1

Explaining Imagination

1.1 Introduction

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

Suppose you awoke one day having lost your imagination.
Some things would be easier. There would be no wavering on what clothes to
wear. You wouldn’t be able to imagine the different possibilities. The creativity of
your work might suffer, however. And you would do well to avoid films and
novels with absurd or devastating plot lines. Unable to imagine the events
­
described, you’d have no choice but to corral them, somehow, into your view of
the real.
Games of pretense would come to an end, confounding your partners in cha-
rades. How can you pretend that you’re a bodybuilder, if you can’t imagine being
one? Worse, your empathy would diminish, as you could no longer imagine what
it’s like to be someone else—no longer stand in anyone’s shoes but your own.
Yet few could rival your honesty. When you can’t imagine things being differ-
ent than they are, you won’t conceive of a lie, much less tell one. This would affect
your personal relationships in interesting ways. Even so, you’d have an enviable
peace of mind. What’s there to worry about, when you can’t imagine the future?
This is all to say that I don’t really imagine it would be you who woke up, hav-
ing lost your imagination. Imagination is too central to who and what we are to
remain ourselves without it. There are animals—crickets, crocodiles, crayfish—
that, arguably, cannot imagine. But that’s not us. If we lost our imagination, we
wouldn’t be around to miss it.
The centrality of imagination to who and what we are hints at this book’s main
thesis: when we imagine, we don’t make use of a distinct faculty of mind or col-
lection of sui generis mental states, quarantined from our actual beliefs, desires,
and intentions. “The imagination” is not something that, like sight, or knowledge
of a second language, could be carved off the mind while leaving our self-defining
commitments and inclinations intact. Instead, when we imagine, we make use of
our most basic psychological states in complex bits of reasoning, planning, and
contemplation. Indeed, imagining is nothing over and above the use of such
states—beliefs, desires, and intentions central among them. To see how this can
be so is to arrive at an explanation of imagination in simpler, more general terms.

Explaining Imagination. Peter Langland-Hassan, Oxford University Press (2020). © Peter Langland-Hassan.
DOI: 10.1093/oso/9780198815068.001.0001
2 Explaining Imagination

1.2 What It Is to Imagine

Despite its importance to who and what we are, imagination remains an elusive
explanatory target—“one of the last uncharted terrains of the mind” (Byrne, 2005,
p. xi). Even in broad outlines, it just isn’t clear what imagination is supposed to be.
Describing our plight without it, I relied on an intuitive notion of imagination as
means for thinking about the world being ways it is not—for considering fictions,
possibilities, and fantasies. And I relied on the fact that, for each of the abilities I

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

imagined us losing—be it for hypothetical reasoning, pretense, empathy, or the
enjoyment of fictions—there are philosophers and psychologists who have held
imagination to be its cognitive engine.1
However, characterizing imagination by appeal to the diverse capacities it
en­ables invites the charge that we’ve lumped together a heterogeneous collection
of quite distinct mental states and processes (Kind, 2013). Why think that what
counts as imagining in the context of enjoying a fiction, or considering someone
else’s perspective, is the same mental phenomenon as imagining during a day-
dream, or during hypothetical reasoning? Indeed there are longstanding concerns
that imagination is an ill-defined notion (Moran, 1994, p. 106; Strawson, 1970,
p. 31). Stevenson (2003) counts no fewer than twelve distinct conceptions of im­agin­
ation at work in philosophy. And P. F. Strawson finds the different uses of
“im­agine” to compose a “diverse and scattered family,” where “even this image of a
family seems too definite” (1970, p. 31).
A natural reaction is to draw distinctions. The current landscape is littered
with them: propositional imagination is contrasted to sensory imagination (Stock,
2017), recreative imagining is distinguished from creative imagining (Currie &
Ravenscroft, 2002), sympathetic imagining from perceptual imagining (Nagel,
1974), enactive imagining from suppositional imagining (Goldman, 2006a),
constructive imagining from both attitudinal and imagistic imagining (Van
­
Leeuwen, 2013), imagining “from the inside” from imagining “from the outside”
(Peacocke, 1985; Shoemaker, 1968), imagining proper from supposing and con-
ceiving (Balcerak Jackson, 2016; Chalmers, 2002), and so on. Yet, somehow, the

1 With respect to imagination’s role in pretense, see, e.g., Nichols & Stich (2000) and Carruthers
(2006); for conditional reasoning, see, e.g., Williamson (2007, 2016) and Currie & Ravenscroft (2002);
for the appreciation of fiction, see, e.g., Walton (1990), Currie (1990), and Stock (2017); for third-­
person mindreading, see, e.g., Goldman (2006a, 2006b) and Nichols & Stich (2003). For the claim that
remembering is imagining, see Michaelian (2016a, 2016b). There is controversy surrounding some of
these putative roles, of course. Matravers (2014), for instance, questions whether we need to appeal to
imagination in explaining our responses to fiction; Debus (2014) and Robins (2020) reject the claim
that episodic remembering is imagining; and, in earlier work (Langland-Hassan, 2012), I have argued
that pretense can be explained without invoking imagination. These controversies will be discussed in
due course.
 1.3 Cats and Bats 3

fog surrounding imagination remains equally thick within each of its slices. As
Amy Kind and Peter Kung comment in their introduction to a recent anthology:

The problem is not simply that philosophers give different theoretical treatments
of imagination but rather that there doesn’t even seem to be consensus about
what the phenomenon under discussion is. Among contemporary philosophers
in particular there is a surprising reluctance to offer a substantive characteriza-
tion of imagination; instead, it is understood simply as a mental activity that is

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

perception-like but not quite perception, or belief-like but not quite belief.
(Kind & Kung, 2016, p. 3, emphasis in original)

I take Kind and Kung’s point to be that we are not in the usual situation where
there is a clear phenomenon to be explained—e.g., temperature fluctuation, or
animal reproduction—and a set of competing theories about its nature and
causes. Rather, in the case of imagination, “there doesn’t even seem to be consen-
sus about what the phenomenon under discussion is,” much less agreement con-
cerning its deeper nature. In trying to characterize “the phenomenon” of
imagination, comparisons are made between imagination and states like percep-
tion and belief; but it’s emphasized that imagination remains quite distinct from
those states. Attempts to specify the precise ways in which it is distinct—and to
thereby distinguish what it is we aim to study—threaten to leave us knee-deep in
theory, before we’ve clearly identified what the theory is supposed to be theory of.
Here is how I aim to move forward. I will follow a common practice in drawing
a distinction between two primary senses of the word ‘imagining.’ Then I will give
superficial characterizations of these two kinds of imagining—more superficial
than is normally offered, in fact. Importantly, they will be characterizations that
mesh with our practice of associating imagination with a range of distinct
abil­
­ ities—from pretending, to daydreaming, to counterfactual reasoning, to
engaging with fictions, and being creative—while remaining neutral on questions
concerning its deeper nature. In short, I aim to say clearly what the phenomenon
of imagination is, such that it can be approached from a variety of different
the­or­et­ic­al standpoints. A deeper account will then follow, the outlines of which
are sketched by chapter’s end. The first step, now, is to distinguish two ways in
which a word or concept might be “heterogeneous.”

1.3 Cats and Bats

There are many kinds of cat. The class of cats is heterogeneous, we might say,
insofar as it contains sub-types. There are Siamese cats, Silver Tabby cats, Maine
Coon cats, and more. However, the very notion or concept of a cat is not
4 Explaining Imagination

heterogeneous or equivocal. Nor is the word ‘cat’ ambiguous in its reference. It

always refers to one and the same kind of thing—namely, cats (if we may overlook
all the jazz cats out there).
The situation is different with bats. There are bats used to hit baseballs; and
there are bats that hang upside down in caves. Is the class of bats therefore het­ero­
ge­neous? Not exactly. It is more proper to say that we have two distinct concepts,
each of which corresponds to the same string of English letters. Assuming that
the meaning of a word is one of its essential properties, we also have two distinct

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

words in play: there is ‘bat’ referring to the cave-dwelling creatures; and there is
‘bat’ referring to the ball-hitting implement. These distinct words have the
uncommon feature of being both homonyms and homographs. There is a kind of
heterogeneity here that is different than what we saw with ‘cat.’ It is a het­ero­gen­
eity of concepts corresponding to one and the same string of English letters. Of
course, with respect to each ‘bat’-concept, there is the same kind of heterogeneity
that we saw with the concept of a cat, relating to sub-types. There are both wooden
and aluminum bats for hitting baseballs; and, among the cave-dwellers, there is
the golden-capped fruit bat, the vampire bat, and many others besides.
‘Imagine’ is a lot like ‘bat.’ There is a heterogeneity of concepts corresponding
to a single string of letters. Further, with respect to each concept, there may be a
heterogeneous collection of states and processes that fall within its extension, as
sub-types. I will briefly explain the heterogeneity of concepts now, as a means to
clarifying this book’s proper subject. Controversial elements of this picture will be
flagged, with their proper defense occurring only later, in Chapters 3 and 4.

