Professional Documents
Culture Documents
by Transaction Books
ISBN 978-3-86838-174-0
9 783868 381740
PhR 9
Fabian Dorsch
verlag
ontos
Fabian Dorsch
Philosophische Forschung
Philosophical Research
Fabian Dorsch
The Unity of Imagining
Philosophische Forschung
Philosophical Research
Herausgegeben von / Edited by
Johannes Brandl Andreas Kemmerling
Wolfgang Knne Mark Textor
Band 9 / Volume 9
Fabian Dorsch
Table of Contents
ix
Acknowledgements
xv
Introduction
23
27
59
79
115
149
191
OShaughnessys View
197
OShaughnessys Arguments
221
Critical Assessment
253
viii
275
281
313
337
365
381
13 Imaginative Agency
385
409
Conclusion
431
Bibliography
443
Index
455
Acknowledgements
xv
Introduction
1
5
7
11
17
23
27
28
32
37
41
42
44
46
47
51
52
53
54
55
x
2
59
61
68
73
79
81
83
85
87
94
96
96
100
103
106
115
116
122
126
128
133
138
142
149
150
152
154
161
168
176
xi
191
OShaughnessys View
197
199
202
213
OShaughnessys Arguments
7.1. The Argument from Origin
(i) Standard Constraints on Cognitions
(ii) Wakefulness as a Further Constraint
(iii) The Origin of Imaginings
Critical Assessment
8.1. The Explanatory Power of Negation Claims
(i) General Limitations
(ii) The Limitations of (NC*)
(iii) The Limitations of (NI) and Similar Claims
(iv) The Limitations of (NC)
221
222
223
232
242
247
253
254
254
256
258
264
266
275
281
282
290
292
294
302
309
xii
10 Visual Imagining As Experiential Imagining
10.1. Martins Argument: Perspectivalness
10.2. A Second Argument: Egocentric Orientation
10.3. Replies to Objections
10.4. The Nature of Experiential Imagining
313
314
322
325
332
337
340
346
351
356
360
365
366
368
373
377
381
385
386
387
391
391
392
394
396
404
409
409
410
413
xiii
14.2. Extensional Adequacy
(i) Non-Imaginative Phenomena
(ii) Imaginative Phenomena
Conclusion
(i) The Agency Account
(ii) The Epistemological and the Dependency Account
(iii) Other Theories of Imagining
(iv) Some Loose Ends
414
414
417
421
423
431
431
434
437
439
Bibliography
443
Index
455
Acknowledgements
I am grateful to the following journals for permission to reproduce material: Enrahonar for parts of my paper Emotional Imagining and Our Responses to Fiction (Enrahonar, vol. 46, 2011; reprinted by permission of
the publisher, Universitat Autnoma de Barcelona); Rivista di Estetica for
parts of my paper The Recreative Imagination (Rivista di Estetica, vol.
54, 2013; reprinted by permission of the publisher, Rosenberg & Sellier);
and Philosophical Explorations for parts of my paper Transparency and
Imagining Seeing (Philosophical Explorations, vol. 13, 2010; reprinted by
permission of the publisher, Taylor & Francis).
My special thanks are due to Malcolm Budd, without whom I would not
have started and continued to work on the imagination; Mike Martin, who
has challenged and pushed me further on the topic than anyone else; Lucy
OBrien, who made sure that I finished my PhD thesis on the imagination
in time; Sebastian Gardner and Gianfranco Soldati, who taught me a lot
about the phenomenological approach to the mind and the imagination; and
Rob Hopkins, who has always been most helpful in his criticism.
In addition, I am very grateful, both for their constructive criticisms and
for their sympathetic encouragements, to Jiri Benovsky, Davor Bodrozic,
Johannes Brandl, Josep Corbi, Peter Goldie, Henning Hahn, David Harris,
Frank Hofmann, Eduard Marbach, Aaron Meskin, Richard Moran, Kevin
Mulligan, Matt Nudds, Martine Nida-Rmelin, Matt Soteriou, Kathleen
Stock, Juan Suarez, Mark Textor, Gian-Andri Toendury, Cain Todd, Lambert Wiesing, Marcus Willaschek, Richard Wollheim, several anonymous
referees, as well as audiences at University College London, the University
xvi
of Valencia and the University of Fribourg.
Finally, I would like to extend my gratitude to Johannes Brandl, Andreas
Kemmerling, Wolfgang Knne and Mark Textor, for their personal assistance and editorial guidance; the Graduate School of University College
London and the Swiss National Science Foundation, for their financial support of my research; and to my family and friends, for their uncomplaining
patience during the last years.
For Evgenia
Introduction
Introduction
The principal aim of this monograph is to present and compare the main
theories of imagining in a systematic and opinionated way.
The presentation is systematic in so far as it is structured by the concern
with the unity and common nature of the many forms of imagining, such as
visualising, supposing or daydreaming. Accordingly, the book involves
both a close study of the distinct kinds of imagining and of their main features, similarities and differences; and a critical discussion of the principal
accounts of imagining put forward in the literature. The book therefore
serves as an introduction to the main ideas and positions that have been adopted with respect to imaginative phenomena. But it does so in a strictly
problem-oriented manner, namely from the perspective of the possibility of
providing an account of imagining that unifies its different incarnations. As
a consequence, the discussion is organised around the possible sources of
the unity of imagining and centres less on particular philosophers and their
specific views.1
The comparison of the corresponding principal theories of imagining is
1
Another reason for this is that with some notable exceptions, such as the works by
Husserl (2006), Sartre (1940), Casey (1976), White (1990), OShaughnessy (2000)
and McGinn (2004) many philosophers have not presented fully developed and focused theories of imagining, but instead embedded their claims in discussions of
other, often more general issues (see, e.g., Collingwood (1938), Ryle (1949/1963),
Scruton (1974), Peacocke (1985), Walton (1990, Hopkins (1998) or Martin (2002a)).
Even of the noted exceptions, who devote whole books or lecture series to the topic,
some have not aimed at formulating a cohesive account of all forms of imagining,
but rather focused on discussing particular kinds of imagining, or on developing a
list of more or less loosely connected ideas and observations.
Introduction
opinionated, on the other hand, in so far as a more specific goal of the book
is to defend the Agency Account of imagining as the best unified account of
imagining available. The main claim of this theory is that imaginings are,
essentially and fundamentally, mental actions of a certain kind. Accordingly, imagining is something that we actively and voluntarily do.
A unified account of imagining is characterised by the fact that it captures the common nature of the central cases of imagining (e.g., visualising, supposing, or daydreaming) in fundamental terms and is able to distinguish them from the central cases of non-imaginative mental phenomena, notably cognitive representations (e.g., perceptions, judgements, or
memories). The minimal goal of any theory intended as a unified account
of imagining should therefore be to achieve extensional as well as explanatory adequacy: it should be valid for the paradigm cases of imaginative
and non-imaginative phenomena; and it should correctly describe and explain the distinctive and fundamental nature of the former. A main theme of
the discussion is therefore the elucidation of primary examples of imagining, and of what distinguishes them from other, non-imaginative mental
phenomena. Another and closely related guiding issue is whether it is possible to account for the specific character of these imaginings by identifying a set of features distinctive of them and responsible for their imaginative status (i.e., their property of being instances of imagining). That is, the
book addresses the issue of whether it is possible to specify and elucidate
this status in terms of individually necessary and jointly sufficient conditions for something to be an imagining, at least with respect to the
paradigm cases.
The issue of formulating a unified account of imagining has often been
neglected in the philosophical tradition. Many of the discussions of imagining in the past and the present have focused, not on the nature of imaginings, but on their role in our mental lives and our interactions with other
people and the world. It has been widely acknowledged that imagining is
very prominent in and significant for various parts of our lives, ranging
from our emotional engagement with other people (e.g., Goldie (2000):
194ff.) and our moral evaluation of actions (e.g., Johnson (1993)) to the
aesthetic appreciation of artworks (e.g., Walton (1990) and its many fol-
Introduction
lowers) and even the acquisition of knowledge about the external world.2
Moreover, even when philosophers address the question of what it
means for a mental phenomenon to be imaginative (i.e., possess the property of being an imagining), they usually concentrate exclusively on specific forms of imagining, notably on sensory or visual imaginings. 3 Indeed,
the neglect of the issue of what all imaginings have in common as imaginings may be closely linked to the neglect of non-sensory or complex kinds
of imagining. Finally, philosophers writing about the imagination have not
always succeeded, or been particularly interested, in developing their ideas
into a proper theory of what is distinctive of imaginings.
As a result, most discussions of imaginings have been concerned either
2
Although imaginings are typically held not to constitute knowledge about the external world or provide relevant epistemic justification by themselves (though see
Chapter 4 for an opposing view), it seems beyond doubt that they are often involved
in other ways in the acquisition of knowledge. See, for instance, the discussions on
thought experiments (e.g., Sorensen (1992) and Gendler (2000a)), the link between
conceivability and possibility (e.g., Gendler & Hawthorne (2002)), the role of mental imagery in geometry (e.g., Giaquinto (1992)) and in counterfactual reasoning
(e.g., Williamson (2008)), or the involvement of visualisation in memory-based
cognitive projects, such as the project of determining whether frogs have lips, or
how many windows are in ones own house (e.g., Kosslyn (1980): 1, or Pylyshyn
(2002): especially 164).
See Collingwood (1938): chs. 9f., Sartre (1940), Peacocke (1985), Hopkins (1998):
ch. 7, and, it seems, Wittgenstein (1984c): vol. II, sec. 63-147, to name just a few of
those who focus more or less exclusively on sensory or visual imaginings. Kind
(2001) does the same: while she argues that all instances of imagining involve
sensory imagery, she distinguishes the resulting sensory imagining from intellectual supposition. Scruton (1974), White (1990), Casey (2000), OShaughnessy
(2002), Currie & Ravenscroft (2002) and McGinn (2004) address both sensory and
intellectual imaginings, but no emotional imaginings (in contrast to emotional responses to imagining), and imaginative projects only in passing and not in relation to
their commonality with imaginative episodes. Besides, although McGinn discusses
various forms of imagining, he does not aim to provide a unified account of them.
Instead, he argues merely that they form an imagination spectrum which extends
from the most simple and temporally and conceptually prior imaginative phenomena
(e.g., those involved in sensory representation) to the most complex and developed
ones (e.g., those involved in creativity; see McGinn (2004): 13). See also the discussion of the five main forms of imagining in Section 1.2 for further references to accounts which focus on particular kinds of imagining.
Introduction
See Scruton (1974): chs. 7f., Casey (1976), OShaughnessy (2000): chs. 11f., and
McGinn (2004) for clear examples. But none of them discusses imaginative projects
(fantasies, daydreams, etc.) in any detail. Whether other proposals are intended or
able to constitute a unified account of imagining is less clear. Humes account of all
imaginings as forming a certain sub-class of ideas suffers from its treatment of the
differences between sensory and intellectual representations and between imaginative and cognitive representations as quantitative (i.e., a matter of vivacity) rather
than qualitative (see Chapter 9 for discussion and references). Although Ryle discusses mainly sensory cases, his account of imagining as a form of internal pretending or pretending in ones head might be applied to intellectual imagining as
well (see Section 12.3, and Ryle (1949/1963): chs. 7f., especially sec. 8.6). White
analyses both visualising and intellectual imagining in terms of thinking of the possible, but does not explicitly connect the two analyses (see Section 5.4, and White
(1990): 122f; 184). Despite the fact that Currie and Ravenscroft treat both sensory
and intellectual imaginings as simulations of their respective cognitive counterparts
(see Section 12.2, and Currie & Ravenscroft (2002): 11 and 49), their main concern
is with the involvement of the imagination in adopting a perspective on the world
different from ones current one, rather than with the general nature of imagining
(ibid: 8f. and 11).
Introduction
See, for instance, the skepticism expressed in Budd (1992), Hopkins (1998): ch. 1,
Wollheim (2003) and Dorsch (2012c) about the existence of the specific form of
imagining seeing the depicted, which Walton refers to in his account of pictorial experience (see Walton (1990): ch. 8, and (2002)).
Introduction
not the only motivations for the search of a unified account. The expectation that the central cases of imaginings share a common nature arises also
from the need to explain two facts about our actual treatment of such representations.
The first is simply that, even after discounting the less obvious cases,
we do group together a large variety of mental occurrences in the class of
imaginings, while excluding many others. Thus we accept visualising, supposing, daydreaming, being engaged with fictions, empathizing, and so on,
as paradigm instances of imagining, or at least as essentially involving
such instances; but not seeing, judging, deliberating, or feeling an emotion
or desire. If such imaginings had nothing in common with each other, but
shared features with the non-imaginative phenomena, this tendency in how
we categorise mental phenomena would be rather mysterious: there has to
be something about the imaginative mental phenomena which causes us to
treat them but not other mental phenomena as members of one and the
same class (see Scruton (1974): 91f.).
The second relevant observation is that our corresponding classifications are stable, and that we usually have a good grasp of whether though
not necessarily of why a given mental state is imaginative or not. Three
observations are particularly relevant here. First of all, we repeat the same
categorisations and do not locate certain representations today on one side
of a divide and tomorrow on the other. There is no change in opinion about
whether, say, visualising is an instance of imagining. Then, it normally
does not take us much effort or thought to come to a classification of a given mental phenomenon as imaginative or non-imaginative. Identifying an
instance of daydreaming as an example of imagining, say, is typically very
straightforward. Finally, we are seldom unsure about how to treat a certain
mental phenomenon. This may happen with borderline cases (e.g., spontaneously occurring images and thoughts), or when we are confronted with an
unusual or rare kind of representation (e.g., Macbeths awareness of Banquo, or of the dagger). But, in most cases, we are not in doubt about whether some mental episode is an instance of imagining. These three aspects of
the stability of our relevant classification, too, strongly suggest the existence of a certain kind of unity among imaginings, to which we furthermore
have easy epistemic access, but which is still in want of further elucidation.
Introduction
Otherwise, it would be very difficult to explain the firmness and ease with
which we take the sometimes very different instances of imagining to be of
the same kind, and why it is that they, but no other phenomena, count for
us as imaginative. Any satisfactory analysis of imagining needs to provide
an elucidation of this unity of the paradigm instances of imaginings.
The idea pursued in this book is that the unity under consideration is due
to some features shared by and distinctive of imaginings. This is the
simplest and most straightforward explanation of their homogeneity; and it
is adopted by all theories of imagining that have the ambition to provide a
unified account. But it is not the only possible account of the two noted
facts and the resulting apparent unity of imagining. One could, for instance, maintain that our grouping together of the variety of phenomena
described as instances of one and the same type of phenomenon is merely
accidental or conventional, without any substantial grounding in shared
features of the mental phenomena concerned. But such a claim would be
difficult to support in view of the facility and assurance with which we categorise mental states as either imaginative or non-imaginative. Until it has
been confirmed that such a strong form of skepticism is inevitable, the
realistic hope for a more positive theory of the common nature of imaginings should outweigh any doubts about the possibility of a unified characterisation of imagining. It is thus reasonable to demand from a theory of
imagining that it account for the fact that we classify a large variety of phenomena as imaginative; and the fact that this classification is not a pure coincidence or convention. A unified theory of imagining promises to provide
such an explanation by treating these facts as a consequence of the presence of certain features shared exclusively by all imaginings.
Introduction
principal ways in which the common and distinctive nature of the core instances of imagining may be elucidated: in terms of their irrelevance for
our epistemic interaction with the world, in terms of their dependency on
perceptions and (occurrent) beliefs, or in terms of their origin in a specific
kind of mental agency.6 The resulting theories of imagining may be labelled the Epistemological, the Dependency and the Agency Account. The
prominence in the literature of these approaches and their central ideas
about the nature of imagining is indicative of the fact that they are the main
contenders for a unified theory of imagining. Accordingly, I concentrate
my discussion primarily on these three proposals and devote to each a
whole part of the book.
The Epistemological Account specifies the difference between imaginings and cognitions by reference to the idea that only the latter are concerned with the representation of reality at least in a particular sense still
to be specified. Our minds interact with the world by means of cognition
and action. Both forms of interaction are primarily the domain of cognitive
representations, such as perceptions, memories or beliefs. In accordance
with this observation, the versions of the Epistemological Account to be
found in the literature identify a lack of cognitive concern with reality and,
especially, the failure to play a substantial role in cognition as the distinctive feature of imaginings. Different versions of this approach may vary in
how they characterise the assumed lack of cognitive concern.7
It is conceivable that a unified account of imagining may also be formulated in terms of their specific insignificance for our active engagement
6
Introduction
with the world. The idea is that imaginings in contrast to, say, desires, intentions, beliefs or perceptions cannot motivate us to act or guide us in
our actions (e.g., by providing us with information about our relevant environment, or about adequate means to achieve our goals). However, the
claim that imaginings lack a guiding role in agency can be traced back to
the idea that they lack a cognitive concern with the world: they cannot
guide us in action because they do not provide us with knowledge about
the relevant aspects of reality (i.e., the environment and the means). The
idea that imaginings cannot move us to act, on the other hand, is not only
controversial (see Note 16 in Chapter 1); it also does not distinguish them
from many non-imaginative phenomena. Hence, it is not very promising to
formulate a unified account of imagining specifically in terms of their
seeming unimportance for our actions; and not very surprising that the
focus in the literature on imagining has rather been on the apparent lack of
cognitive concern.
The Dependency Account, by contrast, puts emphasis on the asymmetric
dependency of imaginings and cognitions. It argues that imaginative phenomena are constitutively dependent on and, hence, have to be elucidated
by reference to their cognitive counterparts. While cognitions exist and can
be made sense of independently of imaginings, the opposite is false. Proponents of the Dependency Account disagree about the nature of the kind
of dependency concerned.8 Philosophers in the Humean tradition, for example, focus on the idea that imaginings depend on cognitions in so far as
they represent them (possibly due to a suitable causal link). Another prominent approach is to assume that, in some sense, imaginings imitate or
mimic cognitions for instance, in the shape of off-line simulation or inner
pretense. Besides, it is widely accepted that there is also some form of semantic dependency: that we acquire much of our ability to represent something in an imaginative way by means of engaging in cognition (i.e., in per8
Hume (1739): 1.1.1.6ff. and 1.3.7.7), and OShaughnessy (2000): 363, endorse the
representational version of the Dependency Account for all forms of episodic imagining; while Peacocke (1985): 22, Martin (2001): 273 and (2002): 404, and Dorsch
(2010c) accept this view for sensory imaginings; and Dorsch (2011b) which is
more or less identical with Chapter 11 for emotional imaginings. Currie &
Ravenscroft (2002): 49, and Ryle (1949/1963): 250ff., may be read as defending a
version of the Dependency Account formulated in terms of simulation or imitation.
10
Introduction
ceiving, sensing, judging, and so on) and developing the relevant recognitional and conceptual capacities.
Finally, the already introduced Agency Account proposes mental activity
as the characteristic element of imagining. It maintains that imaginings are,
in a particular way, essentially active; while cognitions are either passive,
or active in a different way. The general idea is thus that imaginings are
mental actions of a special kind: they are imaginative determinations or expressions of the will. They differ in this respect from mental passions,
among them: judgemental thoughts or occurrent beliefs that are formed in
response to epistemic reasons; episodic memories that are determined by
past experiences; perceptions and sensations that are determined by their
external objects; and deviantly caused instances of the each of the preceding cognitive kinds of episode. Again, distinct versions of the Agency Account vary in how exactly they individuate the active character distinctive
of imaginings.9
The particular version, that I intend to put forward in this book, maintains that imaginings are mental actions, that aim at the active and direct
formation of mental representations with specific contents. The requirement of representational specificity demands that the underlying motivational states (e.g., tryings, desires or intentions) determine which particular
features are to be represented as being instantiated by which particular objects; while the directness requirement is meant to ensure that the motivational states end up determining the content of the formed representations
without making use of epistemic or merely causal mechanisms of content
determination and, notably, those mechanisms involved in the actively
triggered manifestation of mnemonic or other mental dispositions, or those
relied on during the voluntary formation of beliefs on the basis of evidence
as means. My defense of this version of the Agency Account of imagining is paired with a rejection of the Epistemological and the Dependency
Account.
9
The Agency Account (or something very much like it) has been endorsed by, among
others, Wollheim (1973): 69, Scruton (1974): 95 (see in general 94-100), Levinson
(1998): p. 232, n. 3, Kind (2001): 90ff., and McGinn (2004): 12ff.; 131f. Most of
these endorsements have been inspired by Wittgensteins comments on sensory imaginings (see, for instance, his (1984b): vol. II, sec. 63 and 627; see also Budd
(1989): 104ff.).
Introduction
11
However, since the present book is not only finite in its length, but also
primarily centred on the issue of the unity of imagining and aimed at giving a balanced overview of the various approaches to this issue, the discussion of the Agency Account has to be confined to roughly the same number
of pages as the two other main theories (or, in the case of the Dependency
Account, as its most promising namely representational version). This
means, unfortunately, that there is not enough space for a comprehensive
elaboration and vindication of the Agency Account, which definitely merits
a monograph on its own. The chapters devoted to this theory therefore concentrate just on the principal aspects of this theory, such as its characterisation of imaginings as mental actions of a certain kind, the essentiality of
the imaginative activity involved in voluntary imaginings, or the special
status of spontaneously occurring images and thoughts in relation to imaginative agency. Other necessary elements of a full defense of the Agency
Account such as its application to daydreams and other imaginative projects, or a discussion of pictorial experience, association, guessing, and
other problematic cases have to be left for another occasion.
12
Introduction
Introduction
13
things actually are, but about how things might possibly be.
The resulting modal theory of imagining is, in fact, nothing but a specific version of the Attitude Account, according to which the difference
between imaginative and cognitive episodes is fundamentally attitudinal.
Minimally, the idea is that cognitions involve a cognitive attitude that is,
the commitment that reality is just like they represent it to be while imaginings lack it. Versions of the Attitude Account may stay neutral on the issue of whether this means that imaginings possess instead a distinctively
imaginative attitude. However, the need to distinguish imaginative episodes not only from cognitive episodes, but also from other kinds of episode
has moved proponents of the Attitude Account to assume that imaginings
are non-neutral as well, and to specify the commitment distinctive of imaginings in more positive terms, such as by reference to an act of positing
entities as non-existent, absent or, indeed, non-actual.13
Next comes the Spontaneity Account which maintains that imaginings
differ from cognitions in their specific origin in the voluntary or involuntary spontaneity of the mind, rather than in the impression of the external world onto the mind. In its most promising incarnation, the view
comes close to the Agency Account especially in its insistence that the
origin of imaginings is not merely causal, but essentially involves some
subjective response or doing. But the two theories differ in that the Spontaneity Account extends the realm of the imagination to the involuntary,
thereby loosely drawing on a distinction between perceiving and imagining
that is similar to the Kantian distinction between receptivity and spontaneity.
The Subjection to the Will Account is also similar to the Agency Account. While the last-mentioned takes imagining to always involve voluntary control, the first-mentioned requires merely that there is the possibility
of such control, without the need for its actual exertion. The Subjection of
the Will Account has the advantage of being able to capture what seem to
be involuntary instances of imagining, while distinguishing them none the
less strictly from, say, perceptions and impression, the passivity of which
13
The idea of a specifically imaginative attitude has been put forward notably in the
writings of the phenomenologists, such as Husserl (2006), Sartre (1940) and Casey
(1976).
14
Introduction
Introduction
15
on particular forms of imagining, notably visualising and belief-like imagining.16 What this may reflect is that the mental category of imagining is
perhaps not a very interesting category with respect to the subpersonal
level; while, as noted above, it plays an important role on the personal
level. Indeed, at least to my knowledge, there has been no serious attempt
so far to spell out, in terms of the neurofunctional entities and processes
recognised by the cognitive sciences, what the sensory, affective,
intellectual, experiential and more complex forms of imagining have
fundamentally in common.
Moreover, it is unclear whether it is reasonable to expect to be able to
formulate such a theory. Perhaps there are several distinct and fairly independent neural or otherwise subpersonal correlates of the feature distinctive of all imaginings on the personal level for instance, one correlate for
each of the basics form of imagining. That is, it is unclear whether such
correlates need to share a scientifically interesting feature in virtue of
which they count as correlates of one and the same personal level phenomenon (i.e., imagining). For instance, green or, more generally, coloured objects arguably do not have any physical property in common that
is significant from the perspective of physics. 17 Perhaps the same is true of
Tye (1991), Thomas (1999) and (2010) provide an overview of recent empirical
studies on sensory imagining, while Nichols (2006) and especially Gendler (2011)
do the same with respect to different kinds of intellectual imagining and related phenomena, such as pretense, empathy or delusions. Harris (2000) offers a developmental perspective on intellectual imagining and pretense, while Currie &
Ravenscroft (2002), Heal (2003) and Goldman (2006) look at it from a simulationist
perspective. More specifically, Currie & Ravenscroft (2002) and Gendler (2011) discuss evidence linking intellectual imagining to empathy, autism and delusion; while
Weinberg & Meskin (2006) develop an empirically-based account of intellectual
imagining partly based on the work by Nichols & Stich (2003) and apply it to
various philosophical problems namely the problems of emotional responses to
fiction (see Chapter 11), of imaginative resistance, and of distinguishing imagining
from supposing (see Note 18 in Chapter 1). By contrast, the empirically informed
debate about the nature of sensory imagining has been largely dominated by the
dispute between Kosslyn (see, for instance, his 1980) and Pylyshyn (see, for
instance, his 2002).
17
If at all, colour hues of surfaces should be identified with certain classes of reflectance properties, the members of which are individuated and grouped together in
terms of their dispositional power to bring about a certain pattern of stimulation in
16
16
Introduction
the neurofunctional correlates of imagining. Besides, the prospects of securing a satisfactory unified theory of imagining are likely to dwindle further, the more removed from the personal level an account becomes. Perhaps, generality is to be had only very close to or, indeed, only at the
level of theorising about persons.
My second reason for not further investigating the prospects of the Subpersonal Account is that it is not really in competition with the unified theories aiming to capture the personal level of imagining. Rather, the two
kinds of account should be seen as complementing each other. While the
previously introduced theories aim to capture the nature of imagining in relation to subjectively accessible features (e.g., representationality, rational
role, origin, passivity, and so on) and in contrast to other conscious mental
phenomena (i.e., perception, feeling, judgement, action, and so on), the
Subpersonal Account intends to identify the neurofunctional foundation of
whichever aspects of our conscious mental lives are essential to imagining.
That is, the Subpersonal Account is concerned with how imagining is realised in the brain, and not what it means to imagine something from the
point of view of persons. Accordingly, personal and subpersonal theories
that is, philosophy and the cognitive sciences have different goals and
employ distinct notions of fundamentality. The same is true of physics or
biochemistry and their concern with happenings in the mind or brain.
My third and final motivation for not spending more time on discussing
the Subpersonal Account is that the more philosophically minded theories
show a certain priority over it most notably with respect to issues of epistemic authority. If philosophy comes up with a satisfactory account of
imagining that makes good sense of the unity of the various forms of imagining on the personal level, it would be no reason to revise or give up this
philosophical view that the cognitive sciences were to reveal that there are
the relevant opponent processing channels and, hence, to bring about certain phenomenally individuated colour experiences. However, from the perspective of physics, these classes of reflectance properties in contrast to the reflectance properties
themselves are of no scientific interest. Physicists have been concerned with
developing and experimentally testing theories of how much of the incoming light
surfaces reflect, but not of the extent to which surfaces stimulate the opponent processing system or cause phenomenally different colour experiences (see Dorsch
(2009a) and (2010b)).
Introduction
17
no significant or interesting commonalities among all the different instances of imagining. Instead, and in the absence of further reasons to
question the philosophical view, it would be more reasonable to conclude
that imagining, as a personal-level phenomenon, has no single correlate on
the subpersonal level. Similarly, the fact that physics is not really concerned with colours does not undermine the attempt to formulate an adequate philosophical theory of colours (see Notes 17f. above). Of course,
philosophical theories of imagining need to be given up if they contradict
empirically well-supported versions of the Subpersonal Account (if there
are any). But, on the whole, the best philosophical theories of imagining
tell the cognitive sciences what to look for (e.g., for the neurofunctional
correlate of a particular kind of mental agency), and not the other way
round.18 Indeed, whether a given version of the Subpersonal Account is
successful should in part be evaluated in respect of whether it captures the
feature, or set of features, identified as being distinctive of all instances of
imagining by the best philosophical unified theory.
See Dorsch (2010b) and Dorsch (2010a) for a similar division of labour between
philosophy and the empirical sciences in the case of colours and of hallucinations,
respectively.
18
Introduction
promises the best explanation of the quasi-observationality of sensory imaginings (see Chapters 2f.).
Second, Part One is concerned with the relationship between imagining
and knowledge. In particular, it argues that certain instances of sensory
imagining can constitute knowledge in a way very similar to that of perceptual experiences and, in response to this observation, asks in which
sense (if any) imaginings can still be said to be uninformative with respect
to the external world (see Chapter 4).
Third, Part One finishes with a discussion of the less important unified
accounts of imagining listed above (see Chapter 5). The only exceptions
are the Causal Account and the Subjection to the Will Account, which are
briefly discussed in the Chapters 9 and 13, respectively; as well as the Subpersonal Account, which is, as mentioned above, not further discussed at
all.
Part Two is concerned with the characterisation (see Chapters 6f.) and
rejection (see Chapter 8) of the various versions of the Epistemological Account of imagining, formulated in terms of the alleged inability of imaginings to be reliable or rationally integrated with our picture of reality, to inform us or provide us with knowledge about the external world, or to possess the commitment, justificatory power or epistemic function essential to
cognitions. The central component of this discussion is a detailed and
scholarly examination of OShaughnessys version of the Epistemological
Account.
Part Three deals with the various versions of the Dependency Account.
Chapter 9 highlights the relevant causal and representational elements in
Humes conception of imaginings and reveals its continuity with more contemporary neo-Humean views that understand the assumed dependency
of imaginings on cognitions in purely representational terms. Together with
the following two chapters, it is also devoted to the applicability of the resulting Representational Account that is, the representational version of
the Dependency Account to intellectual imaginings (see Section 9.5),
imaginative projects (see Section 9.6), sensory imaginings (see Chapter 10)
and emotional imaginings (see Chapter 11). The conclusion to be put forward is that, while the Representational Account is probably true of sensory and emotional imaginings, it fails to pay justice to the nature of the
Introduction
19
other two forms of imagining. Part Three concludes with a negative assessment of the prospects of devising a unified account of imagining by reference to the idea that imaginings simulate or otherwise imitate cognitions
(see Chapter 12).
The final Part Four develops and defends the Agency Account of imagining. It does so, first of all, by spelling out the main thesis of the theory,
distinguishing it from the Subjection to the Will Account and presenting an
argument in favour of the endorsement of the Agency Account (see
Chapter 13). The central claim of this account is that imaginings are to be
identified with mental actions that have the purpose of producing mental
mental representations and of applying direct control over which states of
affairs are thereby represented. And it is crucial for the truth of this view is
that the kind of mental agency involved is essential to the formed
imaginative representations, thus ruling out the possibility of passive
imaginings. In addition, Part Four illustrates how the Agency Account can
meet the two desiderata for unified theories of imagining established
earlier and deal with potential counterexamples, notably the imaginative
counterparts to non-representational phenomena, as well as spontaneously
occurring images and thoughts (see Chapter 14).
Many of the parts, chapters and sections of the book are more or less
self-contained and can be read fairly independently of the others. Engagement with Part Two and the Epistemological Account requires mainly
knowledge of Chapter 1 and Sections 2.1f., although reading at least the
first half of Chapter 4 will be helpful as well. In addition, the reader may
decide to skip most of Chapter 6 and all of Chapter 7 if (s)he is not interested in the details of OShaughnessys theory of imagining and his more
general approach to conscious awareness. For Chapter 6 begins with a brief
synopsis of the main claims and arguments of OShaughnessys, which
provides all the information necessary to proceed directly with Chapter 8.
Part Three (the Dependency Account) and Part Four (the Agency Account)
presuppose even less than Part Two, namely really not much more than acquaintance with the content of Chapter 1, as well as perhaps Sections 2.1f..
Both Chapter 10 and Chapter 11 on sensory and on emotional imagining,
respectively can be read largely on their own, which is in part due to the
fact that they are based on previously published material. Similarly, the dis-
20
Introduction
Part One
The Nature and Variety of Imagining
This opening part of the book has two main functions. First, it is intended
to provide an overview of three different things: (i) the different forms of
imagining; (ii) their main features; and (iii) various unified accounts of
imagining (i.e., those that differ from the three main theories to be discussed later on in being less well-developed, less plausible, or less prominent in the literature). In short, it is meant to serve as an introduction to the
philosophy of imagining. Second, this part also aims to set the stage for the
rest of the book. In particular, its purpose is to specify the main desiderata
for a unified account of imagining and to highlight some features of imaginative episodes that are to become crucial during the upcoming discussions.
In the first chapter, I spell out the two main desiderata and list the five
central forms of imagining. The next two chapters are devoted to a description of the main differences between imaginative and cognitive episodes
and, in particular, their sensory kinds. In Chapter 2, I highlight the differences in attitude and epistemic function between imaginings and cognitions and argue that both may refer to aspects of reality, though in virtue of
different referential mechanisms. Chapter 3, by contrast, focuses on important characteristics of sensory and, especially, visual imaginings,
which distinguish them from perceptual experiences. The following
chapter addresses the relationship between imagining and the acquisition
of knowledge. More specifically, I argue that sensory imaginings can, under the right circumstances, lead to knowledge in a way that is very similar
to the way in which perceptual experiences may result in knowledge. And I
24
I also intend this not to imply disjunctivism about episodes or states of knowledge.
In particular, I stay neutral on whether belief is a proper constituent of knowledge
(see Williamson (2000) for further discussion), or whether perceptions and
hallucinations share some or all aspects of their nature (see Martin (2006) and
Dorsch (2012a) for further discussion).
25
CHAPTER 1
Unified Accounts of Imagining
For the assessment of the main candidates for a unified account of imagining, it is necessary to say first a bit more about what is required of such a
theory. An important part of this task is to clarify which phenomena should
definitely be captured by a unified account of imagining, and which definitely not. Accordingly, the three sections of this chapter are concerned with
the particular demands on, and the mandatory scope of, theories of imagining.
In the first section, I specify two desiderata for a unified account of imagining: (a) extensional adequacy with respect to all central imaginative and
non-imaginative phenomena; and (b) explanatory power with respect to the
distinctive nature of imaginings. The following section lists the paradigm
instances of imagining and their non-imaginative counterparts. In particular, I describe five main forms of imagining which any unified account has
to capture: (i) sensory imaginings; (ii) affective imaginings; (iii) intellectual imaginings; (iv) experiential imaginings; and (v) imaginative projects.
In addition, I introduce the main examples of non-imaginative phenomena,
notably cognitive states and projects. The third and final section briefly
discusses the status of unclear, controversial or borderline instances of imagining (e.g., dispositional imaginings, or representations symptomatic of
psychological disorders) and their relative irrelevance for the evaluation of
unified accounts of imagining.
28
29
30
ings independent of their imaginative character (e.g., their being conscious, or their
being representational).
See OBrien ((2007): ch. 8) for a defense of the claim that actions form a natural
mental kind or are basic elements of the psychological repertoire, much as beliefs
are (ibid: 130).
31
See, for instance, the various attempts to define knowledge in terms of justified true
belief; or Wittgensteins related skepticism and his alternative idea of family resemblance concepts.
32
stead, it may simply consist in the capacity to use, in a certain way, one or
more of our more basic mental faculties (e.g., to form mental representations). In any case, the question of whether there is a faculty or capacity involved in all and only cases of imagining need not be answered by a unified account of imagining.5 Similarly, a theory of what it means to imagine
something may lead to a better understanding of the imaginativeness exhibited by creative or inventive people. But again, it is not clear whether
there is a particularly close link between imaginativeness and instances of
imagining. Accordingly, a unified account of imagining does not have to
address the issue of the imaginativeness of people or their works. It needs
to be concerned solely with states, acts and projects of imagining.
There has been a long tradition of assigning the faculty of imagination a particular
role in our cognition of the world (see Hume (1739) and Kant (1781); for further
discussions, see Strawson (1970), Warnock (1976): Part I, and Sellars (1978)). However, the faculty in question is not exclusively related to the formation of imaginative representations, but usually said to be also centrally involved in the occurrence
of, say, perceptions or memories. Hence, it cannot be interpreted as the faculty or capacity to imagine; and its discussion need not be part of theory of imagining.
OShaughnessy ((2000): ch. 11) provides a general discussion of the relationship
between imaginative representations and the capacity to imagine.
Representations may also have a referent, and be correct relative to whether how
they represent entities as being satisfies a certain standard of correctness (e.g., the
relevant facts in reality, or ones intentions). But I would like to stay neutral here
concerning the issue of whether all types of representation can refer or be correct.
33
My contrast between sensory and intellectual episodes is meant to track the basic
difference in representationality between perceptions and judgements (or beliefs),
given that they may be about the same objects and features. Although I would like to
stay neutral towards the issue of whether the difference in question is one of degree
or of kind, I assume that there is some distinction to be drawn, as it has been accepted in the tradition at least since Kant. In addition, I take this difference to be more
fundamental than but presumably overlapping and perhaps even coextensional
with some of the various differences to be found in the more recent literature,
which have been characterised in terms of the difference between conceptual and
non-conceptual (e.g., Peacocke (1992): ch. 2; Crane (1992)), digital and analogue
(Dretske (1981): ch. 6) or propositional and non-propositional (or object-directed)
representation (OShaughnessy (2000): ch. 10).
My notion of a representation is thus very close to Lockes notion of an idea,
Humes notion of a perception, Kants notion of a representation (i.e., Vorstellung), or Husserls notion of intentional act. It is perhaps less close to certain contemporary uses of the term representation, in particular to those employed in cognitive sciences.
34
Its content may be taken to consist in what it represents, that is, in the represented entities and features. Accordingly, the judgement that it rains, the
supposition that it rains, the desire that it rains and joy about the fact that it
rains all have the same content, while the judgement that it snows has a
different content. But we sometimes also refer to the type of representation
in question when we talk about the contents of representations for instance, when we note that the visual content of the perception that it rains
differs from the intellectual content of the corresponding judgement. I follow both usages of the term content and hope that the context will clarify
which one I have in mind.
In either case, the content of a mental representation strictly contrasts
with any potential attitude towards what is represented (e.g., whether it is
believed or desired), as well as with any affective element (e.g., any emotional feeling) possibly involved. Besides, I understand the representational
element of a mental representation as precisely that concrete aspect of it
which realises its content and its type of representation. Two representations, which represent the same entities and features and are of the same
type of representation, involve the same representational element for instance, a certain visual image, or the entertainment of a particular proposition.9
Now, that certain kinds of mental phenomena are central cases of imaginative or non-imaginative phenomena, respectively, means that their possession or lack of an imaginative character is not subject to serious disagreement, neither in daily discourse nor in philosophical debate. It is important to note, however, that whether something is an instance of imagining is not merely and presumably not even primarily a matter of whether we label it as such. Our use of terms like imagining can vary greatly.
9
My broad use of the term visual image (or of the related expression visual imagery) is perhaps unorthodox, given that others reserve it for visual memories and
imaginings alone and do not apply it to visual perceptions (see Sartre (1940)). In addition, it is not meant to imply the presence of any kind of mental or internal pictures, as they have been criticised by Sartre (1936): part 2, Sartre (1940): ch. 1, Wittgenstein (1984d): sec. 620ff., especially 638 and 642, and Ryle (1949): ch. 8, esp.
sec. 3. Perhaps some of the positions which endorse what I call an echo thesis
(roughly, the claim that imaginings depend on cognitions in such a way that they are
both similar to and different from them) may be taken to embrace the idea that visualising involves internal pictures.
35
And they are not always used to denote imaginative representations. For
instance, the statement could you really have imagined that can, in some
situations, question the imaginative skill of someone (e.g., in response to
hearing a great story, or to the assertion of a person that she had no problem visualising a complex scene); but in others, it can cast doubt on the capacity to have previously believed or expected something (e.g., a surprising turn of events). Hence, the agreement about the paradigms of imaginative and non-imaginative phenomena, that I have in mind here and aim at,
concerns the nature of the respective mental phenomena, which is taken to
be largely independently of how we refer to them.
There are very many mental phenomena which are clearly not imaginative. Most prominent among them (especially in discussions about imagining) are cognitive phenomena: on the one hand, cognitive states, such as
perceptions, judgements10, beliefs or memories; and, on the other hand,
cognitive projects, such as trying to come to understand another person, or
finding the best answer to the mind-body problem. Other central cases of
non-imaginative phenomena are desires, emotions, intentions, moods, inclinations, the practical projects of deciding on a course of action, or of
calming oneself down, and so on. Any theory of imagining which cannot
distinguish such cognitive and other non-imaginative phenomena from
imaginative ones should not count as providing a unified account of imagining.
What I take to be central cases of imagining, on the other hand, is perhaps best illustrated by considering an example. Suppose I ask you to imagine that Rome is covered by a thick layer of snow. In case you intend to
follow my instruction, it is quite likely that you start to visualise how certain parts of the city might look, their appearance transformed by the
masses of snow. You may thus picture the Piazza Navona, or one of the
narrow streets in the old centre, as brightly white, and void of most of the
people and all kinds of traffic. You may also auditorily imagine how the
10
36
snow swallows and mutes the sounds, or how it makes a distinctive sound
when pressed down by your feet while you begin, in your imagination, to
walk through the city. You are thereby perhaps able to imagine feeling the
resistance of the snow to your feet, the texture and thickness of its layer,
the harshness and freshness of the cold, how it slightly hurts your almost
numb skin, and the pleasure and exhilaration at experiencing this rare and
beautiful scenery. Or you may begin to imagine what you would like to do
in such a situation, or where you would feel like going.
But instead of, or in addition to, summoning up certain experiential
qualities in your mind, you may also simply begin to hypothesise about the
consequences of the highly unusual setting for the citys completely unprepared inhabitants, on the basis of imagining that the streets and squares of
Rome have disappeared beneath large amounts of snow. Thus, you may
wonder how they might cope with the unexpected situation and, in particular, how they might manage to move around and keep the basics of daily
life running. You may imaginatively think through different options of
dealing with the rare and surprising circumstances, and you may imagine
how it would feel for the inhabitants of Rome to handle the situation,
whether they would be taken up with the same enjoyment that you have
imagined yourself to feel (or that you may actually feel while imagining),
or instead would despair with or be annoyed by the unexpected disruption
to their normal life which leaves them stranded in their homes. Finally, you
may engage in imagining these things about Rome and its inhabitants, not
in response to my invitation to do so, but in response to reading a work of
fiction in which a scene is, in some detail, described in terms similar to
those used above.
This complex example of imaginative engagement on your behalf contains all five cases of imagining that I consider to be paradigmatic: (i) sensory imaginings; (ii) affective imaginings; (iii) imaginative thoughts; (iv)
experiential imaginings; and (v) imaginative projects. Being a member of
one of the five groups should not be taken to necessarily exclude being a
member of another one. As will become clear, it is very likely that there are
considerable overlaps between the five different types of imagining. In addition, this list is not necessarily meant to be exhaustive: it might not capture all cases we would agree on as central cases of imagining. But to my
37
This idea Berkeleian in spirit is defended by Peacocke (1985) and, at least for
some central cases, in Chapter 10 of this book and by Martin (2002a): sec. 3 and
Martin (2001): sec. 3, as well as Dorsch (2010c). It comes under attack in Williams
(1966), White (1990): chs. 13-15, Hopkins (1998): ch. 7, Noordhof (2002) and Currie & Ravenscroft (2002): sec. 2.2. Note that it may seem natural to one to describe
visualising as imagining seeing without being committed to the claim that seeing is
part of what is imagined (see Wollheim (1973): 55; Hopkins (1998): ch. 7). Besides,
there is also the (independent) debate about whether sensory imagining or experiential imagining, for that matter sometimes or even always involves an intellectual
element, such as an imaginative thought or intention-in-action. This debate is closely
related to the question of how best to understand the multiple use thesis, according
38
ments in support of this claim in Chapter 10. But, for the time being, I intend to stay neutral on the possibility of non-experiential sensory imaginings, that is, of sensory imaginative episodes that do not have perceptions
or similar episodes as part of their contents.
Many discussions of sensory imaginings have focused on visual instances (e.g., see Peacocke (1985); Sartre (1940)), and I mainly follow
them in that respect. Instances of visualising a face or a landscape thus figure as paradigm cases of sensory imagining. Of course, there are important
differences between, say, visualising something and hearing or touching it
in ones imagination, as there are important differences between seeing it
and hearing or touching it in real life. But I presume that these differences
do not bear much on the issues linked to the possibility of formulating a
satisfactory unified account of imagining. Also, it is interesting to ask
whether there are kinaesthetic or proprioceptive imaginings; whether these
are similar in structure to cases like visualising; and whether they amount
to imagining kinaesthetic or proprioceptive experiences. But my hope is
that they do not differ in important respects from the much more investigated example of visual imagining. So far, there does not seem to be a good
reason to doubt that a unified account of imagining capable of capturing
visual imaginings is also able to capture other types of sensory imagining.
Let me thus say a bit more about the nature and variety of visual imagining comments which I hope will apply, to the relevant extent and with
the necessary modifications, to instances of imagining in other modalities.
One way of visualising something and presumably the easiest is to stop
seeing (e.g., by closing ones eyes) and to concentrate on the production of
the visual image. But we can also visualise something while seeing (as
well as thinking about) something completely different (see Wittgenstein
(1984c): vol. II, sec. 65). For example, while walking down the street, we
may visualise the face of a friend whom we plan to surprise with something in the near future. However, this may be more difficult than the first
way of visualising something (e.g., the result may be less rich in detail or
in determinacy), and require more effort in particular, in that we have to
to which the same imaginative visual representation may figure in different instances
of imagining (see Noordhof (2002), from whom I borrow the label, as well as Martin
(2002a)). I address this question and the respective debate in the literature in Section
3.6.
39
sufficiently shift our attention from the perception (or the thinking) to the
imagining.
