on those themes in unrealistic ways.  When we appreciate this psychological background, it typically becomes easy to see them as bits of reasoning about possibilities, or the tokening of occurrent desires, or the making of plans and decisions.  Reductions of the sort I pursue are then available. 

However, this still leaves the door open to cases that, for whatever reason, bear no such relation to that psychological background.  So, what can I say if someone (Alon!) insists: “No, no, there is no background to the bizarre imagining I just enjoyed.  I simply spontaneously and for no reason I am aware of imagined that the cheesecake will taste like a burger.”  My concern with such a claim is the epistemological assumptions it masks.  How do we know when such imaginings occur?  I think we would be overestimating what introspection can reliably tell us were we to say that, through introspection alone, we can discern the type of mental state we are in (viz. a sui generis imagining) and its content, even in cases where the state has no causal bearing on our current perceptions, actions, or inferences.  It assumes that we would know just what an imagining would look like if we saw one (introspectively).  This is one of the common assumptions about imagination I want to resist.  At bottom, this may mark a methodological difference between Alon and myself.  I think that theories of imagination should work their way from the outside in and not be constrained by (putative) data that can only be known from the inside.  This may leave us in a staring contest with respect to the bizarre spontaneous imaginings Alon has in mind.

I move on now to the challenge Alon raises for my account of what occurs when we enjoy a fiction.  The challenge rests on two observations.  First, a fiction may present things as true that it later reverses (e.g., an apparent ally turns out to be an enemy).  Second, on a subsequent reading or viewing, we may respond emotionally to events at the earlier stage of the fiction in ways that would suggest that we don’t yet know what is to come.  (This is how I interpret Alon’s claim that “We may be well aware that, say, a certain fictional character turns out to be nasty, but in rereading the early chapters, which describe him as pleasant and considerate, we have no difficulty imagining him as such.”)  This might seem to be a problem for me, given my claim that the states that generate fiction-directed affect are typically beliefs and desires about what is true in the fiction.  How can I delight in the (merely apparent) friendliness of a character if I believe that, in the fiction, he is a scoundrel? 

My response is that the question Alon is asking might be asked—and has been asked—of our responses to fictions in general.  Slightly rephrased, it is the question behind the famed “paradox” of fiction: why do we ever react emotionally to fictions, given that we know what they present is not really happening?  (And which I address in Chapter 11.)  If we can explain why we ever respond emotionally to fictions despite knowing that events they depict never occurred, we can by the same means explain why we might respond emotionally to fictional events despite knowing that they conflict with what the fiction will present as true later on.  

For instance, suppose that someone asks, “Why are you responding to the fiction with fear and anxiety, given that you know everything turns out fine in the end?”  Here we may say, “Yes, I know how it turns out, but at this point the characters are presented as in danger and have no idea how they will survive.”  Likewise, if someone asks, “Why are you responding this way to the fiction, given that you know none of this is really happening,” we may say, “Yes, I know this is a fiction, but, in the fiction, the characters are in real danger.”  If we feel satisfied with this sort of response in the latter case (as I do), we should feel satisfied in the former.  Further, as I argue in Chapter 11 of the book, saying, instead, that we are imagining the events of a fiction wouldn’t help in either case.  For we often imagine the events of fictions that leave us entirely unmoved.  The key is to explain why we generate desires about courses of events we deem fictional.  If we can answer this, then we can, by the same means, answer Alon’s question about why we form desires about the way a fiction currently presents things.


  1. Alon Chasid

    Peter, thank you for the (wise, as usual) replies. The argument will probably continue, but here is a short question: You classify my question about why we respond emotionally to events we know to be non-fictional–as a sub-question of the more general question regarding why we respond emotionally to events that we know to be unreal. But these questions are different. I think you’ll admit that, in the relevant cases, there must be some correlation, vis-a-vis emotional response, between what happens in the fictional world (‘in the fiction’) and similar real events. All else being equal, you wouldn’t expect someone to fear (or i-fear) of a fictional happy event; rather, you’d expect one to be happy (or i-happy). It is one thing to ask why we fear of fictional events–events we know to be unreal, and another thing to ask about the lack of correlation between the the nature of the fictional event and our expected emotional response. In short, on your reductionist account, if we know that, in the fictional world, the heroine will survive, our emotional response, or concern (i-concern), that she will die is unexplained. It remains unexplained even if we explain the general question of why we emotionally respond to merely-fictional events, since the fact that the heroine dies is not even fictional…

  2. Thanks, Alon. Yes, I see your point. We should expect some strong correlation (or “congruence” in Livingston & Mele’s term) between what happens in the fictional world and one’s emotions. And if we explain this in terms of one’s emotions fitting what is true in the fiction, then we need some other explanation for cases where our emotions don’t fit what is true in the fiction (and instead fit what is non-fictional). So, two separate questions, it seems.

    I think what I want to say here (and which I didn’t say in the book) is that, in all cases—whether we know an event or proposition to be non-fictional or know it to be fictional—we respond emotionally to what we take the fiction to present as true at the stage of the fiction we are encountering. Now the two questions become one again: why do we (ever, and typically) respond in emotionally congruent ways to how a fiction presents things at the state of the fiction we are encountering?

    On the one hand, I think this kind of account makes some sense because, intuitively, we often know that we can’t be sure how things will turn out at the end of the fiction—i.e. what will finally be fictionally true—and so we focus not on what is true in the fiction, full-stop, but instead on the here-and-now of what the fiction presents as true at each stage. So there’s independent reason to see this as the default mode. On the other, I’m a little uneasy with this because I’m not sure how robust the phenomenon is that you mention—of continuing to respond as if p is fictional when we know that not-p is (ultimately) fictional.

    For instance, think of how Jack Nicholson’s character is presented at the beginning of The Shining. At the beginning, he is a reasonably normal family man (if memory serves…), and we respond to him as such. But, once we know how things go later on, will we really respond to him in the same way at the beginning of the film on a subsequent viewing? In short, I think there are cases where knowing what is (really) true in the fiction interferes with or changes our responses on a second viewing. So, perhaps we need some sort of account where there’s a default tendency to respond to how the fiction currently presents things that can be overridden by knowing that things will be presented as quite different later on. You might be able to then press me on why this overriding ever occurs. But that will probably also be a question for standard imagination-based accounts. I’ll have to think more about this! Curious if you have other thoughts.

  3. Alon Chasid

    Even if we don’t always respond emotionally in the second reading (watching etc.) in precisely the same way as we did in the first reading, we sometimes do respond in that, or at least in a similar, way. In principle, knowing how the plot evolves doesn’t necessarily thwart our emotional engagement with the content.
    I think our emotional engagement depends on the degree of immersion. I know you define immersion in terms of emotional engagement, but I disagree (as you perhaps saw in my new Synt paper on immersion). Maybe if you revise your definition of immersion, your account of norms of emotional responses to fiction may solve this problem… I have to think about it though.

Comments are closed.

Back to Top