1.4 Imagistic Imagining and Attitude Imagining

Just as there are distinct concepts corresponding to ‘bat,’ there are at least two
distinct concepts of philosophical interest corresponding to the term ‘imagine.’ I
will refer to them as imagistic imagining and attitude imagining, respectively.2
While I will define them in ways that make them my own terms of art, they align
closely with other conceptions of imagination in the literature (e.g., Van
Leeuwen, 2013; Kind, 2016b).3

2 The one salient sense of ‘imagine’ I will not discuss is the one that means, roughly, “to believe
falsely and without good reason,” as in: “He imagines himself the Canadian Casanova.” This sense
corresponds to definition six for ‘imagine’ in the Oxford English Dictionary: “to form an idea or notion
with regard to something not known with certainty; to believe, fancy ‘take into one’s head’ (that).
Often implying a vague notion not founded on exact observation or reasoning” (Oxford English
Dictionary, 2009).
3 A particularly close fit is Van Leeuwen’s (2013, 2014) distinction between imagistic imagining—
which closely tracks my notion of imagistic imagining—and both his and Kind’s (2016b) “attitudinal
imaginings,” which correspond roughly to my attitude imaginings. However, my understanding of
attitude imagining is importantly different from theirs, in ways I will discuss.
 1.4 Imagistic Imagining and Attitude Imagining 5

At a first pass, imagistic imagining (or “I-imagining”) refers to the use of

endogenously generated mental states that appear image-like, or to have sensory
character, to the people having them. This meshes with the first sense of ‘imagine’
recognized by the Oxford English Dictionary: “to form a mental image of . . . to pic-
ture to oneself (something not present to the senses)” (Oxford English Dictionary,
2009). This sense of ‘imagining’ is at work in the platitude that imagining involves
image or picture-like mental states. We describe ourselves as visualizing, or as see-
ing an image in our mind’s eye, or, yes, as imagining, where the ‘image’ in ‘imagine’

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

is emphasized. Whether the mental states so described really are image-like is a
matter of debate (Block, 1981; Kosslyn, 1994; Pearson & Kosslyn, 2015;
Pylyshyn, 2002); but there is no controversy surrounding the claim that most
people make occasional, or even frequent, use of mental states that seem to them
to be image-like, or to have sensory character—where such states arise not from
an external stimulus impinging on a sense organ but from endogenous causes of
some kind.4 Making use of such mental states is equivalent, in my terms, to using
mental imagery. I-imagining, then, is simply the use of mental imagery in
thought.5 Conjure a mental image, no matter the reason or context, and you are
imagining, in the I-imagining sense. So understood, I-imagining is not a dis­tinct­
ive type of mental process (at least, not obviously) but, rather, any sort of cogni-
tion or mental process at all that involves a mental image. To trigger a memory of
this morning’s breakfast is to engage in I-imagining, provided that the memory
involves a mental image. (Whether episodic memories are imaginings in some
other sense of ‘imagine’ is more controversial—see fn. 1). I will offer a few refine-
ments to this characterization of I-imagining in Chapter 3. For now, this general
definition will suit our needs.
The second sense of ‘imagining’—attitude imagining (henceforth,
“A-imagining”)—has it that imagining is a kind of thought process that allows us
to step outside of what we really believe to consider mere possibilities. It is in this
sense of ‘imagine’ that the things we imagine are, well, imaginary. Here the
emphasis is on the capacity of imaginings to enable rich, elaborated thought
about the possible, fictive, pretended, and fantastical, without any attached stipu-
lation that the thoughts are image or picture-like. Of course, delusions and hallu-
cinations are also rich and elaborated ways of thinking about the possible, fictive,
and unreal—as are sequences of false judgments generally. This highlights another
important aspect of imagining in the attitude sense. A-imagining is a way of
engaging in rich, elaborated cognition about the possible, fantastical, pretended,
and so on, that is epistemically compatible with things not really being the way
they are being thought about, and with one’s not believing things to be that way.
People who are imagining in the A-imagining sense are not epistemically at fault

4 Reports of “aphantasia” are the exception (Zeman, Dewar, & Della Sala, 2016).
5 More thorough characterizations of mental imagery and I-imagining are developed in Chapter 3.
6 Explaining Imagination

or at risk—are not being unreasonable—when they engage in elaborated cognition

about a situation that does not (and never did) obtain, or about an object they
believe does not (and never did) exist.6 A-imaginings allow us to safely step out-
side of what we really believe, without subjecting ourselves to epistemic scrutiny.
The Oxford English Dictionary recognizes several senses of ‘imagine’ that mesh
with this characterization. Its second entry for ‘imagine’ is: “to create as a mental
conception, to conceive; to assume, suppose”; its third: “to conceive in the mind
as a thing to be performed; to devise, plot, plan, compass”; its fourth: “to consider,

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

ponder, meditate” (Oxford English Dictionary, 2009). Note that none of these
­definitions invoke mental images or the act of “picturing to oneself ”; yet all of
them allude to thought about imaginary things. On the face of it, we have two
different notions.
To foreshadow: A-imagining aligns roughly with the idea, common in philoso-
phy, that there is a propositional or cognitive attitude of imagining, or a sui gen-
eris psychological mode of imagining. This is why I have resorted to using the
term ‘attitude imagining’ in naming these imaginings. I aim to be talking about
the same basic phenomenon as these other theorists. However, in my usage, the
notion of an “attitude” does not occur within the definition of A-imagining itself.
This is important. Cognitive (and conative) attitudes are common theoretical
constructs within philosophy. As we will see, in Chapter 2, what it is to bear an
attitude toward a proposition is itself a matter of controversy. When we define
imagination, in one of its primary senses, by appeal to an attitude of some kind—
one with a certain force or “direction of fit” (Searle, 1983)—we have moved into
explaining imagination before we have said what it is we aim to explain.7 We
should instead remain as neutral as we can in our initial characterization of
im­agin­ing. The definition of A-imagining as “rich, elaborated, epistemically safe
thought about the possible, fantastical, unreal, and so on,” aims for that kind of
neutrality. It is akin to characterizing believing as “taking to be true” or desiring
some state of affairs as being impelled to its attainment. Such definitions are con-
sistent with a wide range of more substantive views about the deeper nature of the
states. Yet they have real value. Asked why we find it intuitive to say that such dif-
ferent activities as pretending, reading fiction, reasoning counterfactually, day-
dreaming, and writing a poem all involve imagination, we can respond that they
all involve one’s engaging in rich, elaborated, and epistemically safe thought about
the possible, fantastical, unreal, and so on. Even in light of the diverse contexts in
which imagining occurs, the class of A-imaginings retains a kind of unity.

6 Cf. Currie & Ravenscroft: “Belief, however weakly characterized, is normative in that an agent
who has contradictory beliefs . . . is in a less than ideal epistemic situation. It is no defect in any agent’s
epistemic condition that she imagines things contrary to what she believes” (2002, p. 17).
7 This point is elaborated and substantiated in Chapter 3.
 1.5 The Relation of I-imagining to A-imagining 7

One might worry, however, that this maximally neutral definition of

A-imagining is too broad—that it risks including acts like supposing and conceiv-
ing that, arguably, are not cases of imagining at all. For that matter, it may seem to
include some ordinary acts of reasoning about possibilities that we might not
want to describe as imagining. Yet this broadness is a feature of the definition and
not a bug. The strategy is to begin with a maximally broad characterization and
then, to ensure that imagination proper does not slip through our fingers, tether
our investigation to imagination’s instances in contexts where common sense tells

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

us it typically occurs. Familiar platitudes tell us that people who are pretending are
imagining, that when we daydream we are imagining, that when we consider dif-
ferent possible plans of action, we are imagining different situations, that when we
make up a story, we do so by imagining, that when we enjoy a fiction, we are
imagining the story it tells. And so on. These generalizations and platitudes are
essential guides. Asked for uncontroversial cases of imagining in the “thinking of
imaginary things” sense, this is where we should look. If we can then give an
explanatory account of the kind of thought that occurs in each context, we can
justly claim to have explained imagination, in the A-imagining sense. No harm is
done if our net pulls in and forces us to explain other things as well.