Furthermore, there are cases in which the content of a perception
provides us with the material for the sensory imagining. Visualising a perceived face as it would look in old age, or when looked at via a distorting
mirror, are examples. But the two episodes are still metaphysically independent of each other: each could occur without the other, and their contents are distinct. This remains true even if there is for us an explicit link
between the objects of the two representations for instance, if we think
them to be spatially linked by visualising an object as being a part of our
currently perceived environment (what Hopkins (2011c) calls visualising
into seen space). This is the case, say, when we see a white wall before us
and visualise how it would look like if a certain painting were hanging
there. Again, the two representations and their contents are distinct and do
not merge in any sense. In particular, we still see the whiteness of the wall
exactly where we visualise the painting to be. The case would be different
if we stopped seeing the white wall (e.g., by closing our eyes) and recalled
its appearance in order to visualise it together with the painting. 12 These
considerations are in fact related to the sometimes endorsed idea that we
cannot have a visualisation with the same sensory content as a simultaneous seeing. For instance, Wittgenstein writes that:
While seeing an object, I cannot imagine it. (Wittgenstein (1984c): vol. II,
sec. 63)13
The idea is that when we already see the National Gallery in front of us we
That perceptual and imaginative representational elements do not or at least not so
easily merge also constitutes a problem for imagination-based accounts of pictorial
experience which take our visual awareness of what pictures depict to be imaginative and to be merged with our perceptual awareness of the depicting surface. For the
proponents of this view have problems to explain why perceptual and elements
merge in the pictorial case, but not in examples like the one discussed here (see
Hopkins (1998): 3 and Dorsch (2012c)).
13
The translation is mine. The original reads: Whrend ich einen Gegenstand sehe,
kann ich ihn mir nicht vorstellen. The same point is stressed by Mulligan (1999):
sec. 3.
12
40
cannot also visualise it at the same time (and, presumably, from the same
point of view). Similarly, the claim rules out the possibility of first visualising the National Gallery and then coming to also and simultaneously see
it (again, presumably from the same perspective). According to one possible way of understanding this thesis, it seems to be true: the sensory representation involved in seeing the National Gallery cannot also be the representation involved in visualising this building; the two do not simply
share a token visual representation. It may also be true in another
interpretation: it may be that (as already suggested above) visualising
something while seeing something else presupposes sufficient attention to
the task of visualising, and that we may lack this required degree of attention when we, say, observe very carefully how the National Gallery really
looks. But a third reading is at least questionable, if not false: namely that
we cannot simultaneously have two independent visual episodes with the
same content and point of view, when one is perceptual and the other imaginative.14 If we can visualise a face (or anything else) while looking at the
National Gallery, it should also be possible that we can instead visualise
the building while continuing to have our perception. For if there is any
feature of visualising which prevents it under all circumstances from occurring at the same time as seeing, it should better not depend on what is
the particular content of the imaginative episode.15
See Hopkins (2011a). Hopkins also observes that visualising something is incompatible with simultaneously visually remembering something (else), as well as that we
cannot have two simultaneous, but separate episodes of visualising (or visual recall)
that is, two episodes that occur at the same time, but visually present their objects as
part of distinct imagined spaces or worlds. His explanation of these observed facts is
that visual recall is a form of visualising that is constrained by the determination of
what is visually represented by what has been visually perceived in the past. My alternative explanation is that both visualising and visual recall involve the same kind
of experiential representation of visual perception (see Section 9.4 and Chapter 10).
15
For example, it might be argued that we generally cannot enjoy an episode of seeing
and an episode of visualising at the same time because the same neurons or regions
in the brain would be involved in the realisation of the two episodes. But claiming
that such a conflict on the neural level prevents us only from simultaneously seeing
and visualising the National Gallery, but not from simultaneously seeing the National Gallery and visualising a face would mean claiming that it is possible to
identify neurons or regions in the brain that are responsible for the visual representation of very specific objects or features in such a way that no other neuron or region
14
41
42
43
44
hearing a story does not bear on the issue of which basic category of mental episode the intellectual representation in question belongs to, even if its
formation may be constrained in different ways, depending on its mental
context.21
45
Jena and thinks of Josephine, or about the next tactical moves (see Wollheim (1984): ch. 3, for an extensive discussion of such cases).
Apart from such cases of imagining experiencing something or imagining being someone, experiential imagining also includes cases of imagining doing something, assuming that it is imagined by means of imagining
the respective experiences which one would have while doing it (see Peacocke (1985): 22; Budd (1989): 100; Walton (1990): 29f.). Experientially
imagining (the experience of) playing the piano involves, for instance, imagining seeing some hands moving over the keyboard, imagining hearing
the corresponding sounds of a piano, and linking the imagined episodes to
each other and to oneself. It may also involve intellectually imagining that
one plays the piano, but cannot be simply reduced to it (see, for instance,
the discussion of Waltons view on pictorial experience in Section 11.2 and
in Dorsch (2012c)).
Sensory and, in particular, visual instances of experiential imagining
have been the subject of extensive discussions which have addressed at
least two important issues: what it means to imagine an episode, and to
imagine it to be of a certain kind and belong to a particular person; and
whether it is possible to imagine an episode (especially if it involves a perspectival representation) without also imagining it as being the episode of a
particular person (who, for instance, is imagined as occupying the respective point of view; for references, see Note 10 above). For my purposes, it
does not matter much how these two issues about the nature of experiential
imagining are settled. But I return to the first question in the Chapters 10
and 11. What I defend there are two ideas. The first is that experiential
imagining is an instance of object awareness, rather than propositional
awareness. That is, it makes us aware of concrete objects and their features
(i.e., mental episodes and their phenomenal characters), instead of the bearers of a truth-value (i.e., propositions). The second idea is that experiential
imagining may but need not be accompanied by an imaginative thought
(or thought-like element) that identifies the type or nature of the imagined
episode and, if necessary, relates it to a particular person as its bearer, or to
other imagined episodes (see also Section 3.6, as well as Peacocke (1985):
25f., and Noordhof (2002)).
46
As already noted (see Notes 3f. in the Introduction), most comprehensive discussions of imagining focus exclusively on imaginative episodes or (simple) experiential imagining, while neglecting imaginative projects. Notable exceptions are Currie
& Ravenscroft (2002: 38ff.) on fantasies and Walton (1990) on our extended imaginative engagement in games of make-believe, or with representational artworks.
OShaughnessy (2002: ch. 5) discusses daydreams, but without linking this discussion to his account of imagining. Scruton (1974), Casey (2000) and McGinn (2004)
mention imaginative projects (e.g., daydreams, reveries, or fantasies) only in
passing. By contrast, there are also some texts which focus more or less exclusively
on imaginative projects, rather than imaginative episodes. Williams (1966), Wollheim (1973), Wollheim (1984: ch. 3) and Velleman (2006) are concerned with the
various ways in which we can imagine (or daydream) something about ourselves, as
well as with the question of what it means and to which extent it is possible to
imagine being someone else. Relatedly, Goldie (2000: ch. 6) provides a detailed
elucidation of empathy (which he understands, roughly, as experientially imagining
oneself as having the thoughts, feelings and emotions of another person) and its relatives, such as imaginatively putting oneself in the situation or the shoes of another person.
47
experientially imagining the sensations, perceptions, and so on, of Napoleon in specific situations; and empathising with another person involves
her thoughts and feelings. But not all imaginative projects include experiential imagining, or at least not as their primary ingredient as, for instance, in the case of inventing in ones mind, or of imaginatively telling, a
third-personally narrated story. Besides, the simplest instances of experiential imagining such as imagining seeing (or visually experiencing) a tree
are mereologically too simple to count as projects (see Section 10.3f.,
Hopkins (1998): ch. 7, and Dorsch (2010c)).
It has been denied that there is any such thing as phenomenal consciousness (e.g.,
see Dennett (1988)). If this turned out to be true, my comments should be regarded
as concerning the phenomenal characters that the respective mental phenomena
seem to have. I discuss the nature of phenomenal consciousness in more detail in
Dorsch (2012a). For characterisations of phenomenal character similar to mine, see,
for instance, Block (1994), Carruthers (2000): 13f., Siewert (2000): sec. 3.10, and
OShaughnessy (2000): 534.
48
49
50
we read a novel bit by bit, or explore a city or landscape over several days
or months). Such experiences are unified by their common object, plus other factors (e.g., the appreciators intentions and memories; see Dorsch
(2000): ch. 1). Another, here more important type of mental complexes are
mental projects such as the imaginative ones discussed above which are
unified by their intrinsic purpose and the corresponding underlying motivational states (see Section 14.2). However, mental episodes can
temporarily overlap, occur simultaneously or form temporal series with
other mental episodes without also thereby forming mental complexes,
such as aesthetic experiences or projects.
Applying these considerations to the five central forms of imagining
identified above, we get the following results. The first three forms of imagining are clearly episodic in character. Intellectually imagining that it
rains is a mental episode; as is visualising a tree, imagining an itch or imaginatively feeling pleasure. Imaginative projects, on the other hand, are
mental complexes: they are unities made up of mental episodes. Experiential imaginings are not so easily classified, partly because their nature is
still unclear. The more sophisticated instances (such as the case involving
imagining being Napoleon) are presumably mental complexes as well.
They involve more than one imagining of a mental episode, and typically
also more than one imaginative thought. But elementary examples of experiential imagining (such as imagining seeing a tree, or even visualising a
tree if the latter is to be identified with the former) may very well be episodic.
Whether simple instances of experiential imagining are indeed episodes
rather than complexes of episodes depends, for instance, on whether
imagining experiencing something may fail to involve an imaginative
thought-like element; and if not, whether this element could occur independently of imagining of the experience (see Section 3.6). If experiential
imagining involves an independent episode of thought, then it is complex,
consisting of the thought and some other, presumably non-intellectual imaginative element (e.g., a sensory imagining). If, on the other hand, experiential imagining involves no or just a dependent thought, then a good case
can be made for the claim that it is episodic, given that it forms an indivisible whole. But irrespective of how this dispute is to be settled, it is true
51
that all the central cases of imagining put forward are phenomenally conscious.
I discuss two other candidates for non-central cases of imagining in much more detail in Dorsch (2012c) and Section 14.4: namely, respectively, pictorial experiences
and spontaneous representations. The reason for spending more time on the former
is that the postulated imaginativeness of pictorial experiences has been very prominent in recent discussions (see the impact of the theories expounded in Scruton (1974)
and Walton (1990), as witnessed, for instance, in Stock (2008)); while the reason for
spending more time on the latter is that they are very similar to imaginative phenomena, despite not being instances of imagining themselves.
52
53
belief for instance, the belief that the writings describe Sherlock Holmes
as living at 221b Baker Street; or the belief that they are such that a suitable appreciation of them requires us to imagine that Sherlock Holmes
lives at 221b Baker Street; or simply the belief that we ought to imagine
that Sherlock Holmes lives at 221b Baker Street.
Here is not to the place to settle this debate. 29 But it is important to note
that many cases, that initially may appear to be instances of dispositional
imaginings, are in fact rather examples of occurrent imaginative thoughts
that stay in the background of our consciousness or attention, after having
been formed in the foreground. When I am daydreaming that I am rich, I
continue to have the respective premiss in mind, even if I stop constantly
paying attention to it; just as my thought that I find the person opposite me
very attractive stays in my consciousness during the whole duration of my
rendezvous with her, despite often receding to the background.
Independent of this, even if dispositional imaginings turn out to exist,
they are bound to come into being only as a consequence of first having
had the respective imaginative thought notably in response to some other
existing entity (a story, an artwork, a game of make-believe, a daydream,
and so on) that determines what to imagine in the situation in question. At
least, it is difficult to conceive of other ways in which we might be said to
form and to have reason to form dispositional imaginings. Hence, it is
natural to expect that any satisfactory account of dispositional imaginings
would be dependent on an account of episodic imaginings; and that any
unified theory of imagining capturing the latter would also be able to capture the former.
There are good reasons to be skeptical about the possibility of dispositional imaginings. For instance, Nichols (2004a) argues against their existence; while Walton
(1990: 17f.) is undecided, but with some skeptical undertones.
54
55
people concerned as imagining something (e.g., when someone imagines that everyone around him is part of a conspiracy against him, or
when someone imagines that the objects he sees are becoming fluid).
But it is not always clear what we mean by this; and if it is clear, we seldom have an instance of imagining in mind. In particular, we use the term
imagining to say that the subjects concerned make something up against
their will; that they wrongly believe it; that they do not realise their own involved (imaginative) activity; or that they are simply in a state which is
different from our normal thoughts, feelings, perceptions, judgements and
memories.
As a consequence, some of the examples such as delusions are better
taken to constitute forms of belief, rather than forms of imagining (see
Bayne & Pacherie (2005)). In contrast, other cases such as inserted or
compulsive thoughts are perhaps more like imaginings in that they involve the active formation of representations, although the subject does not
fully acknowledge them as the product of their own doing (see Roessler
(2001)), or cannot easily banish or terminate them. I briefly return to obsessive and similar representations in Section 13.3 and the Conclusion,
where I indicate which of them should count as imaginative, and how the
Agency Account of imagining can accommodate them.
56
acted in an unexpected way, for which he or she does not seem to have any
apparent reason. These cases are closer to imagination than to cognition in
so far as they are concerned, not with real or close possibilities supported by ones evidence, but with possibilities unlikely to be realised, or more
or less unrelated to (what one knows about) reality. But, as I argue further
below in Section 5.4, their concern with possibilities whether they are
realistic or fanciful is as such not enough to mark them as instances of
imagining.
However, what may signal that they do involve imagining in one form
or another is that the states at issue are usually accompanied by a belief
about their fancifulness. If this were not the case, they would come turn out
to be delusions of some sort and would have to be treated accordingly (see
the preceding section). It is clear to most of us that only a handful of specific people have a realistic chance of becoming the next monarch of the
United Kingdom. Hence, if one of us wishes to become the next king or
queen without taking this possibility to be far-fetched, he or she will, at
least to some extent, be deluded. Now, that we take our own unrealistic expectations, hopes, and so on, to be fanciful may indicate that they are partly
imaginative in so far as they involve, as one of their elements, one or more
of the five central forms of imagining described. Wishing to become the
sovereign of the United Kingdom, for instance, is likely to involve some
imaginative thoughts about what it means to be in that position; while the
excitement involved in the desire for Switzerlands victory may be founded
on a vivid imaginative representation of how it would be like to experience
such an event.
To recapitulate, a unified account of imagining has to satisfy two important requirements: (i) it has to get the extension right with respect to the
central instances of imagining and with respect to the paradigm cases of
non-imaginative phenomena (notably cognitive ones); and (ii) it has to be
able to elucidate the imaginativeness essential to all imaginings and, possibly, also some of their other shared features. In the next few chapters, I
spell out the second desideratum in more detail by highlighting important
properties of imaginings that distinguish them from cognitions, and which
are in need of explanation by a unified theory of imagining. After that,
beginning with Chapter 5, I start to to present and assess the various ways
57
CHAPTER 2
Key Features of Imaginative Episodes
The debates about imagining in the literature have focused primarily on the
opposition between imaginative and cognitive phenomena.32 Since this particular contrast is of great significance in the following chapters notably
for the discussion of the Epistemological and the Dependency Accounts
it is helpful to identify and elucidate important differences between imagining and cognition, at least in their episodic form. The elucidation of
these differences also gives further support to, and deepens the understanding of, the distinction between central cases of imaginative phenomena and
central cases of cognitive phenomena, as outlined in the last chapter.
Two particular differences in kind between imaginings and cognitions
have often been noted in the philosophical literature and are of special importance in the subsequent discussion of the various accounts of the unity
of imagining. This is due to the fact that they are relatively fundamental
and concern all episodic forms of imagining. Moreover, the second difference parallels closely that between imaginative and cognitive projects.
The first difference is phenomenologically salient and consists in a difference in attitude, that is, in a difference in the stance or commitment
which the episodes in question adopt towards their content. The idea is that
only cognitive states or cognitions, as I also often say involve a cognitive attitude: only they make a claim about how things actually are.
The second difference becomes manifest in the different justificatory
and motivational power of imaginative and cognitive episodes: while we
32
See Budd (1989): 100, Currie & Ravenscroft (2002): 11, and OShaughnessy
(2000): 362ff. for the claim that the various kinds of imaginings have cognitive
counterparts.
60
usually are entitled and moved to endorse cognitions in belief (and, relatedly, in action), the same is normally false of imaginings. This difference
in rational role is a reflection of a difference in epistemic function or aim:
cognitive episodes possess and imaginative episodes lack the function
to ground or constitute knowledge (and, subsequently, to influence action).
The second difference is therefore comparable to the fact that cognitive
projects but not imaginative ones have the purpose to lead to the acquisition of knowledge (see Section 14.2).
The two differences have in common that they are strict differences, that
is, separate all imaginings from all cognitions. Moreover, they are perhaps
closely connected to each other in the sense that the phenomenologically
salient difference in attitude may count as a conscious reflection or manifestation of their difference in epistemic function (see Section 2.2 below).
For example, that there is a subjective difference in endorsement between
seeing and visualising something may indicate that only the first, but not
the second, has the function to inform us about reality. Such a link between
the phenomenal and the epistemic dimension of the episodes concerned
may also help to explain why we typically rely on cognitions, but not on
imaginings, when forming beliefs and intentions. The thought is that we
(do not) do so because we are phenomenally aware of the (lack of an) epistemic function (see Dorsch (2012a) and (2013)). In any case, given that
the two strict differences characterise all forms of episodic imagining, any
unified account of imagining should as part of the second desideratum introduced in the last chapter be expected to be able to accommodate and
illuminate them.
In the first two sections of this chapter, I zoom in on the specific contrast between imaginative and cognitive episodes and, in particular, on the
significant differences in phenomenal character and epistemic function,
which play an important role throughout this book. The third and final section of the chapter is then concerned with the fact that imaginative episodes
none the less share with cognitive ones the capacity to refer to real (past or
present) entities although it is still true that the two kinds of episode possess this capacity for different reasons, and that a lack of such reference is
the default case for imaginings, but not for cognitions.
61
That I speak of perceptions, memories, judgements or beliefs as making a phenomenologically salient claim about how things actually is not intended to imply
that they represent the modal status of the relevant states of affairs (i.e., that they obtain in the actual world), and that the formalisation of their content requires the use
of a modal operator. Rather, what is meant is that their representation of the relevant
states of affairs (i.e., that they obtain) purports to be concerned with the actual world
and, hence, is to be assessed with respect to that world. The same applies to imaginings and their commitment to some imaginative (or possible) situation or world being a certain way. Rather than saying that cognitive and imaginative episodes differ
in whether they make a claim about how things actually (in contrast to possibly or
imaginatively) are, I may perhaps also have said that they differ in whether they
make a claim about how things are (in contrast to how they could be). One advantage of the way of talking chosen in the main text, however, is that it stays neutral on
whether a commitment to some imagined situation being a certain way needs to be
or even can be spelled out in terms of possibility (see Section 5.4).
62
how things actually are (i.e., which colour my car in fact has). Correspondingly, it is to be evaluated with regard to whether my car is actually blue:
my judgement is incorrect if the car is in fact red. Similar considerations
apply to the corresponding sensory episodes (see Martin (2002a): 391).
Visualising my car as being blue represents it as being blue, but is not committed to its actually being blue. Seeing my car as being blue, on the other
hand, involves such a commitment about the colour of my car. Accordingly, while perceptions are straightforwardly veridical or not with respect
to the actual world, sensory imaginings need not count as faulty if they fail
to match reality.
This difference between imaginative and cognitive representations is a
difference in their attitude or stance, commitment or non-neutrality towards their content.34 More specifically, it can be described in terms of the
attitude distinctive of cognitive states: a mental representation involves a
cognitive attitude if and only if it makes a claim about how things actually
are. Accordingly, the two types of mental representation differ in that cognitive states involve a cognitive attitude, while imaginative episodes do
not.35 I take it that Husserl and Sartre intend to refer to the same difference
when they say that only cognitive states involve a thetic quality (see
Husserl (1901): part V) or a positing attitude (see Sartre (1940): ch. 1).
In the case of judgements, their cognitive commitment or attitude may also be described in terms of their purporting to be true, or their involvement of an endorsement or acknowledgement of their propositional content as true, or of the corresponding state of affairs as being the case. Whether this way of talking extends naturally to perception partly depends on whether perceptual experiences are themselves
propositional, or at least possess a content with a structure very similar to that of
propositional contents (see Crane (1992)). OShaughnessy seems to be among the
very few who do not assume that cognitions and imaginings differ necessarily in
their phenomenal character or their attitude (see OShaughnessy (2000): 345; 349;
355). I discuss this apparent aspect of his view in the last part of Section 7.1. Even
Hume acknowledges that cognitions and imaginings differ both phenomenally and
in their rational or functional role (see Chapter 9).
35
This would remain true even if it would turn out that there are some non-representational imaginings (see Section 14.3 for a discussion and rejection of this possibility): since such imaginings would not have any content, they could not involve any
attitude or commitment relating to a content. In contrast, there should be no doubt
that cognitions are, indeed, always representational. Otherwise, they could not possess their epistemic function or play their actual role in cognition.
34
63
64
tion 5.1).
See Husserl ((1912): 172f.) on the idea that imaginings involve the attitude of
Quasi-Setzung (i.e., quasi-positing). A very similar kind of commitment is involved in pictorial experience and, in some sense, also in depiction itself this time,
with respect to the depicted situation or world (see Dorsch (2010c)).
65
66
not as part of the very same imaginative project), as well as with the demand to believe that there is an elephant present in the imagined situation
(if suitably prompted). Both demands, however, depend on the respective
episode of visualising being non-neutral about the presence of an elephant
in the imagined situation. For example, there is a tension between visualising an elephant in the city and, as part of the very same imaginative project, imagining that there is no elephant in the city precisely because both
imaginative episodes are committed to the imagined situation being a certain way (and because these two ways are incompatible with each other).
Similarly, it is right to believe that the imagined world contains an elephant
(rather than a tiger) because this is how the episode of visualising presents
the imagined situation to be like.
Finally, the third fact in need of explanation is that imagining may be
central to our acquisition of knowledge say, about possibilities or conditional truths.41 In his most recent book, Timothy Williamson argues that
visualising is one of the many empirical capacities that we may employ in
order to acquire modal knowledge or, in his concrete example, knowledge of some conditional, which forms the first step to modal knowledge.
Considering a situation in the mountains, he describes the largely non-inferential process involved in coming to know that a certain rock would
have landed in a lake, if its path had not been blocked by a bush:
You notice one rock slide into a bush. You wonder where it would have
ended if the bush had not been there. A natural way to answer the question
is by visualising the rock sliding without the bush there, then bouncing
down the slope. You thereby come to know this counterfactual: [...] If the
bush had not been there, the rock would have ended in the lake. (Williamson (2008): 142)
The explanans has also to include the assumption that imaginings can refer directly or indirectly to aspects of the actual world. Besides, I argue in Chapter 4 that
imagining may also be central to the acquisition of knowledge about actual and contingent facts (e.g., whether a sofa is smaller than a door) for instance, when mentally rotating a piece of furniture in a shop in order to see whether it is possible to
get it through ones front door at home (see also Martin (2002a): 412 and 414).
67
gined situation. But this requires, again, that our relevant episodes of imagining are committed to their presence or at least to the presence of some
objects instantiating the respective properties in that situation.
The issue of which property instances our episodes of sensory imagining can actually represent and be non-neutral about is thereby controversial. One option is to maintain that we are able, say, to visualise (and not
just think of) an object as being a rock; another to insist that we are instead
limited to merely visualising it as having an approximately round shape
and grey colour (and to taking it to be a rock in additional imaginative
thought). The very same issue arises with respect to perception and depiction for instance, whether we can see or visually depict also the property
of being a rock, or only the properties of being (approximately) round and
grey. Here is, however, not the place to settle this issue. 42 In what follows, I
assume that sensory representations including sensory imaginings can
represent and be non-neutral about the instantiation of certain higher-level
properties, such as being a bottle, a tiger, a lemon, or a lake. If the respective examples turn out to be problematic, it should be easy to replace them
with less controversial cases, such as the sensory representation of triangularity or sphericality.
Now, according to the preceding considerations, what perceiving,
judging and imagining have in common is that they are non-neutral about
how things are. They just differ in whether they are concerned with the actual world or, alternatively, with some imagined situation. In other words,
they differ in which particular theoretical attitude they posses: a cognitive
or an imaginative one. By contrast, conative and similar mental representations (such as desires and intentions) show a practical attitude. Although it
might perhaps be suggested that they, too, are non-neutral about how
things are in some non-actual (i.e., desired) situation or world, this is false
in the light of the fact that they are normally concerned with the actual
world. When desiring an ice-cream or that there is peace, we are not concerned with what is true in some possible situation. Rather, we want reality
to be a certain way: that is, we want real ice-cream and real peace. But, of
course, conative representations do not make a claim about how things are
in reality. Instead, they involve a claim about how things should be in the
42
See, for instance, Millar (2010), Hawley and Macpherson (2011) and Dorsch
(2012b) for further discussion.
68
actual (or perhaps also some other) world. That is, they possess a practical
rather than a theoretical attitude.43
Imagined (or, if one prefers, fictional) worlds resemble possible worlds,
not the least in consisting of propositions or states of affairs, and in being
non-actual or not matching reality. But there are two reasons why it is perhaps better to resist identifying imagined worlds with possible worlds.
First, imagined worlds may be incomplete or indeterminate in ways that
possible worlds cannot. For instance, an imaginative project concerned
with the adventures of Sherlock Holmes (as described in Conan Doyles
stories) may stay neutral on the colour of his eyes. As a result, the imagined world contains the state of affairs that he has eyes, and even the state
of affairs that his eyes have a determinate colour (e.g., this may be an implicit assumption of the stories and, hence, the project that responds to
them), but no state of affairs linking his eyes with the instantiation of a particular colour. All possible worlds that contain Sherlock Holmes, on the
other hand, are determinate with respect to which colour his eyes posses.
Second, it seems possible to imagine impossibilities if not within a single
episodic representation, then at least as part of an imaginative project or
another complex of imaginative representations (see Section 5.4).44
69
70
need not lead to the belief that it is raining. But it disposes us to form this
belief. Hence, as soon as you ask me how the weather is like, my perception of the rain will move me to believe that it is raining (and to express
this belief to you by saying it is raining). Visualising rain, on the other
hand, does not have the same effect: it does not by itself incline us to believe that it is raining. More generally, imaginings do not on their own motivate us to endorse them in belief.
The two aspects of the epistemic function of cognitions are intimately
connected. That cognitive episodes incline us to form certain first-order beliefs is a manifestation of their epistemic authority over us in theoretical
deliberations. But this does not mean that we cannot sometimes fail to endorse a cognition. For cognitive episodes enjoy this authority only as long
as we take them to provide us with reasons for the respective beliefs. Once
we start to seriously doubt the justificatory power of cognitions or the suitability of the circumstances of their occurrence (e.g., because we receive
evidence about their non-veridicality), we stop following them as our
guides in belief formation. For example, we cease to rely on our perceptual
experience as of rain when we begin to question its status as veridical perception.47 The same applies to a belief, the rational grounding of which we
mistrust for whatever reason. On the other hand, the difference in epistemic
function does not rule out the possibility that we may endorse some imaginings in belief and subsequently rely on them in action in roughly the
same way in which we endorse and rely on perceptual experiences (I discuss this possibility in Chapter 4).
Hence, an imaginative episode, which lacks any proper epistemic function, may none the less move and entitle us to form certain beliefs under
special circumstances; while a cognitive episode, which possesses the epistemic function described, may sometimes fail to live up to its function and
not motivate or justify us. The corresponding difference between cognitions and imaginings is still strict, however: every cognitive episode but
47
71
72
cies to successfully interact with the world and survive in it. Compare
again the analogy with assertions: that we usually trust the assertions of
other people may be due to the fact that it is their underlying (social) function to express reliable and sincere views which ensures that (or which the
assertions possess only if) they usually are trustworthy. If most or all
people began to lie, assertions would lose both their trustworthiness and
their (social) function to express sincere opinions. But the introduction of a
biological function of cognitions also faces some difficulties, notably the
problem of specifying precisely when a cognition fails to fulfil its function
(e.g., which environments are relevant for the acquisition and utilisation of
the function). Hence, I aim to stay neutral on whether the epistemic function of cognitions may be accounted for in evolutionary terms.
Finally, it seems natural to assume that the strict difference in epistemic
function is closely linked to the difference in cognitive attitude. Otherwise,
their coextensionality would remain unexplained. This issue arises, in the
first place, because the two differences are not identical. And they are not
identical because the epistemic function of an episode and its cognitive attitude are different kinds of properties. Looking outside through a window
and seeing rain has the power both to move us to believe that it is raining,
and to justify the resulting belief. But this power need not become realised
under all circumstances. We may not form the belief because a trustworthy
friend may inform us that there is someone pouring water from above (e.g.,
for a film which is currently shot outside of the building), or because we
simply fail to consider the issue of what the weather is like (e.g., all our attention may be occupied by a problem or a reverie). And, even if we form
the belief that it is raining, it may not be justified by our perception because we ignore the defeating testimony of our friend, or because we have
formed the belief completely independently of the perception (e.g., in response to hearing the weather forecast in the radio). Making a claim about
how things are, on the other hand, does not allow for the same distinction
between a power and its realisation. Our perception tells us that it is raining
outside independently of the context for instance, independently of
whether we consider the issue, and of what other evidence is available to
us.
But if the two differences are not identical, we need to identify another
73
intimate connection between them. The most promising proposal is perhaps to maintain that the difference in attitude renders the difference in epistemic function to some extent phenomenologically salient. We already encountered a way in which part of the epistemic function becomes manifest:
namely in its concrete realisation, that is, in the actual formation (and justification) of a first-order belief in response to the cognitive episode concerned. But this need not be the only way in which an epistemic function
of an episode may become salient. Another possibility is that it contributes
to the determination of the phenomenal character of the episode. As a result, episodes with different attitudes will also differ in phenomenal character. In the case of cognitions and imaginings, this phenomenal difference
may be identical with the difference in attitude. That is, the possession of a
cognitive attitude may be a phenomenologically salient reflection of the
possession of the function to constitute knowledge.50
74
ies and use it for, or include it in, his imaginative representation. Perhaps
one might want to maintain that only those aspects of the content of his
representation, which are not remembered or known, constitute the imaginative content. However, the imaginative attitude and lack of epistemic
function involved characterise the representation as a whole, and not only
some parts of it. This also fits the fact that, in our practice and talk, we take
the complex representation to constitute a single and imaginative mental phenomenon: we assume that the person imagines his friend as being at
the opera.
Furthermore, there is the possibility of cases in which an imagining is
made up entirely of material provided by past memories or knowledge. 52
The result of, say, using ones visual memory of the friend in order to picture him sitting in an equally remembered opera house is clearly not a visual memory, but a visual imagining (e.g., it differs from the case of the two
memories occurring separately, but temporally close to each other). If,
however, the mnemonic parts of the representation would not count as constituting the imagined content, the imagining would not have any, given
that the imaginative combination of the two remembered contents does not
add any further sensory material. Besides, it is not merely the combination
which is imaginative, but also the resulting episodic representation not
only the act of combining the image of the friend and the opera house, but
also the resulting sensory image of the friend in the opera house. Hence,
the complexity of imaginings, or their involvement of cognitive elements,
is not a good reason to deny their potential reference to reality.53
In addition, we can imagine existing objects as having features which
they really have. The subject may imagine a truth by imagining that his
friend is visiting the opera; and he may visualise the friends and the opTo some extent, this is in fact true of all sensory imaginings though only with respect to each of the represented features taken on its own, and not with respect to the
objects which these features may characterise or constitute. That is, we may not be
able to visualise something red without having seen red before; but we can visualise
a red cloud, even though we may have not already seen one.
53
Compare also the discussion in the Sections 3.1 and 3.3, in particular of the point
that the obtaining of a direct causal link between the referent and the representation
is not a necessary condition on reference or factivity: while this may be true in the
case of perceptions, it is not true in the case of thoughts, judgements, or feelings of
desire or emotion.
52
75
76
77
10.4).
Besides, sensory or at least visual imaginings are special in that they
allow for two kinds of reference to objects, namely not only to particular
objects (e.g., that friend), but also to types or sorts of objects (e.g., some
man or another):
If asked to visualise a green apple, you may well succeed in bringing to
mind an image of an apple. But, in many contexts, it is simply inappropriate
to press the question which apple you have imagined. In visualising an
apple, there need be no particular apple which is imagined.
This is not to say that particularity cannot be injected into our imagery.
One can, if one wants, imagine the very green apple now nestling in A. A.
Gills pocket. For we can use imagery for particular imaginative projects,
for particular tasks; and in such cases it is entirely appropriate to take the
imagery to be the imagining of the particular objects or events that one sets
out to have in mind. (Martin (2001): 275)
Indeed, the first type of sensory imagining is more fundamental than the
latter54, given that the default case is to imagine something about entities
which we do not assume to have a specific identity (let alone to exist).
Simply visualising an apple means visualising some apple of a certain type,
but no apple in particular. By contrast, visualising a specific apple such
as the one in A. A. Gills pocket requires some additional representation
of numerical identity of the visualised apple. This raises the question of
how precisely particularity becomes injected into our imagery and,
more generally, how the reference of our sensory imaginings is fixed.55 I
return to this issue in Section 3.6 on the presence of an intellectual element
in sensory imagining and its significance for the latter. Before that, it is
helpful to discuss two other, closely related features of sensory imaginings
that distinguish them strictly from perceptions (and, to some extent, also
sensory memories).
Just as depictions of types of object are more fundamental than depictions of
particular objects (see Wollheim (2003) and Dorsch (2012c); see also Note 11 in
Chapter 9). The related idea that sensory imaginings are representations (or
echoes) of types of perception and, therefore, also of types of perceived objects, is
discussed in Chapter 10.
55
The same question is not pressing in the case of intellectual imaginings since they do
not differ from other kinds of thought in their referentiality.
54
CHAPTER 3
Key Features of Sensory Imaginings
The features noted so far are central to and, to some extent, also distinctive
of episodic imagining, independently of whether it is sensory, affective or
intellectual. But there are also important differences between sensory imaginings and perceptions, which have notably been identified and described
by philosophers in the phenomenological tradition. These differences have
no direct relevance for the formulation of a unified account of imagining,
since most of them do not pertain to intellectual or non-episodic forms of
imagining, and since some group sensory memories together with sensory
imaginings, rather than with perceptions (see below; and, for an explicit
example, Sartre (1940): 10).56 None the less, it is worthwhile to gain some
understanding of them not only because they are philosophically interesting in their own right and figure prominently in the debates on imagining,
but also because the ability to capture and explain these features may separate otherwise equally adequate unified accounts of imagining and identify those among them eventually to be preferred over the others.
Perhaps following Hume in this, Husserl has noted two respects in
56
I stay largely neutral on the issue of whether the features to be discussed distinguish
sensory memories from sensory imaginings, or rather from perceptions. Sartre
(1940: ch. 1), for instance, treats memories as one kind of image, and imaginings
as another, contrasting both with perceptions; and this seems to be true with respect
to vivacity or immediacy (see Section 9.4). But it is more controversial to assume,
say, that we are immune to error through misidentification with respect to the identity of sensorily recalled objects, or that sensory memories are quasi-observational
(e.g., do not allow for knowledge-increasing changes in perspective onto the remembered objects). However, since nothing in what follows depends on settling
these issue, I do not pursue them here any further.
80
Interestingly, both Hume and Husserl do not always seem to clearly distinguish the
differences in vivacity and immediacy perhaps because of the possibility of explaining the former in terms of the latter (see below), or perhaps also because of
their implicit acknowledgment that the difference between imaginings and cognitions is qualitative, and not quantitative. Accordingly, a difference in vivacity can
distinguish the two kinds of episode only if it does not come in degrees, or if it can
at least be traced back to a difference which does not come in degrees, such as immediacy (see Section 9.1f. for further discussion).
81
82
83
3.2. No Immediacy
That sensory imaginings (as well as sensory memories) lack the immediacy
of perceptual experiences means, first of all, that they do not present their
objects as being there before us in our actual environment. When we see a
tree, it seems to be right there before our eyes. But when we visualise a
tree, we do not have a similar impression of its presence in our actual environment. This is true even if we imagine the tree to be part of the real
landscape in front of us: the perceptual experience of the landscape and the
imaginative experience of the tree do not merge in such a way that the tree
appears to be actually present in the same way as the landscape (see
Dorsch (2012c)). In Husserls words, the tree is not bodily present, but
instead merely sensorily re-presented (see Husserl (1905): sec. 42; see
also Martin (2001)).
It is important to distinguish immediacy from the possession of a cognitive attitude. While memories involve a cognitive attitude, they lack immediacy: when I recall the appearance of the tree that used to stand in our old
garden, but got cut down in the meantime, the visually represented tree
does not thereby begin to appear to be present in my actual environment.
Similarly, judgemental thoughts make a claim about how things are in the
actual world, but do not relate us in an immediate way to the external objects concerned. Hence, the possession of a cognitive attitude does not require immediacy. The explanation of this is that while the last-mentioned
feature concerns the issue of whether we are in direct contact with the represented entities, the first-mentioned feature pertains to the issue of whether the mental representation in question involves a claim about how things
actually are.
Whether immediacy presupposes or implies a cognitive attitude is more
controversial. While some think that immediacy gives rise to an epistemic
84
commitment about how things (really) are (see Martin (2002a)), or at least
contributes to such an attitude in so far as it identifies the world, that the
commitment is concerned with, as the actual world (see Dorsch (2010a)
and (2010c)), others deny any such link to theoretical non-neutrality (see
Husserl (1901): 442). One underlying issue is whether immediacy is part of
the attitude (i.e., phenomenologically salient commitment) of the mental
episode concerned, or instead part of its content (i.e., the phenomenologically salient manner in which things are presented). In this respect, its status
is similarly unclear as that of, say, the visual or auditory character of sensory representations (see Section 1.2).
Completely independently of this issue, the difference in immediacy is
perhaps best understood in terms of the idea that sensory imaginings are
re-presentations or re-productions in so far as they inherit their sensory
character and content from an underlying perception (i.e., a presentation
or original production), to which they stand in some constitutive and representational relation. Indeed, the just noted participants in the debate concerning the relationship between immediacy and commitment agree on this
point (see Husserl (2006) and Martin (2002a), as well as Chapter 10). That
sensory imaginings are re-presentations of perceptual experiences may
mean, for instance, that they are instances of experiential imagining: that
is, indirect representations of objects and their features in virtue of being
direct representations of (types of) perceptions with a corresponding content. This representational indirectness can then explain why, or in which
sense, imaginings do not mark their external objects as actually present in
our environment. I discuss this proposal in more detail in the context of the
discussion of the representational version of the Dependency Account and,
notably, its idea of a dependency of imaginings on cognitions.58
In addition, the idea of a difference in immediacy may be used to account for the typical difference in vivacity. The core idea is that, because
sensory imaginings and perhaps also sensory memories, for that matter
are taken to be some kind of reproduction of the original sensory impres58
85
sions, they may be expected to be normally less vivid (in either of the two
senses distinguished above) than the latter. This may, again, be said to be
due to the fact that sensory imaginings do not literally possess the kind of
experiential qualities pertaining to perceptions and bodily sensations (e.g.,
the phenomenal qualities constitutive of being a painful experience, or an
experience of something red), but instead merely represent instances of
these qualities (see Sections 10.1 and 11.5). Besides, the idea of immediacy
is also important for the ability of the Agency Account to accommodate
imaginings which might seem to be non-representational (see Section
14.3).
3.3. Indeterminacy
Casey proposes another though, as it seems, closely related strict difference between sensory imaginings and cognitions. He maintains that imaginings are necessarily indeterminate in the sense that they can never be
perfectly determinate, nor allow for an improvement in determinacy
(see Casey (1976): 104-106, 110). Although Casey does not explicitly restrict his comments to sensory imaginings, all his examples concern only
sensory cases. Indeed, it would be difficult to try to apply his ideas to intellectual representations. His description of what he means by indeterminacy is not always entirely clear, despite its extensiveness. Most of his examples concern representational and possibly also non-representational 59
aspects of sensory imaginings that are phenomenologically salient, and
which involve differences in vivacity (see Casey (1976): 36f., 105-109).
This suggests that indeterminacy may be the same as a lack of vivacity.
But then, his claim that sensory imaginings are always indeterminate becomes untenable in the light of the previous considerations. Moreover, Caseys only apparent argument in support of this claim (see Casey (1976):
105f.) turns out to be unsuccessful.
He acknowledges that perception can be as indeterminate as sensory
imaginings for instance, if occurring, in the visual case, under bad lighting conditions. His example consists of a perception of a statue of a woman
59
For instance, being blurry is not obviously representational (see Casey (1976): 106;
Peacocke (1983): ch. 1; Soldati & Dorsch (2011)).
86
at dusk which does not represent where exactly her arm emerges from her
torso. But, he maintains, we can none the less investigate the perceived object further (e.g., in improved conditions) to find out about its determinate nature or at least to get a determinate perception of it. That is, we
can see (or feel or otherwise discover) at which precise place the statues
arm begins. In contrast, the same is not possible with respect to a given
sensory imagining and its object: neither can be made more determinate
as it is. Caseys main idea is here, of course, that we cannot wait for better
conditions, or use other cognitive means, to further investigate the nature
of a sensorily imagined statue.60 Therefore, as Casey concludes, an
imagined object is inherently indeterminate and can never be apprehended
as perfectly definite (see Casey (1976): 106).
Apart from the problematic switch between the representations and their
objects as the bearers of the feature of indeterminacy, the main problem
with this argumentation is that it does not sufficiently distinguish the claim
that we cannot make an already indeterminate sensory imagining more
determinate from the thesis that there cannot be determinate sensory imaginings. The latter is what he really needs to argue for; but his argument
can at best support the former. In particular, it still remains a possibility to
deliberately try to improve the determinacy of ones sensory imaginings.
For instance, we can decide that the visualised statues arm should begin at
a certain point of its torso and then act on the resulting intention by sensorily imagining it to start there. We may perhaps not always succeed in such
a task for some contingent reasons for instance, because we are distracted, or because the task is too ambitious for our ability to visualise. But we
can at least in principle fill in the details missing in a given image. Casey
might presumably rightly insist that the alteration brings about a new
episode and hence cannot count as the determination of a previously indeterminate imagining. But he does not show that we cannot produce determinate sensory imaginings in the first place at least not if determin60
Caseys indeterminacy is therefore perhaps also similar to Sartres quasi-observationality. In this context, Casey actually quotes Sartre as writing: A hare which is
vaguely perceived is in itself a determinate hare. But a hare which is the object of a
vague image is an indeterminate hare (see Casey (1976): 106; see Sartre (1940):
16). But again, this observation does not show that there cannot be precise or determinate images.
87
3.4. Quasi-Observationality
Two further characteristic features of sensory imaginings, which play a
central role in Sartres and in Wittgensteins writings, differentiate them
strictly from perceptions (and, arguably, sensory memories). The first feature is that sensory imaginings are, as Sartre calls it, merely quasi-observational (see Sartre (1940): 8ff.; see also Wittgenstein (1984d): sec. 63, 80
and 88).61 Sensory imaginings appear to be observational in so far as they
61
As with his general discussion, Sartre focuses solely on cases of visualising. But I
assume here that his considerations may apply equally to other form of sensory imagining.
88
One thing to note in this passage is that, since Sartre refers with [turning] a cubeimage in thought to the mental rotation of a visualised cube, thought includes conscious operations of the sensory imagination. The other thing to be highlighted is
that he uses the term image to establish a contrast both with perceptual experiences
and with the perceived objects (see also his comment that image is ambiguous in
simultaneously [signifying] a consciousness and its transcendent correlate (Sartre
(1940): 148)). That is, image refers both to sensory imagining and to its object,
thus reflecting Sartres view that the sensorily imagined objects are entirely determined (or even constituted) by our imaginative awareness of them (see Section 3.6). It
is also worthwhile to point out that he generally uses the term image to cover not
only sensory imaginings, but also sensory memories (and that the latter are also
quasi-observational and immune to error through misidentification; see Sartre
(1940): 8f.).
89
whole, for what it is, in its appearance. If you turn a cube-image in thought
to amuse yourself, if you pretend that it presents its various faces to you,
then you will not be more advanced at the end of the operation: you will not
have learned anything. (Sartre (1940): 8f.)
90
same when seen from a particular point of view. Hence, coming to know
whether the seen object is one or the other requires the adoption of a
different perspective onto it.
It is important to note that the quasi-observationality of sensory imaginings extends beyond features that are open to sensory representation and
the related sensory non-neutrality. Real lemons and perfect wax replicas of
lemon share the same visual appearance, at least from the outside. What it
means for the replicas to be perfect is exactly that there is no point of view
from which we can visually discriminate them from real lemons by looking
at their surfaces. There should be no doubt that we can perceive the property of having the visual appearance of a lemon, given that the latter
amounts to nothing more than a (complex) set of of colours, shapes, sizes,
textures, and so on (see Martin (2010) and Dorsch (2010d)). Whether the
property of being a (genuine) lemon is perceivable, that is, visually or otherwise sensorily representable, on the other hand, is debatable (see Section
2.1)). But even presuming that we can indeed see the lemon-hood of an object, the property of being a perfect replica of a real lemon should not be
assumed to be visually perceivable.
The reason for this is that the visual appearance shared by both the lemons and their replicas can be indicative of the presence of only one of
them. That is, recognising this appearance can help us to visually discriminate at best one of the two kinds of object (see Millar (2010), Martin
(2010) and Dorsch (2010d)). If you encounter an object with a light yellowish colour and an egg-like shape, of the size of an apple and with a porous texture, seeing this visual appearance may be enough to see it as being
a lemon primarily because this appearance is characteristic of lemons in
the sense that most objects bearing it are in fact lemons. But it cannot also
be enough to see it as being a perfect replica of a lemon partly because
hardly any objects with this appearances are such replicas, but also partly
because one and the same visual appearance cannot be indicative of two incompatible properties (i.e., one and the same perception cannot present an
object as having two properties that exclude each other).
Now, these considerations about perceivability apply also to imaginabiltially in whether they provide us with (indefeasible) justification for perceptual belief (see Dorsch (2011a)). While Sartre seems to reject this kind of disjunctivism
(see the quoted passage), McDowell (1994) endorses it, for instance.