1.5 The Relation of I-imagining to A-imagining

The conceptual distinction between I- and A-imagining brings with it no assump-

tions concerning whether, or to what degree, each notion picks out the same class
of phenomena. This follows from our starting with a theoretically neutral def­in­
ition of each. Some I-imaginings may also be A-imaginings, and vice versa. (Here
we have a difference with bats, insofar as fruit and vampire bats are not, to my
knowledge, used to hit baseballs.) For all I have said, it may be that all I-imaginings
are also A-imaginings and that all A-imaginings are I-imaginings, just as all
renates are cordates and all cordates renates, despite the distinction in the con-
cepts relating to each. If, for instance, all uses of mental imagery are also cases of
rich, elaborated, and epistemically safe thought about the possible, unreal, or fan-
tastical, then we could say that A-imaginings include I-imaginings as a sub-set.
The view I will defend, in Chapter 3, is that the concepts’ extensions only partially
overlap. Some, but not all, I-imaginings are also A-imaginings; and some, but
not all, A-imaginings are also I-imaginings. For the time being, distinguishing
­A- from I-imagining allows me to clarify that my main project in this book is to
explain A-imagining. But it would be wrong to conclude from this that I am not
also explaining paradigmatic cases of I-imagining (or “sensory imagining,” or
“perceptual imagining”) as well. Many I-imaginings are simply A-imaginings that
involve mental imagery; so there can be no comprehensive explanation of
8 Explaining Imagination

A-imagining that is not also, in part, an explanation of I-imagining. In particular,

some have argued that there is a kind of imagining that both involves mental
imagery and constitutes a sui generis mental image-involving process, procedure,
or attitude (see, e.g., Arcangeli, 2019; Kind, 2001). This view is just as much a tar-
get for my reductive approach as is the more general claim that A-imagining
involves use of sui generis imaginative states. These matters are sorted out in more
detail in Chapter 3.

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

1.6 Explaining in What Sense?

There are many kinds of explanation. The sort of explanation I want to offer works
by breaking a complex phenomenon into simpler, more general parts. The parts
are more general in the sense that they are found both within and outside of the
phenomenon they are called on to explain.
An example: the fact that H2O composes water explains many things about water.
Yet it would explain little if we had no understanding of hydrogen and o ­ xygen
outside of their roles in composing water. If, for instance, hydrogen and oxygen were
found nowhere but in water, and if their relations to other molecules were obscure,
then water’s being composed of H2O would explain little. Of course, hydrogen
and oxygen have rich theoretical lives outside of their roles in constituting water.
That is why they can appear in explanations of water and its properties.
This kind of explanation works by unifying. In a classic paper in the philosophy
of science, Philp Kitcher (1981) argues that a theory’s explanatory value lies in its
ability to unify a diverse set of phenomena under a single set of principles or “argu-
ment patterns.” It is explanatory unification, so achieved, that brings understanding.
Scientific theories provide this kind of understanding-through-unification insofar
as they uncover a certain core pattern or style of argument that can be “used in the
derivation of descriptions of many, diverse, phenomena” (Kitcher, 1981, p. 514).
Newtonian mechanics and Darwinian evolutionary theory are two of Kitcher’s
examples. The theories provide understanding of a diverse set of facts by showing
them to be instances of a single core pattern. Understanding is provided even
when the theory offers few means for predicting which specific phenomenon (e.g.,
which specific species, in the case of Darwin’s theory) we should see next. Indeed,
part of Kitcher’s project is to distinguish mere prediction from ex­plan­ation, as
there can be reliable predictive devices—e.g., barometers—that fail to explain what
they predict. Explanatory argument-patterns typically apply a small set of primi-
tive terms—such as ‘force,’ ‘mass,’ and ‘acceleration,’ in the case of Newtonian
mechanics—to a broad set of phenomena. These unifying terms may themselves
be primitive terms, in the sense that they are not unified under yet more general
principles or argument patterns. Explanatory theories simply aim to “reduce, in so
far as possible, the number of types of facts we must accept as brute” (1981, p. 530).
 1.6 Explaining in What Sense? 9

Without any suggestion that the theory developed here has the significance of
Newton’s or Darwin’s, I want to ask how we might break imagination into smaller,
more general parts. How can imagination be unified within a broader framework
for understanding the nature of the mind? We don’t have to endorse Kitcher’s
precise account of scientific explanation to see that, in answering such questions,
we move toward a better understanding of imagination.
One way to provide an explanatory unification of imagination would be to
show how imagining involves the use of particular neural states and processes—

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

neural states and processes that we have some independent understanding of and
that are also used in cognition outside of imagination. A theory that invoked such
neurobiological states and processes might thereby unify imagination with other
modes of thought. However, that is not the approach I will take here. While I
think it holds promise, there is important prior work to be done. To break im­agin­
ation into neural parts at this stage would be getting ahead of ourselves. It would
be like trying to understand the world’s biodiversity by appeal to the molecular
structure of DNA, in advance of evolutionary theory. (My argument for this point
occurs across the entire book to come.)
Another possibility would be to show how imagining can be broken into
smaller cognitive components, where these components are understood as parts
of a functionally specified cognitive architecture. Cognitive architectures—which
seek to detail the actual information-processing algorithms, representational
structures, and data stores exploited in human cognition—abstract away from the
specifics of neural implementation to describe a computational processing archi-
tecture that could, in principle, be instantiated in non-neural cognitive systems.
This “boxological” approach to explaining imagination has proven popular in
recent years (Doggett & Egan, 2007; Nichols, 2006c; Nichols & Stich, 2000;
Schellenberg, 2013; Weinberg & Meskin, 2006b). I’ll have much to say about it in
due course; yet it offers another path I will avoid. In my view, recent attempts to
analyze imagination in boxological terms—as in many of the essays in Nichols
(2006c)—only obscure its nature. My skepticism with respect to those approaches
doesn’t stem from any general misgivings about cognitive science or the practice
of understanding cognition in computational or functional terms. As elaborated
in Chapter 2, my concern is rather that, in the case of imagination, the move to
the level of box-and-arrow diagrams brings with it questionable assumptions
about the representational format of cognitive states, and needlessly forecloses
dialogue with those who don’t share those assumptions.8

8 To endorse boxology in its canonical form (Fodor, 1987; Nichols & Stich, 2000; Quilty-Dunn &
Mandelbaum, 2018) is not equivalent to being a functionalist about mental states; boxology requires,
in addition, the existence of mental representations that are tokened “in” the various boxes that are
posited. A (mere) functionalist about mental states need not be committed to mental representations
at all. These matters are discussed in Chapter 2.
10 Explaining Imagination

What options then remain for breaking imagination into smaller parts? My
approach will be to break imagination into independently understood folk psy­
cho­logic­al states and processes. By “folk psychological states and processes” I
mean the kinds of mental states and processes that ordinary adults know and talk
about. Beliefs, desires, and intentions are examples of folk psychological states—
as are hope, gratitude, and resentment. Folk psychological processes include men-
tal events like thinking, judging, deciding, remembering, and—yes—imagining.
Mental states, capacities, and processes known only to those with a background

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

in empirical psychology or neuroscience—such as working memory, semantic
memory, or feed-forward neural networks—are not folk psychological kinds,
however real they may be. Because folk psychological states like belief, desire, and
intention play a prominent role in our discourse about the mind outside of situ­
ations having anything to do with imagination, they are the right sort of pieces
with which to explain imagination.
These pieces will not themselves receive an explanation or deep analysis here.
The project is instead to reduce questions about A-imagining to questions about
states like beliefs, desires, intentions, decisions, judgments, and the like—mental
kinds that have a life outside of imagination. One folk psychological process,
imagining, will be explained in terms of a collection of others. At the same time, I
won’t propose any adjustments to the most superficial, platitudinous accounts
one might give of these other states and processes. To believe something is to take
it to be true; to desire something is to be impelled to its attainment; to intend
something is to have it in mind as something to be done; and so on. Views about
the deeper nature of these states—concerning, e.g., their representational format
and realization in the brain—are relevant to only some of the specific arguments
in this book. Chapter 2 is an extended meditation on the nature of folk psy­cho­
logic­al states and the question of when and why their format and neural realiza-
tions are relevant to the project of explaining imagination.

1.7 What We Do When We Imagine

The idea that imagination can be reduced to other kinds of folk psychological
states is roundly rejected by most philosophers and psychologists now working
on imagination. In fact, “most” is, to my knowledge, an understatement. I do not
know of anyone else who proposes the sort of reduction I pursue here. In her
introduction to The Routledge Handbook of Philosophy of Imagination, Amy Kind
presents the claim that “imagination is a primitive mental state type (or group of
types), irreducible to other mental state types” as one of “four basic claims about
imagination that enjoy near universal agreement” (2016, p. 2). According to this
consensus, if we listed a person’s current set of beliefs, desires, intentions, judg-
ments, decisions, hopes, wishes, fears, and so on, and fully described their
 1.7 What We Do When We Imagine 11

on­going use of those states in practical and theoretical reasoning, we will have left
open whether they are imagining. For the facts about what, if anything, a person
is imagining are thought not to be entailed by any facts about these other kinds of
folk psychological states one might be in.9 This is the sense in which imagination
is thought to involve an irreducible or “primitive” type of mental state—one’s sui
generis imaginative states, for lack of a better term. I will be in conversation with
this view throughout the book.
My view is that A-imagining is a complex folk psychological process that can

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

be broken down into, and explained in terms of, more basic folk psychological
states and processes. I won’t, however, be arguing that A-imagining is just the
same thing as believing, or judging, or desiring. My argument is not that im­agin­ing
that p is the same thing as believing that p—or even as weakly believing that p. It
is rather that some uses of beliefs, desires, judgments, memories, and so on—none
of which may have the precise content p—constitute cases of imagining that p.
Further, whether they constitute cases of imagining that p will at times turn on
matters extrinsic to the states themselves, such as the reason for which the
­judgments are made, or the social context in which they occur. By loose analogy,
J. L. Austin (1975), in How To Do Things With Words, emphasized that some vocal
utterances constitute acts of naming, dedicating, taking a vow, and so on, depend-
ing on the context of the utterance. Similarly, I’ll argue, some uses of beliefs and
desires, judgments, intentions, and decisions constitute instances of imagining,
and whether they do depends in part upon the context in which they occur.10
To some, the kind of reductive explanation pursued here will seem to involve
an “elimination” of imagination—a kind of denial that imagination really exists.
But that misinterprets my view. Showing that a phenomenon—water, say—can be
explained in more basic terms—molecular composition, say—doesn’t write the
phenomenon out of existence (not on my metaphysics, anyhow). People really do
imagine things, and that ability brings with it the important capacities mentioned
at this chapter’s opening. My project is to explain what we do when we imagine,
not to establish that there is no imagining.
Why has no one else pursued this sort of view? Mustn’t it be crazy, by dint of its
novelty alone? While the view is indeed novel, the approach is not as idiosyn-
cratic as it might at first seem (and perforce not as crazy). The most common
explanatory strategy in the philosophy of imagination over the last twenty years
has been to characterize imagination in terms of its similarities to other folk
psy­cho­logic­al states—imagination being said to be belief-like (Arcangeli, 2018;
Currie & Ravenscroft, 2002; Nichols, 2006a; Van Leeuwen, 2014), desire-like

9 Most will allow that we can sometimes make reasonable inferences as to what, if anything, a
person is imagining, from facts about the other mental states she is in. The point is that, on the ortho-
dox view, facts about what a person is imagining are neither logically entailed nor metaphysically
determined by facts about the other folk psychological states the person is in.
10 Thanks to Amy Kind for suggesting the analogy to Austin.
12 Explaining Imagination

(Currie, 2010; Doggett & Egan, 2007), or perception-like (Currie & Ravenscroft,
2002; Kind, 2001) in its instances. Implicit in these proposals is the thought
that we can gain a better grasp on the nature of imagination by appreciating its
similarities to other, less mysterious folk psychological states—and, indeed, that
we have a good enough idea of what these other states are like to make it worth
highlighting imagination’s similarities to them. I share those guiding assumptions
and thus am engaged in much the same explanatory project—albeit with more
enthusiasm.

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

There are, nevertheless, clear reasons others have not followed me as far as I
wish to go. Some of those reasons can only be explained and addressed over
the course of several chapters. Yet many of what initially seem to be the most
powerful objections to my approach are also the most easily defused. Doing so
will be my project in the balance of this chapter, where my aim is to create some
breathing-room for this book’s larger thesis. I will, in the process, preview some
of the central arguments to come in later chapters. This chapter is, in effect, an
extended trailer for the book as a whole—some spoilers included.

1.8 Simple and Complex Attitudes

To see why a reduction of imagination to more basic folk psychological states

seems so implausible to so many, it helps to contrast cases where such reductions
are more obviously available. Thankfulness, regret, and suspicion are good
ex­amples. Each is a folk psychological state that can be invoked to explain or pre-
dict someone’s behavior. Yet few are inclined to hold that thankfulness, regret, or
suspicion are sui generis folk psychological states. It seems that we can explain
each in more basic folk psychological terms. In that sense, we can say they are
each complex attitudes, insofar as they can be understood as combinations, or
particular types, of more simple attitudes (Schroeder, 2006). For example: I am
thankful that my university gave me a full academic year free of teaching to work
on this book. What does this thankfulness consist in? Here is a sketch: I believe
that I have the full academic year free of teaching; I desire to have the year free
of teaching; I believe that I could have, without injustice, not had the year free of
teaching. I have feelings of relief when I recall that I have the year free of teaching.
And so on. Perhaps I have not completely nailed thankfulness with this charac-
terization. But it seems that we could get there if we really tried. Once we give a
full specification of my current beliefs, desires, and intentions—and, perhaps, my
dispositions to go into certain affective states—we will have captured what it is
that qualifies me as being thankful to have the year free of teaching. The same
goes for regret: to regret that p, I need to believe that p. I probably also need to
desire that not-p and to believe that I could have done something to prevent it
 1.8 Simple and Complex Attitudes 13

from being the case that p. Perhaps, too, I must experience some negative affect
when I recall that p. If something along these lines is correct, regret is a complex
psychological state with these more basic parts. Similarly for suspicion: if I believe
that p with less than full certainty, or if I believe it is somewhat probable that p, it
seems fair to say that I suspect that p. My suspecting that p is nothing over and
above my having such beliefs. If that is right, we need not include suspicion
among the sui generis folk psychological states.
In each of these reductions of complex to simple states, we find a characteristic

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

asymmetry: it is possible to believe that p without being thankful that p and with-
out regretting that p. But, arguably, it is not possible to be thankful that p, or to
regret that p, without believing that p. And it is possible to believe that p without
suspecting that p (as when I am certain that p). But it is not possible to suspect
that p without believing that it is somewhat likely that p. This apparent asym­
metry is essential to belief ’s being more basic than these other mental state kinds.
It also alerts us to a possibility: perhaps being thankful (or regretting) that p is
simply a matter of believing that p with a few accoutrements (including, perhaps,
relevant desires). Likewise, perhaps suspecting that p is nothing other than
believing it is somewhat likely that p. After all, why should believing that p (or
that it is likely that p) be necessary to being in these other states, if these states
were not going to decompose into simpler parts, one of which was belief itself?
Matters are different with imagination and its ilk (viz., conceiving, entertain-
ing, supposing, assuming, considering). Not only is it possible to believe that p
without imagining that p; it is also possible to imagine that p without believing
that p (and without believing that it is somewhat likely that p). Or, at least, so says
common sense—and so I will agree. Similarly, just as one can desire that p with-
out imagining that p, one can also imagine that p without desiring that p. A sim-
ple reduction of imagination to the two most distinguished folk psychological
states appears stopped in its tracks.
It is crucial to see, however, that such observations only stand in the way of the
most simplistic, homogeneous reductions we might pursue. The fact that a person
can imagine that p without believing that p only shows that not every case of
imagining that p is a case of believing that p. But it is quite compatible with some
cases of imagining that p consisting in one’s believing that p. It is also compatible
with all cases of imagining that p consisting in one’s believing something other
than p. The same goes for desire: not every case of imagining that p can be equated
with one’s desiring that p, sure. But this does not, by itself, show that none can. It
is only if one assumes, at the outset, that every instance of a complex attitude
must reduce to simple attitudes in just the same way, if it is to reduce at all, that
the project of reducing imagination is defeated by platitudes such as that we can
imagine that p without believing that p. Fortunately, there is no reason to limit
our investigations with an assumption that any reduction of imagination must be
14 Explaining Imagination

homogeneous in this manner. After all, as we saw at the outset, a starting point
for many theorists is to claim that the term ‘imagining’ picks out a diverse and
“scattered” family of states. Even after we have distinguished two concepts of
imagining—I-imagining and A-imagining—we may find that each concept picks
out a heterogeneous disjunction of different, more basic kinds of folk psy­cho­
logic­al states.
Compare: within philosophy, many apply the phrase “entertaining the prop­os­
ition that p” to any of a heterogeneous set of occurrent mental episodes during