91
ity (and depictability, for that matter). We may be able to imagine an object
as being a lemon simply by visually representing it as having the visual appearance of a lemon (i.e., as being yellow, egg-shaped, apple-sized, and so
on) meaning that the resulting representation of lemon-hood would be
entirely visual. But we cannot represent something as being a perfect replica of a lemon in a similarly visual manner. Instead, what we have to do is
to visually represent it as having the appearance of a lemon and then intellectually representing it to be merely a replica, rather than the real thing
(see Dorsch (2010c)). None the less, visually imagining a replica of a lemon is as much quasi-observational as visually imagining a lemon. In both
cases, our knowledge of the kind of object that we are visualising that it
is a wax replica or, alternatively, a real lemon is not open to defeat, especially not by evidence gained through visualising the object from a different perspective. To some extent, this claim is trivial: given that genuine
lemons and their replicas look exactly the same from the outside, we cannot tell, by changing our spatial point of view on it, whether something belongs to one kind of object or the other. But the claim is true even with respect to cases in which we, say, imaginatively slice open the objects and
begin to visually represent their insides. I return to this example in Section
3.6.
However, perceptual observation is not limited to looking at objects
from different points of views. We may also perceptually observe an object
by paying more attention to its details, which may happen without any
change in our perspective onto the object. For instance, we may come to
notice the slight cracks in a wall while continuing to stare at it from one
and the same point of view. Correspondingly, Sartre assumes that our
knowledge about which kinds of object we are sensorily imagining is indefeasible not only with respect to changes in perspective, but also with respect to paying closer attention to detail. That is, he assumes that there is a
further way in which sensory imaginings count as quasi-observational,
namely relative to changes in attention. For instance, our belief that what
we are visualising is a wall without any cracks could not be undermined,
say, by us happening to discover some cracks in the visualised wall. For,
according to Sartre, we could not make such a kind of discovery with respect to visualised objects. Here is how he describes this contrast between
92
In the last paragraph, Sartre also hints at an explanation of the indefeasibility of our knowledge of which kinds of object we are sensorily imagining:
it is due to the fact that sensory imagining involves an intellectual element
which provides us with indefeasible knowledge about what is imagined. 65 I
return to this proposal in Section 3.6, which is specifically devoted to the
discussion of the nature and role of the intellectual element involved in
sensory imagining.
For the time being, it is worthwhile to point out that there is good reason
to think that the quasi-observationality of sensory imaginings does not extend to shifts in attention (contrary, say, to what Hopkins (2011b) argues).
In Sections 4.6 and 7.2, I argue that sensory imagining allows for the involvement of object-directed attention and for the acquisition of new evidence concerning the nature of the represented objects in the same way as
65
This difference between visualising and seeing is, for Sartre, furthermore reflected in
the different metaphysical status of the two kinds of represented objects. The link is
established by Sartres view that sensorily imagined objects are constituted (i.e., determined both in their nature and in their existence) by our imaginative awareness of
them with the result, for instance, that they are not presented as being in the same
space or world as perceived objects (see Sartre (1940): 11ff., and also Sections 3.6
and 5.3 below).
93
94
might be accounted for with respect to imperceivable features will be addressed towards the end of Section 3.6.
The absurdity, the irrelevance of likeness and the example about Kings College
have been introduced in Wittgenstein (1984b): 68. They are discussed and supplemented with the irrelevance of causal origin in Peacocke (1985): 20 and 26f.. See
also Sartre (1940): 8ff., although he does not always clearly distinguish between the
immunity to error and the quasi-observationality of sensory imaginings.
95
Hertford College.
The same is obviously not true of perceptions (as well as sensory
memories). First, we often are not sure about the identity of the objects that
we are seeing (e.g., in the dark, or in heavy rain). And second, even if we
are sure about their identity (e.g., perhaps, when they seem to tell us
straightaway what they represent), it is not guaranteed that we can never be
mistaken. This is due to the fact that which objects we are perceiving is determined by which objects have caused our perceptions, or by which objects possess the properties that we are perceiving as being instantiated by
the perceived objects (see Dretske (1986)). Hence, given that we may err
about which external objects have actually caused our perception to occur,
or which external objects actually possess the features that we see the perceived objects as having, we may also go wrong in our identification of the
perceived objects. Such misidentifications occur, for instance, when we fail
to recognise: (i) that we are merely hallucinating the presence of a particular person; or (ii) that we mistake a particular person for her twin (or fall
victim to a cleverly disguised imposter); or (iii) that we simply misperceive
the properties of the person in front of us (e.g., her height, hair colour,
shape of face, etc.) and, as a result, take her to be someone else.
It is important to note that sensory imaginings are immune to error
through misidentification of the represented objects only to the extent to
which they do represent particular objects. If it is indeed true that sensory
imagining is, by default, concerned with the representation of types of objects rather than specific tokens (i.e., possesses a general rather than a particular content; see Section 2.3), then the issue of whether we can err with
respect to the identity of the sensorily imagined objects does not arise in
default cases (i.e., with respect to the sensory imagination as such). Instead, it becomes an issue only in non-default cases, in which the sensory
representation is supplemented by some intellectual element which determines the particular identity of the imagined objects. When we visualise
some man, but no man in particular, it is a trivial observation that there is
no question of whether we might be mistaken with respect to the identity
of what we are visualising there is nothing to be wrong about in the first
place. Only once we visualise a particular man perhaps because we think
of what we are visualising as being a specific man does it become a sub-
96
stantial claim to maintain that we are immune to error through misidentification with respect to which object we are visualising.
3.6. Intellectuality
According to both Peacocke and Sartre, this immunity to error through
misidentification and the related absurdity of asking whether one might be
mistaken about what one is currently sensorily imagining can be properly
accounted for only by assuming that sensory imagining involves an intellectual element specifying what is sensorily represented.67
97
resented object (e.g., that it is Kings College, or a suitcase with a cat behind it).
Another aspect of S-imagining is that it ascribes if not explicitly, then
at least implicitly (see ibid.: 26) the sensory content of the real imaginative episode to an imagined perceptual experience of a certain kind (e.g., a
veridical or hallucinatory one; one had by a person different from oneself;
etc.). Moreover, it may but need not further identify or describe this
imagined experience, as well as the subject which is imagined to undergo
it. For example, it may be S-imagined that the imaginative sensory representation of Kings College is in fact identical with one of Wittgensteins
perceptions of the college that he (could have) had in the past. But the two
noted aspects of S-imagining the identification and description of the
sensorily imagined objects and the identification and description of an imagined experience of these objects are independent of each other. One
may, for instance, agree with Peacocke that the identity of sensorily imagined objects is due to some additional imaginative thought, without assuming this thought to be concerned with some imagined perceptual experience.69
The claim that sensory imagining involves imagining a perceptual experience in virtue of S-imagining the sensory content of the real imaginative episode to be the content of an imagined perceptual experience might
seem closely related to or even identical with the claim that sensory
imagining involves imagining a perceptual experience in virtue of being an
instance of experiential imagining (see Chapter 10). But such an impression would be misleading and Peacocke, who endorses both, distinguishes them clearly (see Peacocke (1985): 22f. and 25f.), More specifically, the two claims are independent of each other (i.e., one may be true or
accepted without the other being true or accepted), not the least because
they concern different aspects of sensory imagination: while the first states
something about the content of the intellectual representation supposedly
69
It would be much more difficult, however, to eschew the de re aboutness and the
consequential dependence of the thought element on the episode of sensory imagining. For there is no other obvious way of ensuring that the sensory and the intellectual representations are concerned with the same object (and the same possible or
fictional world) and, indeed, that the sensory representation is particular in the first
place.
98
involved in sensorily imagining something (i.e., that it propositionally represents the identity of a real with an imagined experience), the second
states something about the content of the sensory representation involved
in sensorily imagining something (i.e., that it experientially or objectually
represents a perceptual experience with a certain sensory content). For instance, Noordhof (2002) is sympathetic to the idea that sensory imaginings
may include or be accompanied by imaginative thoughts that assume them
to share their sensory content with some imagined perceptual experience,
but denies that the sensory representation itself is a matter of imagining or
representing a perceptual experience.
Returning to Peacockes view, it is not absolutely clear whether Peacocke maintains that all instances of sensory imagining involve S-imagining even if they are all instances of experiential imagining.70 The crucial
issue here is what, for him, imagining having a visual experience as of a
tree, say, requires. According to Peacocke, both imagining seeing a tree and
imagining visually hallucinating a tree have to involve a dependent imaginative thought specifying that the imagined experience is, respectively,
perceptual or hallucinatory. By contrast, when we imagine having a visual
experience as of a tree, the status of the imagined experience is left open
(ibid.: 25). But Peacocke does not say whether he takes this neutral stance
to require a corresponding imaginative thought characterising the imagined
experience simply as a (possibly veridical or hallucinatory) perceptual experience; or whether he instead takes this neutral stance to be the default
case of experiential imagining (i.e., what happens when we experientially
imagine an experience without any additional dependent thought about it in
mind).
This unclarity about Peacockes view is closely related to a similar unclarity about Noordhofs multiple use thesis, according to which [t]he
same mental image may be used to fulfil different imaginative projects
(Noordhof (2002): 428). Indeed, Noordhof introduces this thesis in order to
70
Remember that the first function of S-imagining namely to identify the sensorily
imagined objects does not ensure that all instances of sensory imagining involve
S-imagining, given that sensory imagining may be concerned merely with types of
objects, rather than particular objects (see Section 2.3). Hence, any universality of
the involvement of S-imagining would have to be due to its second function
namely to identify the (experientially) imagined perceptual experience.
99
100
the case of visualising a closed cube. For the time being, it suffices to note
that the multiple use thesis is best understood as the claim that the same
(mereologically simple) episodes of visualising that is, self-sufficient
imaginative visual representations may figure in different (complex) imaginative projects.73 It is therefore to be distinguished from the claim that
different (simple) forms of sensory imagining, which always combine a
imagistic with a thought-like element, may involve the same
(mereologically dependent) imagistic element. Similarly, the thesis that all
instances of sensory imagining or of experiential imagining, for that
matter involve S-imagining seems to be too strong.74
101
ledge, which is indissolubly linked to the intention, specifies that the object
is such or such, adds determinations synthetically. To constitute as an image
in oneself a certain consciousness of the table is at the same time to constitute the table as an object of imaging consciousness. The object as imaged
is therefore contemporary with the consciousness that I have of it and it is
exactly determined by that consciousness: it includes in itself nothing but
what I am conscious of; but, inversely, everything that constitutes my consciousness finds its correlate in the object. My knowledge is nothing other
than knowledge of the object, knowledge concerning the object. In the act
of consciousness, the representative element and the knowledge element are
linked in a synthetic act. The correlative object of this act is therefore constituted as a concrete, sensible object and at the same time as an object of
knowledge. (Sartre (1940): 11)
102
103
104
105
Of course, the identification of the intellectual element with an intention-in-action presupposes that sensory imagining is an instance of agency.
But, as I argue in Part Four of this book, this assumption should be accepted because it enables us to account best for the unity of imagining. The
Agency Account defended there is, however, also compatible with the alternative proposal put forward by Peacocke: namely to take the intellectual
element to be an imaginative thought (or the imaginative employment of a
concept) which accompanies or even constitutes or penetrates the instance of sensory imagination in question. This compatibility obtains
especially if the imaginative thought is itself assumed to be determined (at
least normally) by the imaginative agency concerned, that is, by the operational intention-in-action.
Whether the intellectual element is understood as being an imaginative
thought or an intention-in-action as well as whether it is understood as a
genuine constituent or a mere accompaniment of the sensory imaginings
concerned also bears on the issue of whether sensory imagining with an
intellectual element may still be episodic in character, or instead constitutes
a (relatively simple) mental project (e.g., one of imagining having a certain
experience; see Section 1.2). The intentions-in-action involved in mental
actions are constituents of the latter, but do not undermine their episodic
character. Actively thinking a thought does not constitute a (non-episodic)
mental project, despite involving both an intellectual representation and an
intention-in-action or a similar element of activity (see Section 13.4).
Hence, equating the intellectual element with an intention-in-action still allows the instance of active imagining concerned to be a mental episode. By
contrast, if the intellectual element is identified with a merely accompanying thought, the sensory representation and the intellectual element form a
complex mental project, rather than a single episode: the thought is simply
added to the episode of sensory imagining (see Dorsch (2010c)).81
intention-in-action or volition). For if it is assumed that what we sensorily imagine is determined by our own mental agency (rather than by our (active or passive)
thoughts), there is no reason to privilege prior intentions over intentions-in-action.
81
This is true even if the thought element depends on the sensory representation in so
far as the former is a de re representation of the latter (see below, as well as Peacocke ((1985): 25f.) on S-imagining, and Dorsch (2012c)). Note that the situation
under discussion is different from the case of a cognitively (or imaginatively) penet-
106
107
108
tual element may equally be applied to the particularity of sensory imaginings, as well as to their capacity to refer to actual entities; but also that it
does not matter for the success of these further applications whether the intellectual element is assumed to be an imaginative thought or an intentionin-action. But differences in explanatory power appear to surface once the
focus is shifted to the quasi-observationality of sensory imaginings that
is, to the fact that (some of) our knowledge of which kinds of object we are
sensorily imagining from a given perspective is not open to defeat by sensorily imagining the same objects from different perspectives. More specifically, the explanatory difference becomes apparent with respect to cases
where our indefeasible knowledge is concerned with the instantiation of
imperceivable features.
Consider the case in which you are visualising a lemon: you intend or
take yourself to imagine such a kind of object, and your imaginative episode involves the visual representation of the outside of a lemon. The question is now what happens if, after imagining cutting open the visualised object, you end up with a visual representation of the inside of a wax replica
of a lemon. Due to the quasi-observationality of sensory imaginings, your
awareness of this change in visual appearance can neither change the fact,
nor defeat your respective knowledge, that, originally, you started off with
visualising a lemon. It is not as if you suddenly discover that what you
took initially to be a lemon in fact has been a wax replica all along. But
there are still two possibilities with respect to which kind of object you are
visualising after switching from visually imagining the outside to visually
imagining the inside of whatever you are visualising. According to the first
possibility, you continue to imagine the object to be a lemon, albeit one the
inside of which looks like that of a wax replica. According to the second
possibility, you are now visualising a wax replica of a lemon that is,
either you are now visualising a numerically different object, or the object
that you have been imagining all along has, in the mean time, changed in
kind from being a lemon to being a wax replica of one.
Now, in both cases, your knowledge of which kind of object you are
now visualising is still immune to defeat. In the first case in which what
you are imagining changes in appearance, but not in kind you continue to
know that you are still visualising a lemon. This indefeasible knowledge is
109
110
See Dorsch (2005) for a detailed discussion of the nature of association, and of how
it differs from imagination.
111
See Wollheim (1987) and Hopkins (1998) for the distinction between (our experiences of) the pictorial and the symbolic or otherwise non-pictorial content of pictures.
112
What is of importance here is that the agency involved in imagining is direct in the
sense of not relying on epistemic or merely causal means (see Section 13.2 and
Dorsch (2009b)). For we typically do not have the same kind of indefeasible access
to the epistemic or merely causal consequences of what we are directly doing.
113
CHAPTER 4
Imagination and Knowledge
In the last chapter, I argued that one of the two main strict differences
between imaginative and cognitive episodes is that only the latter possess
the epistemic function to exemplify knowledge. This difference is reflected
by the fact that we typically rely on our cognitions when making up our
mind about what to believe and do with respect to the actual world, but
usually do not take our imaginings into account during such deliberations. I
also noted that cognitions can fail to fulfil their epistemic function: not all
perceptual experiences, say, are veridical; and not all judgements or beliefs
amount to knowledge. Relatedly, if we have sufficient doubts about whether a given cognition lives up to its function, we normally stop relying on it
in the formation of beliefs and intentions.
In this chapter, I would like to argue that something similar applies to
sensory imaginings. Although they do not possess any epistemic function,
they may sometimes play a role in cognition and ground or constitute
knowledge in a manner very similar to that of sensory cognitions. It is true
that imagining is not by itself, or intrinsically, constrained by how reality is
like. But it may still be so constrained in an extrinsic manner, namely by
being put to a certain use in the context of our active engagement with a
wider mental project. More specifically, we may actively ensure that certain sensory imaginings are reliable and non-arbitrarily veridical in their
representation of which properties certain objects in the external world instantiate.
The possibility of knowledge-constituting sensory imaginings is central
to my main line of argument and is therefore implicitly or explicitly ad-
116
The different ways in which perceptions and occurrent beliefs are actually restricted
by reality are described in Section 7.1, where I turn to OShaughnessys account of
imagining.
117
118
ing images of him on the basis of ones memory and adjusting them in response to whether one visually recognises the imagined person as being the
same in appearance as the remembered person, now only much older. Once
satisfied that there is such a correlation in content between the visual images and the original memory, one may use the visual image in question to
try to recognise the friend among the real crowd in the station. Perhaps the
imaginative engagement involved constitutes an extended mental project.
But again, it is the final image which is in the same way as a comparable
visual memory might be in other circumstances of primary cognitive importance in the recognition of the friend, or the formation of a corresponding judgement about his appearance. For it is this image which contains the
information required for recognising the friend, or forming the judgement.
What these descriptions of the two examples indicate is that they lead to
visual imaginings which seem to have the potential to ground knowledgeconstituting judgements or beliefs and, in the second case, also the acquisition of certain recognitional capacities. Indeed, the two final images share
important features with each other, as well as with the corresponding perceptions or memories, which should be sufficient to confirm their status as
cognising states and, more specifically, as grounds for intellectual knowledge.
First of all, the sensory imaginings described can non-accidentally and
reliably match reality and lead to practical success. Thus, one may correctly decide on the basis of ones image that the sofa will fit through the
door, and hence buy it; or one may be able to recognise ones friend by
means of ones image of him. The potential veridicality of the visual image
will thereby be a matter of, for instance, the issue of whether one rightly
remembers the size of the door or the appearance of the friend, whether
one does not change the perceived size of the sofa while visually rotating it
in ones mind, or whether ones impression of how the appearance of
people alters with age is adequate (at least with respect to the person in
question). If the imaginative representations turn out to be veridical, it is
therefore to be expected that they typically do so non-accidentally. Because
of their dependency on information provided by perception and memory,
and because of the importance of ones skill to properly visualise the respective scenarios by means of mental rotation or extrapolation, the result-
119
ing sensory imaginings will match reality (if at all) not merely by chance,
but in virtue of the reliability both of the underlying cognitive representations and of ones control over the mental capacities involved in visualising.
Not everyone can always or even sometimes act reliably on his intention to sensorily imagine something. But it seems plausible to assume
that this is at least a possibility, in particular with respect to the first example. There is no reason to doubt, for instance, that we can mentally rotate simple objects (or shapes) without changing their sizes relative to other
simple objects (or shapes), at least within a practically permissible margin
of error. Hence, if the imagining person is sufficiently skilled, the mental
processes of actively producing sensory imaginings can be reliable if not
to the same extent as the comparable perceptual mechanisms, then at least
enough to enable practical reliance on them.
Furthermore, one may be aware of the potential reliability and non-arbitrary veridicality pertaining to the imaginative representations, in virtue
of the introspective knowledge of the nature of ones imaginative project
and in virtue of the memory of practical success in the use of visualising of
this kind in the past. For instance, one may be aware that one uses perceptions and memories and tries to come to an appropriate image by not altering the sizes of the represented objects while visualising them together;
and one may be aware that one had success with previous projects of a
comparable nature. Ones awareness of the potential trustworthiness and
non-accidental veridicality of ones imaginative project becomes also apparent in the fact that one may explicitly decide to rely on ones visualising
in order to find out whether the sofa will fit through the door, or whether
the person over there is ones friend.
Finally, the visual imaginings represent and refer to the respective objects in the world (the sofa, the door, and ones friend). Three considerations speak in favour of this idea.
First, visual imaginings may represent real objects if not particular objects (such as the sofa in the shop), then at least particular types or sorts of
objects (e.g., some kind of sofa which has the same size and shape as the
sofa in the shop; see also Section 2.3). Accordingly, it seems reasonable to
say that the two visual imaginings under consideration represent the re-
120
See Dretske (1986) on the difference between natural functions and assigned functions to represent; and Lewis (1980) for the idea that it is essential to the representationality of perceptual experiences that their contents are nomologically dependent
on the respective aspects of the environment.
121
The idea that there is a close link between misrepresentation and full intentionality is
further defended in Dretske (1986).
122
123
One is that there does not seem to be a plausible candidate for an alternative ground. The two cases described involve certain cognitive elements
such as perceptions or memories of particular objects, and ones general
knowledge or recognitional capacity concerning certain ways in which objects behave (that they remain constant in size, or how they change appearance with the progress of time). But none of these elements carries the information which is contained in conceptualised form in the resulting instances of knowledge or recognition. For instance, neither the perception of
the sofa nor the memory of the door, nor the belief that objects keep their
relative sizes with respect to each other, can give rise to the judgement that
the sofa is larger (or smaller) than the door. Accordingly, the person trying
to find out whether the sofa fits through the door does not remember (or
perceive) that the sofa is larger (or smaller) than the door; nor does he
manifest some previously acquired knowledge of this fact. Equally, neither
being able to visually recognise how the appearance of persons change
with age, nor remembering what ones friend looked like when young, can
ground the acquisition of the capacity to visually recognise the friend, or
the corresponding intellectual knowledge about the current particular appearance of the friend. In particular, the person trying to meet her friend
does not recall his current appearance; all she remembers is his appearance
in the past.
It might, to some extent, be possible to infer the appearance of the friend
on the basis of ones beliefs about how people age, perhaps in conjunction
with ones beliefs about how one remembers him to have looked years ago.
For instance, someone may come to believe on these grounds that her
friend will have some wrinkles, or grey hair. But such beliefs are not involved in the second example. Instead, this example involves the use of
ones capacity to visually recognise age-related changes in the appearance
of people. Moreover, beliefs about these changes will not be rich enough to
determine all the biometrical details (the location, size and shape of the
wrinkles, etc.) with which the visual image may provide us on the basis of
the imaginative modification of the memory of the friend. What becomes
important here is that the example has to involve a visual ground for knowledge, and not merely an intellectual one. For the imagining subject may be
able solely on the basis of this ground to visually recognise the friend. In
124
normal circumstances (e.g., when the friend has no distinguishing characteristics, such as a scar on his cheek), a judgement or belief about the appearance of the friends face should not suffice for this task.
Hence, in both examples, what the underlying cognitive states represent
has to be imaginatively combined and modified by using ones ability to
visualise in order to extract the information needed for grounding the respective judgements or recognitional capacities. This is not to deny that
certain aspects of the cognitive elements concerned with the representation
of particular objects may still pertain to the final visual images. For instance, the image of the sofa and the door will presumably represent these
entities with their visually remembered sizes, shapes, colours, and so on.
And the image of the aged friend will presumably represent him as still
having certain of the facial features which one remembers him as having
had in the past (e.g., the distance between the eyes, the shape of the head,
etc.). Similarly, it is still possible that the visual images may be said to include or manifest aspects of ones general knowledge or recognitional capacity concerning how objects behave. That the sofa has not changed in
size, or that the friend is visualised as having wrinkles or grey hair, may
perhaps be taken to be instances of such an inclusion or manifestation. But
what is not true of the two visual imaginings is that they are cognising because they include or manifest aspects of perceptions, memories, beliefs,
or recognitional capacities.
To stress the point again, one does not perceive, remember or bring otherwise back to attention that the sofa is larger than the door, or that ones
friend will have particularly shaped wrinkles on his cheeks. The reason for
this is that ones acquisition of knowledge is based on both the perceptions
or memories of the particular objects and the general knowledge or recognitional capacity concerning how real objects behave. What is thus needed
is an additional process or activity which combines these cognitive elements in such a way as to lead to the extraction of previously unknown
pieces of information. In the two examples, this required additional activity
is imaginative and results in cognising visual imaginings.
In this respect, the two examples differ from the following case. 88 I may
88
See Kosslyn (1980): 1. Pylyshyns case of finding out how many windows ones
house possesses by imagining walking through it thereby visualising each of its
rooms and counting the windows in them is basically of the same kind (Pylyshyn
125
wonder whether frogs have lips and, being unable to form a judgement on
this matter by merely thinking about it, may decide to form a visual image
of a frog in order to reach an answer. I may thereby form the image by falling back on my episodic visual memories of a particular frog, or by relying
on my visual memory or knowledge of how frogs generally look linked to
my respective recognitional capacity. This procedure may be veridical and
reliable with respect to the visual representation of the mouth of the
particular frog, or of frogs in general. Hence, it may allow me to form on
the basis of the resulting visual image the knowledge-constituting judgement that, say, frogs indeed have lips. But it can be argued that those elements of my image, which represent the visualised frog as having lips, are
instances or manifestations of my visual memory of the appearance of the
particular frog, or of my general knowledge or recognitional capacity concerning how frogs look. After all, it seems natural to say that, by forming
the image, I come to recall that (some) frogs have lips. 89 Hence, although
the overall visual representation may still be imaginative, its cognising aspects seem to be cognitive. That is, the resulting visual imagining appears
to be cognising only in so far it contains mnemonic representational elements which are cognising: what ultimately grounds the judgement are
mnemonic and not imaginative elements of the visual representation.
In contrast, in the two previous examples, there are no cognitive elements involved (over and above the visual imaginings) which could reasonably be said to ground the respective knowledge-constituting judgements
and, in the second case, the acquisition of the capacity to visually recognise
the friend. The main difference seems to be that, in the case of the frog, we
come to pay attention to previously unnoticed, but none the less consciously experienced aspects (similar to the moment when we suddenly
realise that someone is sad or angry: something which we have suppressed
or ignored before); while in the case with the sofa and the door, or with the
friend, we unearth new information which we yet have not experienced, or
been able to notice, in its entirety, given that it so far has been scattered
89
126
See, e.g., OShaughnessy ((2000): 1ff.; 68ff.; 113ff.) and the discussion of his relevant ideas in Section 7.4.
127
condition on knowledge may be that it has to be ultimately based on perception. But the two visual imaginings fulfil this requirement: they ultimately derive from the respective perceptions of the objects concerned (the
sofa, the door, the friend), as well as from the perceptions underlying the
acquisition of the relevant general knowledge or recognitional capacity
(e.g., perceptions of objects with constant sizes, or of changes over time in
peoples appearances). Of course, the two visual imaginings do not provide
direct access to reality as the corresponding perceptions would do. But
judgements or sensory memories are not immediate in this sense, either,
and too require past or present perceptual mediation, too.But neither prevents them from having the potential to constitute knowledge. Hence, there
is no reason to worry about the indirectness of the imaginings, and their ultimate dependency on perceptions.
Similarly, no particular problem should arise pertaining to the visual
imaginings, if it is assumed that knowledge-constituting states have to be
partly constituted by (a link to) the respective aspects of reality. For the
idea that such imaginings are factive and differ in kind from phenomenologically indistinguishable non-cognising imaginings should not be more (or
less) problematic than the idea that cognising judgements or sensory
memories are factive and differ in kind from phenomenologically indistinguishable non-cognising judgements or sensory memories. In all three
cases, the factivity cannot be realised by means of a direct causal link to, or
immediate awareness of, the respective aspects of the world (as it happens
in the case of cognising perceptions). But while this has the consequence
that the idea of a constitutive link between the cognising visual imaginings
and the world would need further elucidation (which might not be easily
provided), the situation is no different with respect to cognising judgements or sensory memories which are said to be constitutively linked to
reality. Requiring a constitutive link to reality would thus be problematic
with respect to all non-perceptual forms of knowledge; while giving up on
this requirement in the case of memories and judgements would also permit giving up on it in the case of imaginings.
Finally, it is not difficult to identify the cognitive prototype or ideal
which the visual imaginings may be said to realise (if one wants to speak in
this way at all). OShaughnessy is right, it seems, that sensory imaginings
128
cannot exemplify the same cognitive prototypes namely successful perception and intellectual knowledge as perceptual experiences or beliefs
(see Chapter 7). But this is also true of sensory memories. And, just like
them, sensory imaginings can have their own cognitive prototype: namely
that of a reliable and veridical representation of reality formed by means of
appropriate imaginative activity. This, too, should count as a prototype,
given that it constitutes the ideal to which the respective imaginative engagement aspires to, and given that it combines the central elements of intentionality, veridicality, reliability, rational integration with ones picture
of reality, and (if also required) a constitutive link to reality.
129
Compare also OShaughnessys view that perceptions can figure as grounds for
knowledge only if they are rationally integrated which is another form in which
the justificatory power of such states may partly depend on the nature of the underlying beliefs.
130
their nature, but also because of their contingent use to which we put them.
However, some additional qualifications are requisite here.
First, the use of sensory imagination is presumably often less reliable or
less precise in its provision of visual information than the use of perception; and we will refrain from judgement more often with respect to the
former than to the latter (e.g., if the issue of the sofa fitting through the
door is one of centimetres). But this does not prevent imaginings from being sufficiently reliable and precise as it is reasonable to often rely on
perception, even though the use of calibrated tools of measurement or of
photographs might be more trustworthy and accurate.
Second, the acts of visualising in the examples do not provide us with
new evidence. Instead, they extract information, which has been previously
unknown to us, from the portions of evidence that we already have had
(e.g., in the form of our perception of the sofa, the memories of the door or
the friend, or the knowledge about the ageing process). This is unproblematic since having new evidence (i.e., involving the epistemic intake of new
information) is not a condition on the provision of knowledge as illustrated by the possibility of a priori knowledge. Indeed, this may be taken to
suggest that the imaginative activity fulfils a cognitive role similar to that
of a priori reasoning, given that the latter (at least according to one common understanding) also leads to the gain of information without the acquisition of new evidence. For example, if we had previously measured the
piece of furniture and the door with some kind of tool, instead of our eyes,
we could have just calculated in our heads and on the basis of our
memory of their sizes whether the former would fit through the latter.
And just as such an example of a priori reasoning would make it superfluous to perform the overt action of solving the problem by means of a sliderule or a geometrical drawing, the act of mentally rotating the sofa replaces
the much more demanding task of actually moving it to ones place and determining whether it fits through the door.
Third, it is also important to note that the sensory imaginings can fulfil
their cognitive role only on the basis of information that is gathered, processed and stored by cognitive means, such as perception, memory and belief. This, together with the fact that they do not provide new evidence,
seems to be reflected by the fact that imagining does not constitute an inde-
131
132
133
partly because sensory imaginings do not refer to real objects by themselves, but depend for their referential capacity on the wider context, such
as their particular use in some mental project. Consequently, the immunity
to error through misidentification concerning what we are visualising is as
unproblematic for the cognising status of the two examples as the quasiobservationality of visual imaginings. In fact, these two aspects of visual
imagining may stem from the same fact, namely that the referents of our
visual imaginings are ultimately determined by our accompanying desires,
intentions or thoughts about our visual imaginings (see Sections 2.3, 3.6
and 14.1f.).
To conclude, the significant differences between the sensory imaginings
and their perceptual or mnemonic counterparts do not matter for the fact
that the former can, under the right kind of circumstances, cognise the
world in a manner very similar to that of the latter. The final visual imaginings in the two examples should therefore be accepted as potential grounds
for knowledge.
The N in (NI) stands for negation claim (see the introduction to Part Two). Note
that what is at issue here is the stronger claim that imaginings cannot provide us
with new information; and not merely the weaker and presumably uncontroversial
claim that imagining is not a source of evidence, and that imaginings therefore
cannot provide us with new evidence (see Section 3.4 for further discussion of this
distinction). That imagining is not a source of evidence still leaves room for the imaginative acquisition of new knowledge on the basis of the extraction of information
previously hidden (i.e., unknowingly contained) in already acquired information
(see the two examples discussed above) and is thus incompatible with (NI). Besides,
the uninformativity of sensory imaginings even if being essential to them is
surely not among their fundamental features, but rather derives from one of their
more basic characteristics (e.g., their special origin in the mind or in mental agency).
134
It is natural to assume that mental states and, in particular, non-intellectual mental episodes can constitute knowledge of reality only if they
provide us with information about it. 93 Seeing the greenness of a tree, for
instance, contains information about the colour of the leaves of the perceived object, as well as about the fact that it is a tree and has leaves. If it
did not convey this information, then it could not help us to acquire the
corresponding knowledge about the nature and other features of the perceived object. Hence, if imaginings are, by their very nature, uninformative, they cannot constitute knowledge.
This means that either (NI), or my interpretation of the two examples
presented in the last chapter is false. In Section 8.1, I argue that it is (NI)
which should be rejected and partly because of its inability to accommodate the two examples. Here, by contrast, I just aim to undermine different ways of providing support for (NI) that are linked to the features of
sensory imaginings discussed in the last chapter, and which may also at
least to some extent be identified in the writings of Sartre and Wittgenstein, both of whom have endorsed (NI) in one form or another. 94 The conclusion of the remainder of this chapter is that (NI), in its strictness, is
false; and that it can at best be maintained that imaginings typically differ
from cognitions in being uninformative with respect to the external world.
Sartre is not explicitly concerned with the question of whether sensory
imaginings may provide us with information about aspects of the external
world (rather than with information about themselves, such as their content
or origin). But he is very clear about his belief that sensory imaginings differ from perceptions in that [we] can never learn from an image what [we
do] not know already (Sartre (1940): 9). Moreover, Sartre proposes a
Here, I use the term information in a very non-technical sense, mainly because this
is how Wittgenstein seems to employ the expression. In particular, I do not mean to
capture the technical uses introduced by Dretske (1981) and Evans (1982), although
some of the following considerations might apply to their notions as well.
94
OShaughnessy is another philosopher who endorses claims which ultimately entail
(NI). I discuss his reasons for endorsing them in Chapter 7, and object to these reasons in Chapter 8. The discussion here is more restricted in that it focuses on more
direct endorsements of (NI).
93
135
straightforward explanation of why this is the case: namely because the information carried by sensory imaginings is put there by us in the process
of producing these imaginative representations, namely in the shape of an
intellectual element which specifies (Sartre (1940): 9).95
Although his concern is primarily with knowledge about the sensory
imaginings themselves and, more specifically, about what they sensorily
represent, there is no obvious reason why Sartres thesis and his explanation could not be extended to knowledge about the external world. If how
our sensory imaginings represent their objects as being is determined by
us, rather than the world, and if, as a result, we cannot learn anything new
about the content of our sensory imaginings, then we should also be unable
to acquire new knowledge about corresponding aspects of external reality
that might be represented by means of the possession of that content. For
example, if I know that what I am visualising is a cube and, moreover, the
cube on the desk of the mathematics teacher next door, precisely because
this is what I am trying or taking myself to do, then I am in no position,
simply on the basis of my episode of visualising, to acquire new knowledge about the real object in question for instance, that it is indeed a
cube that is sitting on a desk, or that it is the cube of the mathematics
teacher next door. I knew this all along, that is, even before engaging in my
imaginative project.96
However, the impact of this line of reasoning is limited. More specifically, it merely rules out the acquisition of three kinds of new knowledge: (i)
knowledge of the (imperceivable) identity of the sensorily imagined object;
(ii) knowledge of the imperceivable nature of that object; and, if applicable, (iii) knowledge of perceivable facts about that object, the imaginative
representation of which is not determined by corresponding intentions-inSee Section 3.4 for a discussion of the longer quotations from which these citations
are borrowed.
96
Information (or knowledge) thereby counts as new for us if we do not already have
access to it by means of other mental representations such as the thoughts or intentions involved in the formation and determinations of the sensory imaginings
which store it in fairly easily recognisable form. Hence, what is claimed is that, for
Sartre, sensory imagining does not lead to a gain in information. That is, all the information, which we can become aware of after the occurrence of sensory imaginings, has already been available to us before this occurrence.
95
136
137
ror and the quasi-observationality (see Section 3.6, especially the last subsection). But none of these features of sensory imaginings puts any further
restrictions on what we might come to know on the basis of engagement in
sensory imagining.
One consequence is that sensory imaginings may, in some sense, count
as manifestations of knowledge that we were already in possession of before entering into the imaginative projects at issue. My imaginative representation of the cube on the desk is an expression of my knowledge that
there is such a cube on the desk in the room next door. The same happens if
I suppose, as part of a hypothetical argument, that the Earth is flat in response to knowing that the Earth is flat, and to intending to imagine something about this planet as it actually is (e.g., I would like to discover some
metaphysical or scientific truth about the Earth). Or, to take another example, I may use visualising to exhibit my generic knowledge of the visual
appearance of frogs (see Section 4.2). But in none of these cases does the
sensory imagining do any justificatory work, or lead to new knowledge.
Moreover, it is unclear whether the manifestation or expression of knowledge in the way described amounts to the constitution of knowledge (as
defined in the introduction to Part One): that is, it is unclear whether the
(sensory) imaginings concerned should count as instances of knowledge. I
return to this question further below, when discussing Wittgensteins view
on the uninformativity of sensory imaginings.97
In any case, what remains unaffected by the line of reasoning presented
above is the possibility of sensory imaginings providing us with new information about perceivable facts about external objects if their imaginative representation is not determined by prior knowledge of those facts, or
by the prior intention to imagine precisely those facts. This is the case, for
instance, in the two examples discussed in the preceding sections: they
concern perceivable facts namely facts about relative size or visual ap97
Note that this question is different from the question of whether episodes of memory
may constitute knowledge. The answer to the latter should surely be yes, at least in
the sense that memories are ways of preserving and recalling knowledge (see Martin
(2001)). Accordingly, assuming that that the visual representation of a generic frog
should count as an episode of remembering (though not of recalling) rather than an
episode of visualising, this example is to be distinguished relative to its status as
knowledge from the others presented in the main text.
138
139
140
102
141
perspective, given to us in the same passive manner as perceptual experiences, they allow, again, for a perception-like kind of attentive exploration
and discovery. That is, we may find out how many columns we have actually visualised by paying attention to the passively filled-in details of our
image of the Pantheon.
Further examples are deliberate attempts at sensorily imagining something, which are not entirely successful (i.e., involve some form of intentional failure), but still result in the production of some sensory representations (see Peacocke (1985): 26). That is, voluntarily triggered imaginings
may end up representing, not what we have intended them to represent, but
something else. For instance, I may decide to visualise a circle, but end up
visualising a shape which is not absolutely round and which I consequently
do not take to be a circle. Or I may fail in my attempt to daydream about
my visit to Lisbon simply by becoming distracted by thoughts about Rome
and by subsequently starting to daydream about this city instead. Again,
these examples illustrate the possibility of cases in which our will does not
fully determine our sensory imaginings and, hence, of cases in which attention may deliver new information about the passively determined aspects
of what is imagined.
The conclusion so far is that, if sensory representations occur involuntarily or include passively determined aspects of their content, they may
provide us with new information. Moreover, it is in principle possible that
the relevant pieces of information may actually be capable of grounding or
constituting knowledge about aspects of external reality namely if the
passive processes responsible for the determination of sensory content happen to be suitably constrained by how the external world is like. However,
the main candidates for such processes seem to be mnemonic ones (e.g.,
processes that lead to the conscious manifestation of episodic or generic
memory) which means that the resulting (parts of the) sensory
representations should count as instances of memory, rather than imagination. This is the case, for instance, when we succeed in acting on our intention to remember or recall something: we deliberately trigger some passive
mnemonic processes that determine how we sensorily represent the objects
concerned (e.g., consider again the frog example). Hence, while the involuntary origin of sensory representations does not suffice to prevent them
142
from being informative, it still supports the claim that there are no sensory
imaginings among the involuntary information-providing episodes. To this
extent, Sartres explanation of (NI), that is, of the uninformativity of sensory imaginings seems to be right.
103
143
144
with respect to the external world, in contrast to Sartres weaker claim that
they just cannot provide us with new information about the world. Indeed,
Wittgenstein understands the property of being informative about reality as
the property of carrying correct or incorrect information about it: that is, as
the general property of carrying truth-evaluable information, or having
veridicality conditions.105 Wittgensteins claim is therefore that sensory
imaginings are uninformative in the sense of lacking veridicality conditions because of the voluntary determination of how they represent objects as being. If we act successfully on our intention to visualise a cube as
sitting on the desk next door, it is, for Wittgenstein, inadequate to ask
whether the resulting visual representation is veridical or not in respect of
where it presents the cube as being located.
But Wittgenstein fails to take into account that the two examples of cognising visual imaginings discussed in the preceding sections show that our
voluntary control may actually be responsible for the extraction of new information (e.g., about differences in size, or about likely changes in appearance) on the basis of already stored visual information (e.g., information about the size or appearance of objects, as stored in the content of perceptions, memories and beliefs). The final sensory imaginings of the imaginative projects at issue carry their information precisely because the subject in question very carefully controls their formation and deliberately selects the cognitive contents that figure as informational inputs to the projects (i.e., the perception of the sofa and the memories of the door and the
friend). Moreover, she decides to let her imaginative agency be guided by
her knowledge of how middle-sized objects behave when moved around,
or by her knowledge about how peoples appearances change with growing
age. While she voluntarily determines which states of affairs her sensory
imaginings represent, she sets herself certain constraints (i.e., to respect
The application of the idea of truth to sensory representations is controversial not
the least because it is not unreasonable to take them to be non-propositional instances of object awareness (see Dorsch (2010c)). But since they are representational and non-neutral with respect to the represented situation (and, in the case of
perceptual experiences, come with a cognitive attitude), there is certainly a sense in
which they can (fail to) match whichever situation they represent and, hence, be
(in)correct or (non-)veridical with respect to that situation (i.e., reality in the case of
perceptual experiences; see Siegel (2010)).
105
145
how real objects behave) and bases her decision about how to visualise the
sofa, door or friend as being on the relevant knowledge (i.e., concerning
how real objects behave). As a result, the subject is in control when visualising the sofa as fitting through the door, or the friend as having some
wrinkles and greying hair. In particular, her sensory imaginings are not the
passive manifestations of episodic or generic memory, or of some similarly
passive mechanisms, such as association.
The capacity of episodes of sensory imagining to provide us with new
knowledge about reality is therefore due to rather than, indeed, ruled out
by the fact that, in these imaginative projects described, we do not simply
follow our inclinations about how to modify our visual representations, but
try to adhere to certain principles: such as not to alter the size of the objects
when mentally rotating it or changing our perspective on it; or to conform
to our knowledge gained by experiencing the changes over time in the appearance of objects. Indeed, principles like these are constitutive of the
imaginative projects in question: the latter would be different imaginative
projects if they were not aimed, in their purpose to produce specific visual
images, at taking into account how the respective kinds of entity actually
behave. In particular, the imaginative projects could not be embedded in
the wider cognitive project of acquiring knowledge about real objects, given that acting differently in visualising the objects (e.g., by following principles that describe a very different and unrealistic spatio-temporal behaviour of objects) would fail to be conducive to non-arbitrary veridicality.
The underlying thought is, of course, that sensory imaginings possess
the ability to refer to real objects and be non-accidentally correct about
their perceivable features if they are suitably constrained (see Section 7.1).
The point made here is basically that such constraints need not obtain independently of us, but may in fact be established by our own (mental)
agency.106 Note that the sensory imaginings concerned may even count as
object-dependent in the same sense as, say, episodic memories: that is, as
depending for their existence or, at least, for their representationality on
106
Indeed, Budd (1989: 113) understands Wittgenstein as allowing for this possibility
and, hence, for the possibility of knowledge-constituting sensory imaginings in the
way described. But this reading has difficulties to make sense of the strictness of
Wittgensteins remark that imaginings do not teach us anything about the external
world (see the passage quoted above).
146
107
147
CHAPTER 5
Other Theories of Imagining
The main result so far has been that there are two important differences
between the central cases of imagining and cognition a difference in their
attitude towards what they represent, and a difference in their epistemic
function with respect to the formation and justification of first-order beliefs. Any satisfactory unified theory of imagining should be able to accommodate these differences and, ideally, further elucidate them in terms
of a more fundamental feature present in imaginings and absent in cognitions. The different unified theories vary in their approach to this task. and,
notably, in their choice of the basic feature distinctive of imaginings.
As outlined in the Introduction, most of this book is taken up by the discussion and evaluation of three views that take imaginings to be characterised by the fact that they are not concerned with the cognition of the world
(the Epistemological Account), or by the fact that they constitutively depend on their cognitive counterparts (the Dependency Account), or by the
fact that they are essentially active (the Agency Account). Before discussing each of these three main accounts in detail, I intend to present other approaches to the unity of imagining and the differences between imaginative
and non-imaginative phenomena. Accordingly, I use this chapter to look at
alternative ways of accounting for the differences between (certain kinds
of) imaginings and cognitions, and to assess them both in their own lights
and in their relation to the general task of providing a unified account of
imagining.108
108
One alternative account the Subjection to the Will Account is not to be discussed
before Section 13.5.
150
One general problem for almost all of these accounts is that they do not
really apply to forms of imagining which are not (obviously) episodic, notably experiential imaginings and imaginative projects. But each of them
also faces its own more specific difficulties. My conclusion is that they are
all less promising than the Cognitive, the Dependency or the Agency Account.
151
claim about how things really are; or it does not. There is no third option,
and there is no continuum between these two poles. The same is true with
respect to whether a mental episode has the function to ground first-order
knowledge or not: such functions do not come in degrees.109
The claim that the phenomenologically salient difference between imaginative and cognitive episodes is qualitative gets further support by the
observation that we are not subject to a particular kind of error with respect
to our awareness of these episodes. Although we may err with respect to
the nature of our mental episodes, we do not seem to be able to err with respect to the kind of state which we take them (rightly or wrongly) to be at
least not if our conscious and first-personal access to them is concerned.
For instance, we may very well mistake one of our episodes of visualising
for a perceptual experience.110 But it seems absurd to say that we may be
mistaken about the fact that we take the episode of visualising to be a perceptual experience, and not, say, a visual recollection.
This, I think, is at least part of what Sartre and Wittgenstein may have
intended to capture when they stated that we cannot confuse the various
kinds of mental episodes and, in particular, perceptions and sensory imaginings.111 Now, it seems impossible to properly account for this observation if we assume that the phenomenologically salient difference between
imaginings and cognitions comes in degrees. For there would then be episodes (e.g., near the middle of the continuum), concerning which we might
very well be in doubt about whether they are subjectively given to us as
imaginative or cognitive.