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

which the proposition p is “before the mind.” On this usage, we entertain the
proposition that p when we judge that p; and we also do so when merely wonder-
ing whether p, or when deciding that p. The fact that entertaining the proposition
that p is not strictly the same thing as judging that p does not stand in the way of
reducing entertaining that p (as a mental state type) to a heterogeneous class of
other occurrent states.
“Fine,” comes the response, “but, in the case of imagination, what on earth
could this heterogeneous array of other folk psychological states be?” Well, I will
come to that. That is what this book is about. But let’s not underestimate the
importance of the point just made. It is no barrier to one instance of imagining
being identified with some more basic folk psychological state that the same type
of state cannot be identified with all imaginings. Put otherwise, showing that
imagination is a sui generis mental state kind requires more than establishing that
there is some particular imagining that is not reducible to a specific combination
of more basic folk psychological states. It requires showing that there is no collec-
tion of more basic folk psychological states with which the token imagining can
be identified. And that is not so easily done.
I will aim to make this point more concrete, and more plausible, by applying it
to specific cases below. For now, two summary conclusions to keep in mind:

1. Don’t assume content-mirroring: In order for a token mental state of φ-ing

that p to consist in one’s being in some more basic token state, that more
basic state need not also have the content p. For instance, we saw that sus-
pecting that p is not precisely reducible to believing that p; but it is plausibly
reducible to believing that q, where q is the proposition that it is somewhat
likely that p.
2. Don’t assume homogeneity: An instance of φ-ing that p may consist in one’s
being in some particular set of more basic mental states ∆, even if another
instance of φ-ing that p does not consist in one’s being in ∆. To assume
otherwise is to presume a kind of homogeneity to the class φ-ing that may
not exist. The case of entertaining the proposition that p was offered as an
example of a class of mental states whose instances are heterogeneous. This
possibility is especially salient when theorizing about a kind, such as
im­agin­ation, which even on its face appears heterogeneous to many.
 1.9 What Do I Mean by “ More Basic ” ? 15

1.9 What Do I Mean by “More Basic”?

Before applying these points to specific examples, I’d like to make a last clarification.
I have said that I’ll reduce A-imagining to more basic folk psychological mental
states and processes—and so provide an explanatory unification of im­agin­ation
with those other states. Above, I briefly explained the sense in which belief is
more basic than suspicion in terms of a certain asymmetry: for any situation
where we might attribute a suspicion to a person, we could alternatively, and

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

equally plausibly, attribute a certain belief; by contrast, there are many cases where
we attribute a belief where we could not alternatively, and equally plausibly, attribute
a suspicion. For instance, I believe that I am sitting at my computer, typing. I do
not suspect—or even strongly suspect—that I am sitting at my computer, typing.
Nor do I suspect that my name is ‘Peter,’ that I am a human being, or that I am
thinking now; though I certainly believe those things. Thanks to this asym­metry—
depicted in Fig. 1.1a—we can say that belief is more basic than suspicion.
Because, in my view, imagination reduces to a heterogeneous set of folk psy­
cho­logic­al states, the sense in which these states are more basic is not as straight-
forward as with suspicion and belief. It is not the case that, for any situation
where we attribute to a person an imagining that p, we could alternatively, and
equally plausibly, attribute a belief with a certain content. I will argue instead
that, for any situation where we attribute an imagining to someone, we could
alternatively, and equally plausibly, attribute either a belief, desire, or intention—
or one of their occurrent counterparts (viz., a judgment, desire, or decision).
Further, for each of belief, desire, and intention, there are many contexts where
we attribute one of those states where we could not alternatively, and equally
plausibly, attribute an imagining. A second diagram—Fig. 1.1b—helps to clarify
the relationships I have in mind.
We can interpret each rectangle in Fig. 1.1b as containing the set of situations
where we can plausibly—and, for explanatory purposes, profitably—attribute to
someone a belief, desire, or intention (or one of their occurrent counterparts).
Likewise for the circle, with respect to imaginings. If the diagram is correct, then
belief, desire, and intention are collectively more basic than imagining. This is
similar to the way in which belief, desire, and certain basic emotions are collectively
more basic than regret or thankfulness, as earlier discussed. A difference, however,
is that the reduction of thankfulness and regret to this more basic collection of
states is relatively homogeneous: every case of being thankful that p is identifiable
with the same kind of collection of beliefs, desires, and emotions (or so I suggested).
In the case of imagination, there is not a single reductive re­cipe of this kind, even
if, on a case-by-case basis, each imagining is identifiable with some collection of
more basic states.
The most obvious way to object to this picture is to hold that there are cases of
imagining that do not fully overlap with any combination of the other three
16 Explaining Imagination

(a)

BELIEF

SUSPICION

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

(b)
BELIEF DESIRE

IMAGINING

INTENTION

Fig. 1.1a The relationship of belief to suspicion

Fig. 1.1b The relationships among belief, desire, intention, and imagining. Note that
the rectangles are intended to include both dispositional and occurrent versions of
these states. Thus, the ‘belief ’ rectangle includes judgments and the intention
rectangle includes decisions.

boxes. Here I want to mention two subtler forms of objection, rooted in concerns
about the notion of basicness at work in my accounts of reduction and ex­plan­
ation. First, some may grant that belief, desire, and intention are collectively more
basic than imagining in the way I have suggested, yet deny that this shows im­agin­
ing to be reducible to—or even explainable in terms of—those states. Second,
some may object that the relations mapped in Fig. 1.1b are misleading, because
the same kind of reduction I propose for imagination (in terms of three other,
more basic states) is possible for one or more of the other states as well. For
instance, they may propose that every context where we ascribe a desire is one
where we could alternatively, and equally plausibly, ascribe a belief, intention, or
imagining. If that were correct, then there would be no sense in which desire is
more basic than imagining; the two would be on a par, with each notion or state
being analyzable into three others. Because these more subtle worries cannot be
adequately addressed without first distinguishing different views one might have
on the ontology of folk psychological states generally, I table their discussion to
Chapter 2 (sections 2.6 and 2.7).
 1.10 The Delicious Mud Pie 17

1.10 The Delicious Mud Pie

I turn now to applying the above points to explicit arguments commonly given
for viewing imagination as irreducible to other folk psychological states. These
arguments appeal to broad generalizations in how imaginings differ from states
like belief or desire. Echoing claims endorsed by many others (including
Sinhababu (2016), Nichols (2006a), Stock (2017), and Picciuto & Carruthers
(2016, pp. 316–17)), Shannon Spaulding (2015) catalogs a number of differences

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

between imagination and belief, as a means to establishing that imagination is
“not reducible to belief.” While the point is tangential to her main argument in
the paper, her list is a helpful compendium of common objections that need
addressing by an account like mine. Noting that imagination bears important
similarities to belief, Spaulding emphasizes that:

Imagination is not reducible to belief. Imagination guides action differently than

belief. Imagining that a mud pie is a delicious treat guides my action differently
than believing it is. Imagination is subject to conscious, voluntary control,
whereas belief is not. Imagination is less restrictive than belief insofar as one can
imagine many false and absurd propositions that one in no way believes.
Imagination-induced affect typically is less intense, less durable, and sometimes
quite different than belief-induced affect. The upshot of these considerations is
that imagination . . . is not reducible to belief. (Spaulding, 2015, pp. 459–60)

I will consider each point on Spaulding’s list in turn. But, first, let’s reflect on the
structure of the argument and its aims. In order for observations of this sort to
count as evidence for imagination’s irreducibility, they cannot assume what is in
question. If imagination just is a species of belief—or even if only some im­agin­
ings are beliefs—then imagination does not guide action differently than belief, is
not related differently to the will, is not less restrictive than belief, and does not
trigger affect differently (at least, not always). In short, it begs the question against
the advocate for imagination’s reducibility to offer such differences as evidence for
its irreducibility. And yet, the platitudes Spaulding lists are hard to deny; indeed, I
don’t deny them, properly understood. So we need a way to take this putative
evidence for imagination’s irreducibility on board without assuming what is
at issue.
Here is how: the generalizations Spaulding lists can be neutrally characterized
as differences in believing that p and imagining that p, respectively. A person who
imagines that p will behave differently than a person who believes that p; a person
can imagine that p at will but cannot believe that p at will; a person can imagine
that p while believing that not-p; the affect a person experiences in response
to imagining that p is usually different than what she will experience in response to
18 Explaining Imagination

believing that p. These differences between the person who believes that p and the
person who merely imagines that p can then be considered evidence for the claim
that imagining is irreducible to belief tout court—without, in fact, assuming
the point.
This is not particularly good evidence for imagination’s irreducibility, however.
It leaves open the possibility that many, even all, cases of imagining that p will
reduce to beliefs of some kind or other—just not the belief that p. It also leaves
open the possibility that all cases of imagining will be reducible to some collec-

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

tion or other of more basic folk psychological states, many of which are beliefs.
To see concrete possibilities of this sort, let’s look at each putative difference
in turn.

1.10.1 Imagination and Action

Spaulding observes that imagination “guides action differently than belief.” A

well-worn example, tracing to Walton (1990), is that “imagining that a mud pie is
a delicious treat guides my action differently than believing it is” (Spaulding, 2015).
Sinhababu also highlights this difference in action-tendencies when arguing that
imagination is a distinct state of mind from belief, observing that “daydreaming
about being Spider-Man typically doesn’t result in actually trying to shoot webs,
and imagining that one is Harry Potter while reading of his adventures doesn’t
usually result in trying to cast spells” (2016, p. 113). Nichols concurs that it is a
“central fact about the propositional imagination” that “imagination and belief
generate different action tendencies” (2006, pp. 6–7). I have noted that, for these
observations to not beg the question with respect to imagination’s reducibility, we
need to view them as assertions about the different behavioral dispositions asso-
ciated with imagining that p and believing that p, respectively. Granting those
differences, the question now before us is whether the dispositions to action asso-
ciated with imagining that p can nevertheless be ascribed through the use of
other, more basic folk psychological terms—terms other than “believing that p.”
Enter Uncle Joe, who believes that he is playing a pretense game with his
nephew, where the point is to act like a mud pie is a chocolate pie. He judges, and
thus comes to believe, that holding the pie to his face while saying “Mmm, tasty!”
is a good way to act like the mud pie is a chocolate pie. This is, after all, how one
might behave around a real chocolate pie. Given that he has these beliefs, and
wants to continue playing this game, how is he disposed to act? It seems to me
that he is disposed to act exactly like someone who is imagining that a mud pie is
a delicious treat (and who has a desire to play the game). After all, he’s not eating
the mud pie; the disposition to do that only holds of people who believe the mud
pie is delicious. He is stopping short of doing anything that would put him at
 1.10 The Delicious Mud Pie 19

digestive risk. He is only doing things you would do if you were, well, imagining
that the pie is a delicious treat and wanted to play along with your nephew. This is
a case where imagining that a mud pie is a chocolate pie generates the same action
tendencies as an ordinary judgment—not the specific judgment that a mud pie is
a chocolate pie, of course, but judgments about how to act like a mud pie is deli-
cious. (I will take judgments to be occurrent mental processes through which one
comes to have a certain dispositional belief.) So, in this case, the mere fact that
imagining that p has different associated action-tendencies from believing that p

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

gives no reason to think that imagining is generally irreducible to other, more
basic mental states. For it turns out there is another collection of beliefs and
desires that does give rise to the same dispositions to action as imagining that p.11
One response in favor of imagination’s irreducibility might be that a sui generis
imaginative state is what enables a person to generate the beliefs and judgments
just mentioned. For instance, it might be thought that Uncle Joe needs to have a
belief-distinct (imaginative) mental representation with the content “this is a
chocolate pie,” in order to see what the appropriate actions would be if the mud
pie really were delicious. (This idea mirrors Nichols & Stich’s (2003) and Currie &
Ravenscroft’s (2002) account of the role of imagination in pretense and hypo­thet­
ic­al reasoning.) This is indeed a possibility for how we arrive at such judgments;
it’s one I reject in Chapters 5 and 6, on conditional reasoning. For the time being,
note that the debate has now shifted to whether having and acquiring certain
beliefs is best explained by one’s having a particular sort of mental representation
that is not itself a belief. Gone is the platitudinous, undeniable claim with which
we began—that imagining that p guides action differently than believing or judg-
ing that p. It has been replaced with a more controversial proposal about the men-
tal states necessary for arriving at certain other beliefs. If we accept that claim, it
must be for reasons other than that we accept the platitude that imagination
guides action differently than belief. The platitude on its own—rendered neutrally
as the claim that imagining that p and believing that p have different associated
behavioral dispositions—is little evidence for imagination’s irreducibility.
Another objection may be to grant that in this special case the dispositions to
action associated with imagining that p are the same as those associated with hav-
ing certain beliefs, desires, or making certain judgements, while objecting that
there are many other cases where no such translation will be available. Again, this
might be so. But there is nothing special about the case just considered. The mud
pie pretense is a standard example, handed down through the generations. More

11 One might object that the present example is simply one where a pretense does not involve
imagination and where we would normally ascribe an imagining. Granting the possibility, I am only,
at this point, explaining a general strategy—one that is applied to a full variety of pretenses in
Chapters 7 and 8.
20 Explaining Imagination

importantly, the objection again shifts the argument for imagination’s irreducibility
from “imagination guides action differently than belief ” to “there are at least
some cases where the behavioral dispositions we ascribe by saying someone is
imagining that p cannot equally well be ascribed through any other plausible col-
lection of beliefs, desires, judgments, intentions, and so on.” Once the objection is
put this way, it hardly seems obvious. Its truth, or falsity, will be a delicate matter.

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

1.10.2 Imagination and the Will

The next reason Spaulding gives for thinking that imagination is irreducible to
belief is that “imagination is subject to conscious, voluntary control, whereas
belief is not.” Sinhababu again agrees, noting that “it’s easy to perform an inten-
tional action of imagining something that isn’t the case. It’s hard or impossible to
perform an intentional action of believing something that isn’t the case” (2016,
p. 113). Nichols is also on board: “belief is not at the whim of our intentions,” he
observes, “but imagination is” (2006, p. 7).
In order to view these claims as evidence for imagination’s irreducibility, and
not mere assertions of it, we should again see them as noting a difference in
im­agin­ing that p and believing that p, respectively. There are cases where we say a
person has freely imagined that p where we would not say he could have freely
judged or come to believe that p. The truth of this platitude, however, does not
offer much reason for thinking that imagination is irreducible to other folk psy­
cho­logic­al states.
After all, this special freedom of imagination is fully evident in the mud pie
example. There’s Uncle Joe again, holding the mud pie to his face: “Mmm,” he
says, “delicious!” He is imagining that the mud pie is a delicious treat. That is what
we are inclined to say as we watch. This game involving the mud pie, and the
imagining that supports it, are things he does voluntarily. No one put a gun to
Uncle Joe’s head. Of course, he didn’t—and can’t—choose to believe that the mud
pie is a delicious treat. But that is irrelevant. For while we can’t choose our beliefs,
judgments, or parents, we can choose the topics on which we’d like to reason. And
that’s exactly what Uncle Joe has done. In choosing to imagine that the mud pie is
a chocolate pie, he has chosen to reason on the topic of how to act like a mud pie
is a chocolate pie; and he has judged that holding it to his face while saying
“Mmmmm . . . delicious” is a good way to do so. He could have instead chosen to
reason about how to act like the pie is a Frisbee, or how to act like a catfish. He
was free to do so. Had he so chosen and put the reasoning to use in related games
of pretense—arriving at judgments about how to make the pie Frisbee-like, or
how to make himself catfish-like—we would have declared him to be imagining
that the pie is a Frisbee, or that he is a catfish. Such freedom, genuine as it is,
 1.10 The Delicious Mud Pie 21

offers no reason to think that imagination is irreducible to collections of beliefs,

desires, and judgments.
Again one may object that, in order to make the judgments in question (about,
e.g., how to act like the mud pie is a delicious treat), Uncle Joe must (voluntarily)
token a sui generis imaginative mental representation with the content “The mud
pie is a delicious treat,” or “The mud pie is a Frisbee,” or “I am a catfish.” And,
again: maybe that is required. But here is another possibility: to voluntarily pre-
tend that some X (e.g. a mud pie, or Uncle Joe) is a Y (e.g. a delicious treat, or a

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

catfish), you simply need some beliefs about what Ys are generally like and a
desire to make X saliently Y-like (Langland-Hassan, 2014b). Uncle Joe knows
some things about catfish: they’ve got whiskers, they’re feisty, they make barking
sounds when out of water. During a pretense, he can draw on this knowledge to
make himself catfish-like in various respects, without ever thinking something he
disbelieves (such as: I, Uncle Joe, am a catfish). This is a possibility that must be
foreclosed if the voluntary nature of imagination is to offer reason for thinking
that imagining is irreducible to other folk psychological states.
Again we may hear the objection that the mud pie example is a special case and
that there are very many freely chosen imaginings that will not fit this explana-
tory mould. Two points in response: first, in line with the possible heterogeneity
of A-imagining, my claim is not that the freedom of all imaginings is to be
explained as a freedom to engage in reasoning on a topic of one’s choice; other
cases, such as idle daydreams, may be explained in other ways. (More on this in a
moment.) Second, with this objection the argument has again changed shape. It is
no longer: “imagination is subject to conscious, voluntary control, whereas belief
is not,” but rather: “there are at least some cases in which we ascribe a freely
­chosen imagining where we could not have alternatively, and equally plausibly,
ascribed any other collection of more basic folk psychological states.” That is an
interesting claim, but not an obvious one. We’ll just have to see whether it’s true,
by examining—in Chapters 5 through 12—a wide variety of paradigmatic con-
texts where A-imagining occurs.