In Section 9.1, I note that, for Hume, differences in vivacity come with differences
in attitude and function. Assuming that the latter are differences in quality, Hume
may therefore be able to avoid the two noted problems for the Quantitative Account.
110
Or one of our perceptual experiences for an episode of visualising: which seems to
have happened, for instance, in the Perky experiments at least under the assumption that pictorial experience is a special kind of perceptual experience (see Dorsch
(2012c)). See Hopkins (2011a) for a discussion of what the Perky experiments do
and do not show.
111
See Sartre (1940): 4, Wittgenstein (1984c): II, sec. 85, and Scruton (1974): n. 13, p.
100. See also the comments on the different language games pertaining to seeing
and visualising in Wittgenstein (1984c): II, sec. 63, and Wittgenstein (1984d): sec.
621.
109
152
153
154
sideratum of explanatory force (see Sections 1.1 and 2.1). Hence, if other
unified theories succeed in providing a satisfactory account of why imaginative and non-imaginative phenomena differ essentially in their phenomenal characters, the Phenomenal Account should not be assumed to capture
the unity of imagining on the most fundamental level possible.
155
located (see Sartre (1940): 52, 86ff., 127ff. and 149); while Casey takes
them to differ in their determinacy (see Casey (1976): 36f. and 104ff.;
and Section 3.3 above). It is tempting to understand these claims in terms
of an ontological difference between two categories of things.
It is important to distinguish the resulting Ontological Account from the
ontological reading of the Phenomenal Account: that is, from the view that
there is a phenomenal difference between (sensory) imaginings and cognitions with respect to whether they present their objects in consciousness as
real (or cognised) or as unreal (or imagined) that is, as being non-neutral
about the actual or some other possible world (see Section 2.1). When
Macbeth becomes aware of the dagger, he does not experience it to be part
of reality and to be literally in front of him: the dagger is not given to him
as a real and perceived dagger, but rather as an unreal and hallucinated (or
imagined) one (see Dorsch (2010a) and (2012a)). The view that there is
such a presentational difference between imaginings and cognitions is
compatible with the idea that imaginings can refer to real objects, or that
there are indeed no unreal objects. The truth of this view may, however,
also support the endorsement of the Ontological Account, given that it
might be thought that the postulation of a categorial difference provides the
best explanation of the phenomenal difference at issue. I return to this issue
shortly.
In some passages of their writings, Sartre and Wittgenstein come close
to accepting the main claim of the Ontological Account. In particular, they
maintain that sensorily imagined objects are part of an entirely different
kind of space (or stand in entirely different spatial relations to each other
and further entities) than perceived ones (Wittgenstein (1984d): sec. 622
and 628; Sartre (1940): 8ff.). McGinn accepts this characterisation of the
distinction between perceived and imagined objects and analyses it as a
difference in the richness of spatial representations: sensorily imagined objects are represented as spatial objects (e.g., as extended), but not as spatially located (i.e., with a specific spatial location; see McGinn (2004):
58f.). However, it is unclear whether any of the three intends this observation to extend to supposed and judged objects; or whether this difference in
space (or spatial relations) implies, or is meant to imply, that we cannot
sensorily imagine the very same objects which we can perceive. Indeed,
156
Sartres comments may perhaps also be interpreted as pointing instead to the idea of
a difference in attitude or immediacy (see Sections 2.1 and 3.2, as well as Martin
(2001): 270f.). This understanding may gain further support from Sartres considerations about the quasi-observational character of sensory imaginings (see Sartre
(1940): 8ff., and Section 3.4 above).
157
ings may represent mere types of objects, rather than individual objects.
For example, we may visualise some man without thereby visualising any
particular man. Indeed, this is likely to be the default case of visualising,
meaning that any reference to specific individuals has to be due to some
additional intellectual element (e.g., an accompanying thought or underlying intention; see Section 3.6). But non-particular imaginative representations are to be expressed by means of (indefinite) descriptions involving
existential quantification, with respect to which the problem of reference to
fictional or non-existing entities does not arise, and which therefore do not
require the postulation of unreal or non-existing objects (see Russell (1905)
and Bodrozic (2009)).
A second source of support for the idea of two fundamentally different
ontological categories of objects might be phenomenological evidence.
That is, there might be a difference in phenomenal character between imaginings and cognitions, the satisfactory explanation of which requires the
postulation of an ontological difference with respect to the represented entities. As noted in Section 2.1, the phenomenal difference cannot be located
in which properties and state of affairs are represented, given that imaginings and cognitions can at least in this respect have the same content. It
also cannot be located in the phenomenologically salient features pertaining to the mode of representation, given that imaginings and cognitions can
be of the same type (e.g., visual or intellectual). Accordingly, the proposed
phenomenal difference has to be part of the attitudinal difference between
imaginings and cognitions for instance in the shape of the difference in
non-neutrality about the actual or some merely possible world which may
manifest, or be constituted by, the difference in ontological category.
But it is far from clear why the explanation of this phenomenal difference should require or at least strongly suggest the postulation of a categorial difference between real and unreal entities. First of all, there are alternatives explanations available that are at least equally promising. In particular, the phenomenal difference may be due to a difference in immediacy (see Section 3.2 and Chapters 9ff.), or to a difference in voluntary control over what is specifically represented (see Section 14.1). Then, the general metaphysical and epistemological worries about non-real entities (see
Bodrozic (2009) and Reicher (2010)) may in fact render the alternative ex-
158
planations better than, and preferable over, the proposed account in terms
of a categorical distinction between real and unreal entities, given that they
do not involve a commitment to some otherwise unneeded and controversial category of objects. Finally, it remains unanswered why other important attitudinal differences for instance, between conative and cognitive
representations, or between perceptions and judgements, on the one hand,
and memories, on the other are not likewise due to ontological differences.
But independently of the possible motives for embracing the Ontological Account, the main problem for this view is that, although many instances of imagining do not refer to real entities and their real features, at
least some do make reference to reality and can even be veridical with respect to it. I may visualise the new car of my friend and thereby happen to
correctly visualise it to blue, without actually possessing any evidence of
its blueness. As already previously noted in Section 2.3, defenders of the
Ontological Account might reply that the imaginative reference to real entities involved in such cases is merely mediated by the more direct imaginative reference to unreal entities similar to the way in which our perceptual experiences of portraits refer indirectly to the portrayed persons by
means of referring directly to the perceived portraits themselves. But this
suggestion presupposes that the unreal objects of the imagination are picture-like in that they refer to aspects of the external world in the same way
as, say, portraits do. Given that imagining does not relate us to pictures external to the mind, unreal entities therefore have to be pictures internal to
the mind.
Sartres view (which is, ultimately, very close to Husserls early view) is
particularly interesting in this context.116 While he denies that there exists
116
See Sartre ((2004): Part II), as well as the excellent and detailed discussion of his
view in Hopkins ((1998): ch. 7). Sartres view is directly derived from Husserls
early account of sensory imagining ((1905/2006): ch. 2 and chs. 4f.).
Both claim that, in visualising a tree, we are not only aware of a tree (i.e., a threedimensional, tall, leafy object that does or could exist as part of a real forest in the
external world), but also of an inner analogon of such a tree (i.e., something
internal that could not form part of a real forest). Moreover, our visual awareness is
completely exhausted by our awareness of the inner analogon (or Bildobjekt, as
Husserl says). By contrast, our awareness of the real tree is mediated by, or a matter
of, our visual awareness of the analogon and the fact that the latter stands in for the
159
160
161
tions are concerned with particular objects (i.e., possess a particular content), while imaginings deal only with types of such objects (i.e., possess a
general content), may be more promising in so far as the difference
between particulars and universals alluded to is perhaps indeed ontologically fundamental. But the resulting view could at best distinguish sensory
imaginings from perceptions, given that thoughts whether judgemental or
imaginative can be particular (e.g., indexical or name-involving) or general (e.g. descriptive).
162
Compare the similar worries with respect to the phenomenon of truth and the corresponding concept or expression is true noted by Frege (1892): 34f. and (1918): 60,
and Soames (1999): ch. 2, among others.
163
states of affairs in some possible world, but neutral about their obtainment
in the actual world. In particular, there are two reasons why the attitude
version of the Modal Account cannot simply identify imaginings negatively as those mental episodes that represent some state of affairs (i.e.,
some objects as having certain features) without involving the claim that
this is really how things are.
The first rationale for a more substantial characterisation is that there are
many representations of states of affairs that lack a cognitive attitude, and
which are not or at least not obviously imaginative. That is, the range
of mental episodes concerned with the non-actual crosses the border
between imaginings and non-imaginings (see Goldie (2002)). One example
are desires or conative feelings: they are precisely concerned with states affairs that they do not yet assume to obtain in the actual world. Similarly,
hopes, expectations or regrets may involve thoughts about features that
they do not take to be actually instantiated. Finally, although wondering
about whether something is the case may lead to the endorsement of a certain proposition as true of the actual world, it need not do so. But it is
doubtful that any of these examples should count as (involving) imaginings
(see Sections 1.2f. and 14.2, as well as the concluding chapter). This does
not mean that the distinction between cognitions and imaginings cannot be
as clearly cut as envisaged in the first two chapters. But it indicates that
there are many kinds of mental representations that are neither cognitive,
nor imaginative.
The second and related reason for requiring a less minimal characterisation of imaginings is that the absence of a cognitive stance towards the actual does not suffice to establish the presence of a concern with possibilities. What is in addition needed is an imaginative stance towards the possible. This is related to the observation that imagining, too, involves a theoretical attitude, that is, some form of non-neutrality towards how things
are (see Section 2.1). Just as cognitions make a claim about how the actual
world is like, imaginings make a claim about some possible world is like.
In other words, imaginings should be characterised, not merely in terms of
the lack of some theoretical attitude, but also in terms of the possession of
another. This is necessary, among other things, to distinguish them from
desires, hopes, wonderings and similar non-imaginative and non-cognitive
164
mental phenomena.
One of the few unified theories of imagining, that have been recently
proposed and defended, embraces an instance of the Modal Account which
comes very close to the attitude version just sketched. 118 In his book The
Language of Imagination (1990), Alan White characterises both the sensory and the intellectual imagination in terms of their presentation of possibilities (note that imagining is here understood to be distinct from, and
not to include, visualising; see, e.g., White (1990): 108):
To visualise something is to think of what it does or would look like. [...]
We can, and often do, visualise, when we are imagining, remembering, recalling, expecting, hoping for or fearing something. In visualising we have
an experience which is like that which we have in actually seeing something, though it is an experience which we commonly, though not always,
produce in ourselves. To visualise something is to think of how it does or
would look, it is to picture it to ourselves or to produce a picture of it.
(White (1990): 122f.)
When I am imagining something I am thinking of it as something that could
be happening, whether or not it did, whereas when I am remembering it or
recalling it I am thinking of it as something that actually did happen. (Ibid.:
171)
To imagine something is to think of it as possibly being so. It is usually also
to think of it in a certain way. Thus, even when what we imagine is simply
an object, such a someones face or a logical difficulty, what we are imagining is what it is or would be like. Hence, we can be asked how we imagine
or think of it or what we imagine it to be. This is why what we imagine is
not usually just some object, whether a face or a difficulty, but something
118
165
First of all, some clarificatory comments on Whites sometimes rather unorthodox terminology are necessary. Imagining, for instance, covers only
intellectual imaginings and imaginative projects with intellectually imaginative components, but not any forms of the sensory imagination (see the
claims and examples in White (1990): 185f.). It is in line with this usage of
imagining that White strictly distinguishes sensory imagining from
visualising: while the former amounts to intellectual imagining (i.e., imaginative thought) about perceivable objects and features (see, e.g., White
(1990): 108), the latter constitutes a form of visual representation. It is true
that he describes visualising sometimes in terms of thinking. But his
other characterisations in terms of experiencing, picturing, and so on,
make clear that he has sensory rather than intellectual representation in
mind (see, e.g., the first passage quoted above). That is, he takes thought
to go beyond the intellectual and to be roughly equivalent with conscious
mental representation.
A more substantial element of Whites view is his identification of
visualising with the visual representational element shared by thinking,
remembering, imagining, and so on, rather than with visually imagining
something (see the first passage quoted above, as well as the examples discussed at White (1990): 122 and 168f.). This means that, strictly speaking,
he endorses the attitude version of the Modal Account solely with respect
to intellectual imagining, which is said to differ from judging and remembering in involving a claim about how some possible world (i.e., what
could be happening) rather than the actual world (i.e., what actually does
or actually did happen; see the second passage quoted above). By contrast, his claims about a comparable difference between seeing and visualising do not distinguish sensory cognitions from sensory imaginings, but
rather perceptions from all other sensory representations (whether they are
mnemonic, imaginative, or something else). However, Whites underlying
assumption seems to be that visually imagining something consists in
visualising it and intellectually thinking (or imagining) of it as possible,
whereas visually recalling something consists in visualising it and intellectually thinking of it as actual and past (see White (1990): ch. 18, especially pp. 168f.). Hence, his restricted Modal Account may still count as a
166
unified account at least with respect to sensory and intellectual imaginings since it distinguishes sensory imaginings from sensory cognitions in
terms of the difference between intellectual imaginings and intellectual
cognitions.
Independently of whether proponents of the attitude version of the Modal Account follow White in his views on sensory imagining, their theory
needs to be further qualified. First of all, as already noted by White in the
quoted passages, what we imagine may coincide with what is actually the
case. For instance, how I visualise the car of my friend as being may happen to be identical with how it actually looks like. But this is unproblematic since it does not (wrongly) entail that my visualisation of the car or
any other instance of imagining involves a cognitive commitment towards how the actual world is like.119 In addition, since the actual world is,
or corresponds to, one of the possible worlds, making a claim about how
things are in the actual world implies making a claim about how they are in
some possible world. Accordingly, the Modal Account might have to assume that the (stronger) cognitive attitude includes a (weaker) imaginative
attitude. Hence, what is distinctive of imaginings might not be the possession of an imaginative commitment, but rather the possession of nothing
but an imaginative commitment.120
Therefore, if Whites claim that imagining something (in my sense) consists in
[thinking] of what it does or would look like, or of what it is or would be like (see
the first and third of the passages quoted in the main text) should instead be read
as meaning that imagining may at least sometimes involve a commitment to how
things actually are (i.e., a cognitive attitude), it would clearly turn out to be false
(see Section 2.1) and also undermine his strict distinction between imaginings and
cognitions. But his additional statement that imagining something means thinking
of it as something that could be happening, whether or not it did (my emphasis; see
the second passage quoted in the main text) indicates that he does not take imagining
to involve a cognitive attitude and merely allows for the case that what we are imagining may happen to be actually the case.
120
This qualification of the attitude version of the Modal Account seems even more
plausible if it is accepted that possible worlds do not really exist, or at least not in
the same substantial sense as the actual world. According to this view, possible
worlds are entities of a completely different ontological category than the actual
world: while the latter enjoys substantial material existence, the former are most
likely sets of propositions or statements that describe things as being a certain way
(see Divers (2002): ch. 2). Hence, cognitions may have in common with imaginings
119
167
But the Modal Account gets even more complicated. The difference
between judgementally and imaginatively entertaining a modal proposition
cannot simply be identified with a difference in attitude towards the same
non-modal proposition, since this would be unable to deal with iterated
cases. For instance, there would be no difference between imagining that it
rains, imagining that it is possible that it rains and imagining that it is possible that it is possible that it rains. It is therefore more promising to relativise possibility to particular worlds of evaluation (see, e.g., Priest
(2008)). Accordingly, there is a (formal) difference between something being possible in (or from the perspective of) one world and the same thing
being possible in (or from the perspective of) some other world, given that
which alternative possible worlds are relevant for the possible truths in one
world is assumed to be independent of (though not necessarily different
from) which alternative possible worlds are relevant for the possible truths
in the other world. Imagining that it rains can then be differentiated from
imagining that it is possible that it rains in the following way: while the
former endorses the proposition it rains as being actually true in some
possible world, the latter endorses the same proposition as being possibly
true in some (e.g., the same) possible world. Further iterations allow for a
similar treatment: imagining that it is possible that it is possible that it
rains, for instance, endorses the proposition it rains as being possibly possibly true in some possible world (i.e., as being possibly true in some world
that is possible from the perspective of the (distinct) imagined possible
world).
Finally, since the best version of the Modal Account (i.e., its attitude
version) identifies the attitudinal difference between imaginings and cognitions as their most basic difference, it in fact constitutes an instance of the
more encompassing Attitude Account. I evaluate the general prospects of
this broad approach to the unity of imagining and, hence, also of its explicitly modal version in the subsequent section. But, before that, I would
that they are concerned with a description of how things are, but differ from imaginings in that they take this description to be true of or, more neutrally, to apply to
the actual world. This does not turn the attitude version of the Modal Account into a
version of the Ontological Account, though, since imaginings and cognitions would
still represent and refer to objects of the same ontological category (i.e., trees or cars
that may exist in the material actual world).
168
121
169
tion of the Modal Account constitutes a specific version of the Phenomenal Account and therefore also faces the general objection raised above
against this approach to imagining (see Section 5.2). The worry was that
essential phenomenal differences such as the one in attitude are unlikely to be basic. Instead, it should be expected that they can be further
elucidated in terms of some underlying features in this case, perhaps, by
reference to the difference in epistemic function. Indeed, the phenomenologists, who have identified the strict difference in commitment between
imaginings and cognitions, need not necessarily be opposed to such an explanatory approach, despite their focus on phenomenological description.122
The main thesis of the Attitude Account is that it is distinctive of, and
fundamental to, imaginings that they possess an imaginative attitude (as
well as a cognitive one). For the same reasons as in the case of the Modal
Account (see the preceding section), the Attitude Account has to specify
the imaginativeness of representations in terms of the presence of an imaginative attitude, rather than merely in terms of the absence of a cognitive
attitude. For, otherwise, imaginings would not be distinguished from mental episodes such as conative feelings or episodes of wondering that are
neither cognitive, nor imaginative. Similarly, at least some versions of the
Attitude Account have to characterise imaginings also by reference to their
lack of a cognitive attitude namely if the possession of an imaginative attitude is weaker than, and included in, the possession of a cognitive attitude. This qualification applies, for instance, to the modal version of the
Attitude Account (see the preceding section).123
It is interesting to ask whether phenomenology understood as a set of methodologies is purely descriptive, or also explanatory. I tend to favour the second option,
which assumes that phenomenological observations should also aim at making sense
of phenomenal aspects in terms of non-phenomenal features (which still count as
phenomenologically salient in so far as they are reflected by phenomenal aspects;
see Dorsch (2012a) and Soldati & Dorsch (2011)). The debate in question is also
closely related to the issue to which extent phenomenologists like Husserl were realists about the external world (see Soldati (1994)).
123
Whether a mental episode may simultaneously possess a cognitive and an imaginative attitude depends on what the postulated imaginative attitude specifically amounts
to. The Modal Account is a case in point: since it characterises the imaginative attitude in terms of a commitment about how things are in some possible world, and
since the actual world is or corresponds to such a world, the view has to maintain
122
170
171
172
126
173
Note that the quoted passage about the suspension of belief applies to sensory representation and, hence, is not literally concerned with intellectual belief and its suspension. Rather, what Sartre seems to have in mind here is the cancellation of the
cognitive attitude (or thesis) that is common to both conscious perceptions and occurrent beliefs (i.e., judgemental thoughts).
174
cognitive attitude. So, Sartres view does not provide a fully unified
elucidation of what renders an attitude imaginative, rather than cognitive;
and it is unclear how his characterisation could be completed. 128 Husserls
postulation of a single imaginative attitude obviously does not face a similar challenge.
Second, the unity of imagining is also under threat by the fact that it is
unclear how his account could be extended to non-sensory imaginings
(something which Sartre himself does not aspire to, though). Intellectual
imaginings seem to share the same imaginative attitude with each other.
But none of the four listed by Sartre is a natural candidate for it. Imagining
that the Earth is flat does not involve a commitment to the absence of that
state of affairs from our actual environment. In fact, the representation is
neutral on what is really the case. This is reflected by the fact that imagining that the Earth is flat is in rational tension neither with believing that the
Earth is flat, nor with believing that it is not flat. Accordingly, the imaginative representation under consideration also does not make the committal
claim that the represented state of affairs does not exist or obtain. Similarly, imagining that the Earth is flat does not purport to be about a state of
affairs that exists elsewhere in the actual world (e.g., on the other side of
the hemisphere, or beyond the Moon). Last, switching attention to the neutralising attitude does not help since suspension of belief involves the neutralisation of judgemental endorsement and, hence, would be wrongly classified as an instance of imagining. Again, Husserls proposal is immune to
this problem. But with respect to the following two objections, Sartres and
Husserls views are on equally bad footing.
Third, episodic memories and other non-imaginative representations
may involve exactly the same commitments. When I remember the appearance of a friend, who now lives on the other side of the planet, my memory
is likely to present him or her as absent or existing elsewhere. Similarly,
memories of people who have died may very well involve an impression of
their non-existence. This is also reflected in the fact that such episodic
It does not matter here whether Sartre also thinks that each sensory imagining constitutively involves a thought (see Section 3.6), which may furthermore be responsible for its exemplification of one of the four types of imaginative representation.
For such additional thoughts could at best be taken to constitute the various commitments, but not to unify them in one general kind.
128
175
memories just like sensory imaginings do not move us to the same kind
of actions towards the represented people, which perceptions of them
would incline us to. The natural reply is, of course, to say that memories
but not imaginings are still non-neutral about the existence of the represented entities in the actual past. So, the first of the four attitudes seems to
fare best because of its ability to accommodate episodic memories. But
there are still other mental representations such as our visual experiences
of the pictorial content of pictures, hallucinations that are recognised by us
as such, or spontaneously occurring images and thoughts that involve
some or all of the proposed imaginative attitudes, despite failing to be instances of imagining (see, for instance, Dorsch (2010a) and (2013), as well
as Section 14.4). Accordingly, Husserls and Sartres views do not succeed
in singling out what is distinctive about imaginings.129
The last problem (and possibly also the second) persists even if only one
of the four attitudes identified by Sartre is chosen to be definitional of imaginings. Not all of them manage to distinguish imaginings from memories;
and none of them succeeds in distinguishing imaginings from the representational element involved in pictorial experience that is concerned with
the depicted. This points to a general problem for the Attitude Account,
which arises independently of how the attitudinal difference between imaginings and cognitions is characterised: namely the problem to accommodate non-imaginative mental representations that are, none the less, committal in a very similar (non-cognitive) way as imaginings. Hence, at least in
its current state, the Attitude Account is untenable as a unified theory of
imagining.
Even the modal version of the Attitude Account has difficulties to distinguish sensory imaginings from pictorial experiences or spontaneous images, without wrongly identifying them as being imaginative themselves.
129
176
177
that, while perceptions originate in the perceived external objects, imaginings originate in the subjects spontaneity. The Spontaneity Account
maintains that this fact constitutes a basic and essential difference between
imaginative and cognitive episodes.
The main challenge that this view faces is, however, to come up with a
satisfactory characterisation of this special way of originating in the subjects mind. Sartre starts off with contrasting the spontaneity of sensory
imaginings with the passivity of perceptual experiences (note that Sartre
uses consciousness in roughly the same sense in which I use episode):
A perceptual consciousness appears to itself as passive. On the other hand,
an imaging consciousness gives itself to itself as an imaging consciousness,
which is to say as a spontaneity that produces and conserves the object as
imagined. It is a kind of indefinable counterpart to the fact that the object
gives itself as a nothingness. The consciousness appears to itself as creative,
but without positing as object this creative character. (Sartre (1940): 14)
178
179
salient, a more substantial difference between internal and external origin: while perceptual experiences are due to factors external to the subject
or his or her conscious mind (whether these factors are perceived external
objects, processes in the brain or something else), sensory imaginings arise
from within the subject or mind (such as mental agency, rationality, association, and so on).134 This assumption would further stress the presumably
intended resemblance between Sartres distinction between passivity
and spontaneity and the Kantian distinction between receptivity and
spontaneity: that is, the distinction between the causal impressions that
we suffer and our own subjective contributions in the forms of mental acts,
responses, and so on.135 But the assumption of an underlying difference in
real origin fits also well with Sartres identification of two kinds of spontaneity (with image, he refers here to an instance of sensory imagining):
We may be reproached for loading the dice by choosing a voluntarily produced representation as an example of a mental image. In most cases, no
doubt, the image springs from a deep spontaneity that cannot be assimilated
to the will. It seems that the involuntary image appears to consciousness as
my friend Pierre might appear to me from around the corner in a street.
(Sartre (1940): 18)
[I]nvoluntary and voluntary images represent two very closely related types
of consciousness, of which one is produced by a voluntary spontaneity and
the other by a spontaneity without will. (Sartre (1940): 19)
It is not clear how best to describe this contrast. Among the options are mental/nonmental, subjective/objective, or personal/impersonal. There is also the problem
of where to place mental dispositions, unconscious mental states, and so on. See also
the objection below (and in Section 14.4) to the effect that unbidden (but non-perceptual or non-judgemental) images and thoughts resemble perceptions more than
mental actions in their origin.
135
I return to this issue in Section 5.6, in the context of discussion the Spontaneity Account of imagining. The distinction presented in the main text is not really Kants
own distinction (see his (1993)), since he does not take receptivity to be causal
(i.e., causality is a relation between empirical entities, and not between the subject
and the transcendent entities impinging on it), and since he assumes the subjective
contribution to be transcendental and pertaining exclusively to reason and understanding.
134
180
181
See Casey (1976): 33-35, and ch. 3. Caseys choice of terminology seems to be
partly motivated by the thought that voluntary and involuntary formations of mental
representations do not have anything significant in common. OShaughnessy, too,
takes these two possible kinds of origin to differ essentially from each other. But he
is fully aware of the consequences of this essential difference for a unified account
of imagining that allows for involuntary imaginings: namely that voluntary and involuntary imaginings do not belong to the same mental kind (see the last part of Section 7.1).
In what follows, I generally accept Caseys characterisation of spontaneity and,
especially, the contrast that he draws between involuntary spontaneity and voluntary
mental agency (see Section 13.4), as well as the observation that perceptions are as
spontaneous as unbidden images or thoughts (see the discussion below in this
section). This continuity in understanding is the main reason why I have decided to
signal the distinct and independent notion of spontaneity to be found in Sartres
writings by using single quotation marks.
Casey also suggests there that all imaginings are in principle controllable, that is,
subject to the will. But it is not clear to which extent spontaneous images or thoughts
which he takes to be imaginative are indeed subject to the will (see Section
14.4). Besides, describing imaginings as being subject to the will is neither very
clear, nor very helpful for the formulation of a theory of imagining (see Section
13.5).
182
To these three features identified by Casey, a fourth may be added: spontaneous representations are usually fleeting in character and disappear as
quickly as they have occurred (see Section 14.4). This is closely related to
the fact that they serve no specific function in our mental lives (see Dorsch
(2010a)).
However, perceptual experiences, too, may show these marks of spontaneity. They are not actively formed and therefore arise without effort and,
typically, also instantaneously and unexpectedly.139 Furthermore, at least
some of them are rather transient because they are concerned with rapidly
changing or moving objects (e.g., cars passing by) or involve rapid changes
in perspective (e.g., because the subject is falling, running, or sitting in a
train). In the light of these observations, it is not surprising to see that Casey tries to say a bit more about the nature of (involuntary) spontaneity and
how it differs both from perception and from imaginative agency:
[A] truly spontaneous phenomenon initiates itself rather than being initiated
by other phenomena: it is autogenous. [...] [A] spontaneous imaginative experience does not present itself as embedded in any [well-ordered, causally
concatenated series of entities or events]. Instead of emerging from a nexus
of causally efficacious factors, it arises suddenly, in a psychical vacuum. To
be sure, upon reflection the imaginer may succeed in locating the spontaneous imaginative appearance within a certain causal context, but this context does not as such enter into the specific content of his experience. What
is experienced is rather an imaginative act-cum-presentation, which presents
itself as unconnected with preceding or surrounding circumstances, including even those circumstances provided by former acts of imagining. This
act-presentation appears of and by itself, as quite independent of the imaginers immediate practical situation (a situation that may even include particular plans and projects for imagining). It is as if this situation, though undeniably present, were irrelevant to the spontaneity of imagining. (Casey
(1976): 68)
This is not to say that perception is completely passive. Perceptual experiences may
involve the activity of attention and may also form the core of certain kinds of mental action, such as watching something or looking for something to happen (see the
work of Crowther, such as his (2010)). But even in the latter cases, the perceptual
experiences concerned occur unbidden, and without any voluntary determination of
what is represented.
139
183
The distinction that Casey draws in this passage concerns, first of all, the
phenomenal character of the kinds of mental episode concerned. His first
observation about the character of spontaneous episodes is negative: they
are not subjectively given as being causally (or otherwise) related to other
mental phenomena or entities external to the mind (see the second sentence
in the quoted passage). In addition, Casey puts forward a second, more
positive description: spontaneous representations are phenomenally presented as being causally (and otherwise) unconnected from everything else
(see the fifth and the sixth sentences). Not appearing to stand in any causal
connection is clearly different from appearing to stand in no causal connection: while the former corresponds to the absence of a certain phenomenal
aspect (i.e., the aspect present in perception, judgement, and so on), the latter pertains to the presence of a different phenomenal aspect (i.e., one that
would be distinctive of imaginings). Accordingly, Casey takes the phenomenal character of spontaneous representations to include an aspect which
positively marks them as lacking any causal (or other) links to other elements of the mind and the world.
But, in contrast to Sartre, Casey also makes some assumptions about the
underlying actual origin of imaginings. First of all, he accepts the natural
thought that spontaneous images and thoughts may have causes that are external to, and independent from, their spontaneous effects. These causes are
just not phenomenologically salient. More important, Casey notes that
spontaneous representations occur of and by [themselves] and [initiate
themselves]. This cannot mean that spontaneous images and thoughts are
their own causes, given that this would not only involve a potentially incoherent and very strong notion of self-causation, but also because it would
contradict the accepted possibility of underlying external causation. Rather,
what Casey seems to have in mind is that spontaneous representations do
not occur in virtue of the impact of other phenomena that are accessible
from our subjective, personal perspective. Both mental agency and perceived external objects are such phenomena: we are consciously aware of
them, either through phenomenal awareness and introspection, or through
perceptual awareness (see Dorsch (2012a) for more on these three distinct
kinds of awareness).
Accordingly, what is said to be distinctive of spontaneous representa-
184
tions is their phenomenologically salient absence of an origin in some conscious mental episode or some consciously represented external object. Together with the specification of voluntary representations in terms of mental agency, the presented characterisation of the involuntary form of spontaneity (in Sartres sense) enables us to provide a more precise formulation
of the main claim of the Spontaneity Account: imaginings are differentiated from all other mental phenomena in their origin either in mental
agency or in nothing that we are consciously aware of.
The resulting view promises to be able to distinguish spontaneous representations from most other mental phenomena. For instance, while perceptions are determined by salient aspects of the external world, judgemental thoughts and conscious decisions are rational responses to consciously represented facts or considerations, and voluntary imaginings and
other mental actions originate in some conscious tryings or intentions-inactions (see Dorsch (2013)). Moreover, the Spontaneity Account is compatible with the basic contrast that Sartre appears to track with his phenomenological distinction discussed above between spontaneity and perceptual passivity: namely the contrast between what merely happens to us
and what we are ourselves doing, or between what we suffer as mere objects (e.g., in the shape of impressions or reflexes) and what counts as our
own response as subjects (e.g., in the shape of rational responses, associations, manifestations of mental dispositions, and the like). 140 The idea is
that spontaneous representations do not come into existence due to the impact of external objects, but originate in some associative, dispositional or
similar aspects of the mind.
In addition, that we experience sensory imaginings as creative (and as
arising from within the mind), while we experience perceptions as passive
(and as arising from reality), may be used to establish an intimate connection between the attitudinal and the origination difference. The thought is
that perceptual experiences involve a commitment to reality because they
The Kantian tradition understands this distinction largely in rational terms: namely
by differentiating the rational in the minimal sense of a responsiveness to epistemic, practical or other kinds of reason from the non-rational or merely causal
(see McDowell (1994) for a good example). A comparable though far less strict
and divisive Humean contrast is perhaps that between impressions and ideas (see
Chapter 9).
140
185
are brought about by (or nomologically linked to) parts of the world; while
sensory imaginings lack such a commitment because they are brought
about by (or otherwise closely linked to) elements in the mind. Hence, the
phenomenal difference in commitment may be constituted partly by the underlying, phenomenologically salient difference in origin. The Spontaneity
Account therefore promises to explain at least some of the principal features of imagining.
But one major objection to the Spontaneity Account is that it cannot distinguish (sensory) imaginings from certain non-imaginative phenomena,
notably hallucinations and episodic memories. In particular, neither hallucinations, nor involuntary recollections occur due to the influence of external objects or conscious manifestations of theoretical or practical rationality (i.e., in response to noticed evidence, motivational states, and so
on).141 Indeed, the subpersonal processes, which are responsible for the occurrence of hallucinations and unbidden memories, may not be very dissimilar from those involved in (non-perceptual and non-mnemonic) spontaneous representations.
The other main objection to the Spontaneity Account is that it wrongly
assumes two basic forms of imagining namely voluntary and involuntary
ones rather than one. This threatens the unity of imagining in so far as
spontaneity (in Caseys sense) and mental agency do not have anything
significant in common, which they also do not share with, say, the rational
origin of judgemental thoughts in perceptions or conscious inferences.
Judgements originate as much in the mind rather than in the world as
voluntary imaginings or spontaneous representations. That is, all three
kinds of mental episode are not impressions on the mind, but belong to our
subjective doings or responses. Hence, Sartres notion of a spontaneous
origin is to wide to single out only imaginings; and it is unclear how it
could be narrowed down to exclude rationally formed judgemental
thoughts without becoming simply a disjunction of two kinds of origin
(i.e., mental agency and Caseys spontaneity).
141
See Dorsch (2010a). Voluntary memories are no problem since they do not allow for
the same kind of direct agency that is characteristic of voluntary imaginings (see
Section 13.2 and Dorsch (2009b)). In Section 14.2, I distinguish the different mental
origins of (voluntary) imaginings and voluntary memories with respect to whether
they involve a specifically imaginative kind of mental agency.
186
This second objection is closely related to the phenomenological observation that the way, in which spontaneous representations occur, seems to
be subjectively much closer in nature to that of perceptions and sensations
than to that of judgemental thoughts or mental actions. Talk of spontaneity should not cloud the fact that unbidden images or thoughts are not
subjectively experienced as coming from ourselves, or as subjective doings or responses. Indeed, as Casey has already noted in the quoted passage, the only thing that we can come to know from our first-personal perspective is where spontaneous representations do not come from: namely
neither from the world, nor from our will or other conscious mental phenomena. But we do not become subjectively aware of where they do come
from: they are simply given as unrelated to all entities that we are consciously aware of. Not surprisingly, we take them to come from nowhere,
to occur out of the blue or, indeed, to appear by and of [themselves].
Phenomenological evidence therefore indicates that spontaneous representations are not among our (involuntary) doings or responses, but enjoy
rather a similarly passive origin as perceptions and sensations. 142 But it also
suggests that spontaneous representations in contrast to perceptions and
sensations do not originate in the world (including our body). Hence,
they seem to differ significantly, not only from voluntary imaginings and
judgemental thoughts, but also from perceptions and sensations. It is therefore perhaps best to treat them neither as being imaginative, nor as being
cognitive, but instead as belonging to an entirely different, third category
of representation (see Section 14.4).
It might be rightly insisted that the four characteristics of spontaneity
listed above constitute aspects of the phenomenal character of the respective spontaneous representations. But, as already mentioned, these aspects
are not sufficient to distinguish spontaneous representations from (certain)
perceptions. Hence, they also do not suffice to phenomenally mark the difference in origin between the two kinds of mental representation. PercepInterestingly, Sartre himself seems to make a similar observation when he states in
the passage quoted above that the involuntary image appears to consciousness as
my friend Pierre might appear to me from around the corner in a street (Sartre
(1940): 18). But, as also illustrated above, he still thinks that involuntary images differ significantly from perceptions due to their spontaneous origin (see Sartre
(1940): 14 and 33).
142
187
See Dorsch (2009b) and Dorsch (2013). Note that the kind of first-personal awareness concerned is sometimes erroneous. Perfect hallucinations are not relational perceptions; but we still experience them as such (see Dorsch (2012a) and Soldati &
Dorsch (2011)). And judgemental thoughts are given to us in consciousness as responses to epistemic reasons, even if they have been irrationally caused by some
other mental phenomena (e.g., an emotion or preference), or by some drug or pathology (see Dorsch (2009b)).
188
Part Two
The Epistemological Account
The three chapters of this part of the book are devoted to the presentation
and assessment of the Epistemological Account of imagining and, primarily, of its exemplary version, the unified account of imagining developed
by Brian OShaughnessy. In the first chapter, I start with providing a succinct summary of OShaughnessys position, before expanding in more detail on his main claims and overall argumentative strategy. The middle
chapter is then dedicated to a thorough study of OShaughnessys complex
considerations and arguments in favour of endorsing his theory of imagining. The evaluation of the Epistemological Account, and of OShaughnessys account in particular, will have to wait until the last of the three
chapters.
The insertion of a short synopsis in addition to an extended discussion is
motivated by the wish to enable the reader to follow the main argument
against the Epistemological Account without having to acquaint him- or
herself with many of the details of OShaughnessys particular version of
it, or with the textual exegesis that the presentation of these details require.
Reading the middle chapter and much of the first chapter is therefore not
required to engage with the objections to the Epistemological Account
presented in the last chapter of this part of the book.
The main idea of the Epistemological Account is to characterise imaginings as non-cognitive in one form or another. This does not mean that all
theories of imagining which assume episodes of imagining to be different
from cognitive phenomena are versions of the Epistemological Account.
For the Epistemological Account goes beyond the thesis that imaginative
192
episodes are not cognitive episodes, and imaginative projects not cognitive
projects. What it also claims is that it is distinctive of imaginings that they
are non-cognitive, and that their non-cognitivity is among the basic features in virtue of which they count as imaginative. Hence, any version of
the Epistemological Account of imaginings makes essential reference to
their non-cognitivity: it assigns their non-cognitivity a central place in the
specification of their nature (and not merely, say, in the specification of the
implications of their nature).
That the Epistemological Account takes the non-cognitivity of imaginings not only to be essential to them, but also to be among their fundamental essential features which cannot be explained by pointing to even
more basic characteristics, distinguishes it from other unified theories that
also accept the essential non-cognitivity of imaginings. One example is the
Agency Account of imagining, or at least some versions of it. It differs
from the Epistemological Account partly because it does not take the noncognitivity of imaginings to be basic, but characterises and explains it as a
consequence of their status as instances of mental agency (see Sections 4.7
and 14.1). Similarly, proponents of the Dependency Account may very well
accept the necessary non-cognitivity of imaginative phenomena, while deriving it from their more fundamental feature of being representations or
echoes of cognitive phenomena (see Part Three). None the less, the Epistemological Account may be combined with another theory about the fundamental essence of imagining, as long as the two accounts are concerned
with different and independent aspects of their basic nature. In other words
is, the negative and epistemic conception of imaginings as essentially noncognitive may be supplemented with a more positive and non-epistemic
characterisation of the other part of their nature.
Indeed, such a supplementation may be deemed necessary because of
the fact that the Epistemological Account characterises imaginings primarily in terms of what they are not, rather than in terms of what they are. In
Section 8.1, I return to the issue of whether some versions of the Epistemological Account have the resources to provide an exhaustive unified account of imagining, or whether they need to be supplemented by some other fundamental theory about the nature of imaginings.
The worry about the comprehensiveness of the Epistemological Account
193
may also help to explain why there is no actual defender of its exhaustiveness. In the preceding chapters, I highlighted the fact that, while Sartre and
Wittgenstein accept the essentiality of the non-cognitivity of imaginings,
they deny the fundamentality of their non-cognitivity and try instead to explain it in terms of more basic features such as a specific origin in the
mind or the will (see especially Sections 4.7 and 5.6). OShaughnessy, the
chief proponent of the Epistemological Account, on the other hand, combines the thesis about the fundamental non-cognitivity of imaginings with a
claim about their equally fundamental representational connection to cognitions. That is, he also endorses the following echo claim about imaginings (see Part Three for more on such claims):
(ER)
OShaughnessy supplements the Epistemological Account with a representational version of the Dependency Account partly because of the difficulty of the Epistemological Account to characterise the nature of imaginings in more positive terms (see Section 8.1). 1 Therefore, in what follows, I
treat the Epistemological Account primarily as one key element in a more
complex unified theory of imagining, rather than as constituting a self-sufficient unified account in its own right. Besides, my discussion is concerned as much with the fundamentality of the claim that imaginings are
essentially non-cognitive as with its truth.
As already suggested, the Epistemological Account can come in various
forms, depending on what is understood by the claim that imaginings are
non-cognitive. But the central claims of these different versions of the Epistemological Account share a structure, because of which they may be la1
OShaughnessys writings do not completely rule out the possibility that the noncognitive aspect of the nature of imaginings may be traced back to, and explained by
reference to, the representational one. It is therefore perhaps also possible to read
him as defending a pure version of the Dependency Account, which does not require
the simultaneous endorsement of an Epistemological Account. But it would then be
difficult to make sense of his argumentative strategy of supporting the negation
claims endorsed by him without any reference to the echo claim equally accepted by
him (see Section 6.3).
194
195
196
Lastly, and perhaps of most interest in the current context, it constitutes the
main and possibly sole manifestation of the Epistemological Account.
For these reasons, it is apposite to devote two whole chapters (Chapters 6
and 7) to the presentation of the view of OShaughnessy and his arguments
for it. However, since many of the details of his view are not crucial to the
assessment of the Epistemological Account, I start the first of these
chapters with a brief summary of OShaughnessys principal claims and his
foremost reasons for endorsing them, thus rendering engagement with the
rest of that chapter and the subsequent one entirely optional. The last
chapter (number 8) is reserved for my objections to the Epistemological
Account and, notably, OShaughnessys version of it.
CHAPTER 6
OShaughnessys View
198
OShaughnessys View
199
Again, I ignore here that OShaughnessy formulates his view in terms of necessary
and sufficient conditions for being an imagining, rather than in terms of an essence
common to all imaginings (see Sections 1.1 and 6.2).
200
Indeed, OShaughnessy does not mention memory at all in his discussion of imagining. Moreover, he talks about memory only very briefly in his book, and only in the
context of its role in the continuity of conscious experience and agency (see
OShaughnessy (2000): 55ff.).
OShaughnessys View
201
sensory imaginings.
One thing to note about this omission is that it is not entirely warranted.
Sensory imaginings seem to be much closer to sensory memories than to
perceptual experiences both in their phenomenal character and in their
causal history (see Martin (2001) and Hopkins (2011a)). So, it seems
reasonable to assume that (if at all) sensory imaginings are more likely to
satisfy the cognitive constraints on successful memory rather than those on
successful perception. This fits well with the fact that it seems more
promising to treat sensory imaginings and episodic memories as belonging
to one and the same mental kind (i.e., imagining) than to do the same with
respect to sensory imaginings and perceptual experiences.10 But the
omission is none the less understandable in so far as it may be reasonably
argued that the special origin of imaginings in the will or in a breakdown
of normal informational processes also prevents them from satisfying the
requirements pertaining to knowledge-constituting memories. That is, there
is no reason to question the extension of (NC*) as well as
OShaughnessys Argument from Origin in favour of it to memory as a
third cognitive prototype. The open formulation of (NC*) and the focus
on this negation claim rather than on OShaughnessys more restricted one
is meant to reflect the possibility of this extension and to allow my
discussion to stay neutral on the issue of how many cognitive ideals there
are, and what their specific nature is.
As already indicated, OShaughnessy supplements his versions of the
negation claims (NC*) and (NC) with a more positive thesis that postulates
a particular dependency relation between imaginative and cognitive episodes: the former are said to be essentially representations of the latter. This
claim which is just the thesis (ER) is at the heart of the representational
10
Hume might be read as endorsing the second view (see Chapter 9, also for objections to his underlying picture of the mind). Hopkins (2011a) defends the first view
by arguing that episodic memories are contextually constrained sensory imaginings:
namely sensory imaginings that are veridical and, in the right way, determined by
the past. OShaughnessys reply to Hopkins could be that imaginings and memories
belong to essentially different mental kinds because of their differences in functional
role, phenomenal character, and determination by the past (see Martin (2001)).
Hence, for him, (non-perceptual) visual experiences involving a claim about the past
and being constrained by the past should not count as imaginative.
202
OShaughnessys View
203
In fact, OShaughnessy thinks that cognition is the only means by which our mind
can really be linked in contrast to merely seeming, or merely presenting itself, to
be linked to reality (see the introduction to OShaughnessy (2000): for instance 6f.
and 12f.). Correspondingly, he also defines the common cognitive significance of
knowledge and perception in terms of the fact that they [represent] Reality (ibid.:
204
This quote makes clear that, for OShaughnessy, all imaginings are to be
elucidated by reference to either perception (and its modal variants) or
knowledge (and its variants, such as self-knowledge, etc.). However, it also
suggests that he thinks that there are no other cognitive prototypes than
these two: after all, they are the fundamental mental cognitives. This is in
line with the fact that OShaughnessy does not mention in his book the existence of any other cognitive prototype. But even if there were other prototypes, they would not according to what OShaughnessy claims be
relevant for imaginings: the latter pertain uniquely and only to perception
and knowledge. His definition of imaginings thus expresses the view that
all instances of imagining are related to one or other of the two prototypes.
The relation in question is that of quasi or seeming instantiation (ibid.:
363). For OShaughnessy, there are two aspects to this relation of quasi exemplification. The first is that, although imaginings resemble their cognitive counterparts in many respects, they cannot really instantiate perception
363).
OShaughnessys View
205
or knowledge.
As OShaughnessy notes, there are certain similarities (or a certain resemblance, as he calls it) between instances of the relevant two kinds of
mental state (ibid.: 363-6). While sensory imaginings are like perception,
intellectual imaginings are like knowledge.12 For not only are both the imaginings and the cognitive prototypes representational, but the imaginings
can have the same intentional objects and be of the same types of representation as the respective prototypes (ibid.: 166f.). Thus, both sensory imaginings and perceptions represent perceivable objects and features relative
to a certain modality; and both intellectual imaginings and states of knowledge involve concepts and take propositions as their objects (ibid.: 365f.).