1.10.3 Imagining What We Disbelieve

Similar points apply to Spaulding’s observation that “imagination is less re­strict­

ive than belief insofar as one can imagine many false and absurd propositions that
one in no way believes.” Schroeder & Matheson give cognitive-scientific dress to
this platitude: “Imagining that p is most obviously distinguished from believing
that p, in that imagining that p does not lead one to automatically store in one’s
memory that p” (2006, p. 25). (Sinhababu (2016, p. 112) and Nichols (2006a, p. 6)
echo this claim, noting that we can imagine that p while not believing that p.)
22 Explaining Imagination

However, on the account I have suggested, it remains correct to describe Uncle

Joe as having imagined a false and absurd proposition—that the mud pie is a deli-
cious treat—that he doesn’t believe. Yet, in line with points already made, we
could have alternatively ascribed the particular judgments and desires mentioned
above. Thus, when asked what it is for him to imagine the (disbelieved) prop­os­
ition that the mud pie is delicious, we can say that it amounts to his having made
those judgments and having had those desires—just as we can say that someone’s
regretting that p consists in his having certain beliefs, desires, and emotional dis-

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

positions. The truth of the platitude that we can imagine what we don’t believe
remains compatible with an imagining’s being reducible to other kinds of mental
states—beliefs included.

1.10.4 Imagination and Emotion

The last consideration Spaulding gives for thinking that imagination is irreducible
to belief is that “imagination-induced affect typically is less intense, less durable,
and sometimes quite different than belief-induced affect.” Nichols finds it “com-
mon wisdom in psychology that imagining scenarios can have significant af­fect­
ive consequences.” While he is more impressed with the similarity of the emotions
felt in response to imagining that p and believing that p than their differences
(2006a, p. 8), Nichols agrees that emotional responses to an imagining are often quite
different than they are when we believe the same content (2006b). A comparable
connection between imagination and affect—where imagination has some, but
not all, of the same relationships to affect as belief—is proposed by many o ­ thers
(see, e.g., Schroeder & Matheson, 2006; Meskin & Weinberg, 2003).
We can again grant the superficial, platitudinous phenomenon: when we say
that someone has imagined that p, we don’t expect them to experience the depth
of emotions they would if they were to have judged that p. But neither are we
surprised if they feel some semblance of those emotions. Admittedly, it is not
obvious what emotions, if any, Uncle Joe experiences in imagining that the mud
pie is a delicious treat. But consider a different example: imagining that your fam-
ily is inside a burning house. Imagining this may cause unpleasant affect. This
affect would be much different, however, were you to judge that your family is
inside a burning house. Yet this is no reason to think that imagination is ir­re­du­
cible to more basic folk psychological states. Imagining that your family is inside
a burning house could very well have the same emotional impact as some other
related judgment—such as the judgment that your family could someday be
caught inside a burning house and that, if they were, terrifying events would
unfold. Although it is only a judgment about what could happen, dwelling on the
possibility may be enough to raise a lump in your throat. Once again, so long as
 1.11 Introspection and Mental Imagery 23

there is some judgment, desire, intention, or decision with the same associated
dispositions to generate affect as the imagining, we can make a case for identify-
ing the imagining with those more basic states. The mere platitude that imagining
that p has different associated emotional dispositions than believing that p does
little, by itself, to establish imagination’s general irreducibility.
So much for the most common reasons given for thinking that imagination is
irreducible to more basic folk psychological states. I want now to consider a
slightly different form of objection—one grounded in introspection.

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

1.11 Introspection and Mental Imagery

It might seem obvious that we know, just through introspection, that the states we
enter into when imagining are not occurrent judgments, desires, or intentions of
the kind I have so far proposed. Just as a matter of first-person phenomenology, it
might seem clear that (1) we know when we are imagining and that (2) we can tell
that our states of imagining are not some other kind of state—judgments, beliefs,
desires, intentions, or whatever. For instance, you may find that you are now
imagining a ninja eating popcorn and that this episode is no belief, desire, or
intention. What do I say to this?
My response is that the argument is question-begging. If you are indeed aware
of an imagining that is not any other kind of state, then, sure, that imagining is
not any other kind of state. The question is why we should think a person is well
placed to introspectively discern that an occurrent mental episode is an im­agin­
ing and nothing else. I see no reason to think people are authorities on this. After
all, if it were obvious to introspection that we can discern what is an imagining
and what is not, there would be no need to consider the arguments made by
Spaulding, Sinhababu, Nichols, and others in favor of imagination’s irreducibility.
We could simply recede into the comfort of our own minds, notice that our
im­agin­ings are one thing, our beliefs, desires, and intentions another, and move
on with our lives. Some readers may have done just that. But if you have taken the
trouble to follow the argument this far, you probably agree that we have no such
ability; arguments of a different kind will be needed to determine which mental
states are basic, and which are not.
It may help to observe that in many ordinary cases of folk psychological
ex­plan­ation, we don’t expect the attributions to tell us much about the person’s
conscious life. When Andrew breaks away from the televised soccer game to grab
a beer from the fridge, we explain it by saying he desired a beer and believed that
there is beer in the fridge (such is the classic example). But we don’t thereby
assume that he said to himself “there is beer in the fridge” or “a beer would be
great right now,” or that he consciously reflected on the question “where should
24 Explaining Imagination

I go, if I want a beer?” before doing so. We don’t expect the belief/desire
­attributions to have obvious phenomenological implications. The same goes for
the attributions beliefs and desires relevant to explaining Uncle Joe’s pretense
behavior, which, on reflection, we may identify with his token A-imagining.
One place where introspection does seem to get a grip, however, is with respect
to mental imagery—especially as I have defined it. For we certainly can tell, intro-
spectively, whether we are currently making use of mental states that seem to us to
be image-like in nature. A separate, empirical question—the subject of historical

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

debate (Block, 1981; Pearson & Kosslyn, 2015; Pylyshyn, 2002; Tye, 1991)—is
whether the representational format of these putative images is indeed picture-
like in some important respect. Let’s assume, for the moment, that the empirical
question is settled: the mental imagery we are aware of through introspection
does indeed occur in a pictorial, iconic, or analogue format. A common, but by
no means universal, view in philosophy is that “propositional” folk psychological
states like beliefs, desires, intentions, and so on, do not occur in an imagistic,
­picture-like format (see, e.g., Fodor (1975)). If that view were correct, and if it
were indeed the case that mental imagery occurs in a picture-like format, then we
could tell, just through introspection, that one of our current mental states was
not what I have termed a “basic folk psychological state” (viz, a belief, desire, or
intention) just by noticing that it involved a mental image. And, of course, in
many of the paradigmatic situations where A-imagining occurs we do find our-
selves making use of mental imagery. This would entail that at least some
A-imaginings (i.e., those involving mental imagery) are not reducible to more
basic folk psychological states.
My response is to deny the thesis that mental images never form proper parts
of “propositional” folk psychological states like beliefs, desires, intentions, and so
on. (Nor am I alone in rejecting this thesis (see, e.g., Kaplan, 1968; Kind, 2001;
Martin, 2002; Van Leeuwen, 2013).) The mental images we introspectively dis-
cern may instead be proper parts of more basic (propositional) folk psychological
states, such as judgments, desires, and decisions (Langland-Hassan, 2015). This
can be true whether or not such images actually occur in an imagistic format—a
point I will explain and defend in Chapters 3 and 4. For now, the shape of my
response to the objection from introspective-awareness-of-imagery should be
clear: given that many basic folk psychological states—including beliefs, desires,
judgments, and decisions—have mental images as proper parts, introspective
awareness of a mental image cannot serve as evidence that we are in some state
that is irreducible to those folk psychological states.

1.12 More Case Studies as Prelude

I will conclude this bird’s-eye view of the book to come with a few more case
studies in reducing A-imagining to more basic folk psychological states. The aim
 1.12 More Case Studies as Prelude 25

of these examples is not to convince one of the overall account but to advertise
the shape of things to come.