Imaginings imitate and are like their cognitive prototypes and may
hence also sometimes be mistaken for them (ibid.: 365).
None the less, OShaughnessy maintains that imaginings necessarily
differ from, and do not fully amount to, their cognitive prototypes (ibid.:
359ff.). Given that the similarity claim does not play a very significant role
in his account,13 it is this negation claim about imaginings which constitutes the first important part of what he means when he says that imaginings are only quasi the cognitive prototypes, but never the real thing
(ibid.: 361). More specifically, OShaughnessy endorses the weak negation
claim that no imagining can cognise reality by instantiating one of the two
Here, I largely ignore a third kind of imagining that OShaughnessy identifies,
namely the imaginative perception supposedly involved in pictorial experience (but
see Note 13 below, Note 7 in Chapter 8, Note 13 in Chapter 11, as well as Dorsch
(2005): sec. 6.5., and Dorsch (2012c), for some further discussion). OShaughnessy
also does not speak of sensory and intellectual imaginings, but instead of directobject and propositional imaginings. His main reason for this is that he believes
that all sensory imaginings are literally imaginings of experiences and are therefore
always directed at objects (i.e., the experiences), while only intellectual imaginings
can have propositions as their objects (see OShaughnessy (2000): 342f.). However,
since this extra claim about the nature of sensory imaginings is controversial (see
Dorsch (2010c)), I use the more neutral terminology.
13
Apart from elucidating the representational nature of imaginings, the similarity
claims primary significance is that it may give some support for the endorsement of
OShaughnessys version of (ER) to be introduced below. The idea is that only this
echo thesis the claim that imaginings have cognitive prototypes as their directly
represented objects can explain why the similarities hold. I return in Part Three to
the issue of how strong a support this provides for (ER) or similar echo claims.
12
206
The only notable difference is, again, that OShaughnessys account is concerned
with providing a definition of imagining, while my concern is with characterising its
essence. This difference in aim is reflected in the concrete formulation of OShaughnessys specific version of (NC*), as well as in the formulation of his specific versions of (ER) and (NC) to be introduced further below. The issue of contention is not
whether each instance of imagining possesses an essence: it surely does, as do all
other entities. Rather, we disagree on whether all imaginings share one and the same
essence (see Section 1.1).
OShaughnessys View
207
episode. Accordingly, OShaughnessy assumes that all sensory and intellectual forms of imagining are instances of experiential imagining: they
consist in imaginings of types of mental episodes (ibid.: 359). What he has
in mind is that we imagine something by imagining our relevant cognitive
access to it: that is, we visualise a face by imagining seeing (or seem seeing) one, or we make-believe that it rains by imagining believing (or
seem believing) the corresponding proposition (ibid.: 363f.). 15 Second,
the imaginings and the imagined cognitive prototypes share the same inten15
In fact, the situation is (again) a bit more complicated. As already noted, OShaughnessy insists that there is no such thing as imagining per se: different forms of imagining possess different essences (see OShaughnessy (2000): sec. 11.6 and ch. 12).
We can visually, auditorily or propositionally imagine something; but we cannot
simply imagine it. This is why he writes that there is no such thing as merely imagining an object, and that imagining a seeing acquires its identity not under imagining, but under visual imagining (ibid.: 364; see also 366). Indeed, it is structurally misleading to treat imagining as a fundamental mode in which we may be
consciously related to entities (e.g., types of experiences and their external objects);
rather, the fundamental modes of awareness include imagining seeing, imagining
hearing, and so on, which are intentionally directed at external objects, and not at
types of experiences (ibid.: 364).
Noordhof (in a talk presented at the Perceptual Memory and Perceptual Imagination conference in Glasgow 2011) takes this to mean that OShaughnessy rejects the
Dependency Account: that visual imagining is just another kind of visual experience,
the content of which is exhausted by the external object visualised. But OShaughnessys insistence that that which is being imagined [...] in the case (say) of visual
imagining is identical with the visual perceiving of an object (ibid.: 363; this passage is also quoted above) undermines this interpretation. Instead, OShaughnessys
point here is, again, that the different forms of imaginings do not share a common
essence (and, especially, a common mode of awareness).
More specifically, the imaginative element in question that imaginings constitutively depend on, and non-intentionally represent, types of cognition (in a way similar to the representation of sense-data) does not correspond to a detachable basic
mode of awareness precisely because the non-intentional nature of the representation
ensures that different forms of imagining are constitutively linked to different types
of cognition.
This also fits well OShaughnessys observation that perceptual and imaginative
awareness are not two independent first-order forms of awareness that may have the
same objects of awareness, but instead a pair of independent first-order and dependent second-order awareness: [t]here are not two distinct ways of relating to an
identical object-content, the perceptual way and the imaginative way: there is one
208
tional objects and, presumably, represent them as having the same features
(e.g., as being a green apple).16
In addition, OShaughnessy understands the relation between imaginings and the imagined prototypes as non-intentional (ibid.: 363 and 371).
He also maintains that imaginings of a certain kind (e.g., auditory imaginings) could not exist if the corresponding kind of prototype (in this case,
auditory perception) did not exist (364; 377). Consequently, imaginings are
said to be functions of the cognitive prototypes in two senses: the former
have to be described in terms of the latter (as indicated in the quote above);
and the former are constitutively related to the latter. In my terminology,
this implies that imaginings are echoes of the cognitive prototypes (see the
introduction to Part Three). It is precisely this specific echoing relationship
between imaginings and the respective prototypes, which the second aspect
of the relation of quasi exemplification introduced above consists in. It can
way, and a shadow of that way (ibid.: 365).
There is, however, still the problem to make sense of OShaughnessys idea of a
non-intentional, relational form of representation that has, as its second relatum, not
a particular mental episodes (or other object), but just a type thereof. Indeed, proponents of an intentionalist understanding of the representational relation between
imaginings and cognitions may point out, for instance, that their view also implies
that imaginings are relations to types of cognitions, given that the general content of
intentional representations constitutively depends on the represented properties (e.g.,
we cannot have an experience as of something red, if the universal redness does not
exist). This may very well constitute one motivation of preferring the intentionalist
reading of (ER) over the relationalist interpretation.
16
Note, however, that the first aspect of OShaughnessys claim that imagining is
merely a [second-order] function only and uniquely of its prototype (OShaughnessy (2000): 366) implies that imaginings do not directly (or on their own) relate to
the intentional objects, but only via the imagined cognitive representations. For the
immediate objects of imagining (i.e., that which is being imagined (ibid.: 363))
are possible instances of one or the other cognitive prototype. As a consequence,
there are not two distinct ways of relating to an identical object-content, the perceptual way and the imaginative way: there is one way, and a shadow of that way
(ibid.: 365). It is in this sense that imaginings possess a second-order nature and are
like replicas or shadows of the corresponding cognitive states (ibid.: 363ff.; see
also Section 3.2 on immediacy, and Chapters 9f. on Humes view and on the Representational Account). Moreover, this explains, according to OShaughnessy, why
imaginings can be and sometimes are confused with the respective prototypes
(ibid.: 365).
OShaughnessys View
209
be captured by the echo thesis that all imaginings are non-intentional representations of one of the two cognitive prototypes, successful perception
and intellectual knowledge.
This echo claim constitutes a specific version of (ER) which characterises the postulated representational link as non-intentional. As illustrated
above, OShaughnessy takes it to be necessary for imaginings to be necessarily only quasi instances of the two cognitive prototypes. This means, in
particular, that he takes the echo thesis and the weak negation claim introduced above to be necessary for imaginings, since they specify what it
means to merely quasi exemplify the prototypes.
But OShaughnessy intends to provide a full definition, that is, one
which also specifies sufficient conditions for imaginativeness: he is looking [...] for the defining marks of being an imagining (ibid.: 342). Although OShaughnessy is not explicit about when sufficiency is reached,
he thinks that his definitional project is completed once the relation of
quasi exemplification is understood in the manner just described: adding
the conception of imaginings as imaginings of cognitive prototypes to the
necessary conditions mentioned earlier [concludes] the analytical characterisation of imagining (ibid.: 366). Hence, it is reasonable to assume that,
for OShaughnessy, the satisfaction of his definition of imagining is both
necessary and sufficient for something to be an imagining if the relation of
quasi exemplification between imaginings and the respective cognitive
prototypes is understood in terms of the weak negation and the echo claim.
This interpretation is in line with the fact that all of OShaughnessys
comments, which emphasise the fact that the conditions so far specified by
him are only necessary, precede his introduction of the non-intentional representational link between imaginings and the cognitive prototypes (ibid.:
359 and 362). Furthermore, this reading gets additional support from the
fact that OShaughnessy argues that imaginings are unique in being constitutively linked to other mental phenomena by having them as their nonintentional objects: such a measure of dependence of one psychological
phenomenon upon another is without precedent in the mind (ibid.: 371). In
contrast, higher-order thoughts, intentions or desires at best take mental
phenomena as their intentional objects; and their identity as distinctive
types of mental phenomena does not depend constitutively on the mental
210
OShaughnessys View
211
I discuss his arguments for his version of (NC*) in more detail in the
subsequent sections. At the moment, it suffices to stress that this weak
negation claim is an essential part of OShaughnessys definition of imagining; and that he therefore concludes that different kinds of imagining
conform to the definition for different reasons: some because of their origin
in the will, others because of their origin in some mere mental causes, and
some in addition because of their constitution (ibid.: 359ff. and 367). Accordingly:
[T]he property of being an imagining is neither a relational nor constitutive
property, even though relational and/or constitutive properties determine its
existence. What matters is that a certain necessity is binding, namely that
imaginings cannot exemplify their cognitive prototype. (Ibid.: 361)
This means that, for OShaughnessy, being an imagining that is, satisfying the definition of imagining cannot be reduced to, or explained in
terms of, a more fundamental feature shared by all imaginings: there is no
feature common to all imaginings, over and above their satisfaction of the
definition in virtue of which they count as imaginings. In particular, there
is no proper imaginative essence: the class of imaginings does not constitute a natural mental kind (ibid.: 361 and 368ff.). Moreover, his conclusion
that there is no underlying intrinsic or relational feature common to all
imaginings, which can account for their imaginativeness, has the consequence that OShaughnessy takes the property of satisfying the definition
of imaginings to be explanatorily fundamental. Given that it is also explanatorily illuminating because it tells us in substantial terms what imaginativeness consists in, his theory of imagining is designed to satisfy the
second desideratum for a unified account of imagining as well.
However, OShaughnessy does not stop here. For him, his analysis of
what it means for something to be an imagining entails that imaginings
cannot cognise the world even though they otherwise resemble cognitive
states in many respects. His main idea is that the imaginings cannot cognise reality because they cannot really exemplify their respective cognitive
prototypes. Indeed, he links the failure of imaginings to exemplify the two
cognitive prototypes with their failure to be able to cognise reality by
Dorsch (2012c)).
212
identifying the same cause for both, namely, the specific origins of imaginings in the mind (ibid.: 345 and 359). In addition, OShaughnessy also
does not seem to think (as already illustrated above) that there are any other cognitive prototypes which imaginings might instantiate. For him, there
do not seem to be more than two fundamental mental cognitives knowledge and perception (ibid.: 363). Accordingly, he concludes with respect
to all forms of imagining:
The fact that the mind acts [...], not as a representative of Reality but in direct opposition, guarantees that imaginings must be cognitively void. [...]
[F]ailure to realise the [cognitive] prototype emerged as a universally necessary feature of imaginings. (Ibid.: 359)
Hence, he does not take imaginings to possess the capacity to cognise reality. This aspect of his theory, which is presented not as part of his definition of imagining, but rather as a consequence of it, can be expressed by
the strong negation claim that no imagining can cognise reality a claim which is
in fact identical with (NC) (with the qualification mentioned above in Note
10 in mind).
To conclude, his versions of (NC*), (ER) and (NC) adequately reflect
OShaughnessys account of imagining: while the weak negation claim and
the echo thesis constitute the two main ingredients of his definition of imagining, he takes the strong negation claim to be one particular consequence
entailed by this definition. For him, these three claims apply to all imaginings; and there are no non-imaginative mental phenomena of which all
three claims are true. Accordingly, he takes imaginings to be precisely
those mental phenomena of which these three claims are true. Since
OShaughnessy thinks that certain entailment relations hold between the
three claims, his theory can presumably be presented in a simpler form
(e.g., by reference to only the first two theses, or to his own statement of
OShaughnessys View
213
his definition). But it becomes clear further below that partly due to some
apparent ambiguities in OShaughnessys notion of the cognitive prototypes it is better to keep these three aspects of his account apart.
With respect to the project of providing a unified account of imagining,
OShaughnessys theory is meant by him to satisfy the two desiderata, notably the demands for extensional adequacy and explanatory fundamentality. As it stands, the account can also reasonably hope in particular due to
his version of (ER) to elucidate the nature of imaginings in positive
terms. Moreover, due to the nature of the three main claims involved,
OShaughnessys theory is a version both of the Epistemological Account
and of the Dependency Account: its key idea is to characterise imaginings
as non-cognitive in various ways, as well as representationally dependent
on cognitions.
214
OShaughnessys View
215
Accordingly, both perceptual experiences and beliefs are meant to constitute knowledge; and they may or may not succeed with respect to this ambition.18 Both fulfil this function if they instantiate their respective prototypes. This indicates that OShaughnessy has two forms of knowledge in
mind when he speaks of perception and knowledge as the two cognitive
prototypes: what he refers to are perceptual and intellectual knowledge.
Moreover, he seems to embrace the traditional picture of intellectual knowledge, according to which belief is a constitutive part of this form of knowledge.19
Strictly speaking, OShaughnessy thinks of successful perception, not as a form (or
mode) of knowledge, but rather as a way of acquiring knowledge (OShaughnessy
(2000): 551). That is, he restricts the constitution of knowledge to beliefs. Here, I ignore this aspect of his position and continue with my general practice in this book to
treat what others take to be merely grounds of knowledge such as successful perception or memory as instances of knowledge as well (see the introduction to Part
One).
19
This appears to be in tension with OShaughnessys idea that there are certain (imaginative) beliefs which cannot constitute knowledge: Propositional imaginings can
be real beliefs [...]. And yet it is clear that their truth must be accidental. No propositional imagining can be a knowing (OShaughnessy (2000): 345). But what he
should have in mind is simply that (non-imaginative) beliefs cannot constitute
18
216
OShaughnessys View
217
218
That there is an essential difference between sensory imaginings and perceptual experiences does not on its own suffice to prevent the former from
instantiating the same cognitive prototype as the latter. What is in addition
needed is the assumption that the nature of the sensory cognitive prototype
is such that only perceptual experiences can exemplify it. Hence,
OShaughnessys Argument from Attention relies on a conception of this
prototype which already excludes sensory imaginings and other non-perceptual experiences as candidates for its realisation. This is reflected,
again, by his identification of the sensory cognitive ideal with successful
perception.
But, now, more or less the same question about OShaughnessys argumentative strategy arises as above. For the weak echo thesis entails the
weak negation claim on the assumption that only perceptual experiences
and beliefs can, by definition, constitute instances of perceptual or intellectual knowledge. The truth of the weak echo thesis ensures that imaginings
differ essentially from perceptual experiences and beliefs. The definitional
link between the cognitive episodes and their respective forms of knowledge then suffices to establish the weak negation claim. Hence, there is no
need to look for any further arguments in favour of this claim. In particular,
the issue of how imaginings are actually produced becomes completely ir-
OShaughnessys View
219
relevant for the issue of whether they can exemplify the cognitive prototypes. This explains why what I have called the Argument from Origin
which traces back the truth of the weak negation claim to the specific ways
in which imaginings arise from our minds would turn out to be completely superfluous. It thus becomes, again, puzzling why OShaughnessy
spends to much time on developing and defending this argument, while not
explicitly mentioning the short and simple line of argument just outlined.
As spelled out in Chapter 2, it is not very promising to give up on the
idea that there is an essential difference between imaginings and cognitions. Instead, the best solution to this puzzle seems to drop the assumption
that the cognitive constraints, which define the sensory and intellectual
forms of knowledge mentioned in the weak negation claim and govern all
attempts at their realisation, include the requirement that the mental representations concerned have to be perceptual experiences or beliefs. This
means that the cognitive constraints should be understood as just being
concerned with more centrally epistemic conditions like veridicality, reliability, and so on. As a consequence, nothing about the nature of the two
forms of knowledge under consideration prevents mental representations
other than perceptual experiences and beliefs from realising them. In accordance with this, the truth of the weak negation claim that is, the failure of imaginings to constitute perceptual or intellectual knowledge has
to be due to some other reasons, such as the inability of imaginings to be
veridical or reliable representations of the external world. This is exactly
what the Argument from Origin is meant to show: that imaginings cannot
exemplify the two cognitive prototypes because they occur in ways which
do not involve their causal or rational determination by reality (see Section
7.1).
Whether this would really be OShaughnessys preferred reply to the interpretative problem raised above is, to some extent, a mute question. His
argumentative strategy and, in particular, his insistence on the significance of the Argument from Origin in the derivation of his weak and the
strong negation claim makes sense only if he is regarded as endorsing the
weak rather than the strong echo thesis, and as characterising the cognitive
ideals primarily in epistemic terms, rather than also in terms of specific
kinds of mental episode.
220
Before I turn to the discussion of this argument, it is important to highlight an important consequence of the preceding considerations. If
OShaughnessys two negation claims would in fact be derivable from his
echo thesis in the way sketched above, they would stop singling out a basic
aspect of the nature of imaginings. As a consequence, all that would fundamentally matter would be the echo claim; and OShaughnessys theory of
imagining would just be a version of the Dependency Account, but not also
one of the Epistemological Account. But his actual argumentative strategy
suggests, rather, that the two negation claims need to be established independently of the representational dependency of imaginings on cognitions.
Hence, the echo claim should better not be read as implying the weak and
the strong negation claim; and OShaughnessys unified account of imagining should indeed better be understood as a combination of the Epistemological and the Dependency Account.
CHAPTER 7
OShaughnessys Arguments
222
and that this difference between the two kinds of sensory episode marks an
essential difference, given that the object-directed attention in question is
constitutive of perception. My aim is to show that the relevant examples do
not support the conclusion that sensory imaginings do not involve perceptual attention in the way envisaged by OShaughnessy.
This argument consists of two chief claims. The first links the capacity to
exemplify one of the two forms of knowledge to the capacity to satisfy certain cognitive constraints (or Reality constraints, in OShaughnessys terminology). The satisfaction of these constraints is necessary and perhaps
also sufficient for the exemplification of the respective cognitive prototypes. The second thesis of the argument states that imaginings cannot satisfy the respective cognitive constraints because of their special origin in
the mind. Cognitions, by contrast, possess this capacity since they can and
typically do originate in the world in the way required for that potential
satisfaction. His overall strategy is thus to derive the difference in cognitive capacity from a difference in origin, thereby making essential use of the
idea of cognitive constraints. Before presenting OShaughnessys defense
of the second thesis concerning the specific origins of imagining, I begin
with his first claim that cognising representations have to satisfy certain
cognitive requirements.
OShaughnessys Arguments
223
224
propositional (ibid.: ch. 10). For the current purposes, however, it is possible to set aside these parts of his account, as well as the general issues
concerned. Instead, I would like to focus on several features of perceptions
in general, and of successful perceptions in particular, which seem to be
less controversial. The first is that perception is intentionally directed to
mediated physical objects (people, sky, trees, and the public occurrence of
secondary qualities) (ibid.: 16, 19 and 166f.). Furthermore:
A material object can come to visual attention only if one of its qualities
does for no physical item can be just seen. (Ibid.: 537)
Consequently, perceptions always represent material objects as being certain ways. They can do so either correctly or incorrectly: the perceptions
can agree or disagree with reality (328). 21 OShaughnessy thereby dis21
OShaughnessys Arguments
225
226
OShaughnessy describes the natural veridicality of perceptions (in contrast to the at best accidental veridicality of hallucinations) as the natural
correspondence between the internal [object] and the outer object (ibid.:
566), or as a general matching correspondence between the internal object
[...] and the outer [...] material object (ibid.: 567), or as the fact that we
harbour internal objects which naturally so match up with the physical environment (ibid.: 568). This means in the light of what OShaughnessy
takes internal and external objects to be that he treats perceptions as
veridical just in case how they represent their intentional object to be reflects how it really is which presupposes, of course, that there is an external object at which the perception is intentionally directed. In the case
of veridical perception, there is a match between the content of the perceptual experience and the represented reality. Furthermore, the natural
veridicality of perceptions is the result of its reliable causation by the perceived object in the world (ibid.: 567):
Perception has the function [...] of generating knowledge of the physical
here-and-now. Now perception can discharge that function only if the internal objects of the perceptual experience tend in general to match the outer objects. (Ibid.: 565)
[I]n seeing the colour and contour of the light and thereby also of the object,
and in that way experiencing the visual appearance of that object, we are
dependent upon a reliability that is situated, not just within our perceptual
apparatus, but externally to our bodies in the environment. (Ibid.: 450)
light (ibid.: 454 and 457). We come to know that we perceive light only by means of
(philosophical) reflection (ibid.: 457). This might also explain why the perception of
light does not seem to involve a respective internal object (ibid.: 569). In any case,
it is important to note that the sense-datum (or sensation) is among the perceived
objects; and that its perception involves a content including a corresponding
internal object. This means that internal objects and sense-data should not be
mixed up with each other (ibid.: 467 and 569).
OShaughnessys Arguments
227
For OShaughnessy, the reliable link between the perceived object and our
visual experience which represents it as being a certain way is realised by
suitable causal links between the two items (ibid.: 451). What he presents
is thus a causal theory of perception which promises to tackle an important
issue for any such theory, namely to take due note of deviant and nondeviant physical causal chains both within and without the perceiver
(ibid.: 450).23 Thus, OShaughnessys general picture of perception (though
23
For OShaughnessys causal theory of perception and its reliability, see OShaughnessy (2000): chapters 16 and 20 (especially pp. 451ff., 537ff., 542ff., 549ff. and
565ff.). Whether his theory succeeds is another matter, which need not interest us
here. But an uncommented sketch of his thoughts may be helpful.
His general idea is that the reliable causation happens in four stages (ibid.: 542ff.
and, especially, 569). The first stage comprises the causal chain from the perceived
object and its respective perceivable parts (e.g., its surface) via light to the
stimulation of the retina. The second causal chain bridges the gap between the
physical and the mental by linking this stimulation to the resulting sensation (or
sense-datum). The third and fourth stages involve only mental causation: one leads
from the sensation to the visual experience in its most minimal form (i.e., as a
representation of a two-dimensional ordered coloured expanse), while the other
occurs within the visual experience and establishes the other internal objects (e.g.,
the surface and the side of the mountain, and finally the mountain itself).
Now, the first two causal chains are reliable because of their subjection to
respective laws. The suitability of the causal links in question is due to their
conformity to laws, or law-like regularities, which govern the interaction of light
with objects and the retina, the processes in the visual system and its connections to
the mental realm (ibid.: 451, 453, 537, 544 and 549ff.). In contrast, the other two
causal chains are reliable, not because of some nomic regularity (ibid.: 544), but (as
it seems to me) because the interpretative steps from one internal object to the
next are small enough to typically adequately reflect the corresponding links
between the perceived external objects (ibid.: 565ff.). The transition from the first
interpretation (concerned with a two-dimensional configuration of colours) to the
last (concerned with the perceived object: e.g., a mountain) is thereby a matter of
trial-and-error, always open to correction by experience or, as OShaughnessy puts
it, we see from experience (ibid.: 567).
That the nomic regularity holds only between the external object and the relevant
sensations is thereby a consequence of the minds ability (e.g., in the form of
associations, desires, emotions, etc.) to interfere with the causal link between the
sensations and the resulting experience (ibid.: 544 and 552). However, if such an
interference occurs, the resulting experience is not a perception any more (and thus
228
not its details) seems to be a fairly standard causal theory (see, e.g., Lewis
(1980) for a similar view). In accordance with these considerations, the
cognitive prototype of sensory knowledge has to be characterised minimally in terms of the complex property of being a representation which is intentionally directed at an object, represents it correctly (and thereby refers
to it), and does so in a reliable manner.
One aspect of this cognitive constraint on perceptual states is that, if
something is to count as an exemplification of the perceptual form of
knowledge, it has to be due to an appropriate external causal determination. Perceptions that do not satisfy the condition of being constrained by
reality in this way cannot play the cognitive role of perceptions (e.g., of
constituting knowledge, or at least of grounding judgements or beliefs that
count as knowledge). More generally, perceptual experiences are subject to
the cognitive constraint of being suitably causally linked to external objects
and of correctly reflecting the nature of these objects (if this is not already
also not an experience in OShaughnessys terminology), but an imagining. This is
due to the fact that the mental episode in question is not caused by the external
object (via the sensations) in the normal way distinctive of perception (ibid.:
352f.): that is, by means of the influence of attention and sometimes also of
understanding which structures the sensations in conformity to their received
nature (ibid.: ch. 20). In other words, cases, in which the causal determination breaks
down completely, fail to be perceptual, given that OShaughnessy takes sensory
representations that lack the normal causes of perceptions to be perceptual
imaginings (ibid.: 352).
Consider the example of vision, in which sensations are identical with noticeable
aspects of the structured visual field (ibid.: 500 and 554). Abnormal causation of the
visual experience by a particular determination of the visual field would occur, if
this specific determination (say, involving a small red colour expanse) would bring
about that other mental phenomena (such as the thought about ones having cruelly
shed blood, and ones related feeling of guilt) let the whole visual field in ones
awareness to become red while in fact only parts of it really register redness (for
similar examples, see ibid.: 352f. and 478f.). That the perceptual character of an
experience depends, for OShaughnessy, on its causal origin in this way might none
the less suggest, however, that the nomic regularity holds also between the external
object and a veridical perception (given that it is indeed appropriately caused and a
perception) in particular since the different ways in which the visual field can be
organised or interpreted by the attention and the understanding are restricted by its
particular determination.
OShaughnessys Arguments
229
ensured by the intentional and the suitable causal link). They exemplify
their cognitive prototype (i.e., constitute successful perception) only if they
satisfy this requirement which means, the three demands of representationality, veridicality and reliability.
Now, moving on to intellectual knowledge as the cognitive prototype for
beliefs, it is first important to note that OShaughnessy takes belief to be
constitutive of intellectual knowledge. His picture is that, while some beliefs do not count as constituting knowledge, others do. Moreover, he takes
it to be the epistemological function of beliefs to constitute knowledge:
Belief is the essence of knowledge, it aspires to the condition of knowledge, and constitutes a core phenomenon which can find itself redescribed
as knowledge. (OShaughnessy (2000): 368)
Accordingly, beliefs (or other intellectual representations) have to be nonaccidentally true in order to count as knowledge. However, OShaughnessy
also believes that the ways in which we form (and sustain) beliefs typically
lead to knowledge: that is, they are reliable. And although he is not explicit
about this, it seems reasonable to assume that this reliability ensures that
most true beliefs are non-arbitrarily true. Interestingly, OShaughnessy sees
230
the reliability of our ways of forming beliefs to be an essential part of waking consciousness (i.e., the state of wakefulness). According to him, what
distinguishes people who are awake from those who are asleep, in a trance,
or unconscious (i.e., knocked out or anaesthetised) is precisely that only
the former are in direct contact with reality (see, in general, ibid.: 1ff., 68ff.
and 113ff.):
To be in the state consciousness is to be in the experiential condition of being aware of the World. As we say of the conscious, they are in touch with
Reality as those lost in a trance or dream are not. (Ibid.: 1)
[C]onsciousness is not a mode of epistemological success as if the world
was an object. Rather, consciousness is correct epistemological posture on
the part of an experiencing subject. (Ibid.: 117)
For OShaughnessy, being awake is thus necessary for coming into epistemic contact with reality and acquiring knowledge about it by means of
perceptual or similar experiences (see also ibid.: 119). Moreover, wakefulness involves the possession of the usually successfully employed capacity
to form beliefs in a reliable way:
[I]f consciousness is to exist [...] a cognitive sensitivity to perceptual
experience must inhere, and of the right kind. This cognitive sensitivity, and
the mode of belief-formation concerning the environmental realities, are I
suggest intrinsic to consciousness, which requires in general that beliefs be
well-formed [...] indeed, should normally lead their owner to Reality. Thus,
in the conscious the mode of belief-formation out of veridical perception
should be such as normally to make knowledge of that belief. (Ibid.: 85)
Hence, he assumes (given that we are usually awake) that our ways of
forming beliefs reliably lead to knowledge; and, furthermore, that typically
most of our beliefs constitute knowledge (ibid.: 157).24 For OShaughnessy,
the feature of wakefulness responsible for the reliability of our belief
formations is rationality. Thus, waking people are said by him to differ
24
OShaughnessy allows for the possibility of wide-spread error, but assumes that a
system of mostly or entirely false beliefs would be open to correction by experience
(OShaughnessy (2000): 157).
OShaughnessys Arguments
231
According to this quote, when we veridically perceive something, the respective belief will enter our net of beliefs about the world only if it rationally coheres both with the already existing and largely true beliefs and with
the veridical perceptions on the basis of which it is formed. But the quote
also specifies that, if the new belief satisfies these conditions and becomes
a part of our picture of the world, it is likely to amount to knowledge itself
(to unlock the veridical information contained in the perceptions). Hence,
OShaughnessy presents the following minimal condition for beliefs to
constitute knowledge (whether they are newly formed or already existing):
they have to be rationally consistent both with other beliefs and judgements (most of which should be true) and with any occurrent perceptions
(again, most or all of which should be veridical). In addition, OShaughnessy endorses the view that beliefs are intentionally directed at their objects and can be either true or false; and he requires knowledge-constituting beliefs to be true themselves (ibid.: 19, 84ff. and 328). Therefore,
232
With the general qualification that, say, Gettier-style examples and similar cases still
have to be satisfactorily dealt with. Besides, it should be clear that reliability in conjunction with some form of correct indication would not be sufficient for knowledge,
at least not in the full sense intended by OShaughnessy. If one would by chance be
inclined to visualise a pink elephant whenever one would begin to suffer from an ear
infection, and if one would come to know about this correlation, one could use the
occurrence of such a visual image as an indicator for a beginning ear infection. But
the image of the elephant is not intentionally directed at the ear. In particular, it does
not represent it as being a certain way (i.e., as infected), with respect to which it
could be either correct or incorrect. For the hypothesised correlation between the
image and the infection does not allow for misrepresentation: if there is an infection,
the resulting image indicates that this is the case; but if there is no infection, the
image does not indicate anything at all (except, of course, if it is correlated to some
other fact).
The underlying reason for this is, presumably, that the correlation between the
image and the infection is not linked to some kind of function of the respective
OShaughnessys Arguments
233
234
The first is that wakefulness requires the presence of a stream of consciousness (or of experience; see OShaughnessy (2000): 5, 82f., 123 and
142). It seems plausible enough to maintain that a person not enjoying any
mental episodes should not count as being awake. As OShaughnessy
points out, it seems that only completely unconscious people (such as those
who are anaesthetised or in a coma) lack phenomenal consciousness altogether; and they are clearly not awake. On the other hand, he illustrates the
fact that the presence of phenomenal consciousness is not sufficient for
wakefulness by reference to dreaming and hypnotised people who experience phenomenally conscious states as well (ibid.: 119 and 123).
OShaughnessys idea that epistemic contact with reality presupposes
wakefulness and therefore phenomenal consciousness should not be taken
to imply that unconscious people cannot have any (knowledge-constituting) memories or beliefs. This would be an absurd view, given that we usually do not lose our memories and beliefs as well as our knowledge
during periods of unconsciousness. Rather, it should be understood as
meaning that unconscious people have no conscious access to their cognising memories or beliefs and hence cannot actually use them (e.g., by
manifesting them in consciousness) to get into contact with reality.27
The second relevant aspect of wakefulness is that it involves what
OShaughnessy calls awareness of the outer world. 28 He distinguishes
this form of awareness from perceptual awareness of reality. In particular,
he takes it to be present even when a waking subject has no actual perceptions or similar episodic representations of his environment. He takes it
thus to be prior to, and independent of, any actual perceptual link to reality
(see also ibid.: 117 and 152):
Strictly speaking, this seems to put a limitation on OShaughnessys account of
knowledge. For wakefulness seems to be a necessary condition only on available
knowledge (i.e., knowledge that can be used to get into contact with reality), but not
on knowledge in general at least not if one wants to allow that temporarily inaccessible memories or beliefs may none the less continue to constitute instances of
knowledge. But it is not clear whether OShaughnessy really has such a limitation in
mind.
28
He also speaks of realistic, linked or connective awareness (OShaughnessy
(2000): 119, 122 and 155f.); of w-consciousness (ibid.: 117); and of being in
touch or contact with Reality (ibid.: 1 and 124).
27
OShaughnessys Arguments
235
[W]aking consciousness is consistent with the perceptual awareness of absolutely nothing in the (outer sector of) (the real) world. None the less,
such a perceptually empty consciousness, since it is awareness of the world,
continually orients to the world: it carries the real world with it all the time,
as the framework within which to site anything it happens perceptually to
encounter. (Ibid.: 119)
Now awareness of the outer world does not mean awareness of the present
facts of the environment. (Ibid.: 152)
[Awareness of the outer world] is not the perception of anything, and in particular is not the perception of the contents of [reality]. Rather, it precedes,
and enables where possible, the perceptual and cognitive reading of the
data indicative of those contents. (Ibid.: 155)
236
The main reason why OShaughnessy assumes that non-arbitrarily veridical perception is not sufficient for epistemic contact with reality is that
somnambulists or hypnotised people and, to a lesser extent, dreaming
subjects may enjoy such perceptual experiences, or at least very similar
ones, without thereby being in a position to acquire knowledge about reality (see ibid.: 118ff., especially 123f.). His view is that such people may reliably and veridically perceive the objects in their environment (e.g., where
they are located, or which shape and size they have), but none the less are
bound to fail to form knowledge-constituting judgements about them. This
failure may manifest itself in various ways: in the fact that they do not correctly interpret what they perceive (e.g., they may take the table in front of
them to be an automobile because the hypnotist says so (ibid.: 122f.)); in
the fact that they do not take the perceived objects to be part of reality (but
instead, say, of some unreal world which is dreamt, or imagined in response to the instructions of the hypnotist (ibid.: 120 and 123)); or in the
fact that they do not properly connect the perceived objects to the rest of
reality (e.g., in the case of a somnambulist or a hypnotised person who can
cognise only small and disconnected parts or regions of reality (ibid.: 120
and 122)).
In contrast, waking people can as long as they are not in a psychologically unusual or abnormal state of mind (e.g., when intoxicated or mentally disturbed) avoid these failures and subsequently acquire knowledge
about reality by means of perception. They then enjoy what OShaughnessy calls realistic perception or awareness: the perceptually based contact with reality which locates its particular objects in reality (ibid.: 119,
122, 124 and 155), links them to other parts and objects of reality (ibid.:
122, 124 and 156), and enables us to acquire knowledge about them under
many different interpretations or conceptualisations (ibid.: 124 and 156).
It is hence not surprising that OShaughnessy concludes that there has to
be an intimate link between wakefulness and perceptually grounded knowledge about reality. He observes that what is missing in all the cases of
dreaming or trance and what is present in the case of wakefulness is the
establishment of a (sufficient) rational consistency between how things are
perceived to be and how they are believed to be. Accordingly, he takes the
OShaughnessys Arguments
237
Now, the specific aspect of wakefulness which is responsible for the difference in the epistemic position between waking and non-waking people is,
of course, identified by OShaughnessy as the involvement of awareness
of the outer world: it is this aspect which equips occurrent perceptions
with the potential to lead to knowledge (ibid.: 119 and 155f.).
In accordance with the preceding considerations, OShaughnessy understands the property of being aware of the outer world as involving three
different things. First, it involves having a sufficiently rich and rational
system of background beliefs about the world which we can readily apply
to what we perceive. It is in this sense that the form of awareness under
discussion is awareness of the world [and] carries the real world with it all
the time, as the framework within which to site anything it happens perceptually to encounter (ibid.: 119; see also 156). Consequently, if we perceive
something while being awake, we perceive it as part of the reality and are
disposed to describe the perceived entity in accordance with our picture of
the world. As OShaughnessy writes, the perceived item brings an entire
238
If, however, this awareness of the outer world which pertains to the waking mind as a whole (i.e., this set of underlying beliefs, capacities, inclinations and enabling conditions) is indeed accompanied by an actual perceptual experience, it may and usually will come to pertain to that particular instance of perception as well. According to OShaughnessy, the result
will be full and proper awareness of the outer world or, in other words,
29
OShaughnessys Arguments
239
240
tions resists reduction to the requirement that they have to be reliable and
non-arbitrarily veridical at least as long as one is convinced by
OShaughnessys examples of somnambulists and hypnotised persons
which are meant to illustrate the respective difference in epistemic standing. For him, the subjects in these examples do not really perceive reality,
that is, are aware of objects as part of the real world, even if their perceptual experiences are reliably formed and non-accidentally veridical. Consequently, OShaughnessy takes rational consistency to be an additional aspect of the cognitive constraint linked to the cognitive prototype for perceptions.
His account thus reaches beyond the traditional picture, according to
which a perception is already epistemically sound if it is non-accidentally
veridical and reliable. The traditional view may very well include the claim
that our perceptual beliefs are epistemically appropriate only if they or
the respective perceptions rationally cohere to a sufficient degree with
our other beliefs; and that we endorse our perceptions in belief only if we
do not recognise them as not being rationally consistent with what we otherwise believe (i.e., only if there are no respective defeaters). But it is not
part of the orthodox position that the epistemic soundness of perceptions
depends also on their rational relations to our beliefs, and not merely on
how they are brought about. Indeed, it is traditionally controversial whether perceptions can stand in rational relations. 30 Therefore, OShaughnessys
theory of perceptual grounds of knowledge makes a substantial addition to
the customary view of the epistemic soundness of perceptions.
His position on the cognitive demands on beliefs is also more complex
than the orthodox approach, mainly because the latter usually does not address the issue of the importance of wakefulness (e.g., in connection with
Cartesian dream arguments). The traditional picture requires cognising be30
Coherentists, for instance, have responded to traditional foundationalism with skepticism about the idea that perceptions can enter into rational relations and hence
provide reasons for beliefs (see, for instance, Davidson (1989)). In recent years,
however, it has become more common to endorse some kind of modest foundationalism and to assume that perceptions possess rational status as reason-givers (see
Martin (1992); Pollock & Cruz (1999): chs. 4f.; Pryor (2001)) although sometimes
only in conjunction with the claim that the content of perceptions is conceptual (see
McDowell (1994)).
OShaughnessys Arguments
241
liefs to be true and justified (plus perhaps to satisfy some further conditions
in order to rule out Gettier-style cases). This may involve that they have to
be rationally consistent with other beliefs and ones current perceptions and
be rationally responsive to ones potential perceptions or judgements. But
OShaughnessy demands in addition that the believing subject is awake,
that is, in a position to have perceptual experiences of the right kind
namely, perceptual experiences which become rationally integrated with
ones background beliefs and which hence indeed present objects as being
located in the real world. This is one aspect of what he means by claiming
that beliefs (and, presumably, also memories) can provide knowledge about
the external world only in conjunction with, and because of their link to,
actual or potential perceptions. The other aspect is the empiricist claim that
all kinds of external knowledge have to be ultimately based on perceptual
access to reality a claim which, again, is not necessarily part of the orthodox view on knowledge. Accordingly, OShaughnessy understands the
cognitive constraint on beliefs as involving two requirements over and
above the traditional demand for their rational integration with ones other
cognitive states: they have to occur in a waking subject; and they have to
be ultimately based on perceptual access to reality. In fact, the two requirements appear to be closely related, given that OShaughnessy assumes that
only waking people can get into perceptual contact with the world.
None the less, his view on knowledge-constituting beliefs is not inconsistent with the traditional analysis. OShaughnessys requirement that cognising beliefs about the world have to be generally linked to wakefulness
and, hence, to the presence of phenomenal consciousness and perceptual
experience can be easily accommodated, given that it is satisfied by the beliefs of normal cognising subjects (with the exception of the states of
knowledge or memory which are part of temporarily unconscious or nonwaking minds and which constitute available knowledge during periods of
wakefulness). The orthodox picture is also to compatible with the empiricist idea that all knowledge about the external world is ultimately grounded
in perception. Indeed, it does not seem to be unreasonable to assume that
proponents of the traditional view sometimes implicitly assume that cognising subjects have to satisfy these or very similar conditions, and that
normal waking people typically do satisfy them.
242
To conclude, OShaughnessy adds two further requirements to the traditional account of the cognitive constraints on perceptions and beliefs in
terms of intentionality, veridicality and reliability (as well as rational consistency in the case of beliefs). The first is that both perceptions and beliefs
and, presumably, also other kinds of cognitive states, such as memories
have to be rationally integrated with ones picture of the world. The second
additional demand is that all kinds of knowledge are ultimately based on
perceptual contact with reality. Both of OShaughnessys further requirements have in common that they can be satisfied only by cognitions which
occur in, or are part of, a waking mind.
What seems most important in this quote is that OShaughnessy emphasises the causal role of reality: the suitable causal determination of mental
representations by the relevant aspects of reality either in a direct way (as
OShaughnessys Arguments
243
in the case of successful perception), or mediated by evidential representations and rational mechanisms (as in the case of intellectual knowledge)
is crucial to the acquisition of knowledge. The problem with imaginings is
thus not that they are causally determined by the mind (as OShaughnessy
notes, the mind is causally involved in the formation of cognitions as well);
but rather that they lack any suitable causal link to reality. In other words,
all significant causal impact on imaginings (e.g., determining what they
represent) comes from the mind.
Now, the quote also suggests that this may happen in two ways: either
the will takes over the role of reality in determining the representations; or
the representations occur passively due to some other causal factors in the
mind and are (perhaps because of sharing this passivity with some cognitions) to some extent mistaken for their cognitive counterparts. But let me
discuss the two ways in which imaginings can arise from the subjects
mind in a bit more detail.
On the one hand, the formation of an imagining may be due to activities
of the will, which is directed (e.g., by means of an intention) at producing
an imaginative representation. In other words, the will is the active starting-point of the process of bringing about the imagining, and the result is
the substitution of ones will for Reality (ibid.: 359). The products of this
process are, at least to some extent, subject to the will. Even though we
sometimes cannot influence their occurrence, we have at least control over
their persistence. Furthermore, the resulting imaginings are selfconscious
in the sense that we are aware of, or experience, them as imaginings and as
presenting their objects not as real (i.e., as not making a claim about how
things are; see ibid.: 358f.), and consequently do not trust them or let them
have an impact on our view of the real world. OShaughnessys prime examples of will-susceptible imaginings are engagement in make-believe and
active visualising (ibid.: 343f., 346 and 351ff.).
On the other hand, the formation of an imagining may involve a weakening of ones sense of reality (ibid.: 352). What he seems to have in mind
here is that we come to be unsure or disturbed (ibid.: 352) about what is
real; and that we may even begin to mistake things, which are not part of
reality and normally recognisable as such (e.g., imagined or hallucinated
entities), for parts of reality hence the weakening of our sensitivity (see
244
Hence, although in many cases, the subject still notices a phenomenal dif31
It is not always clear what is presented as unreal. The passages referred here seem
to concern the represented external entities, while other passages seem to concern
the directly imagined cognitive prototypes (see OShaughnessy (2000): 349).
OShaughnessys Arguments
245
246
ideas from his premisses and conclusions and, notably, his comments on
the nature of the cognitive constraints on perceptions and beliefs.
According to OShaughnessy, these constraints consist at least partly in
the requirement that certain epistemically sound processes of formation
as mediated by suitable causal (and perhaps also rational) links precede
the occurrence of the representations in question. Will-impervious imaginings cannot satisfy the respective cognitive constraints because they emanate from the breakdown of exactly these processes. In the case of passive
sensory imaginings (such as hallucinations), the normal causal determination by external entities does not obtain; and in the case of passive intellectual imaginings (such as dream-beliefs), the same is true of the rational impact of our cognitive representations of reality. Consequently, neither kind
of will-impervious imagining can realise a cognitive prototype.
Similar considerations apply to imaginings originating in the will.
OShaughnessys idea is here that the causal determination of the imaginative representations by the will, which is already sufficient for their occurrence or persistence, prevents them from being causally (and rationally)
determined in the appropriate ways by either aspects of the actual world or
evidential representations about it. That is, representations cannot derive
both from the will and from reality in the fashion required for the provision
of knowledge: with respect to will-induced imaginings, the epistemic processes involved in the normal formation of perceptions or beliefs are
simply not involved (see also Dorsch (2009b)).
Considerations like these let OShaughnessy form the more general conclusion that both ways, in which imaginings may occur, rule out the possibility of the satisfaction of the respective cognitive constraints. The issue is
thereby not that imaginings cannot be representational or veridical. As
already mentioned, OShaughnessy assumes that they can be both
(OShaughnessy (2000): 166f., 345 and 363). Similarly, the worry is not
that imaginings cannot occur in a waking mind; they obviously can. Instead, the problem is that imaginings are formed in ways which cannot ensure that they are reliable, rationally integrated with our picture of the
world, and ultimately based on our past or current perceptual experiences.
But since it is essential to instances of the two forms of knowledge that
they do satisfy those constraints, OShaughnessy can infer his weak nega-
OShaughnessys Arguments
247
To ensure that his version of (NC*) can make a substantial contribution to a unified
account of imagining, OShaughnessy has to presuppose that there are no entirely
voluntary cognitions and, especially, no purely actively formed beliefs or judgements. If cognitions could be determined exclusively by the will, instead of at least
partly by reality or evidence about it, they would not differ in this respect from willinduced imaginings and hence would be subject to OShaughnessys Argument from
Origin as well. In this respect, OShaughnessys version of the Epistemological Account is in agreement with the Agency Account to be defended in Part Four. I
provide an independent argument for the involuntariness of judgements and beliefs
in Dorsch (2009b).
248
notice each of the daffodils as such, but only their totality, they are all none
the less part of the visual content. In such a case, we are aware of each
daffodil in one sense, but not in another: we see them, but we do not notice
them. The difference consists in the fact that we notice things that is, see
them only if we specifically attend to them. Noticing thus requires more
than mere seeing: it necessarily involves attention (see also Dretske (1979)
and Martin (1994)).