1.12.1 Daydreaming: Imagining that I Am Rich and Famous

Kendall Walton offers the following as a “paradigm instance of an exercise of the

imagination”:

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

Fred finds himself, in an idle moment, alone with his thoughts. Feeling unsuc-
cessful and unappreciated, he embarks on a daydream in which he is rich and
famous. He calls up images of applauding constituents, visiting dignitaries, a
huge mansion, doting women, fancy cars. But alas, reality eventually reasserts
itself and Fred gets back to selling shoes. (Walton, 1990, p. 13)

The orthodox view of imagination has it that Fred makes use of a sui generis
imaginative state in the course of this daydream. In the face of cases like these, we
have to ask if there are no other, more basic kinds of folk psychological states at
work in disguise. Clearly, Fred wants to take leave of his position at the shoe store.
He wants to be applauded by constituents, visited by dignitaries, housed in a
mansion, pursued by women, driving Lamborghinis. Fred has many unfulfilled
desires flooding his mind as customers wait for him to deliver their loafers in the
correct size. We could describe it as his imagining these things—these objects of
desire. But it would be more perspicuous call it what it is: the conscious uprising
of Fred’s outlandish desires. Some of these desires may have mental images—of
cars, of women, of adoring fans—as proper parts. But they are desires all the same.
Granted, even if some of what get called “daydreams” are simply occurrent
desires, not all of them are. In some cases, we simply tell ourselves a story; in
others, we confront our fears. These and other examples of daydreams are
­
addressed in Chapter 4, on imagistic imagining, and Chapter 12, on the role of
imagination in creativity.

1.12.2 Pretense—a Sketch of Chapters 7 and 8

Bananas not only dominate sales of produce. They are also ubiquitous in discus-
sions of pretense (Friedman & Leslie, 2007; Leslie, 1987; Nichols & Stich, 2000;
Richert & Lillard, 2004). A weathered example is someone holding a banana to
his ear and speaking into one end, pretending that it is a telephone. When we
look under the hood, what, psychologically, does this little caper require?
Inspecting the banana, a man—Carl, let’s say—judges it to be shaped like the
receiver of an (old-fashioned) telephone. He wants to have some fun, to play a
little game. So he decides to treat the banana in telephone-like ways, holding one
26 Explaining Imagination

end to his ear, talking into the other. Carl is able to do this—to temporarily make
the banana telephone-like in various respects—because he knows some things
about telephones. Of course, he does not believe that the banana is a telephone.
That’s why we describe him as only pretending. We might also say that he is
“imagining that the banana is a telephone.” But it might offer a clearer view of his
mind to simply say that he believes that the banana resembles a telephone
receiver, wants to play a little game, and, in order to do so, has decided to make
the banana telephone-like in various respects, while believing it is not, in fact, a

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

telephone.
True enough, this is not a template for explaining all pretenses. Some pretenses
require us to reason hypothetically about what would be the case in some possible
situation or other. Pretending that an airplane engine has landed in my backyard,
for example, might require me to form some judgments about what would hap-
pen if an airplane engine landed in my backyard. Those if-then conditional judg-
ments could then guide my pretense. This raises a question: does evaluating and
making judgments about conditionals require sui generis imaginative states?

1.12.3 Conditional Reasoning—a Sketch of Chapters 5 and 6

We have many beliefs of the form: if p then q.12 How do we arrive at these beliefs,
in cases where we don’t already believe that p? A popular proposal is that we
imagine that p and, with p fixed in imagination, observe what else emerges as
likely in imagination (Nichols & Stich, 2000; Currie & Ravenscroft, 2002;
Williamson, 2016). If q is one of those things, we will then come to believe that if
p then q. Imagining that p enables us to infer the likely consequences of p being
the case, it is said, just because imagination is “belief-like” in its inferential prop-
erties (Nichols & Stich, 2000; Currie & Ravenscroft, 2002).
Here is a different approach. I judge that, if my coffee cup turns over, then the
coffee will spill out. How did I arrive at this belief? I don’t believe that my coffee
cup is now turned over, or that the coffee has spilled out, after all. However, I have
lots of relevant background beliefs about liquids and spills. In particular, I believe
that, ceteris paribus, when a container holding a liquid is knocked over, the liquid
spills out. Asked what would happen if my cup turns over, I access that belief and
infer straightaway: If my coffee cup turns over, then the coffee will spill out. This
is not a deductive inference, of course. Such is the case with most of our condi-
tional judgments. They are based on past experience and our knowledge of how

12 In this chapter, I don’t distinguish between subjunctive and indicative conditionals, though that
distinction is important and prominent in the full discussion of conditional reasoning in
Chapters 5 and 6.
 1.12 More Case Studies as Prelude 27

things normally go. We can make use of that knowledge to infer conditionals,
without ever representing to ourselves something we disbelieve.
In some cases, it may seem there are no relevant past experiences, no relevant
“way things normally go,” to be used in arriving at the new conditional belief. This
may seem to be the case with the airplane engine landing in my backyard. One
question is whether these appearances are correct. Are there really no generaliza-
tions or past experiences on the basis of which I can infer what would happen if
an airplane engine landed in my backyard? Another question is whether sui gen-

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

eris imaginings would offer any help to the inference if there are not (Langland-
Hassan, 2012). The surrounding issues—concerning the truth-conditions of both
indicative and subjunctive conditionals, their relation to the material conditional
of formal logic, and the psychological processes by which we arrive at our beliefs
in each kind of conditional—are complex. I explore them across Chapters 5 and 6,
arguing that we gain no traction on the psychology conditional reasoning by
invoking sui generis imaginings. We can better explain the key inferences at work
by appeal to beliefs alone.

1.12.4 Consuming Fiction: The Barest Sketch

Imagination is often cited in philosophical discussions of fiction (Currie, 1990;

Friend, 2008; Matravers, 2014; Meskin & Weinberg, 2003; Nichols, 2004a;
Stock, 2017; Walton, 1990; Weinberg & Meskin, 2006b). Cindy, let us suppose, is
watching the Steven Spielberg classic, E.T. E.T. levitates Elliot’s dirt bike on their
way to meet the mothership that will return E.T. to his home planet. They are
­silhouetted by the moon. As Cindy watches, we can say she is imagining that
E.T. is going home. She doesn’t really believe that E.T. is going home, after all.
Alternatively, if we are pursuing an explanatory reduction, we could say that she
is judging that, in the film E.T., E.T. is going home. This is something she believes.
This judgment leads her into a certain emotional state—a state of wistfulness. For
she wanted it to be true, in E.T., that E.T. goes home; but she also wanted it to be
true, in E.T., that E.T. and Elliot remain close friends on Earth. Her wistfulness
makes sense, given her beliefs and conflicting desires. We get a clear picture of her
overall cognitive situation if we identify her episode of imagining with these states.
But why does Cindy care about what is happening in a mere fiction? Why
should beliefs about what is happening in a fiction generate any affect at all? These
questions lie behind the well-known “paradox of fiction” in aesthetics
(Friend, 2016; Lamarque, 1981b; Radford, 1975). I won’t venture a summary reply
here; Chapter 11 is devoted to the topic and develops a novel response. Other
puzzles relating to fiction-consumption and imagining—including how we
extract implicit truths from a fiction and how we are to define fictional truth
28 Explaining Imagination

itself—are addressed across Chapters 9, 10, and 11, with Chapter 10 generating a
special challenge for any view at all that tries to find work for sui generis im­agin­
ings in fiction consumption.

1.13 Summary

In my experience, differences in the general platitudes surrounding imagination

Downloaded from https://academic.oup.com/book/36668/chapter/321687586 by guest on 29 November 2023

and belief—that we can imagine that p without believing that p, and so on—are
what underlie the seeming obviousness of the view that imagination is irreducible
to belief (or indeed to any more basic set of folk psychological states). It makes
sense that this would be so. If it is indeed obvious that imagination is irreducible
to other kinds of folk psychological states, its obviousness should lie on the sur-
face. It is that superficial obviousness that I have tried to chip away at here. I hope
that imagination’s reducibility to other folk psychological states now seems an
open, even delicate question.
In the chapters to come, my strategy for explaining imagination will be to iden-
tify contexts and abilities commonly agreed to involve imagination (in the
A-imagination sense) and to show how the mental states and processes at work in
those contexts can be understood as more basic folk psychological states. These
abilities include conditional reasoning (Chapters 5 and 6), pretending (Chapters 7
and 8), engaging with fictions (Chapters 9, 10, and 11), and creativity (Chapter 12).
In each case, my aim is to tell a how plausibly story where the cognition we associ-
ate with imagination is composed of other, more basic folk psychological states.
At the same time, I aim to cast independent doubts on explanations of these abil­
ities that have appealed to sui generis imaginative states.
By the end of Chapter 12, there is no general reductive definition of im­agin­
ation offered—no identification of all A-imaginings with certain specific kinds of
other states or patterns of inference. Instead, there is a collection of strategies for
showing how paradigmatic contexts where imagination occurs can be understood
as exclusively drawing upon a more basic collection of mental states, including
beliefs, desires, and intentions. This is the right form of reduction, in my view,
given that imagination is not a natural cognitive kind but is instead a het­ero­ge­
neous collection of more basic mental states and processes that acquire the label
‘imagining’ on the basis of being cases of rich, elaborated, epistemically safe
thought about the possible, fictional, and fantastical.