But why should we think that we are perceptually aware of all the daffodils, although we are not attentively aware of (most of) them? OShaughnessy presents two reasons. The first is that what we notice is determined
by everything that is part of the content of the visual impression (see
OShaughnessy (2000): 501). In particular, we have to be aware of the
single daffodils in order to be able to notice their totality: we could not see
a host or field of daffodils without seeing each of them (ibid.: 356). The
second reason is that there are physical features, that determine which area
of the content of the visual impression shows which colour-value and
hence that enable us to test whether our visual impression in fact includes
information about each of the daffodils (ibid.: 355f.). Although he does not
say much more about the issue, the underlying idea seems to be that, in the
case of perception, the visual content is completely determined by nonpsychological, physical factors (e.g., the stimulation of the receptors in the
eyes and, presumably, certain areas of the visual cortex), that are part of the
normal causal link between external world and experience, and hence can
be investigated by scientific means. Hence, there is a distinction between
what we merely perceive and what we actually notice in the spotlight of
our attention.
A similar distinction, as OShaughnessy argues, cannot be drawn with
respect to sensory imaginings. His main reason for this claim is that the
presence of unnoticed, but none the less sensorily presented elements in the
content of the imaginative representations cannot be shown by means of a
scientific investigation (of, say, the visual system in the brain). For, due to
the absence or breakdown of the mechanisms that typically link perceptions to the external world, there are no determining physical causes to be
investigated. But if the distinction between being mere sensory awareness
and additional attentive awareness cannot be demonstrated, it should not be
OShaughnessys Arguments
249
Another good example are hypnagogic images. While falling asleep, visual images
may arise without our (conscious) active influence and surprise us with what they
represent, or allow us to attend to previously unnoticed details of the represented objects.
250
having. But I may discover this number by focusing my attention to the respective part of the visual field, or to the imagined scene that it presents to
me. Accordingly, the sensory imagination is open to attention and discovery: we may come to notice previously unnoticed details of a visualised
scene by attending to them.
It might be insisted that, rather than discovering the number of columns,
we determine it by means of imaginative activity when trying to find
out how many columns we are visualising the Pantheon as having. That is,
it might be argued that we shift our attention from one imaginative project
(i.e., visualising the Pantheon with no determinate number of columns) to
another (i.e., visualising it with this determinate number), rather than from
one aspect of the visual field of our episode of visual imagination (e.g., the
stairs) to another (e.g., the columns and their number). As a consequence,
visualising would not, after all, involve the kind of attention that is characteristic of perception, and which enables us to notice previously unnoticed
things. It would just be able to occupy our attention in the same way as all
other mental episodes, including thoughts and mental actions.35
However, two things speak against this reply. First, the proposed alternative explanation of the role of attention in visualising cannot explain the
fact that our recognition of how many columns we are visualising may be
accompanied by a genuine feeling of surprise, or discovery. In such cases,
we find ourselves visualising a certain number of columns, without being
aware of actively determining this number. This is true even if the number
of imagined columns is determined by sub-intentional mental agency: just
as we can discover that we are actively scratching our back, we may find
ourselves visualising a particular number of columns. Second, visualising
shares with seeing the property of being transparent to the represented objects (see Section 2.1). But this just means that, if we try to shift our attention inwards to our episodes of visualising, we end up noticing the external
35
Hopkins (2011b) defends a view on visualising and attending that comes very close
to this reply. The distinction between the two kinds of attention pertaining to perceptual awareness and conscious thought and, notably, the fact that attending to perceptual experiences means, at least in part, attending outward (i.e., to some external objects), while attending to conscious thoughts does not is introduced in
Peacocke (1998), and discussed in more detail in Martin (1998), OBrien (2007): ch.
6, and Dorsch (2012a).
OShaughnessys Arguments
251
objects and features that we are visualising. In other words, visualising involves the same kind of attention as seeing and not the same kind of attention as, say, thinking. It is true that visualising lacks the immediacy of
seeing in the sense of not presenting those objects and features as part of
our actual environment (see Section 3.2). But the kind of attention involved in perception is linked to transparency, and not to immediacy.
These considerations do not show that perceptions and sensory imaginings could not still differ in kind with respect to the involvement of attention. But they demonstrate that OShaughnessy has not presented us with a
good reason to believe the opposite. We should therefore take the Argument from Origin to be his main motivation for endorsing the weak negation claim and, hence, (NC*) also because that argument is not limited to
any particular form of imagining. This concludes the expository part of my
discussion of OShaughnessys theory of the imagination. I now move on
to the assessment of this and other instances of the Epistemological Account.
CHAPTER 8
Critical Assessment
254
Critical Assessment
255
ive claims in such a way that the resulting theory discloses fully what imaginings essentially are. The preceding considerations also show that the
positive claim to be added has to be more substantial than, say, the statement that imaginings are episodic mental representations. For the view that
what is distinctive of imaginings is that they are merely mental, episodic
and representational provides, again, primarily a negative characterisation
of their nature, given that the positively mentioned features are too ubiquitous among mental phenomena. It is much more promising to combine a
negation claim with one of the echo theses that are central to the different
versions of the Dependency Account of imagining. This is, for instance,
what OShaughnessy is proposing, partly in response to the noted negativity of negation claims (see Section 9.4). There is then, however, the worry
that the chosen echo thesis may describe the nature of imaginings in more
fundamental terms than the respective negation claim that, for instance,
the lack of a certain cognitive feature is due to the fact that imaginings are,
in one way or another, dependent on cognitions.
In addition to their negative character and the resulting general limitation of their explanatory power, particular negation theses may fail to be illuminating with respect to the common nature of imaginings. Consequently, they may fail to add anything to a unified account of imagining,
even if they are true of all central cases of imagining and supplemented by
some more positive claims. Three cases are of particular significance here.
First of all, negation claims may lack explanatory power to the extent to
which they focus on features the absence of which is not particularly distinctive of imaginings. The thesis that imaginings are not moods tells us
not much of interest about what imaginings are or are not, even though it is
presumably true of all imaginings. The main reason for this is that many
non-imaginative phenomena lack the specified features as well: most mental states are not moods. Accordingly, the lack of the respective features is
not very specific to imaginings. Even less informative is the claim that
imaginings are not tables, given that no mental phenomenon is a table (and
neither most non-mental phenomena). Negation claims that apply only to
imaginings or at least just to imaginings and to a few other mental phenomena promise to be much more illuminating.
Then, negation claims may be unilluminating with respect to the pur-
256
pose of providing a unified theory of imagining because they describe different kinds of imagining in terms of different kinds of features (or, rather,
lack thereof). To say that visual imaginings are not seeings, intellectual
imaginings not judgements, affective imaginings not real feelings, and so
on, has precisely this result: it does not tell us about the nature of imagining since it does not identify a single feature the lack of which characterises all kinds of imagining. As a consequence, such a negation claim may
be true and illuminating with respect to particular forms of imagining; but
it will not be able to say anything about what imaginings have in common
and what renders them imaginative.
Finally, negation claims may fail to have proper explanatory power because they do not pick out the most fundamental feature(s) the lack of
which is distinctive of imaginings. The idea that imaginings are not tables,
say, is unilluminating in this way, given that it is merely a consequence of
the fact that imaginings are not entities in the external world (which again
may perhaps be traced back to another negation claim, or the positive idea
that imaginings are mental phenomena).
See Chapter 7 and, for instance, Martin (1992), Pollock & Cruz (1999), or Williamson (2000).
Critical Assessment
257
OShaughnessy seems to have been aware of the need for such an interpretation of
his claim that no imaginative episode can realise the particular prototype linked to
its specific cognitive counterpart for instance, when he stresses that imaginings are
bound to remain cognitively void, whatever may happen to them (OShaughnessy
(2000): 359). However, OShaughnessys discussion of imaginings and, in particular, his Argument from Origin does not address the issue of whether there are or
can be any further prototypes, over and above perception and intellectual know-
258
cessary for the occurrence of knowledge in all of its forms; and it has to be
absent in the case of imagining. Only then can it account for the assumed
fact that no imagining can cognise reality in the same single way as cognitive episodes. The result is an interpretation of (NC*) which understands it
in the light of a shift in focus from the particularities of the related cognitive constraints to their commonalities and, indeed, identifies (NC*) with
another, more schematic negation claim.
Critical Assessment
259
260
largely unilluminating because the corresponding non-imaginative affective states lack the cognitive properties in question as well. 39 The affective
imaginings concerned correspond either to non-imaginative episodes
which are non-representational (e.g., certain moods or bodily sensations,
such as anxiety or orgasm), or to non-imaginative episodes the representationality of which is not concerned with cognition, but rather with conation
or affection (e.g., feelings of desire or hope, or object-directed emotional
feelings, such as love). It may be denied that there are non-imaginative affective states of the first kind (see Crane (2001)). But the existence of
emotional states of the second kind should be beyond serious doubt (see
Goldie (2000)).
Now, none of these non-imaginative episodes seem to possess any of the
cognitive properties listed above. For instance, episodes of feeling anxious
or of longing for a cigarette cannot really be said to cognise reality, be reliable in their representation of it, be integrated in our epistemic picture of
the world, be informative about reality, be constitutively linked to it, possess a cognitive attitude, provide prima facie reasons for first-order belief,
or have the function of cognising reality. The explanation of this fact is
precisely that the non-representational, conative or otherwise affective
episodes concerned lack the kind of representationality linked to cognition
and the instantiation of properties related to cognition: they are not putatively cognitive representations, and their representationality (if they possess any at all) is not concerned with cognition. 40 As a result, the discovery
that the corresponding affective imaginings such as imagined feelings of
anxiety, orgasm, longing, hope, love, and so on lack these cognitive properties as well is not very informative. Given that longing for a cigarette
cannot cognise aspects of the world, it is no surprise that imaginatively
In addition, the negation claims remain unilluminating with respect to the non-cognitive affective aspects of affective imaginings that have belief-involving counterparts (such as conscious emotions that involve a belief about an actual situation).
40
See Section 2.1; as well as Velleman (2000): 182 and 248f.. The episodes under discussion may, however, be integrated with our mental life in other ways for instance, as reasons or motives for actions, or for higher-order beliefs. In particular, all
conscious mental phenomena (including imaginings) may be said to be rationally integrated in so far as they can ground introspective judgements concerning their presence (see Dorsch (2013)). But this form of rational integration surely does not render
them cognitive or cognising in any plausible sense.
39
Critical Assessment
261
Indeed, this may mean that, strictly, speaking, the negation claims formulated in
terms of the properties (i)-(vii) fail to contribute to a unified account of imagining in
this way not (merely) because they are unilluminating with respect to affective imaginings, but (also) because they do not even apply to them and, hence, cannot be true
(or false) of them. The problem with these negation claims would then pertain to
their extensional adequacy rather than to their explanatory power.
262
tude, or providing prima facie reasons for first-order belief may pertain
to the episodes making up a cognitive project, but not to the project or its
pursuit as a whole.
One reason for this is that mental projects do not themselves constitute
representations. Rather, they are activities which partly consist of representations. And since cognition and knowledge are exclusively a matter of
representation, it is only the episodic representations involved in projects
that can exemplify the cognitive properties (i)-(viii), but not the complex
projects comprising them. The other reason for the fact that cognitive or
other mental projects cannot reasonably be said to instantiate the cognitive
features under discussion is that many of them may lead to the manifestation or acquisition of knowledge, while none the less containing imaginative elements or non-cognising cognitions. When aiming to acquire knowledge about a certain issue, we may make extensive use of imagined hypothetical cases; and many successful cognitive projects inadvertently involve the formation and later revision of false judgements or beliefs.
But it would be strange if the project as a whole instantiated certain cognitive properties lacked by many of its elements. Again, saying that imaginative projects lack the features (i)-(viii) reveals hardly anything about their
distinctive nature, given that cognitive and other projects are due to their
very nature as mental projects bereft of them as well.
The only exception is perhaps the negation claim that imaginings do not
aim at the cognition of the world. This property related to cognition can
perhaps be instantiated, not only by episodes, but also by projects. Indeed,
it is reasonable to maintain that cognitive projects are to be characterised
and to be distinguished from imaginative projects as having the acquisition of knowledge as their aim. In addition, adopting an Agency Account of
imaginings and identifying them as instances of mental agency with a particular non-cognitive purpose would explain this teleological difference
between the two kinds of mental project (see Section 14.2).
The negation claim that imaginings are not reliably formed, on the other
hand, does not constitute any exception. For the occurrence of a particular
mental project that is, the occurrence of several, suitably linked mental
episodes is reliable only in so far as its episodic constituents are reliably
formed and connected to each other. This is the sense, for instance, in
Critical Assessment
263
which the two projects discussed in Chapter 4 are reliable: they are so because they involve the reliable formation of certain episodes of the sensory
imagination.
It might be thought that imaginative projects can be defined in terms of
imaginative episodes: namely as precisely those mental projects which
contain, aim at, or lead to, episodic imaginings (of a certain kind). And the
hope might then be that the negation claims are able to illuminate the
nature of imaginative projects by illuminating the nature of imaginative
episodes and characterising imaginative projects in terms of those episodes. However, this proposal should be rejected for four reasons.
First, it could still not ensure that the negation claims apply to all kinds
of imaginative projects. For it is possible that there are imaginative projects
which are to be characterised by reference to affective imaginings of the
kind discussed above: that is, to imaginings whose imaginative nature the
negation claims would still have nothing to say about.
Second, the resulting account would not be illuminating with respect to
the ambition of providing a unified theory, since it would account for the
imaginativeness of episodes and projects in terms of two distinct features:
while episodes would be said to be imaginative because of their lack of a
certain cognitive property, projects would be said to be imaginative because they involve imaginative episodes. Hence, the theory would be disjunctive: it would not identify a common feature of all imaginings, but instead postulate two very different (though related) ways of being imaginative.
Third, the idea that imaginative projects can be defined by reference to
their involving imaginative episodes is untenable. Although it seems true
that all imaginative projects have to involve at least some imaginative episodes while cognitive projects need not involve any imaginings, the latter
may involve and even be deliberately directed at the occurrence of imaginative episodes just like the former. Above, I already mentioned cognitive
projects that involve imagining in the shape of hypothetical reasoning. But
an even better illustration of the possibility of cognitive projects, that lead
to and, in fact, are aimed by us at the generation of imaginative episodes of
a certain kind, are the two examples presented in Chapter 4.
Fourth, specifying imaginative projects in terms of the aim or con-
264
Critical Assessment
265
kind (i.e., the essential link of imaginings to the will; see Section 4.7). In
Part Four, I argue that it is advisable to follow his suggestion at least with
respect to imaginings of which (NC) is indeed true and trace back the absence of the capacity of cognise to the presence of another feature, namely
their active character. The cognitive properties (i)-(viii) discussed in the
last section constitute plausible candidates for a more basic feature of the
first kind that is, of a feature that is responsible for the cognising power
of all cognitions, but which all imaginings lack. Indeed, (NC) may perhaps
be understood as the specification of (N y) in terms of the most general
cognitive property, the power to cognise reality, which is derivable from a
more basic specification in terms of one or more of the features (i)-(viii).
But even if (NC) is assumed to be the most basic negation claim available, which cannot be elucidated by reference to a more fundamental thesis
about the nature of imaginings, it is still limited in its explanatory power.
For just like the (other) specifications of the schema (Ny), it is incapable of
differentiating affective imaginings and imaginative projects from their
non-imaginative counterparts, and for the same reasons. Again, the issue of
whether they can cognise reality does not arise with respect to many affective episodes irrespective of whether they are imaginative or not because of their non-representational, conative or otherwise non-cognitive
nature. And, again, mental projects are in contrast to their episodic parts
not of the right kind of entity for the realisation of cognising power, given that the constitution of knowledge is a matter of representational episodes or states. Hence, (NC) does not fare better with respect to explanatory
power than the other negation claims.
To sum up, the various explanatory limitations of the negation theses
discussed should already be enough to disqualify them as substantial components of a unified account of imagining. They may be true with respect
to affective imaginings and imaginative projects, but not distinctively
that is, not in virtue of their imaginativeness since the claims are true of
all affective episodes and mental projects, and not only imaginative ones.
As a consequence, negation claims are severely restricted in their contribution to the elucidation of the common imaginativeness of the different
forms of imagining. Hence, they should not be part of a theory of imagining, or at least not if the aim is to provide a unified account of imagining
266
From OShaughnessys own perspective, there may actually be another counterexample: pictorial experience. For him, such experiences involve an imaginative experience of what is depicted, that is, an imagined perception of it (OShaughnessy
Critical Assessment
267
268
(vii) means also that they cannot be necessary for cognising reality. In other words, mental episodes may realise a (non-intellectual) form of knowledge without possessing a cognitive attitude or function, and without
providing us with prima face reasons for belief. 43 Besides, (NC*) and the
specifications of (Ny) in terms of the features (v)-(vii) are still subject to
the severe explanatory limitations noted above and, hence, cannot make a
substantial contribution to a unified account of imagining.
It seems that the mistake in OShaughnessys argumentation does not
pertain to his defense of (NC*) by means of the Argument from Origin. Indeed, it is reasonable to assume that he is right in arguing that imaginings
cannot satisfy the particular cognitive constraints on perceptual experiences and beliefs; and that they therefore cannot count as cognising reality
in the same manner as these cognitive states. The problematic aspect of his
argumentation is, rather, his derivation of (NC) from (NC*). The crucial
assumption in this derivation is that there are no other cognitive prototypes
which imaginings could exemplify.
His reason for making this assumption seems to be his belief that there
are only two cognitive prototypes: the fundamental mental cognitives
knowledge and perception (363). Accordingly, he appears to have overlooked the possibility of other cognitive prototypes, the exemplification of
which also requires intentionality, veridicality, reliability and possibly other cognitive properties, but which differ from the perceptual and the intellectual one in that they demand a different kind of reliability. Perceptions
have to be reliably caused by the perceived aspects in the world. Beliefs (or
judgements) have to be rationally supported by other relevant cognitive
states. Sensory memories have to reliably depend on our past perceptions
of things. The two visual imaginings in the examples have to be based on
reliable (and veridical) perceptions or sensory memories; and they have to
be the result of imaginative activity which is reliable at least with respect
to the extraction of information already contained in the underlying cognitive states. Hence, there seem to be at least four different forms of knowledge, each with its particular type of reliability, its specific cognitive constraints and its particular cognitive prototype or ideal.
This raises the question of why the examples do not seem to have
43
This may explain why these features do not figure at all in OShaughnessys arguments for the negation theses (NC)*) and (NC).
Critical Assessment
269
270
though there are many cognitive projects which (have to) involve imagining, few of them involve cognising imaginative episodes. Or, to speak in
Dancys terms, the imaginative episodes involved in those projects do not
play a role in the justification of the resulting knowledge-constituting beliefs, but merely enable us to acquire them (see Dancy (2000): 127).
Consider the following three examples: (a) we can find out about the
thoughts and feelings of another person by imagining being in her position
and state of mind; (b) we can discover the logical consequences of a certain proposition by engaging in hypothetical reasoning on the basis of imaginatively entertaining it; and (c) we can find out about the aesthetic qualities of a certain configuration of objects (e.g., whether hanging a specific
picture on a certain wall would look nice; or whether adding a yellow
square to a Mondrian painting would destroy its balance and harmony) by
visualising them and observing our emotional reaction towards them. The
last category may also include emotional reactions to imagined objects
which are not necessarily or clearly aesthetic (e.g., feeling disgust when
imagining the taste and consistency of worms in ones mouth).
In each of these cases, it seems reasonable to assume that we can gain
some knowledge about reality (understood in a wide sense, so as to include
the recognition of response-dependent features): we may learn something
about the minds of other people, about the logical properties of propositions, or the aesthetic qualities of certain combinations of objects. But a
proponent of (NC) has the plausible option to deny that the respective cognising representations are imaginative.
In the example from aesthetics, the relevant ground of knowledge is the
(aesthetic or other) emotion felt towards the imagined object: we judge the
configuration to be beautiful or balanced on the basis of our real (i.e., nonimaginative) feeling towards it (see, e.g., Budd (1995); though see also
Dorsch (2007)). With respect to imaginatively gained knowledge of what
can be inferred from a given proposition, on the other hand, it is not the final imaginative intellectual representation of the inferential train of suppositions which grounds the knowledge, but the inference as a whole. For
the last supposition does not contain any information about its inferential
origin: only (higher-order) reflection on the whole inference can establish
the recognition of the inferential link between the original and the inferred
Critical Assessment
271
272
44
Note that the practical purpose of cognitive projects differs from the epistemic function of cognitive episodes. For the general goal of cognitive projects is not to cognise reality, but to bring about episodes which cognise reality (see Section 14.2).
Part Three
The Dependency Account
The central idea of the Dependency Account of imaginings is that, fundamentally, imaginings are constitutively dependent on cognitions (and similar phenomena, such as emotional feelings). This constitutive claim is often accompanied by a claim about the concept of imagining and its necessary elucidation in cognitive terms. Here, I focus mainly on the metaphysical claim, given that unified accounts of imagining are primarily concerned with the nature of imaginings, and not with how we conceive of
them.
The various versions of the Dependency Account differ, first of all, in
their notion of the assumed fundamental dependency relation. In particular,
imaginings may be said to have cognitions either as their necessary causes,
or as a necessary part of their representational contents, or as the necessary
origin of their representational capacity, or as the necessary target of their
imitative or simulative employment. Depending on the kind of dependency
concerned, there are also differences which cognitive phenomena serve as
the dependency base. While representational and imitative links concern
the whole set of cognitive counterparts of imaginings (e.g., judgements or
beliefs as well as perceptions and sensations), causal and semantic dependency bases consist primarily of cognitions that link us directly to reality
(e.g., perceptions and sensations). The resulting theories may be labelled
the Causal, the Representational, the Semantic and the Imitation Account
of imagining.1
1
Hume endorses both the Causal and the Representational Account, given that he
takes imaginings to be ideas that causally and representationally depend on particular impressions (see Hume (1739): 1.1.1.6ff. for the double echo claim; and 1.3.7.7
276
OShaughnessys echo claim, introduced in Section 6.2, is a more specific version of the Representational Account in that it takes the representational relation to be non-intentional, and the relevant cognitions to be either
perceptions or knowledge-constituting judgements or beliefs. Indeed, his
claim illustrates that versions of the Dependency Account may also differ
in their characterisation of the cognitive phenomena that serve as the dependency base for instance, whether they include only knowledge-constituting episodes or, more generally, all cognitive episodes. But they all
have in common that they specify the nature of imaginings by reference to
perceptual experiences, judgemental thoughts (or occurrent beliefs) and
similar mental episodes (e.g., sensations or feelings).2
Just as in the case of the Epistemological Account, the various endorsements of the Dependency Account show the same structural form. More
specifically, their central claims such as the theses listed above model
imaginings on their cognitive counterparts or (as I will say) understand
them as echoes of the latter. Claims of this kind which I call echo claims
involve at least three components: the idea of dependency, the idea of inheritance, and the idea of counterparts.
The idea of dependency is, of course, that imaginings are dependent in
some way or another on cognitions which explains why they have to be
for the characterisation of imaginings). OShaughnessy (2000) endorses the Representational Account with respect to all kinds of imaginings (see the discussion in Section 6.2), while Peacocke (1985): 22, Martin (2001): 273, (2002): 404, and Dorsch
(2010c) endorse it with respect to (central cases of) sensory imaginings (see also
Note 10 in Chapter 1). Finally, Currie & Ravenscroft (2002): 49, explicitly defend a
semantic dependency (or echo) claim and may also together with Ryle
(1949/1963): 250ff. be read as embracing the Imitation Account. See the Chapters
9f. and 12 for detailed discussions of each of these endorsements.
It would in principle be possible to account for imaginings in terms of a dependency
on non-cognitive phenomena. But no one has, to my knowledge, defended such a
view and for a good reason, given that imaginings are by far most similar to cognitions and, hence, should be modelled upon them (if on anything at all). Note that entertaining a proposition understood as a necessary ingredient to all instances of
thought and possibly also to all sensory representations, if the latter turn out to be
propositional as well does not count as an exception, despite being non-cognitive
and essential to imagining. For entertaining a proposition in this sense does not constitute a full-blown mental episode, but merely a necessary part of one (see Section
14.2).
277
Not dissimilar to the way in which fake lemons are made in order to resemble real
lemons (by recreating their visual or otherwise perceivable appearance), without
amounting to real lemons (i.e., without recreating their DNA; see Dorsch (2010c)
and (2012a)).
278
formulated). What prompts the specific assignments are, again, the similarities holding between the respective relata. The counterpart of visualising,
say, is seeing because both kinds of episode represent the same kind of entities and features (i.e., visible ones), and do so in the same manner (i.e.,
visually). This explains why cognitions are the best candidates for the dependency base for imaginings: the two kinds of episode resemble each other (almost) perfectly in what they are capable of representing, and in which
type of representation they may instantiate (see Section 2.1).
The Dependency Account of imagining may be taken to provide an exhaustive unified account of imagining in the sense that it suffices to completely characterise the distinctive nature of imaginings (as well as in the
most fundamental terms). But, as the example of OShaughnessys view
has illustrated, it is possible to supplement the Dependency Account with
another view. One particular reason for doing this is the existence of sensory memories. The thought is that sensory imaginings are even more similar to sensory memories than to perceptions, notably in respect of whether
they represent their objects as currently present in our actual environment
that is, in respect of whether they represent their objects in an immediate
way, as specified in Section 3.2. To account for this fact about the three
kinds of sensory episode, and to avoid taking sensory imaginings to be dependent on episodic memories, it is perhaps necessary to treat sensory imaginings as echoes of perceptions as well (see Martin (2001)). But, then,
maintaining that imaginings are dependent on cognitions may not suffice
any more to distinguish them from episodic memories. Hence, some further fundamental claim about the nature of (sensory) imaginings may have
to be needed to be added to the echo thesis in question.
I return to this issue in Section 9.4, where I suggest that the purity of
the Dependency Account may be saved by taking imaginings to be dependent on types of cognitions rather than on tokens, as in the case of episodic
memories which serve as recollections of particular past experiences. None
the less, the following discussion and assessment of the Dependency Account should be understood as staying neutral on whether the acceptance of
an echo claim may suffice to provide a complete unified account of imagining, or whether it may just make a substantial contribution to such a theory.
279
The chapters and sections of this part of the book are structured around
the discussion of the various versions of the Dependency Account, as well
as the corresponding echo claims listed above. Three of the chapters
(Chapters 9 to 11) deal with the Representational Account, while one
chapter (Chapter 12) concentrates on other versions of the Dependency Account. The long discussion of the Representational Account is warranted by
the fact that this theory of imagining is the most promising rival to the
Agency Account and, moreover, reveals a lot about the nature of two particular forms of imagining, namely sensory and affective imaginings.
Chapter 9 is devoted to the prospects of the Representational Account as
a unified theory of imagining, as well as to the Humean origins of this account. In the first three sections, I critically discuss Humes conception of
imaginings. His view of imaginings as representations (or copies) of cognitions differs from contemporary versions of the Representational Account by endorsing a specifically causal rather than intentional or relational understanding of the representational link. As a result, Hume ends
up endorsing both the causal and the representational version of the Dependency Account: imaginings are causally and representationally dependent on their cognitive counterparts in so far as they represent precisely
those particular cognitions on which they causally depend. 4 Accordingly,
the discussion of Humes conception includes also a discussion of the causal version of the Dependency Account that is, what might be called the
Causal Account of imagining. My conclusion is that his conception does
not provide us with a satisfactory account of the distinctive nature of imaginings which, arguably, also was not Humes ambition.
The fourth section of Chapter 9 introduces the main intentional and relational variants of the Representational Account and highlights their continuity with the Humean conception spelled out before. In the remaining
two sections, I discuss two forms of imagining namely intellectual imagining and imaginative projects the nature of which cannot be properly
captured by means of a representational echo claim. If my arguments are
4
OShaughnessy is therefore not the only philosopher who combines an echo claim
about imagining with another thesis about their nature. But Humes approach is different from OShaughnessys in that it combines two echo claims rather than an
echo claim and a negation claim and thus still provides a pure Dependency Account.
280
CHAPTER 9
The Representational Account
The Representational Account has its origin in the empiricist idea that all
mental representation starts with, and derives from, sense experience
(where this may include sensation as well as perception). One particular respect, in which Hume stands out from the other British Empiricists, is that
he assigned a very prominent place in his philosophy of mind to the imagination. While this is primarily true of the role of the imagination (understood as something like a mental faculty or capacity) in the distinction,
combination or association of ideas, it also applies to the imaginative complex ideas (understood as mental episodes) that are the particular results of
imagining. Indeed, Hume was one of the first philosophers to discuss episodes of imagining in any detail and, moreover, had a great influence
though mainly by serving as a target of objections on the subsequent discussions of imaginings, notably in the writings of Husserl, Sartre, Ryle and
Wittgenstein.
In addition, he inspired together with his empiricist predecessors the
recent endorsement of the idea that imaginative (and other) episodes are
representations of perceptions, which constitutes a version of the Representational Account of imagining. The first two sections of this chapter are
concerned with, respectively, the presentation and criticism of Humes conception of imaginings (and other non-perceptual episodes) in terms of the
copy principle, which understands representation in essentially causal
terms. In the third section, I briefly argue that the causal element of
Humes view has no role to play in a unified account of imagining and
should therefore be dropped from such an account. The last section illus-
282
trates different contemporary ways of spelling out by reference to intentional or relational conceptions of representation the resulting purely representational version of the Dependency Account.
See Hume (1739): 1.1.1.1, and also some of the passages quoted further below.
Memories are a special case and will be discussed separately further below.
283
distinct heads, and assign to each a peculiar name to mark the difference.
(Ibid.: 1.1.1.1)
The near resemblance of instances of the two types of mental episode and
our subsequent problems to tell them apart are still not meant to imply that
some ideas might be more vivid than some impressions they only approach each other very closely.6
Another point is that Hume takes the vivacity of our mental episodes to
be an aspect of their subjective characters which enables us to tell apart,
from the inside, instances of the various kinds of mental episode. For he
acknowledges that the vividness of mental episodes is part of how they appear to us in consciousness and of what lets us distinguish them from our
first-personal perspective. This becomes apparent in the passages where
Hume describe the subjectively accessible differences among ideas of three
kinds, namely those of judgement, memory and imagination.
For tho it be a peculiar property of the memory to preserve the original order and position of its ideas, while the imagination transposes and changes
them, as it pleases; yet this difference is not sufficient to distinguish them in
their operation, or make us know the one from the other; it being impossible
to recall the past impressions, in order to compare them with our present
ideas, and see whether their arrangement be exactly similar. Since therefore
the memory is known, neither by the order of its complex ideas, nor the
nature of its simple ones; it follows, that the difference betwixt it and the
imagination lies in its superior force and vivacity. (Ibid.: 1.3.5.3)
An idea assented to feels different from a fictitious idea, that the fancy
alone presents to us: And this different feeling I endeavour to explain by
calling it a superior force, or vivacity, or solidity, or firmness, or steadiness.
[...] [I]t is something felt by the mind, which distinguishes the ideas of the
judgement from the fictions of the imagination. (Ibid.: 1.3.7.7)
In Hume (1739): 1.3.5.7, the focus is also on the effect a de- or increase in vivacity
has on what we take a given episode to be. Hume also notes there the possibility that
an imaginative episode may change into a cognitive one that an often enough repeated idea of the imagination may become an idea of judgement or memory. But
again, this makes clear that a sufficient de- or increase in vivacity leads to a different
kind of episode.
284
One may wonder whether the enumeration of different terms really already
helps to explain what vivacity amounts to. But Hume offers more, namely
the claim that a higher degree of vivacity comes with two other important
aspects: (i) an increased sense of presence or reality with respect to the objects and features presented7 to us by the episodes at issue; and (ii) a higher
motivational (or rational, as one would feel inclined to say today) impact
on our beliefs, emotions and actions.
This variety of terms, which may seem so unphilosophical, is intended only
to express that act of mind, which renders realities more present to us than
fictions, causes them to weigh more in the thought, and gives them a superior influence on the passions and imagination. (Ibid.: 1.3.7.7)
In the case of judgemental thoughts, their high degree of vivacity also ensures that they lead to the formation of a more stable and enduring belief.
It gives them more force and influence; makes them appear of greater importance; infixes them in the mind; and renders them the governing principles of all our actions. (Ibid.: 1.3.7.7)
Accordingly, the subjective vivacity of a mental episodes reflects the closeness of its connection to perception and its impact on belief which is, of
course, in line with Humes thought that perceptions are the most vivid
episodes that we enjoy, and that imaginings are characterised by the least
degree of vivacity.
His second and less explicit characterisation of the difference between
impressions and ideas introduces both the idea of a resemblance between
the two and the notion of a causal dependency of the latter on the former.
My use of the term presentation is meant to be neutral enough to allow for both intentional or relational forms of presentation of objects, as well as for the sensory or
intellectual presentations of objects that may be given as past, present, actual, nonactual, and so on. The presentation of an object is, however, always taken to be a
conscious presentation. The expression representation, on the other hand, is intended to highlight the fact that the represented perceptions or judgements are thereby
not themselves present in the stream of consciousness, but instead merely re-presented by the respective episodes.
285
Thus we find, that all simple ideas and impressions resemble each other;
and as the complex are formd from them, we may affirm in general, that
these two species of perception are exactly correspondent. [...] Let us consider how they stand with regard to their existence, and which of the impressions and ideas are causes, and which effects.
The full examination of this question is the subject of the present treatise;
and therefore we shall here content ourselves with establishing one general
proposition, that all our simple ideas in their first appearance are derivd
from simple impressions, which are correspondent to them, and which they
exactly represent. (Ibid.: 1.1.1.6f.)
286
That a given complex idea may be caused merely by several simple impressions, and not also by a corresponding complex one, raises the issue of
which impression(s) it is said to resemble. It seems plausible to maintain
that the complex idea resembles each of the simple impressions in so far as
it possesses parts (i.e., simple ideas) which resemble the latter. On the other
hand, and as already noted, the resemblance at issue here concerns primarily the (non-mental) objects and features presented by the ideas and impressions compared. And this similarity is unlikely to hold between complex ideas and simple impressions, at least with respect to the presentation
of objects and of higher-level properties. For Hume understands simple impressions as presentations of basic features, such as colours, tastes or
smells (ibid.: 1.1.1.2). Hence, although complex ideas may resemble each
of the respective simple impressions, this is not the kind of similarity referred to in the copy principle. Instead, what is meant is the resemblance of
ideas on those impressions that present the same objects and features as the
former in this case, the resemblance of complex ideas on similarly complex impressions. However, it remains unclear which complex impression
could be relevant in cases where a complex idea is not preceded by any
corresponding complex impression. It is true that such ideas would resemble a complex combination of the relevant simple impressions, if any
would actually come into existence. But this is not the same as saying that
the complex idea resembles an already given impression.
The same set of issues does not arise with respect to the proposed representational link between ideas and impressions. On the one hand, in contrast with resemblance, representation need not be genuinely relational in
the sense of requiring an actually given second relatum. And, on the other
hand, a complex idea represents not simply individual simple impressions,
but also their complex structure. Imagining a blue book on a brown table
corresponds to seeing a blue book on a brown table, and not to seeing a
brown book on a blue table, despite both complex impressions involving
the same simple ones. Accordingly, what complex ideas represent are complex impressions, though not necessarily particular ones (e.g., the one I had
yesterday when looking at my friends house).8
8
The preceding considerations liken Humes complex ideas to pictures, given that
both allow for the presentation of some kind of object, without the need to present a
particular and actual instance of that kind; and given that both may perhaps still give
287
This suggests perhaps also a solution to the resemblance issue. The key
thought is that ideas present certain (non-mental) objects and features precisely because they are representations of impressions which present those
objects and features. This has the result that impressions and ideas differ in
immediacy. But it also means that ideas resemble the impressions that they
represent in so far as they present the same (non-mental) objects and features as the latter albeit typically in a less vivid way. For they represent
an impression with a certain property such as a certain content or phenomenal character and thereby acquire themselves this property, or at
least a property very similar to it (see the discussion of immediacy in Section 3.2). A complex idea therefore resembles the complex impression that
it represents. And although the latter need not enjoy actual existence in the
past or present, it is clear which complex impression is relevant. Moreover,
we do not generally find claims about the resemblance between actual and
merely represented entities problematic. We may recognise similarities
between friends of ours and characters in a film; or between a real person
and our mental image of her which we have formed prior to meeting her or
knowing anything about her appearance.
These considerations allow now for a more precise reading of Humes
copy principle. Since the symmetric resemblance between ideas and impressions is probably best understood as a consequence of their asymmetric
representational link, we need not any more mention the former separately.9 According to the resulting interpretation, the principle maintains that
particular ideas are copies of particular impressions in that they (i) causally
depend on the later, and (ii) are representations of the latter in such a way
288
289
of how the two distinct characterisations of the difference between the two
types of mental episode fit together. In the case of photocopies, the contrast
and saturation of the marks on their surface tend to be less than those of the
marks on the original sheets of paper. Similarly, the vivacity of episodes
might be understood as an aspect of their subjective character which is
bound to decrease when mentally reproduced. That is, this reproduction
might be of such a nature that it results not only in the episodes inheritance of the presentation of certain objects and features, but also in a diminishing of their sense of reality or presence and of their impact on beliefs,
emotions and actions.
We have finally reached the point where we are in a position to become
more concrete about Humes conception of imaginings. Both his examples
of imagining and his discussion of the difference between imagining and
remembering something suggest that he takes imaginings to be complex
ideas. Imagining the New Jerusalem, whose pavement is gold and walls
are rubies (ibid.: 1.1.1.4), or winged horses, fiery dragons, and monstrous
giants (ibid.: 1.1.3.4), means combining simpler ideas whether the latter
are still somewhat complex (such as the ideas of horses, wings, and so on)
or indeed among the simplest (such as the ideas of colours, tastes, smells,
and so on). Correspondingly, ideas of the imagination differ from ideas of
memory in whether they preserve the order in which the relevant impressions occurred before (ibid.: 1.1.3.2).11 And this, again, presupposes that
both are complex ideas possessing an internal structure. According to the
proposed reading of the copy principle, Hume therefore maintains that it is
constitutive of imaginings that they represent complex perceptions and
causally depend on the prior perceptual occurrence of the simple aspects of
those perceptions.12
Note that this fact is not directly subjectively accessible (see the passage from Hume
(1739): 1.3.5.3 quoted above). It is interesting to ask whether the also postulated and
subjectively salient difference in vivacity between memories and imaginings might
be said to indirectly reflect this difference of how the two kinds of episodes are
taken to relate back to the original perceptions.
12
I assume here that our perceptions are simple that is, for instance, of a single colour and no other sensible quality only in rare and artificial circumstances. Perhaps
there may also be simple instances of imagining, in which case Humes conception
would have to be slightly modified as to allow for simple ideas of the imagination as
well.
11
290
291
If repetition of an impression is understood here as meaning the literal reoccurrence of the original impression, the explanation of the memories impact on beliefs has the price of rendering them indistinguishable from perceptions. But if it is instead taken to denote, in accordance to the copy principle, the less vivid representation of a perception, the initial problem of
accounting for the memories perception-like link to belief reoccurs. This
illustrates that Hume has difficulties to accommodate the fact that memories are very similar to perceptions in their epistemic function, while also
holding on to their differences in vivacity and immediacy that is, their
differences in how they present us with objects.
Finally, the first characterisation of the difference between impressions
and ideas is in tension with the second one precisely because of this aspect
of Humes conception of vivacity (and despite the explanatory link
between the two mentioned above). For treating ideas as copies of impressions in the sense just specified means treating the two as being different in
kind. Hence, it clashes with the claim that the only difference between the
copies and what they are copies of is one of vivacity (ibid.: 1.1.1.3). In
fact, this inconsistency seems already to be inherent to the copy principle
itself, given that its second clause postulates a resemblance in all respects
except vivacity (i.e., a difference in degree), while the third clause puts forward the idea of a representational link (i.e., a difference in kind).
Now, Humes theory of the mind is subject not only to general challenges to his theory of mental episodes. Some objections are also more
specifically related to the particular incorporation of imaginings in his
overall view of the mind. Hume claims that episodes of imagining are
causal reproductions and therefore representations of their cognitive counterparts. In addition to the difficulties already mentioned, this thesis is
problematic for at least two reasons.
The first is that it is unclear how to avoid the conclusion that all imaginings (just like all thoughts or judgements) involve some sensory or affect-
292
ive element, given that they are or include copies of perceptions with sensory or affective qualities. This idea makes sense in the case of affectively
imagining a pain or of sensorily imagining something red: the character of
the first episode contains some element of painfulness, and the second
some quality of reddishness.13 But we can suppose (or, more generally,
think) that an object is a certain way without any sensory or affective element involved. The underlying problem is, again, that Humes theory cannot properly accommodate the distinction between sensory, affective and
intellectual episodes, especially given that this distinction is one in kind.
The second reason for being uneasy about the proposed claim about
imaginings is that it can at best play only a minor role in an account of
imagining. That imaginings are dependent on perceptions in the way described is not distinctive of them within Humes theory. For the same is
said to be true of episodes of memory, thought and judgement. This leads
back to the worry that, ultimately, reference to degrees in vivacity is his
only means to establish differences among mental episodes, while the more
fundamental or significant ones of these are in fact differences in kind.
Moreover, even if Humes copy principle is read in such a way as to describe a qualitative difference between impressions and ideas, it does not
suffice to strictly distinguish imaginings from other kinds of ideas, such as
judgemental or mnemonic ones. Humes conception of imaginings therefore does not and also was probably not intended by him to provide the
resources to formulate a proper theory of imaginings, given that it does not
really treat them as a distinct kind of mental phenomena.
293
294
295
imagining. In accordance with these thoughts, the Representational Account of imagining (i.e., the representational version of the Dependency
Account) aims to elucidate the nature of imaginings primarily if not exclusively in terms of the following fundamental echo claim:
(ER)
Endorsing (ER) should in fact mean endorsing the view that all instances
of imagining are instances of experiential imagining (see also Sections 1.2,
10.4 and 11.5, as well as Walton (1990): ch. 1, and Walton (1997)). The alternative option of treating imaginings as intellectual representations of
(i.e., thoughts about) cognitions is evidently untenable because sensory and
affective imaginings cannot be reduced to purely intellectual episodes (see
Section 11.3). Moreover, even intellectually imagining that the Earth is
flat, say, is not the same as entertaining the (higher-order) thought that one
judges or believes that the Earth is flat. In particular, supposing that the
Earth is flat does not require the possession or application of the concept of
judgement or belief.
But how does the idea that the objects of imaginative representation
that is, what is imagined are cognitive episodes square with the fact that
we may visualise landscapes, or suppose that the Earth is flat? The idea is
to distinguish between direct and indirect objects of representation. What
imaginings are said to represent directly are cognitions. That is, the representational content of the imaginative episodes consists just of cognitive
episodes. But cognitions are themselves representational: they represent
landscapes, the flatness of the Earth, and so on. Hence, what proponents of
the Representational Account assume is that, by (directly) representing
cognitions, imaginings also (indirectly) represent what the cognitions represent. The imaginative episodes are understood as inheriting the representationality of the cognitive episodes in so far as the latter are together
with their content part of the content of the former. 14 Of course, (ER)
14
See Section 10.4 for a more detailed discussion of how this is supposed to work in
the case of visualising. In the case of intellectual imagining, this form of inheritance
is likely to be similar to the sort of embedding that may occur in the case of higherorder thought. Thinking that one is thinking that it rains literally includes thinking
296
could be reformulated in such a way as to make this assumption more explicit. But this would render it unnecessarily complicated. It seems sufficient to say that imaginings are representations of cognitions to indicate
that they represent external objects not in the same (direct) way, but instead
(indirectly) by representing cognitive representations of them.
Now, simply dropping (EC) and shifting the focus exclusively to (ER) is
not enough to address all the problems for the Humean approach that have
been noted in the previous sections of this chapter. But this general approach to imaginative episodes is open to further improvements, without
having to give up on the fundamental idea that imaginative episodes depend on cognitive ones by being representations of them. Two additional
important modifications are, first, the introduction of the qualitative distinction between sensory and intellectual presentations; and second, the
substitution of the qualitative differences in attitude and epistemic function
for the quantitative difference in vivacity. Humes contrast between impressions and ideas is thus replaced by the opposition of cognitions and
imaginings, as well as the orthogonal opposition between sensory and intellectual episodes. As a result, not only perceptions, but also episodic
memories and judgemental thoughts are taken to be episodes that do not
represent others in the relevant way, but instead can figure as the objects of
such a representation. And only imaginings plus possibly memories (see
below) and spontaneous images continue to count as representations of
cognitions.
These two further modifications suffice to solve several of the other difficulties facing Humes theory of the mind. The resulting version of the
Representational Account can accommodate the fact that the various kinds
of cognition and imagining differ in kind from each other, and along the dimensions outlined above. It also avoids any of the problems linked to the
introduction of vivacity as an important element in an account of imaginings. Intellectual imaginings need not involve any (more or less vivid)
sensory elements any more, since they can now be construed as representations of intellectual cognitions, such as judgemental thoughts or occurrent
that it rains. That is, there are two thoughts and two contents which are such that
one of them contains the other. A similar form of inclusion or overlap can be expected to be present in the case of imagining thinking that it rains, which also entails
thinking that it rains (see Section 9.5).
297
beliefs. Finally, the modified theory comes closer than the original to the
identification of a distinctive feature of imaginings, which separates them
strictly from cognitions and other non-imaginative mental episodes. For
while it is said to be constitutive of imaginings that they are representations
of cognitions, the same is not true of perceptions, judgemental thoughts,
bodily sensations, feelings of emotion or desire, and so on.
The application to episodic memories may appear to remain problematic, however. They still seem to fall in between perceptions and instances
of sensory imagining. While they share their cognitive attitude and epistemic function with the former, they do not present their objects as being
there before us in our environment and, in this respect at least, resemble
the latter. This raises again the issue of how they can actually share their attitude and impact on belief with perceptions, despite their lack of the latters direct connection to reality. And it also generates the question of
which set of features is distinctive of imaginings, if it turns out that episodic memories, too, are best treated as representations of perceptions.
But the Representational Account might have the resources to satisfactorily address these two issues, and thus to properly deal with episodic
memories (see Martin (2001)). If episodic memories are indeed representations of past perceptions, they may inherit the particular content of the latter. That is, they may also be (indirect) representation of the specific objects and features then perceived and, moreover, represent these objects
and features as they were once represented by ones past perception. In this
way, episodic memories may provide us with access to particular aspects of
the past. And this fact may very well explain why they involve a subjectively salient commitment to how things actually were, and why they influence our beliefs in roughly the same way as the original or other perceptions. They would still differ from the latter in that they do not represent
their objects as being there before us, but locate them in the past.
The contrast with sensory imaginings might then be established by arguing that the imaginative episodes although representing some type of
perceptions do not represent particular perceptions (see Martin (2001)).
This would ensure that they do not bring us into contact with the (present
or past) actual world something which is reflected by the fact that their
objects are not given to us as actual, and that they do not show the cognit-
298
ive attitude and impact on belief of cognitions. The Representational Account may therefore identify the representation of non-particular cognitions
as the distinctive feature of imaginings. And this would not only suffice to
distinguish them from episodic memories and other non-imaginative episodes, but would also promise an account of their lack of epistemic features.
Indeed, it should become clear now that separating the representational
echo claim from the causal one and understanding solely the former as being central to the Representational Account does not only promise the reward of a unified account of imagining, but also has the further advantage
of allowing for non-mechanistic interpretations of the postulated
representational link between imaginative and cognitive episodes. While
Humes empiricist picture of the mind leads him to take imaginings to be
causal reproductions of particular cognitions, it is now possible to conceive
of them also as intentional or relational representations of types of cognitions. That is, instead of stressing the causal dependency between imaginative and cognitive token episodes, it may be argued that imaginings constitutively depend on types of cognition in the sense that the latter are necessarily part of the representational content of the former (see Peacocke
(1985), OShaughnessy (2000): 365f., or Martin (2002a): 404-407).15
The result are endorsements of (ER) that maintain, for instance, that
visualising amounts to imagining seeing, or that supposing is nothing but
imagining judging or believing. In other words, visualising and supposing
are assumed to be nothing else than instances of experiential imagining. 16
As already mentioned above, the content of the cognition is thereby taken
to be embedded in some form or another in the content of the imagining. With respect to intellectual imaginings, the conceptual content of the
judgement or belief simply constitutes part of the content of the supposiThe idea of representing types, in contrast to representing tokens, certainly needs
further investigation (see also Section 10.4). But that this distinction obtains can be
independently illustrated by reference to pictures: while some paintings (e.g., portraits) are of particular people, others are of no particular people at all, although they
depict people or, rather, types of people (e.g., in genre paintings; see Wollheim
(2003)).
16
Note that the claim here is not the (relatively uncontroversial claim) that there are instances of experiential imagining, but rather the claim that the basic episodic forms
of imagining (i.e., sensory, intellectual and, presumably, also affective imagining)
can be reduced to another form of imagining, namely experiential imagining.
15
299
tion. In the case of sensory imaginations, the situation is a bit more complex since the sensory content of perceptions may loose some of its aspects
(e.g., its immediacy, or its vivacity) when imaginatively represented (see
Sections 3.2 and 10.4, as well as Martin (2002a)).
Taking imaginings to be dependent on types of cognitions, rather than
on tokens, has the consequence that imagining something does not require
any more the prior occurrence, in the same subject, of instances of the cognitive types concerned. Hence, the proposal can accommodate the possibility (if it is a genuine one) of, say, the visualisation of a previously unperceived shade of blue: we visualise this colour, not in virtue of having previously seen it, but by imagining having a certain type of perceptual experience, namely one of that colour. In addition, the representational view under discussion does not need to restrict (ER) to simple cases or aspects of
imaginative representation. For the view can explain why we are able to
visualise a unicorn without previously having seen one by reference to the
existence of the corresponding type of perception which, after all, could be
instantiated if unicorns would exist. Besides, the view may also be able to
account for the dependency of sensory memories on perceptions in a way
that does not render them imaginative (i.e., that does not render their dependency indistinguishable from that of imaginings). For, as suggested
above, it may be argued that, while episodic memories represent and depend on previously existing tokens of experience, sensory imaginings are
so related to types of experience (see Martin (2001): 279). What these considerations show is that formulating the Representational Account in terms
of a type-based representational echo claim avoids most of the remaining
problems facing Humes causation-centred conception of imaginings.
It might seem most natural to understand the postulated representational
link between imaginings and cognitions in intentional terms. After all, we
are capable of imagining cognitive episodes which no one ever has actually
had (e.g., when imagining the perceptions of the first man on Mars, or the
memories of the pyramids). But that this is not the only option is illustrated
by OShaughnessys account, the core idea of which is much closer to
Humes copy principle than any intentionalist echo thesis.17 In OShaugh17
See Section 6.2 for a more detailed discussion of OShaughnessys view. He hardly
ever makes explicit references to the philosophers who have had a (positive or negative) impact on the presentation and content of his own arguments. But he does so
300
nessys theory, imaginings have similar contents and involve similar types
of representation as cognitions because they are shadows or replicas of
the cognitive prototypes (see OShaughnessy (2000): 363ff.). What he
has here in mind is the idea that imaginings have types of veridical perception or knowledge as their immediate, non-intentional and necessary objects (not unlike sense data in the case of perception) and borrow their
world-directed intentionality from the latter. This is actually what he means
when he speaks of cognitive prototypes for imaginings: the imaginings
must not only be described in terms of the prototypes, but are also necessarily modelled upon them, that is, constitutively related to them in the special way described. It is in this sense that we are said by him to imagine
something by representing the relevant cognitive prototype: that is, visualise a tree by imagining seeing (or seem seeing) one, or make-believe that
the Earth is flat by imagining believing (or seem believing) the respective
proposition (see OShaughnessy (2000): 363f.).18
The closeness to Humes original account manifest in the postulation
of a genuinely relational, and not merely intentional, link between imaginings and their cognitive counterparts has the consequence that, for
OShaughnessy, imaginings of a certain kind (e.g., with a certain content or
of a certain type) could not exist if the corresponding kind of perception
would not exist (see OShaughnessy (2000): 377). It is not clear whether
in this context and, of course, to Hume which suggests that even OShaughnessy
himself sees himself as standing in the tradition of Hume (see OShaughnessy
(2000): 365).
18
As already mentioned in Note 11 in Section 6.2, there is at least one good reason to
prefer an intentionalist understanding of the representational link between token
imaginings and types of cognitions over a relationalist conception: namely the difficulty of rendering intelligible the idea of a genuinely relational form of representation, which has, as one of its constituents or relata, a type of episode, rather than a
particular episode. Since there cannot be genuinely relational forms of awareness
with purely general contents, the solution has to be something like the proposal to
treat the represented prototypes themselves as particulars, rather than as universals
(or their nominalist equivalents). But it is unclear how this might be supposed to
work. Note, however, that OShaughnessy clearly speaks of imaginings as representations of types of cognition although he acknowledges that the imagining
subject is not aware of the meta-representational aspect of imagining, but instead enjoys a mental episode that is, so to speak, transparent to the imagined external objects and their features (see OShaughnessy (2000): 364f.).
301
302
303
304
spect to intellectual imaginings. The first is that this endorsement can fulfil
a certain explanatory role: namely to explain the similarities between intellectual imaginings and beliefs, and yet to elucidate why and perhaps even
to entail that the former cannot instantiate types of the latter (e.g., the respective cognitive prototype). More specifically, one component of the idea
is that intellectual imaginings involve the same contents as beliefs simply
because they are representations of types of beliefs, and hence include the
contents of the latter as part of their own contents. And the other component of the idea is that this representational link to types of beliefs also illustrates how they differ from beliefs, and perhaps why they cannot provide
knowledge in the same way as beliefs.
The second plausible motivation for the endorsement of (ER) in the intellectual case is that it promises to shed more light on the relationship
between these imaginings and the corresponding judgements or beliefs.
That there is an intimate link between the two kinds of phenomena has
been expressed by means of the idea that the elucidation of the nature of
imaginings has to make essential and substantial reference to beliefs (see
Scruton (1974): 100; Currie & Ravenscroft (2002): 32; see also Section
12.1). For OShaughnessy, this is part of the truth of his version of (NC*):
that imaginings are, essentially and fundamentally, incapable to exemplify
the two cognitive prototypes of successful perception and intellectual
knowledge. He concludes from this that any account of imaginings has to
elucidate them in terms of the cognitive prototypes and, in particular, intellectual imaginings in terms of knowledge-constituting beliefs. But
simply saying that imaginings cannot realise the two forms of knowledge
does not say much about the relationship between intellectual imaginings.
OShaughnessy may therefore have felt the need to specify the connection
between imaginings and, especially, intellectual imaginings and the respective cognitive phenomena in more positive terms: namely by means of
(ER) or, more precisely, his particular version of (ER) (see Section 6.2). A
similar desire may have moved other philosophers to endorse their versions
of the representational echo thesis.
However, these suggestions cannot settle the issue in favour of the application of (ER) to intellectual imagining. On the one hand, endorsing this
representational echo thesis is not the only way to account for the similarit-
305
ies and differences between intellectual imaginings and beliefs. On the other hand, (ER) is not the exclusive option of providing a positive account of
what it means to be an intellectual imagining. Let me discuss each point in
turn.
First, the introduction of (ER) is not the only way to explain how intellectual imaginings are both similar to and different from beliefs (or judgements).21 For instance, the alternative picture of intellectual imaginings described above can account for the similarities by reference to the possibility that intellectual imaginings and judgements can involve the entertainment of the same propositions. And it can trace back their differences to
the presence or absence of a cognitive attitude to this content. In particular,
one consequence of this lack of cognitive attitude seems to be that intellectual imaginings cannot stand in the required rational relations to other beliefs (and also to perceptions). Imaginatively (or merely) entertaining propositions cannot provide rational support for the judgemental endorsement
of a proposition, even if the latter can be inferred from the former. If two
propositions, which we endorse, entail a third proposition, we will not imaginatively (or merely) entertain the implied proposition, but endorse it as
well. In accordance with this, the form of normativity or rationality pertaining to intellectual knowledge concerns reasons for endorsement, and
not reasons for entertainment.
This may, indeed, be a satisfactory explanation of why intellectual imaginings cannot constitute knowledge. It may also imply that being an intellectual state of knowledge requires the involvement of a cognitive attitude
which may explain why there do not seem to be any intellectual counterexamples to (NC). It does not, however, necessarily entail that states of
knowledge have to be of the same kind of mental state as false beliefs: disjunctivism may still be an option (see Dorsch (2011a), and Note 3 in
Chapter 8). Nor does it imply that sensory grounds of knowledge have to
involve a cognitive attitude. For such grounds do not have themselves to be
rationally supported. The idea here is just that reasons for belief are always
reasons for endorsement and not that reasons for belief can be provided
only by rational endorsements (see, for instance, Pollock & Cruz (1999)).
Second, a similar strategy can be adopted towards the claim that what
21
For a similar argumentative strategy with respect to sensory imaginings, see Hopkins (1998): ch. 7.
306
speaks in favour of the introduction of (ER) is that it can specify the nature
of intellectual imaginings in positive terms. For the alternative model may
identify other, equally suitable candidates for the feature which marks intellectual imaginings as imaginative (e.g., their involvement of an imaginative attitude; or their being active). In particular, even if it is true that a
satisfactory account of imaginings has to make reference to corresponding
types of cognitions, there is no need to assume that this reference has to
part of the positive side of such an account. Instead, it may pertain to its
negative side such as when (NC*) or a similar negation thesis is taken to
be central to any successful theory of imagining. Besides, it is not certain
that an account of imaginings has to refer to the respective kinds of cognitions. As the discussion above of OShaughnessys motivation to assume
the necessity of such a reference illustrates, this assumption may depend
largely on the controversial idea that (NC*) applies to all kinds of imagining, thus establishing an intimate link between imaginings and cognitions
which is in need of further elucidation. But OShaughnessys main motivation for endorsing his version of (NC*) namely his belief that there is no
intrinsic or relational feature common to all imaginings, in virtue of which
they count as imaginative, because there is too great a variety of origins
and constitutions among imaginings (see Sections 6.3 and 7.1) need not
be true, or at least is not obviously so. Indeed, I will argue in Chapter 13
that it is false.
It might be insisted that there are other differences between intellectual
imaginings and beliefs which can be accounted for solely in terms of (ER).
But the first worry that (ER) itself is unilluminating with respect to intellectual imaginings would still apply. Moreover, it is not obvious which
further differences could demand such an explanation. With respect to
sensory or affective imaginings, it may be argued that (ER) is necessary to
account for some of their phenomenologically salient features (e.g., how an
imagined itch can somehow feel itchy without really being an itch; or how
visual imaginings can possess a perspectival content; see Martin (2002a),
as well as Chapters 10f. and Section 14.3); or for the fact that we can be
immune to error through misidentification with respect to what is imagined
(see Peacocke (1985), and also Section 3.5). But in the case of intellectual
imaginings, there do not seem to be similar aspects in need of explanation.
307
Their phenomenal character is not very complex and comprises not much
more than their lack of an attitude (or the involvement of an imaginative
one) and, presumably, also their entertaining of a specific proposition (see
Soldati & Dorsch (2005) and Dorsch (2009b)). Similarly, our knowledge of
what intellectual imaginings represent can be accounted for in roughly the
same terms in which we explain our knowledge of what our beliefs (or
judgements) represent: namely by reference to the concepts and semantic
mechanisms which the first- and the higher-order representations involve
(see Peacocke (1985)). Hence, intellectual imaginings do not seem to involve any features whose explanation depends essentially on the acceptance of (ER).
However, what needs to be elucidated might not be some feature of intellectual imaginings, but rather their distinctness from other (non-sensory)
forms of imagining. The introduction of an echo claim with respect to sensory imaginings may be motivated by the idea that it can explain the differences between visual, auditory and other modes of sensory imagining (see
Peacocke (1985) for an endorsement, and Hopkins ((1998): ch. 7) for a rejection, of this motivation). In a similar vein, it might be thought that the
representational echo thesis with respect to intellectual imaginings has the
power to explain the differences between, say, imaginatively believing
something and imaginatively desiring it. The idea is that the two kinds of
imagining differ because they involve different types of propositional attitude as part of their contents: the first echoes belief, while the second
echoes desire. But at least two considerations speak against this view.
First, there seems to be an asymmetry between the two phenomena:
while it does not come natural to us to speak of imaginatively desiring
something (there is no obvious desire-like equivalent to supposing, assuming or imagining-that), it is commonly accepted that there are cases of
imaginatively believing something (suppositions, etc.). But this suggests
that the two phenomena should not be treated on a par. It is much more
straightforward to say that, while intellectually imagining something denotes a sui generis way of entertaining a proposition, imaginatively desiring something does not. Otherwise, people would have noticed the desirelike way of imaginatively entertaining a proposition and would have invented a name for it (see Soldati (2006)). Hence, this alternative picture seems
308
to provide a much more plausible explanation of how intellectual imaginings differ from cases of imaginatively desiring something (in whichever
other way they have to be characterised) than the idea that the two phenomena merely differ in what they are imaginings of.
Second, it is not easy to come up with cases of imaginatively desiring
something which cannot be analysed either as cases of intellectually imagining that one desires something (plus perhaps having certain closely related affective imaginings, such as imagining a feeling of wanting or longing), or as cases of experientially imagining (the feeling of) desiring something. Imaginatively desiring something if there is such a thing in the first
place (see Note 16 in Chapter 1) is therefore best understood as an instance of intellectual or experiential imagining. Consequently, the application of (ER) to intellectual imaginings lacks argumentative support.22
Moreover, as noted above, the fact that there is no satisfactory argument
in support of the claim that intellectual imaginings are imaginative representations of types of beliefs is not the only difficulty. There is also a consideration that speaks strongly against the application of (ER) to intellectual imaginings. This consideration concerns the fact that we can, and often
do, entertain propositions without either endorsing them or imagining (or
otherwise representing) them to be endorsed. We may consciously desire
that it will rain; we may hope that it will be sunny; we may worry that we
will be too late to reach the shop before it closes; or we may simply wonder whether Quito is the capital of Ecuador. None of these cases need (and
some perhaps cannot) involve a judgemental endorsement of the enter22
309
tained proposition.23 But it is also false to require that they have to involve
imagining an endorsement of the entertained proposition. At best, it may be
argued that they presuppose having some prior beliefs about the possibilities available (e.g., that it could rain or be sunny in the future). However, if
we can entertain propositions without either endorsing them or imagining
them to be endorsed, it seems arbitrary and unlikely that this should none
the less be impossible in the case of intellectual imaginings. In particular, it
is reasonable to assume that the simple active entertainment of a proposition without endorsing it is already an instance of imagining (see Section
14.2). As long as possibilities like this are not ruled out, (ER) should be rejected if intellectual imaginings are concerned.
The situation appears, again, to be different in the sensory case. The instances of the main non-imaginative kind of non-perceptual sensory representation namely sensory memory possess a cognitive attitude and may
very well turn out to have perceptions as part of their contents (see Section
9.4). Hence, there may be no visual representations that lack a cognitive attitude, and of which (ER) is not true. After-images seem to involve a cognitive attitude and, in this respect, to be perception-like: they involve a
claim about the presence of an experience-dependent patch of colour.
Spontaneously occurring or otherwise unbidden images, on the other hand,
are likely to be covered by (ER) at least if imaginative and mnemonic
images are. This just leaves the visual awareness of the depicted that forms
a necessary part of pictorial experiences which, too, should be taken to
be perceptual and to involve a cognitive attitude, given that it is not imaginative (see Hopkins (1998) Dorsch (2012c).
See Gordon (1987) on epistemic emotions (in contrast to factive ones), which
need or even can not involve any cognitive attitude. Merely entertained thoughts
may be another example, at least if they are not taken to be imaginative themselves
(see Section 14.2).
310
projects, given that only the episodic parts of projects but not the project
as a whole possess by themselves a phenomenal character. At best, imaginative projects may be said to consist of episodes (i.e., imaginative episodes) that represent types of cognitive episodes. But two things speak
against this proposal. First, it has to assume the truth of (ER) with respect
to intellectual imaginings which, as just argued, is highly problematic.
And, second, it has to presuppose either that imaginative projects consist
solely of imaginative episodes, or that cognitive projects do not involve
any imaginative episodes. For, otherwise, imaginative projects would not
differ from cognitive ones in respect of whether or in respect of the extent to which they involve the representation of cognitive episodes. But,
as already noted, imaginative projects may also include non-imaginative
episodes, while cognitive projects may also involve imaginative episodes.
For instance, when we daydream about being wealthy, we make use of our
beliefs about the world of the rich; and we may use imaginative thought
experiments in order to solve a theoretical problem (see also Sections 2.3
and 14.2, as well as Chapter 4).
Very similar considerations undermine the different suggestion that it is
distinctive of imaginative projects that they aim at, and lead to, the formation of episodes which are imaginative representations of types of perceptions, beliefs, and so on. This strategy has already been discussed as a way
of trying to avoid roughly the same problems in the case of negation
claims; and its application to echo claims faces substantially the same objections (see Section 8.1). First of all, given that it is unlikely that (ER) applies to intellectual imaginings, the suggestion leaves out imaginative projects that have the purpose to bring about solely intellectual imaginings
(e.g., the project of imaginatively drawing out the consequences of a certain counterfactual situation in a philosophical context). Then, imaginative
episodes and imaginative projects are assumed to count as imaginative for
very different reasons, hence undermining the unity of imagining not the
least because the imaginativeness of projects is said to be derivative of that
of episodes. Finally, the proposed characterisation of imaginative projects
remains extensionally inadequate. In particular, there are cognitive projects
which specifically aim at, and lead to, the formation of imaginative episodes as well. The two examples discussed in Chapter 4 are cognitive pro-
311
jects of this kind: they successfully aim at the formation of knowledgeconstituting visual imaginings. Accordingly, the suggestion under
consideration would have to give up the ambition to provide an account of
the imaginativeness common to all forms of imagining. (ER) should therefore not be assumed to form part of a unified theory of imagining.
CHAPTER 10
Visual Imagining As Experiential Imagining
That the Representational Account should not be accepted for some forms
of imagining notably intellectual imaginings and imaginative projects
does not mean that it should not be accepted for others. Indeed, in the following two chapters, I aim to argue in favour of the truth of (ER) when applied to sensory and emotional imaginings. Hence, even though the Representational Account does not single out the common nature shared by
episodic and non-episodic, as well as intellectual and non-intellectual, instances of imagining, it still succeeds in elucidating some important aspect
of the nature of sensory and emotional imaginings namely their distinctive representationality. This is, indeed, part of the reason why the Representational Account has been endorsed in the first place even though typically only with respect to visual imaginings (see, for instance, Peacocke
(1985), Martin (2001) and Martin (2002)). Now, while Chapter 11 is focused on emotional imaginings, I intend to use this chapter to show that
(ER) is true of at least some instances of visualising.
Although this does not suffice to establish the conclusion that (ER) is
true of all instances of sensory imagining (or even all instances of visualising), it constitutes an important step towards that conclusion. 24 My argument is, to a considerable extend, a defense of the argument to the same effect that is presented in M. G. F. Martins paper The Transparency of Ex24
314
perience.25 This argument has been often misunderstood (or ignored), and
it is worthwhile to discuss it in detail and to illustrate how it is immune to
the objections standardly raised against it. In addition, I present a second
and independent argument for the claim that visualising is imagining seeing, which is not part of Martins paper (and might not be endorsed by
him), but which none the less makes use of roughly the same example as
Martins argument. Both arguments show that it should better be accepted
that visualising is, at least sometimes, essentially imagining seeing.
The thesis to be defended in this chapter is, again, a more specific version
of both (ER) and the Dependency Thesis, this time restricted just to one
particular kind of sensory imagining, namely visualising.
Now, what we are aware of when imagining a perceptual experience is
just that, some experience. More specifically, we are aware of the first-personal side of an experience, that is, of its phenomenal character. We imagine some experience by imagining some instantiation of its character. Its
third-personal side (if experiences have any) is, so to speak, invisible to
object-directed imagining. Of course, we can have additional thoughts
about it. But it is not presented as part of a case of imagining with an exSee his (2002a). Here, I ignore the fact that Martins argument for the claim that
some instances of visualising are instances of imagining seeing forms part of a wider
argument against intentionalist (and for certain disjunctivist) theories of perception. I
discuss and reject this further application in Dorsch (2010c), from the middle
sections of which most of the material in this chapter is taken.
26
Note that Martin uses experience here as short for sense experience, such as perception or bodily sensation. My own use is less narrow in also including, say, visualising or other imaginative instances of object awareness.
25
315
perience as its object. The reason for this is that this form of imagining is
experiential, in the same sense in which seeing and visualising are visual.
Just as the latter are limited to the presentation of visible entities, experiential imagining is restricted to the presentation of phenomenal aspects of
mental episodes. The non-phenomenal (or structural) features of episodes
lack an experiential appearance, so to speak. Again, experiential awareness does not differ in this respect from, say, visual awareness. When we
see or visualise a lemon, we see or visualise its visual appearance, but not
its biological nature, for example.
Martin uses the Dependency Thesis to formulate an argument against intentionalism about perceptual experiences the view that veridical and hallucinatory experiences share the same intentional nature. Indeed, intentionalists typically select the Dependency Thesis as the main target of their criticism of Martins argument. This is partly explained by the fact that intentionalists treat imaginative experience as involving the same general kind
of intentional representationality as perceptual experience. For them, it is
therefore difficult to understand why, say, seeing and visualising should not
make us aware of the same objects, namely external things. If perception
does not involve an awareness of an experience, why should imagination
do so, if both are assumed to involve the same kind of intentional awareness? This doubt should be taken serious not the least because it simply
confirms that there is in fact some tension between intentionalism and
(ER). While Martin draws the conclusion that the former should be given
up, his opponents question the plausibility of the latter.
Some of the intentionalist objections, however, seem to have misunderstood the nature of Martins argument. What is crucial here is that it concentrates on, and exploits the special features of, cases in which our episodes of visualising involve certain subjective properties. Subjective properties are characterised by the fact that they are experience-dependent: their
instantiation is dependent on the occurrence of a specific experience. Martins focus is on instances of visualising which involve a certain kind of
perspectivalness. Accordingly, the conclusion of his argument should be
understood as being restricted to those cases (or to similar cases, such as
imaginative experiences involving aspects of painfulness or itchiness).
What it therefore claims is that visualising is identical with imagining per-
316
ceiving when it involves the subjective perspectival element at issue. Perhaps sensory imaginings are generally perspectival in this way. But, if not,
the argument fails to establish (ER) for all of them.
So, what is the kind of perspectivalness at issue? By perceiving an object, we may acquire knowledge about the latters specific spatial location.
But our perception does not thereby place the object in objective space.
When we look at a building that is located roughly to the South-East of the
bench on which we are currently sitting, we do not see it as being to the
South-East of that bench. In particular, we do not perceive objects as being
oriented in accordance with the cardinal directions. Instead, we see them as
being oriented towards ourselves for instance, we see the building as being to our left. What this means is that we perceive objects as part of
egocentric space, and not as part of objective (or absolute) space. 27 One
manifestation of this fact is that our perception of the building inclines and
entitles us to judge that it is to our actual left, but not that it is to the SouthEast of the bench. Coming to know the latter requires additional information notably about our own location and orientation in objective space
(see Campbell (1994): especially ch. 4).
None the less, our perceptions of egocentric locations are still as much
concerned with actual space as is our knowledge of objective locations. We
see the building as being to our actual left, as part of our actual environment. If this were not so, our experience would not be able to provide us
with all the information necessary to properly interact with what we see
for example, to succeed in walking over to the actual building. But that
perception does provide us with this information is illustrated by the fact
that such interaction does not require inferring the presence of the building
to our actual left from perceiving it as being to our left and believing that
27
It does not really matter for Martins main argument whether we are concerned here
with two different sets of spatial properties of objects (e.g., one objective, and the
other subjective), or instead with two modes of presentation of one and the same set.
What is relevant here is primarily the fact that our perceptual access to spatially located objects is perspectival and, in particular, presents them as oriented towards us,
rather than in more objective terms. But many of the points involved in the argument
can be described more easily by reference to egocentric properties. Besides, the postulation of subjective orientations is not much different from the postulation of subjective modes of presentation (see the similar issue raised below with respect to the
aspect of leftishness and similar phenomenal aspects).
317
such experiences are (typically) concerned with the actual world. The issue
of which world our perceptions are concerned with simply does not arise
it is our world, the world in which we perceive. Something similar is true
of the temporal relevance of our perceptions: they concern our present environment. We see the building as presently being to our left, and not as
having been there in the past, or as going to be there in the future.28
Part of our perception of the building as being to our actual left is implicit, however. We do not explicitly experience ourselves and our spatial relation to the building when perceiving the latter. We are not among the entities presented to us by our experience. Of course, we can see other perceivers and even ourselves, say, by utilising a mirror or some similar apparatus which turns us into the object of our own perceptions. But normally, when we are simply subjects of perception and perceive the orientation of objects relative to us, we do not see us, but only the objects. Our
own perspective is only implicitly reflected in our perceptual experiences,
namely as the point of view oriented to which objects are presented to us.
As a consequence, what figures explicitly in our experience is not the relational property of being to the left of us, but the monadic quality of being
leftish.
It seems that such a quality can figure in perception in two different
ways. The perceptual experience may instantiate the quality; or it may instead present an external object as having that quality. In both cases, this
has consequences for the phenomenal character of the experience concerned. In the first case, the quality constitutes one of the non-presentational aspects of that character. In the second case, it is a constituent of one
of the characters presentational aspects.29 That the quality of leftishness
figures in our perception of the building therefore means that the latter instantiates a certain character aspect either a non-presentational aspect, or
the presentational aspect of presenting the building as being the monadic
property of being to the left.30
In what follows, I concentrate on the fact that perceptions present objects in actual
space and mention the temporal dimension only when it becomes relevant.
29
It should become clear very shortly that there is a third possibility: the experience
may represent another experience as instantiating or presenting the quality.
30
Using the expression being to the left to denote a monadic property is not ideal,
since this expression clearly has some connotations of relationality. But it is not easy
28
318
Which view is to be preferred in the end does not matter here. Indeed, it
is not so clear whether they actually differ in any substantial way which
might explain why Martin appears to switch between both views in some
of his formulations (e.g., when talking about the quality of itchiness). The
step from acknowledging the presence of a non-presentational aspect of the
character of a perception to projecting this aspect onto the perceived object
is indeed small as discussions about blur or similar phenomena illustrate
(see, for instance, Peacocke (1983), as well as Soldati & Dorsch (2011)).
Moreover, any presentation of something as being leftish would lack the
status and force of the presentation of it as being to our actual and present
left. In particular, we do not see the building as having the monadic property of being to the left; and we are not inclined or entitled to believe it to
genuinely instantiate this property. Of course, we may say the building is
to the left. But when prompted, we will happily clarify that what we really
meant was that it is to the left of us.
In any case, that the character of our perception of the building include
this phenomenal aspect let us call it the aspect of leftishness should not
be doubted. We can attend to it; and we can exploit it when drawing a picture of how the building looks like when seen from our current point of
view. That is, we can depict an object as being to our actual left by drawing
it on the left side of the canvas instead of, say, by drawing both ourselves
and the object.31
But how is the instantiation of the aspect of leftishness linked to the perception of the property of something as being at some specific location to
our actual left? More generally, how does the perspectivalness of an
to come up with another formulation, without altogether loosing the connection to
the perceived property of being to our actual left. I am grateful to one of the referees
for making me aware of this issue.
31
Very similar issues arise, for instance, with respect to the status of the quality of
ovalness another perspectival aspect of perception which figures in our experience when we are looking at objects from an angle and perceive them as round.
Again, we typically draw round objects by tracing elliptical shapes on the canvas.
But it is debatable whether our experiences present round objects as elliptical in addition to presenting them as round (see, for instance, Peacockes discussion (1983)
of what he calls sensational properties). One significant difference from egocentric
orientation is, however, that, while roundness is an objective property, being to the
left of us is not.
319
experience relate to the determination of what is experienced? Martins insightful observation is that the former suffices for the latter (see Martin
(2002a): 410). If an experience of an object exemplifies leftishness that
is, shows a respective non-presentational phenomenal aspect or, alternatively, presents the object as being to the left then it is an experience of
the object as being to our actual left. More specifically, the presence of the
perspectival aspect of leftishness is sufficient to ensure, first, that the experience concerned is an experience of something as being to our left
(rather than to our right) and, second, that it is an experience of something
as being to our actual left (rather than to our left in a merely possible situation). Indeed, Martin claims even more, namely that it also suffices for
having a perceptual experience of something as being to our actual left.
This makes sense since the other two kinds of visual experience, which
may involve the aspect of leftishness, are not or at least not in their
simplest forms concerned with our current environment. Episodes of
visualising present objects as part of imagined situations (see below), while
episodes of visual memory present objects as part of past situations. I return below to the issue of how important this additional claim is for Martins argument.
That the instantiation of the aspect of leftishness turns the respective experience into an experience of something to our actual left is a direct consequence of the implicitness involved in our perception of the spatial relations that objects bear to us in egocentric space. As noted above, we see
objects as being to our actual left (and not, say, as being at an egocentric
location in some merely possible space). But this relational property is typically not explicitly given to us. Instead, what figures in our experience is
solely the monadic quality of leftishness. Hence, we perceive the instantiation of the property of being to our actual left simply by being aware of
the quality of leftishness. When we see the building as being to our actual
left, no aspect of our perception but its aspect of leftishness plays a role in
determining that we experience the building at that specific location in our
actual environment. If the aspect of leftishness is taken to be presentational, this thought becomes even more straightforward: our perception
presents the building as being to our actual left just by presenting it as being to the left; no other presentational element is needed or involved. What
320
321
But even then, there is no real competition between what we see and what
we visualise. For example, when looking at a certain picture on our kitchen
wall, we may visualise with open eyes how things would look if there were
a different picture at the same spot on the wall. But such a complex and
mixed presentation does not amount to a presentation of the impossible
state of affairs of two pictures occupying the same part of space.
So, episodes of visualising may involve the aspect of leftishness without
presenting something as being to our actual left. But due to the subjectivity
of the aspect of leftishness noted above, its instantiation is inseparably
linked to the presentation of something as possessing the relational property of being to our actual left. Hence, the instances of visualising concerned cannot exemplify the aspect of leftishness. This raises the question
of how it is involved in visualising instead. Martins proposal is that, in
visualising, we imagine an experience as instantiating the aspect of leftishness that is, we imagine a perspectival experience of something as being
to the left in the imagined situation. When we visualise a building as being
to the left, our imaginative episode does not instantiate the aspect of leftishness. But it still involves the latter by representing another experience as
instantiating it.
The proposal captures the specific subjectivity of the aspect of leftishness. For it takes the instantiation of that aspect in a certain world to be
sufficient for the occurrence of an experience of something as being located to the experiencing subjects left in that very same world. Actual perspectival experiences concern actual space, while imagined perspectival
experiences concern imagined space. Moreover, what needs to be imagined
is a perceptual experience. As noted above, other perspectival experiences
are not concerned with the current state of the world in which they themselves occur. Instead, they are concerned with the past of that world (as in
the case of visual recall), or with an entirely different possible world (as in
the case of visualising). Hence, neither episodic memories, nor imaginative
episodes can instantiate the subjective aspect of leftishness. If we want to
imagine an experience with that aspect, we therefore have to imagine a
perspectival perception. This conclusion can also be inferred more directly
from Martins additional claim mentioned above, namely that the presence
of leftishness suffices for the presence of perception. Indeed, the reasoning
322
put forward in support of that claim has been very similar to the one rehearsed in the second half of this paragraph. But, strictly speaking, the additional claim does not seem to be necessary for Martins argument.
Martin further illustrates this argument by comparing the subjective perspectivalness of perceptions to the subjective aspects involved in some
bodily sensations. His example are experiences of itchiness; but experiences of pain are equally good candidates (see Sections 11.5 and 14.3). 34
Experiences of itch instantiate the phenomenal aspect of itchiness: they are
sensations of itch. Moreover, having a sensation of itch is sufficient for
there actually being an itch and, hence, for experiencing an actual itch. If
we sense an itch at a certain location on our skin, then that part of our skin
does indeed itch independently of whether there is any skin irritation,
say. By contrast, merely imagining our skin as itching does not involve the
presentation of an actual itch. In particular, it does not induce us to scratch
the respective part of our skin. But this raises, again, the question of how
imagining an itch can still involve the phenomenal aspect of itchiness
which it clearly does, albeit possibly to a lesser degree of intensity and determinacy than real feelings of itch. For, otherwise, imagined itches would
not phenomenally resemble genuine itches and, moreover, be classified as
imagined itches in the first place. As above, the solution is to understand
imagining an itch as imagining a sensation of itch that is, as imagining an
experience which instantiates the phenomenal aspect of itchiness. This
concludes what are, in essence, Martins considerations in favour of the
truth of (ER) limited to (some) instances of visualising.
Nothing here depends on whether the presented view on itches and sensations of itch
(or the comparable view on pains and pain experiences) is correct. The analogy is
merely meant to further illustrate Martins treatment of the involvement of subjective elements in imaginative experience.
323
for the latter. Our skin does not really itch if we do not feel itch. Of course,
other things may distract us so that we do not always notice the itch. But if
we do not feel any itch at some location on our skin, despite being sufficiently attentive to the latter, it does not seem true to say that that part of
the skin in fact itches. In particular, by-standers cannot insist that we sense
an itch by pointing to some irritation of our skin. Such evidence cannot
trump our failure to feel an itch. Accordingly, the presence of an itch requires an experience of that itch and, presumably, as part of the same (actual) world. Hence, imagining an itch has to involve imagining sensing that
itch.
Now, egocentric orientational properties seem to be similarly subjective
opening up the possibility of formulating a similar argument in favour of
the idea that visualising means imagining seeing. Martin does not discuss
this second route to the conclusion; and it is not clear whether he would accept the subjectivity of egocentric orientations, or the argument exploiting
it. But even if not, it is still worthwhile to discuss both. When we see a
building as being to our left, it does not possess this orientation independently of being perceived by us as having it. Certainly, the objective location
of the building comes with the disposition of giving rise to a perception of
leftishness when viewed from a position to its North-West by a normal human being with a normal orientation in objective space (e.g., standing on
his feet, etc.) who faces South. But its perceived property of being to our
left cannot simply be reduced to this objective disposition. Instead, the instantiation of this egocentric orientation seems to depend on our actual perceptual awareness of it.
For one thing, which dispositional property is correlated to the property
of being to the left of us varies with changes in our location in objective
space. Once we being to move or turn around, the building may very well
cease to be to our left though it may also begin to be to the left of another
person who steps in and takes our previous spot. The disposition may
therefore constitute the property of being to the left of whoever occupies
the objectively specified location to its North-West with the respective objectively specified orientation. But it does not amount to the property of
being to the left of us (understood in first-personal terms). This is reflected
in the more general fact that egocentric space cannot be fully specified in
324
objective terms which is why the two are to be distinguished in the first
place. In particular, what we describe with the expression to our left is not
a cardinal direction in objective space (see Campbell (1994): especially ch.
4).
Without this lack of strict correlation between egocentric and objective
spatial features, it would also seem impossible to explain why we cannot
suffer an illusion with respect to perceiving something as being to our actual left. Of course, when facing South, we may perceive a building as being to our actual left while, in fact, it is located to the South-West of our
current location in objective space. But, as the previous considerations
have indicated, the objective orientational properties of the building are
neither sufficient, nor necessary for its instantiation of any subjective orientational property. What happens in cases like this is just that we fail to
track the former by perceiving the latter an error which is due to some
breakdown in our relation to our environment.35
But the subjectivity of egocentric orientations has consequences for our
attempts to visualise objects as having them. Objects can possess these
subjective features only when they are perceived as having them. Furthermore, this is true as much of imagined or other possible situations, as it is
true of actual ones assuming that they all contain the same ontological
kind of objects and properties. Finally, the dependency in question does not
range over different possible worlds, but is confined to a single one: the
perceived object and the perception are always part of the same world.
Hence, visualising an object as having such an experience-dependent property requires imagining a suitable perception of that object. Visualising a
35
The property of being to our left shares both discussed aspects with the property of
being here if ascribed to ourselves. The latter, too, cannot be specified in purely objective terms. And we cannot go wrong in being aware of ourselves as being here.
One difference between our awareness of us as being here and our awareness of other objects as being to our left is, however, that the latter may concern hallucinated
objects. This is not problematic for the present argument since it relies only on the
claim that the instantiation of egocentric orientations requires the occurrence of a
perception of them, but not on the reverse claim. Besides, as already implied in Note
9 above, there are perhaps other ways of reconciling the subjectivity of a property
with the possibility of hallucinating an instance of it for instance, if the instantiation of the features concerned is understood as depending on the mind of some imagined subject, rather than on the mind of the imagining subject.
325
building as being to the left, for instance, has to involve imagining a perception of a building as being to the left.
But it is not clear whether Burges charge really is one of begging the question. For he still seems to briefly discuss and dismiss Martins argument involving the subjective perspectivalness of our presentation of orientational features. Here is how the first part of the passage just quoted
continues:
It is quite true that one could have such a perspective on the object only if
one were to have an experience of the object. It does not follow that if one
imagines something from a perspective that one could have only if such and
such were the case (only if one were experiencing the object from that perspective), then in imagining something from that perspective one must imagine such and such to be the case. (Burge (2005): 63)
326
Currie and Ravenscroft (2002, 28) also do not address the issue of subjectivity when
they briefly sketch Martins motivation for endorsing (ER) with respect to visualising.
327
that we find this thesis which is identical with the Dependency Thesis
more plausible in some cases than in others:
As I have already noted, the Dependency Thesis varies in plausibility depending upon the sensory modality we consider. It is more plausible when
we consider what is involved in imagining the feel of somebodys skin or
the taste of bacon. Our capacity to imagine these things seems to rest upon
our capacity to imagine our experiences of these things. The proponent of
the Dependency Thesis needs to explain why it is more plausible in these
cases given that the Dependency Thesis holds across the board. My guess is
that the proponent of the Dependency Thesis might try to argue that the
variation in plausibility depends upon contingent psychological facts about
what we find easier to consider independent of experience. The issue is
whether we should search for an explanation there rather than in the objects
and properties imagined. If the feel of someones skin or the taste of bacon
imply the existence of corresponding perceptual experiences in contrast
with other objects of imagination, then the Dependency Thesis cannot be
true for all sensory imaginings. (Noordhof (2002): 446)
328
depicted by a painting. The marks on the surface (plus perhaps our general
recognitional abilities) determine whether the painting depicts a blonde
man or a brunette woman. But assuming that it is a blonde man, extrapictorial factors such as the stipulation of a title or the exploitation of
iconographic conventions decide whether it is a portrait of, say, Saint
John or Hercules. The pictorial element puts certain constraints on the extra-pictorial take on the nature of the depicted entities, which cannot be
overridden by the latter. An artist may turn his painting of a man into a portrait of Jean of Arc by labelling it as such. But it will then be a painting of
Jean of Arc in the disguise of a man.
Noordhofs observation is that very similar factors are responsible for
determining the nature of the objects of visualising. On the one hand, there
is the basic visual presentation and, on the other hand, the accompanying
intentions or suppositions which put a certain conceptual gloss on that
visual presentation. To take an example from Peacocke (1985), when we
are visualising a suitcase, we may think of it as a suitcase with a cat hidden
behind it, or as a suitcase which is merely hallucinated by some brain in a
vat. The visual presentation is limited to the presentation of the suitcase,
while the wider imaginative project concerns also other aspects of the imagined situation. And again, how we conceive of the visualised entities as
part of such a project is constrained by how these entities are visually given to us. We may use the visual presentation of the suitcase in order to imagine a car, but only by imagining a car with the visual appearance of a suitcase.37
According to Noordhof, this constraint on what is imagined as part of
the wider imaginative project is problematic for the application of (ER) to
visual imaginings. His starting point is the idea that it is possible to pursue
the project of imagining a certain object, and nothing else. The subject engaged in this project conceives of it in terms of the object, but not in terms
of any experience of that object. Hence, if the project is also taken to involve imagining an experience of the object, this cannot be due to any of
the accompanying intentions or suppositions of the subject. So a proponent
of (ER) with respect to visualising has to assume that the imagining of the
experience is part of the visual presentation of the object. But this seems to
37
329
violate the constraint of the visual presentation on the extra-visual interpretation of it. For we cannot pursue the project of imagining nothing but
an object by means of an episode of visualising which involves imagining
more than that object. The only way out appears to be to consider the extravisual gloss on the visual presentation to be irrelevant or, even worse, misleading:
When we consider what characterises an imaginative project, it is clear that
there are cases where the project is to imagine merely an F. Proponents of
the Dependency Thesis dont have to resist this but, if they dont, they must
insist that certain facts about the mental image override a subjects own
characterisation of his or her imaginative project so that, in fact, what is
imagined is a perceptual experience of an F. This is quite a strong claim to
have to establish. (Noordhof (2002): 430)
330
Again, the problems start once we focus on visualising orientational features that locate objects in egocentric space. Their imagination requires the
imagination of a perceptual perspective onto them. That it is still natural
for us to say that what we are imagining is just the object and its orientation may suggest that we either make use of an elliptical characterisation,
or have no full grasp of the nature of our episode of imagining and, in particular, of the subjectivity involved in the perspectival presentation of egocentric orientations. Indeed, it is to be expected that not all of our conceptions of subjective properties characterise them as subjective. For example,
we can discover that the phenomenal aspect of leftishness and the property
of being to our actual left are in fact experience-dependent. So we should
be able to conceive of those properties without conceiving of them as subjective. But this means that we can engage in imagining without fully
grasping what we are thereby engaging in. In particular, we may imagine
an instance of a specific egocentric orientation without realising that we
are thereby imagining a perception of it.
This helps to resolve another of Noordhofs worries. Granting his opponent, for the sake of argument, that imagining perceiving something constitutes sometimes the most basic episode of imagining available to us, he
still insists that it has to involve an explicit or implicit thought which
conceives of what one is imagining in terms of an experience. But that
visualising is therefore supposed to require the possession of the concept of
an experience casts doubt on whether, say, children under the age of four
can visualise something, given that they may lack the notion of an experi38
See also the similar point against the Dependency Thesis made by Currie &
Ravenscroft (2002): 27f..
331
ence.
If the Dependency Thesis rests on the claim that imaginers, at least tacitly,
suppose that they are imagining a perceptual experience, then it links the
capacity to imagine with possession of the concept of perceptual experience. In which case, the attribution of imaginings to the autistic, young children, and animals, becomes as doubtful as their possession of the concept of
perceptual experience. (Noordhof (2002): 436)
While acknowledging that the evidence seems undecided on this last issue,
Noordhof is right in pointing out that this connection to the empirical question of when children acquire the concept of experience threatens to undermine the Dependency Thesis, and thus also (ER) applied to visual imaginings. For not only does it seem empirically far less controversial whether
young children can visualise something than whether they can conceive of
perceptions. But the issue of when children become able to conceive of experiences should not be expected to have such a strong bearing on the issue
addressed by (ER) namely the constitution (if not the concept) of imaginative experience.
These problems related to an over-intellectualisation of visualising arise,
however, only if (ER) indeed implies that imagining perceiving something
requires the possession or application of the concept of perception. Noordhof assumes that it does. But the proponents of (ER) need not and should
not follow him in this. As already noted, we need not fully grasp the
nature of our imagining when being engaged in it. And this includes the
fact that it may involve imagining an experience. But, more important, this
partial ignorance is possible because we imagine an experience, not by
thinking of it as part of the imagined situation, but by experientially imagining the instantiation of its character. That is, imagining an experience is
a form of object awareness, and not of thinking. What (ER) claims about
visual imagining is that the visual presentation involved in relevant cases
of visualising consists in the experiential presentation of the character of a
visual perception. And this kind of object awareness does not involve the
employment of the concept of a visual perception. At best, it relies on some
discriminatory or recognitional capacities with respect to experiences, as
well as perhaps some basic demonstrative ways of referring to the respect-
332
333
Noordhof is absolutely right about the last point. But the proposed kind of
representation is perhaps less mysterious than it might seem to him and
others. Consider the reproduction of a painting for instance, a postcard
hanging at your wall. This image does not itself amount to a painting and
differs substantially from one (e.g., it does not involve paint and has no
perceivable texture). But it none the less inherits important aspects of the
reproduced painting. Most of all, it depicts the same objects and features,
and from the same perspective, as the painting. Indeed, if the reproduction
is done well, its perspectivalness derives from that of the reproduced painting, and not from the perspectivalness of the photographic process involved in the reproduction. That is, the impact of the point of view occupied by the lens directed at the painting is typically negligible in comparison to the impact of the point of view inherent to the photographed depiction.39 Much more can surely be said about how the reproduction does end
up presenting the same situation from the same perspective as the painting.
But the absence of such further elucidations does not render the kind of
representation involved in photographic reproduction mysterious or completely unilluminating. We accept that this kind of representation exists.
And the description given above gives us some grasp of what it amounts
to. In fact, we know at least that the reproduction represents the painting
partly by representing the visual perspective of the painting; and that it represents the latters perspective by presenting the same objects and features
as they are presented to the point of view of the painting.
Imagining a perception involves the same kind of representation. It represents a perception partly by representing the latters perspective. And it
does this by presenting the same external objects and features as they are
presented to the point of view of such a perception. In imagining a perception, we thus imagine a possible perceptual perspective onto the world. 40
And, as in the case of the reproduction of a painting, the resulting episode
of visualising ends up with a character very similar in its visual, perThe process of photocopying, which does not involve any such perspective onto the
reproduced piece of paper, is perhaps an even better illustration of the kind of representation pertaining to experiential imagination (as proposed by Martin in a personal discussion about how best to understand Humes copy principle).
40
This idea is not new. In particular, Humes copy principle may be read as claiming
pretty much the same if applied to the case of imagining.
39
334
335
One part of the answer to this challenge is to stress that, for a proponent of
(ER), there is contrary to what Currie and Ravenscroft suggest a substantial overlap in content between seeing and visualising. For imagining a
perception of an external object involves the visual presentation of that object as part of the imagined situation just as the reproduction depicts
whatever is depicted by the reproduced painting. In other words, experiential imagination has two objects: the imagined perceptual experience and
the external thing presented by that experience (and visualised as part of
the instance of experiential imagination concerned).42 The other part of the
reply to Curries and Ravenscrofts challenge is that, as already illustrated
by reference to the analogy with the reproduction of paintings, the
presentation of external things involved in imaginatively adopting the subjective perspective of a perception shares many important aspects with the
presentation of those things involved in perceiving them from such a perspective. Accordingly, seeing and imagining seeing resemble each other,
not only in what they make us aware of, but also in how they present it to
us. This is indeed best explained by reference to the truth of (ER) when applied to cases of visualising.
42
See Dorsch (2010c) for more on how both certain disjunctivist and certain intentionalist approaches to perception can accommodate the fact that imagining seeing involves two objects of awareness.
CHAPTER 11
Emotional Imagining
As Experiential Imagining
Discussions about imagining and, in particular, the truth of (ER) normally concentrate on the imaginative counterparts of perception and judgemental thought (or occurrent belief). Other forms of imagining such as
daydreaming, or the imaginative counterparts of bodily sensations and
episodes of emotion or desire are less often considered.43 In this chapter, I
aim to develop an account of emotional imagining as a specific instance of
experiential imagining. That is, I intend to defend the view that (ER) is true
of the imaginative counterparts of emotional feelings: emotional imagining
is essentially imagining emotional feeling. Accordingly, emotional imagining consists in non-propositionally (or non-intellectually) imagining the instantiation of the phenomenal character of an episode of emotion.
I motivate the proposed account in response to the theories of Kendall
Walton and Richard Moran. Waltons view stays too unspecific about the
nature of emotional imagining when it matters; and my own account may
be understood as supplementing Waltons by rendering it more specific.
Morans theory, on the other hand, is in conflict with both my own view
and that of Waltons; and I argue that it should be given up in favour of the
latter. Although, strictly speaking, not all affective imaginings are emotional (e.g., conative imaginings or imagined bodily sensations), it seems
fair to assume that many if not all of them possess the same nature as,
43
As already mentioned in the Notes 3f. in the Introduction, Scruton (1974), White
(1990), OShaughnessy (2002), Currie & Ravenscroft (2002) and McGinn (2004)
are recent examples of this kind of limited focus.
338
say, imagining feeling jealous or sad (see also Sections 11.5 and 14.3 on
imagining feeling pain). Hence, my discussion of emotional imaginings is
meant to apply to other types of affective imagining as well.
Both Walton and Moran discuss the connection between imagination
and emotion in the context of our responses to representational media.
Both pictures and texts, and possibly also pieces of music and other artefacts, portray fictional or in the case of didactic stories or thought experiments hypothetical worlds. One thing that is particularly interesting
about our engagement with such works is that it need not be concerned
with real persons, situations or events to help us to acquire knowledge
about reality. Reading about the adventures of a fictional character, being
confronted with a potential dilemma or envisaging a new possibility may
enable us to gain theoretical or practical insights into the actual nature of
ourselves and of aspects of the world. In the aesthetic and the moral cases,
these types of engagement with representations of fictional or hypothetical
worlds and the resulting instances of knowledge acquisition are often accompanied or facilitated by emotional responses. Some of these responses
constitute episodes of real emotion, while others amount to instances of the
affective imagination. The latter are therefore relevant for both aesthetics
and ethics.44 Although the subsequent considerations are focused exclusively on our aesthetic engagement with representations of fictional worlds,
they should equally apply to our moral assessment of hypothetical situations.45
The consideration of hypothetical scenarios is also central to science and theoretical
philosophy and, to some extent, also to theology and religion. We invent or use stories and models, say, when attempting to make sense of the structure of atoms, or the
nature of the universe. In the case of faith, this is likely to involve emotional episodes as well; in the case of scientific or metaphysical investigation, on the other
hand, probably less so.
45
Indeed, the affective imagination may be relevant for ethics in several respects. First,
as just described, we may come to determine what is morally required of us (or
someone else) to do in a given situation by imagining performing the different available actions and considering our emotional responses to those instances of imagining assuming here that emotional responses provide some indication of the presence of values (but see Dorsch (2007) for the restriction of this idea to subjective
values). Second, in order to assess (or even understand) the actions of another person, it may be necessary to empathise with them which, again, may require imagin44
339
340
341
See Walton (1990) for a discussion of the link between imagination and fictionality.
342
both types of responses belong to the same kind (i.e., the kind of
emotions). Their disagreement is related to the issue of whether being a
genuine emotion is compatible with being imagination-based in the sense
described. Accordingly, while both sides accept that our engagement with
fiction is imagination-based and involves affective responses, they are at
odds with respect to whether these imagination-based responses constitute
real emotions, or merely emotion-like experiences. Here is a nice quote
from Walton describing an example of the central element of the kind of
reaction at issue (denying, in the process, that it constitutes a genuine emotion directed at fictional entities):
Charles is watching a horror movie about a terrible green slime. He cringes
in his seat as the slime oozes slowly but relentlessly over the earth, destroying everything in its path. Soon a greasy head emerges from the undulating
mass, and two beady eyes fix on the camera. The slime, picking up speed,
oozes on a new course straight toward the viewers. Charles emits a shriek
and clutches desperately at his chair. [...] Charless condition is similar in
certain obvious respects to that of a person frightened of a pending realworld disaster. His muscles are tensed, he clutches his chair, his pulse
quickens, his adrenaline flows. Let us call this physiological -psychological
state quasi-fear. [...] Afterwards, still shaken, he confesses that he was terrified of the slime. (Walton (1990): 196)
Quasi-emotions like these are real episodes in our mental lives. Moreover,
they are affective episodes which is reflected in the fact that they resemble
genuine emotions (e.g., fear) in various significant respects. First of all,
they are similar to genuine emotions in their involvement of actual emotional feelings and emotion-related physiological events. Charless experience is, from his subjective point of view, very similar to an experience of
genuine fear notably in that it involves a similar kind of unpleasantness
and makes him aware of similar bodily changes. 50 But quasi-emotions resemble genuine emotions also in being triggered by the same mechanisms.
Both types of emotional response are partly due to dispositions to react af50
343
51
See the essays in Hjort and Laver (1997), and especially Levinson (1997). Walton
allows for the possibility of emotions which do not presuppose any kind of belief,
but still stresses we cannot have emotions towards something that is merely imagined (see Walton (1990): 245). The puzzle may also be formulated in two other
ways, namely in terms of belief in certain relevant propositions, or in terms of belief
in the existence of the entities in question. The first of these alternative option is
more specific than the other two (i.e., it implies, but is not implied, by the latter) and
is not required in all cases. Fearing something may presuppose believing it to be
dangerous, but hoping for something does not presuppose any belief (in particular,
we may be agnostic about whether the hoped-for state of affairs has already been
realised). The second alternative of spelling out the puzzle can allow for emotions
directed at the past or the future only if existence is understood as denoting reality,
that is, actuality. Furthermore, emotions may be said to be dependent, not on what
we believe to be real or existent, but what in fact is real or existent (see externalism
about thought contents).
344
345
nitive element present in emotion: they are concerned with how reality is
like and therefore require a specific take on the latter (see, e.g., Walton
(1990)). Perhaps proponents of this answer are also prepared or forced
to accept the additional claim that we are in some sense wrong or irrational
to treat our emotion-like responses to fictional entities as if they were
genuine emotions.
The second option is to reject (b). This may be done for several reasons.
For instance, it may be claimed that imagining the reality or existence of
objects rather than believing in it may already suffice for having emotions towards those objects (see Moran (1994)). Or it may be assumed that
there are two fundamentally different kinds of emotion, one exclusively
directed at real entities and the other exclusively at fictional ones; and that
(b) applies only to the first, but not the second kind. Or, finally, it may be
maintained that (b) does not express a constitutive dependency, but merely
a rational requirement, so that it is still possible albeit irrational to have
emotions without belief (see Radford (1975)). Again, the denial of (b) may
(have to) be accompanied by the postulation of a systematic form of irrationality inherent to our responses to fiction, in this case concerning the
fact that it is in some sense unreasonable to feel emotions towards entities
which we do not take to be real. At least, to respond with fear seems to be
more suitable when one perceives or believes a lion to be in the room than
when one merely visualises or imagines it to be there just as there is
something irrational about fearing real spiders which one takes to be completely harmless (see Goldie (2000) and Dorsch (2007)).
This is not the place to settle the debate between the two approaches.
Our concern is with the nature of affective imagining, and not with the
nature of our emotional responses towards representational artworks and
the fictional worlds that they portray. But since both Walton and Moran
the first of whom rejects (a), and the second (b) develop their views on
affective imagining in the context of this debate, it is worthwhile to a look
a bit closer on their views on how we react emotionally to representational
art.
346
At least in this case; Walton allows that, in other cases, something else might be
missing (see Walton (1990): 245.
347
is true within the respective fictional world. One of the central ideas of
Waltons account of representational artworks is that the fictional truths related to such works are not only concerned with the represented objects
and their features, but also with us and our access to those objects and features. The worlds of paintings and novels include landscapes and battles.
But they also include our fictional perspectives on those landscapes and
battles for instance, our seeing or thinking about them. According to
Walton, this is part of our conventional rules of engagement with representational art.
Second, and on the basis of the first experience, we intellectually imagine whatever is part of the represented fictional world. Indeed, the demand to imagine the fictional truths created by a representational artwork
in accordance with the relevant conventions is an integral part of our engagement with such art. Walton is adamant that, without this element, our
experience would not really amount to an experience of something as representational art. But imagining the fictional truths in question requires that
we recognise the representational content of the work. 53 We thereby exploit
explicitly or implicitly known principles or conventions which link the perceived material configurations to the represented entities. This enables us
to recognise words and their meanings, or the three-dimensional arrangement of objects in a scenery, by perceiving the specific nature of the marks
on the respective surfaces.
And, third, we imagine, again on the basis of the preceding elements of
our engagement with the work, having a certain epistemic access or
standing in a certain epistemic relation to the imagined world of the
work. For instance, we may imagine seeing the landscape depicted by the
painting before us, or believing the propositions expressed by the sentences
in the novel. Again, this is an essential part of our imaginative engagement
with representational artworks and required by the fictional truths determined by the latter.
Since, for Walton, affective imagining works very similar to pictorial
53
348
It is in this sense that we or, more precisely, our subjective perspectives enter
the fictional world. This fits well with Waltons characterisation of imagining from
the inside as one (but not the only) form of imagining de se.
349
right that, if pictorial experience is indeed partly imaginative, the imaginative element has to be integrated with the perceptual one in a single visual
experience.55 What matters here is the twofoldness of pictorial experience.
Seeing something as a picture of something else involves two instances of
object awareness: our awareness of the picture and our awareness of what
is depicted. And although they are distinct, they are also inseparable from
each other. We can, at least to some extent, shift our attention from one object to the other. But we cannot stop being aware of one of them without
ceasing to have a pictorial experience. Moreover, we are aware of both
objects as part of a single and unified experience.56
According to Walton, our affective imaginative engagement with fiction
is similar to pictorial experience in that it involves the same kind of elements as the latter.
First, we experience some quasi-emotion concerned with some aspects
of what the respective work represents. This real emotional reaction is
thereby triggered by our more basic non-emotional and imaginative engagement with the work for instance, our recognition of the portrayal of
an approaching slime or lion, and our imagination of the danger posed by
the latter. For Walton, the occurrence of quasi-emotions brought about in
this way in conjunction with the conventions of our engagement with
representational art makes it fictional that we feel the respective genuine
emotion towards the fictional entities concerned. The quasi-fear triggered
by imagining a dangerous lion approaching does not amount to real fear of
the fictional lion. But it determines that it is fictionally the case that we are
frightened of that lion.
Second, in response to the general demand to imagine what is part of the
fictional world related to some representational artwork, we intellectually
imagine that we have a certain genuine emotion namely that corresponding to the quasi-emotion towards the fictional entities in question. In our
example, we imagine that we fear the approaching lion.
And, third, we imagine some corresponding form of access to the fictional world, this time an affective kind of access. More precisely, we imaSee Hopkins (1998) and Dorsch (2012c). This also explains why OShaughnessy
(2000) talks about imaginative perception when describing pictorial experience and
its relationship to the imagination.
56
See Hopkins (1998), following the writings by Richard Wollheim.
55
350
gine feeling the genuine emotion towards the fictional entities at issue. We
do so by imagining of our quasi-emotion that it is a real emotional response towards what is represented by the work. That is, we imagine being
frightened by the lion by imaginatively identifying our quasi-fear with an
instance of genuine fear of the lion. Again, the kind of imagining in question is imagining from the inside: the imaginative identification in question
involves an identification of two subjective emotional perspectives, one
real and the other fictional.
Our emotional responses towards fictional entities are not twofold: we
are not emotionally aware of two different objects. In particular, the quasiemotions concerned are not directed at the respective artworks. Indeed,
they do not have any (clear) object. They are triggered by the imagination
of some fictional entities or situations, but are not about them (at least according to Walton). Hence, the problem of guaranteeing that our imaginative response forms a unified experience is less pressing than in the case of
pictorial experience. Feeling the quasi-emotion and imagining it to be a
genuine emotion towards fictional entities need not form a single and unified experience. Instead, the main reason for assuming the third element
over and above the other two seems to be that the intellectual imagination
does not involve any affective elements. Imagining that one is feeling an
emotion does not suffice for having an affective reaction towards it. But
just as our awareness of what is depicted possesses a visual character, our
response to fictional entities possesses an emotional character. Assuming
that we also imagine feeling an emotion promises to introduce the required
affective element into the experience.
However, it is not clear how this supposed to work how imagining
feeling an emotion can really possess an affective character. As Moran
notes, the problem arises because the emotion is assumed to be merely part
of what is imagined.57 In the case of intellectual imagining, this is precisely
57
See Moran (1994). Note, however, that Moran seems to misunderstand Waltons
position by ascribing to him the view that what is central to affective imagining is
intellectually imagining that one has the emotion concerned. The reason for this misunderstanding appears to be the failure to see what is responsible, according for
Walton, for the fictional truth that we are feeling an emotion towards the fictional
entities in question. Moran seems to assume that this is due to intellectually imagining that proposition hence the view which he ascribe to Walton; while the latter in-
351
352
reason, our emotional responses to fictional entities are very similar to, and
as unproblematic as, our more ordinary affective reactions to certain real
objects, situations or events such as those which are in the past or the future, or which constitute unrealised, but real possibilities (e.g., missed opportunities or alternative courses of action).
As discussed in the previous section, Walton rejects all three claims. He
thinks that emotional feelings are part of what we imagine, not of how we
imagine it. He also maintains that our imaginative emotional engagement
with fictional worlds is only prompted by, and about, the relevant quasiemotions, but does not include them as one of its constituents. And finally,
neither the quasi-emotions, nor our imagining feeling an emotion are, for
him, instances of emotion which is reflected by his acceptance of (b).
Again, the aim here is not to settle the debate about the third claim that
is, about how best to reply to the seeming paradox of fiction. But the first
two claims are relevant for the nature of affective imagining. We have
already considered Waltons position. It it now time to look into the details
of Morans view.
Moran understands our affective responses to fiction as instances of
what he calls emotional imagining. This label fits very well with the fact
that the form of imagining concerned is treated by Moran as being genuinely emotional. Episodes of emotional imagining are taken by him to be
instances of real emotion and, hence, on a par in this respect with normal
emotional feelings, in the same way in which episodes of visual imagining
are sometimes taken to be instances of visual experience and, hence, on a
par in this respect with visual perceptions.58 In the course of his paper,
Moran distinguishes emotional imagining from three other imaginative
phenomena: propositional imagining, dramatic imagining, and imaginativeness. While propositional, emotional and dramatic imagining have in
common that they occur in the form of mental episodes or activities, imaginativeness constitutes a mental ability or disposition. He does not explicitly talk about a fifth form imagining, namely sensory imagining. But
there is no reason to assume that he would not acknowledge its existence,
58
I follow Moran in reserving the expression emotional imagining for affective responses to fiction that are genuinely emotional albeit being imaginative, rather
than real, affective states. The debate between Moran and Walton is therefore about
the possibility (or at least actual occurrence) of such responses.
353
354
tudes consists in imagining something with feeling or emotion, in contrast to imagining it dispassionately (see Moran (1994): 90 and 105). Moran concentrates on propositions as candidates for what we can imagine
with feeling. But just as with the existence of sensory imaginings, it is fair
to suppose that he would also allow for the emotional imagining of objects
or events. Morans examples for emotional imagining are imagining something with loathing, anticipation, apprehension or regret (see Moran
(1994): 86, 90 and 93). The affective aspect of the imaginative episode
consists thereby in a real and not merely in an imagined feeling. Accordingly, imagining something with regret involves really having a feeling
of regret. As a consequence, emotional imagining cannot or not exclusively be a matter of propositional imagining. In particular, imagining
something with, say, sadness cannot be reduced to imagining that one feels
sad. While the former involves a real feeling of sadness, the latter does not.
Now, given that the affective aspect of emotional imagining is real, and not
merely imagined, it should as Moran maintains be located in the manner (or mode), and not in the content, of the imagining (see Moran (1994):
90 and 93). The statement that something is imagined with feeling or emotion thus qualifies how it is imagined, and not what is imagined. It is therefore likened by Moran to the statement, say, that something is imagined
visually or auditorily (see Moran (1994): 93).
Both propositional and emotional imagining occur in the form of single
mental episodes. By contrast, dramatic or empathetic imagining is typically
more complex by involving several distinct episodes in a single imaginative project (see Moran (1994): 104). More specifically, dramatic imagining
consists in the imaginative adoption of, and identification with, a certain
point of view different from ones own. The adopted perspectives in question are typically characterised partly by a set of evaluative attitudes and
the related emotional or conative dispositions. Thus, imaginatively adopting such a point of view usually involves imagining having the respective
evaluative and affective responses to given situations, in addition to more
neutral propositional and sensory imaginings about those situations.
Morans description of dramatic imagination renders it very similar to if
not identical with the phenomenon of empathy, or the closely related phenomenon of imagining being in the place or shoes of someone else (see
355
60
This is true even if, say, what is involved is imagining having certain sensory and affective experiences. For our act of imagining need not further specify the perspectivalness of the imagined experiences, or assume by default that it is our own
(see Martin (2002a) and Dorsch (2010a)).
356
Moran presents another motivation for assuming that our responses to fiction and the
involved affective imaginings are really and not merely imaginatively emotional:
namely that we are often held responsible for having or failing to have them. He
notes that we may be praised or blamed (morally or otherwise) in relation to whether
we react to fictional situations, say, with laughter or lust; and that how we react often
reveals something important about our personality (see Moran (1994): 93f. and 105).
Laughing at a racist joke, for instance, may reveal racist tendencies or beliefs. However, as Walton has correctly pointed out (see Walton (1997), what manifests our
convictions and is subject to assessment can equally well be our dispositions to ima-
357
358
See Chapter 13, as well as Scruton (1974) and McGinn (2004). A similar argument
may be formulated against the idea that our awareness of what a picture represents is
imaginative, given that it is usually not up to us what we experience a picture as depicting (with the exception, perhaps, of ambiguous pictures), or whether we experience it as depicting something in the first place (see Dorsch (2012c)). None the less,
Walton (1990) is not the only one who defends an account of pictorial experience in
terms of imagining. The same idea can be found in, for instance, Scruton (1974),
OShaughnessy (2000) and Stock (2008).
359
Not all desires are conative. We may perhaps desire the occurrence of peace, or that
it will rain (in contrast to desiring to actively bring about peace or rain). But such
desires are not motivational states. Whether they are like emotions or preferences,
and whether they involve a distinctive manner of representation, are interesting
questions which, however, need not concern us here.
360
theatre production, or have thoughts about what we feel. But each of the
representational elements is still confined to a single manner of representation. However, as Moran acknowledges, it is possible to visualise or, indeed, propositionally imagine something with feeling. Again, this provides
a good reason to doubt that there is an emotional mode of imagining, in addition to and of the same kind as sensory and intellectual modes.
Moran therefore faces a dilemma in relation to his insistence on the existence of emotional imagining. If he conceives of the emotional element
as something in addition to sensory or intellectual imagining, he cannot ensure that the two components are unified in a single instance of affective
imagining. But if he understands the emotional element as a substitute for
the sensory or intellectual component in other instances of imagining, he
cannot accommodate the fact that the affective element behaves in a different way and, indeed, combines well with sensory or intellectual elements.
The conclusion should be that what Moran calls emotional imagining
that is, imagining something with emotion does not constitute a distinctive form of imagining. At best, it captures the fact that some of our imaginative representations give rise to quasi-emotions.
361
imagining possess an affective character in virtue of representing an emotion, without actually including any real emotional feeling?
Moran has proposed two types of imagining concerned with emotion:
propositional imagining about emotions; and imagining something with
emotion. The first is characterised by the fact that the emotions form part
of the propositional content of the imagining for instance, when we imagine that we have or feel fear directed at an imagined lion in the room. By
contrast, Moran takes the latter to be an episode of imagining something in
an emotional manner. Propositional imagining is dispassionate and therefore no good candidate for affective imagining; while imagining something
in an emotional manner can be ruled out since there are good reasons to
doubt the existence of such a mode of representation. But independently of
what one thinks about the plausibility of imagining something emotionally,
there is at least a third alternative in which emotions may enter imagination: they may be the direct objects of non-propositional imagining.
Consider the case of imagining a pain (or a similarly subjective bodily
phenomenon, such as an itch; see Sections 10.1f.). This imaginative episode differs subjectively from real instances of pain. Most notably, we do
not come to find the former unbearable in the same way as the latter for
instance, we do not cry or faint as a consequence of experiencing it. In this
respect, imagined and remembered pains are much closer to each other
than to really felt ones. Moreover, this difference between imagined (or remembered) and really felt pains is not simply a matter of degree in determinacy or intensity. We sometimes have real pains which are not very intense or determinate, but which we still experience as real pains, and not
merely as imagined ones. And we also can imagine having rather strong
and specific pains, without thereby beginning to really feel pain. None the
less, imagined (and remembered) pains still involve the quality of pain.
This is reflected by the fact that we describe their subjective character in
terms of pain (e.g., that they feel similar to genuine pain) and group them,
from our first-personal perspective, together with real feelings of pain,
rather than, say, with thoughts about pain, or with sensations of itch. In
short, imagining a pain is an experience, but not a real pain experience.
That is, it involves the quality of painfulness, but does not instantiate it.
The best explanation of this situation is to assume that imagining (and
362
perhaps also remembering) a pain is an instance of non-propositional object awareness (or acquaintance) which takes the feeling of pain rather
than the felt pain as its direct object. 65 The idea is that, while a feeling of
pain involves painfulness by instantiating it, the imaginative (or mnemonic) awareness of such a feeling involves painfulness by representing it as
instantiated. As a result, feeling pain and imagining it are subjectively similar in that both their phenomenal characters involve the quality of
painfulness. But they differ from our first-personal perspective in that they
involve this qualitative aspect in different ways: the former is really an experience of pain, while the latter is an episode of representing pain. The involvement of painfulness in the case of imagining pain is thereby not a
matter of the non-propositional manner of representation. Rather, it is a
matter of what is imagined, namely a feeling of pain which instantiates the
quality of painfulness.
The proposed treatment of imagining pain therefore satisfies all the conditions on affective imagining: it is non-propositional; it does not involve
an emotional manner of imagining, but instead is a representation of emotion, and its affectivity does not amount to the real thing, that is, to a genuine instance of emotion. Emotional imagining may and should thus be
understood, along the lines of (ER), as an instance of experiential imagining: namely as imagining an episode with an affective phenomenal character (i.e., an episode of emotion). This guarantees that the affectivity of
emotional imagining arises from what is imagined, and not from how it is
imagined. But it also ensures that there is an affective element involved in
emotional imagining, but no real emotional feeling, only a represented one.
The proposed imaginative and non-propositional manner of representation does not face the same problems as Morans emotional manner. First
of all, it puts a restriction on what can be imagined, namely particulars and
their experienceable features in this case, mental episodes and aspects of
their phenomenal characters. And then, it excludes other manners of representation. This is true independently of how the precise nature and role of
the non-propositional manner involved in emotional imagining is specified.
There are basically three options. First, it may be held that all instances of
object awareness involve the same basic non-propositional way of repres65
See the comparable claims about itchiness and perspectivalness in Martin (2002a),
which I presented in Section 10.1.
363
364
interact with the respective entities at which they are directed. The idea is
that the emotional aspect looses its motivational power, once it is experienced, not in an immediate way, but instead only mediated by a non-propositional representation of it. Thus, while the experience of fear felt towards a real lion has the power to move us to run away, the imaginative
representation of such an experience of fear directed at a fictional lion does
not possess this power any more. Similarly, the intensity and determinacy
of the feeling usually decreases when we move from a real experience to a
represented one. Thus our imaginations (and memories) of fear are typically though not necessarily always less vivid than the comparable experiences of fear in real life situations (see Section 3.2). The claim that our
emotional responses towards fictional entities amount to the non-propositional imagination of having emotions towards those entities may thus help
to explain some important characteristics of our engagement with representational art.
Finally, the provided account of emotional imagining is compatible with
the idea that it is distinctive of imagining that it allows for voluntary control. The occurrence of quasi-emotions and the subsequent establishment of
a fictional truth to be imagined by us may be beyond the direct influence of
our will. But whether we follow this demand and imagine feeling the respective emotion is at least in principle up to us. The truth of (ER) with respect to emotional imagining and other instances of affective imagining is
therefore compatible with the Agency Account of imagining.
Chapter 12
Semantic Dependency,
Simulation, and Pretense
The representational echo thesis (ER) should be given up since it does not
apply to intellectual imaginings and imaginative projects and, therefore,
cannot capture the shared nature of all central forms of imagining. The
most plausible version of the causal echo claim (EC), on the other hand,
has turned out to be irrelevant for the formulation of a unified account of
imagining, given that it takes imaginings to be dependent on cognitions
only in so far as the latter are needed to acquire the necessary means for
representing something in a sensory or intellectual manner in the first place
a claim which is too weak to be distinctive of imaginings, or to link imaginings to cognitive counterparts with the same content and mode of representation.
But what about other echo claims about imagining? In what follows, I
argue that the postulation of a different kind of dependency is equally unsuitable for capturing the nature of all or even most instances of imagining. In the second and the third section of this chapter, I consider the attempt to characterise imaginings in terms of simulation and pretense, respectively. Both views may be understood as endorsing the idea that imaginative phenomena fundamentally imitate cognitive ones:
(EI) Imaginings are essentially imitations (or simulations) of their
cognitive counterparts.
The next section, by contrast, takes a closer look at the idea that imagin-
366
See the extensive debates started by Putnam (1975) and Burge (1979).
367
See Thomas (1999) for apparent empirical evidence to the effect that congenitally
blind people can enjoy a kind of sensory imagining that comes very close to visual
imagery; and Scruton (1974: 104) for the more orthodox opinion that such people
cannot visualise anything.
368
369
370
One obvious target for criticism is their ultimately incorrect rejection of the
Agency Account of imagining, as well as of the Representational Account with respect to sensory and affective imaginings (see Section 10.3 and Part Four). Their
postulation of desire-like imaginings is also problematic (see Note 16 in Chapter 1).
Carruthers (2003) and Soldati (2006) note further difficulties in their critical reviews
of Currie and Ravenscrofts book. Note, however, that the similarity claim is compatible with (ER): it suffices for resemblance that the content of the imaginings contains the content of the perceptions (see Sections 9.1 and 9.4).
371
This would help to explain a certain passage in Currie (2000: 176), where he describes his simulationist view without at all mentioning the asymmetric dependency.
372
73
It does not matter for this characterisation of imaginings whether the perspective,
that we merely possibly adopt, is actually occupied by someone else, or was or will
be by our own past or future self. It is true that perspective-shifting may help us to
get clearer about the actual situation of others in this sense. But it always does so, at
the basic level, by considering a perspective which we do not actually, but merely
possibly adopt.
373
See Ryle (1949/1963): sec. 8.6. See Ishiguro (1966), Shorter (1952) and White
(1990) for critical assessments of Ryles position.
374
375
The underlying reason for why imagining and pretending are distinct
and fairly independent phenomena is, of course, that, while instances of
imagining are mental phenomenon, instances of pretense are just like instances of assertion or stage-acting inseparably linked to publicly observable behaviour and, relatedly, require some actual or at least potential audience (see Ryle (1949/1963): sec. 8.5, and Currie & Ravenscroft (2002):
32). Pretending is aimed at behaving as if one were a particular person, or
as if one were having certain mental states. Moreover and, again, just like
assertion or stage-acting pretense may be used for deception. Consequently, it is no surprise that imagining cannot literally amount to (a kind
of) pretending.
Third, it might be attempted to elucidate imagining by means of drawing an analogy with pretense for instance, by claiming that imagining is a
form of inner pretense, while not really amounting to pretense proper
(i.e., publicly observable pretense). But this last proposal is not much more
plausible than the previous two. This time, the problem is that imagining
and pretending do not have any significant aspects in common, apart from
their being instances of agency.77 In particular, while pretending to do or to
be something is modelled on a specific kind of action, or on some condition or state closely related to specific behavioural dispositions, imagining
is not (see White (1990): 157f.).
This means, first of all, that pretense involves the same, or very similar,
actions as the object of pretense (i.e., the mimicked action, condition or
state), but without becoming the real thing. Pretending to have a fight with
someone involves performing some of the bodily movements pertaining to
real fights, but does not amount to fighting; while pretending to be an aristocrat consists in acting in ways in which real aristocrat behave, without
thereby really becoming an aristocrat. Moreover, acts of pretense can be
appropriately described only by reference to what their object is: sham77
White lists certain other similarities for instance, that both imagining and pretending are rather unspecific with respect to their objects, or that both establish a contrast
with what is real (White (1990): 153ff.). But this is just to say that both lack a cog nitive attitude and may be relatively indeterminate which does not, however, distinguish them from many cognitions (in the case of indeterminacy; see Section 3.3)
or other non-cognitive phenomena (in the case of a lack of cognitive attitude; see
Sections 5.5 and 8.1).
376
fighting has to be elucidated in terms of real fighting. Finally, the pretending person has to rely on her beliefs about the object of pretense and the related ways of acting when engaging in her pretense: she can pretend to
fight only if she has some idea of what it means to really fight, and only if
she lets her respective beliefs inform and influence her behaviour in a manner appropriate for pretense (see Ryle (1949/1963): sec. 8.5; see also White
(1990): 155ff.).
In contrast, acts of imagining are not modelled on certain actions, conditions or states in such a way that they mimic the latter by involving the
same type of activities as the latter; that they have to be accounted for by
reference to the latter; or that the imagining person has to rely on her beliefs about the latter in her imaginative agency. If at all, imaginative episodes should be modelled on cognitive episodes: that is, they should be
taken to be pretend perceptions or pretend beliefs. But the only relevant
activities which may pertain to both imaginings and cognitions are attention and inference; and neither suffices to establish that imaginings mimic
cognitions. Paying attention to what we sensorily imagine (e.g., to certain
aspects of a visual image) already presupposes that the respective sensory
imagining has been formed. Inferring a certain proposition from a set of
others, on the other hand, is not necessary for the active formation of intellectual imaginings.78 Hence, neither form of agency can constitute the kind
of activity at the heart of active imagining.
Relatedly, the elucidation of what it means to imagine something does
not have to make reference to attention or inference. And the imagining
person also does not have to make instrumental use of her beliefs about attention or inference while being engaged in active imagining. Consequently, imaginings are not pretend cognitions. At best, one may perhaps speak of pretend inferences which mimic real inferences by using
and resulting in intellectual imaginings instead of judgements or beliefs;
78
It is maybe also not sufficient. For while inferences on the basis of beliefs seem to
compel us to endorse the resulting propositions, it seems ultimately up to us whether
we (continue to) imaginatively entertain a proposition which we have recognised to
rationally follow from other imaginatively entertained propositions (plus perhaps
some believed ones). This difference may also be related to the fact that there are no
dispositional imaginings which could compel us to imaginatively endorse a proposition simply by becoming manifest independently of our will (see Section 1.3).
377
and of pretend attention which mimics real attention by focusing on sensorily imagined objects instead of really perceived ones. But the fact that
we can pay attention to what we sensorily imagine and employ intellectual
imaginings in inferences appears to be primarily a consequence of the fact
that imaginings are of the same types of representation (e.g., visual or intellectual) as their cognitive counterparts. It does not pertain to a pretend
relation between imaginings and cognitions that is structurally similar to
that holding between acts of pretense and their objects.
378
Part Four
The Agency Account
In the preceding two parts of the book, I discussed the Epistemological and
the Dependency Account. While the first focused on the non-cognitivity of
imaginings, the second highlighted their descendance from, or constitution
by, cognitions. However, I found both theories wanting in respect of their
prospects of providing a unified account of imagining. It is therefore necessary to look for a different feature that is common to, and distinctive of, all
imaginings, whether they are sensory or intellectual, episodes or projects.
The feature that I would like to put forward in this part is their intimate relationship to the will.
It should be uncontroversial that many instances of imaginings are voluntary: they are actively formed and, as such, deliberate expressions of
what we want. In particular, it makes sense to ask or order someone to imagine, and also to answer to such a demand. 1 This aspect of imagining has
already been emphasised in the context of the discussion of Wittgensteins
and OShaughnessys writings on imagining (see Section 4.7 and Chapter
7, respectively), who both take the active character of the respective imaginings to be responsible for their non-cognitivity (e.g., their uninformativity, or their lack of a power to cognise reality). I agree in general with their
explanatory strategy of saying that voluntary imaginings are not in the
position to provide knowledge because of their origin in the will. But I disagree with them in two important respects. First, I assume that all and not
only some instances of imagining are actions. Second, I assume that it is
1
See Wittgenstein (1984c): vol. II, sec. 125; Scruton (1974): 94f.). See also
OShaughnessy (1980): vol. I, 1ff. for a detailed discussion of the link between voluntariness and the applicability of orders.
382
The Agency Account has been endorsed by Wollheim (1973): 69, Scruton (1974): 95
(see in general 94-100), Levinson (1998): p. 232, n. 3, Kind (2001): 90ff., and McGinn (2004): 12ff.; 131f. Hopkins (2011a) also assumes the truth of the Agency Account, though perhaps only for the sake of argument. Most of these endorsements
have been inspired by Wittgensteins comments on sensory imaginings (see his
(1984b): vol. II, sec. 63 and 627; see also Budd (1989): 104ff.).
At first sight, it might seem that Sartre and Collingwood also present versions of
the Agency Account. Sartre characterises sensory imaginings as spontaneous and
383
creative (Sartre (1940): 14) and always involving, or being accompanied by, an
intention (Sartre (1940): 11; 19; 32), which distinguishes them from passive
perceptions (Sartre (1940): 14; 33). And Collingwood claims that imaginings are
active (Collingwood (1938): 195ff., and chs. 9f. in general). But, as noted in
Section 8.2, they both allow for involuntary imaginings (Sartre (1940): 19;
Collingwood (1938): 179; 195); and, as already discussed in the Sections 4.1 and
5.6, Sartre takes many other intentional states to be spontaneous, among them
sensory memories and pictorial experiences (Sartre (1940): 10). Therefore, although
many ideas to be found in Sartres and Collingwoods writings notably the idea
that imaginings are active are congenial to the Agency Account, they do not
really constitute endorsements of it. In particular, Collingwood seems to favour the
weaker Subjection to the Will Account (see Section 13.5).
Finally, OShaughnessy (2004: ch. 11f.) would be a proponent of the Agency
Account, if he would not wrongly assume that hallucinatory, illusory, pictorial and
similar experiences are imaginative, rather than perceptual (see Section 14.4).
CHAPTER 13
Imaginative Agency
See Dorsch (2009b) for a discussion of all these examples of non-imaginative mental actions. The main point raised there is that the connection between the basic action of thinking of something pleasant and its causal consequence of an improved
mood does not significantly differ from the connection, say, between the basic action
of pulling the trigger of a gun and its causal consequence of the death of a person. In
particular, in both cases, we engage in the basic action because we intend to bring
about a certain result (i.e., an improvement in our mood, or the death of the person),
and because we know that we can cause that result to occur by engaging in the basic
action in question. In other words, killing a person is as directly or indirectly linked
to agency as improving ones mood. Hence, there is no reason to deny the latter the
status of an action, if it is assigned to the former. Killing, however, is often treated as
a paradigm instance of agency.
386
Imaginative Agency
387
not absolutely precise. While mental actions at least typically consist solely
of mental episodes (and perhaps also dispositions) and the mental processes linking these together, it might be argued that some mental actions
involve also certain forms of relatively non-interfering bodily actions. For
instance, it does not seem to matter much for the action of finding the best
move in a game of chess whether one scans the position on a chess board
with ones eyes or merely visualises it in ones mind. Arguably, it might
also not matter much whether one actually tries out the moves on the board
(say, during an analysis) or visualises them in ones mind. Only once one
actually makes a committal move in the game is ones action interfering
with the world in a relevant or significant manner with respect to the situation in question that is, in this case, changing the actual course of the
ongoing game of chess in question.
However, I assume in what follows that, in most cases, the distinction
between mental and non-mental actions is precise. I also take it for granted
that imaginative actions as instances of imagining are realised entirely
in the mind. Their pursuit may require the performance of certain bodily
actions as enabling conditions (e.g., we may have to find a comfortable
position in a quiet corner and to close our eyes in order to be able to visualise something). And imaginative actions may figure in wider projects that
are neither purely mental nor purely imaginative (e.g., instances of aesthetic appreciation, or of prop-involving games of make-believe; see
Walton (1990): especially chs. 1 and 6). But imaginative actions do not by
themselves involve any events which occur outside the mind of the imagining person. In what follows, I therefore focus on mental actions consisting
solely of conscious episodes and the mental processes linking them together, and ignore the possibility of other types of mental agency.
388
See, for instance, Pink (1996) and Owens (2000) for detailed discussions of, respectively, the involvement of theoretical considerations in the formation of intentions,
and the involvement of practical considerations in the formation of beliefs.
Imaginative Agency
389
390
is the directness condition: the voluntary determination of what is represented has to happen without any reliance on epistemic or merely causal processes as means. The main thesis of the Agency Account of imagining
should accordingly be expressed by means of the following thesis about
the fundamental nature of imaginings:
(ACT)
Imagining may also serve purposes extrinsic to it. For instance, we may
imagine something in order to improve our mood, or in order to solve a
problem. But what identifies instances of imaginings like this as imaginings are not those contingent and external ends, but instead the necessary
and internal goal of directly producing one or more specific representations. Similarly, this intrinsic imaginative purpose end may give rise to further instrumental (and intrinsic) ends. Imagining something may require,
say, shifting ones attention away from what one is perceiving (e.g., by
closing ones eyes) and concentrating on the imaginative task in question;
or recalling the appearance of something; or imagining separately each of
the parts of something. But, again, what marks the resulting formation of
representations as imaginative are not these contingent or necessary means,
but instead its ultimate intrinsic goal of forming them in a direct way and
with the specific determination of which states of affairs are represented.
This is why (ACT) is concerned with the specification of the ultimate intrinsic purpose of imagining.
The thesis (ACT) is meant to apply primarily to the central cases of imagining described in Section 1.2. It may also be understood as capturing a
core notion of what it means to be an imaginative mental phenomenon. Accordingly, to imagine something may simply mean to directly form one or
more specific representations of it. But the main intention of the Agency
Account is to take (ACT) to be a claim about the nature of the paradigm instances of imagining, and not about our concept of imagining. Although
(ACT) claims that all imaginings share the same essence and, in this sense,
form a particular kind of mental phenomenon, it does not assume that they
Imaginative Agency
391
13.3. Qualifications
(i) Concreteness
One set of qualifications concern the nature of the concrete goals of imaginative agency. First, the purposes of imaginative mental actions or the
relevant underlying motivational states need not determine the type or
mode of the representations to be formed. We may decide to imagine a
square box, say, without any particular intention to do so by means of representations of a particular type (e.g., tactile rather than visual representations, or intellectual rather than sensory representations). As a result, we
are free in our pursuit of the respective project to use appropriate
representations of any modal type.
Second, the purpose pertaining to an instance of imagining may be
more, or less, general in its description of the representations to be formed.
For instance, it is possible to imagine a garden party on the basis of intending to represent one, sensorily imagine one, or visually and auditorily imagine one.
Third, the determination of the states of affairs to be represented may be
more, or less, specific. We may aim to simply visualise a tree, or to visualise a Japanese cherry tree in full blossom, or to visualise such a tree in a
large garden with bridges, lakes and tea pavilions. In the same way, we
may just want to imagine being Caesar; or we may want to imagine being
Caesar during his successful political career in Rome; or we may want to
imagine being Caesar at the moment when he realises that his son is among
his assassins. The degree of specification can thus vary greatly as long as
there is at least some active determination of which properties the represented objects are represented as having.
392
(ii) Voluntariness
Another set of qualifications pertains to the fact that the active and direct
determination of what is represented need not be complete or entirely voluntary. Four considerations are of relevance here.
First, it may very well be that there are instances of mental and, in particular, imaginative agency that we voluntarily pursue without noticing it.
That is, there might be forms of imagining that are sub-intentional in the
same way in which, say, absent-mindedly scratching ones back is (see
OShaughnessy (1980): vol. II, ch. 10; Kind (2001): 91). This possibility is
perhaps more plausible with respect to intellectual representations than
with respect to sensory ones. For instance, we may continue to think about
a certain problem while being occupied by a completely different task; and
we may actually come up with a solution to the problem in this way. Or we
may discover that we have been worrying all along about what a person
has meant by something she said to us, only realising afterwards the effects
that these worries have had on our feelings and behaviour. Or we may find
ourselves imagining about meeting a person again who we encountered the
other day at a party. Perhaps not all types of mental or imaginative activity
allow for such an engagement without attention. But some might do.
Second, there are certain general limits on what we can imagine. I have
already noted that intellectual imaginings especially if they do not occur
in isolated form, but are embedded in a wider imaginative context may
be subject to moral or logical constraints (see Section 1.2). Thus, we may
have difficulties of adopting the moral point of view adopted by the (implied) narrator of a story (Gendler (2000)). In addition, when we imagine
something about a certain character in a particular situation, while the
nature of both is already determined in rich detail by our previous imaginings and other representations, our imaginative activity may gain its own
momentum: that is, we may more or less automatically imagine the person
to behave in a way which follows logically from what we take her situation
and her personality to be (Wollheim (1973): 69f.; see also Eco (1994) and
Carroll (2000): 124f.). But the restrictions on what we may be able to imaginatively represent are not due solely to our moral attitudes or our rationality. They may also arise from limitations in the scope of our past experiences, our conceptual capacities, our imaginative skills, our ability to con-
Imaginative Agency
393
centrate, and many other factors (see also Sections 9.1 and 9.3 on causal,
and Section 12.1 on semantic dependency).
Third, not all aspects of the contents of the representations formed in accordance with the specific imaginative purpose have to be actively and directly determined by the underlying motivational states. For instance, what
we are sensorily imagining is often partly determined by how our mind
passively and spontaneously fills in the details of the images that we bring
about (see Section 14.4 for more on spontaneous representations). When
we act just on the general aim to visualise a building with no specific shape
or number of windows, we may none the less end up visualising it as having a determinate shape and number of windows possibly due to the influence of association or generic memory (see Sections 4.6 and 7.2). But
this does not undermine the imaginative status of the respective episodes
since at least some of the represented states of affairs (e.g., that the imagined entity is a building) remain determined actively and directly.
Fourth, there may be cases of imaginative episodes in which we actively
imagine something, but ultimately against our will in the sense of akratic
or obsessive agency: that is, against what we really want to do or what we
take to be the best thing to do. In particular, we may produce images and
thoughts, which we would in fact prefer to banish from our minds, but cannot help to (continue to) produce. For instance, when we suspect that our
partner has an affair with someone else, we may not be able in our jealousy
to stop imagining the two together despite the fact, for instance, that we
would prefer not to picture or think about them because it would make us
feel much better; or despite the fact that we know our suspicion to be
without any serious evidential foundation. Similarly, our perception of the
physical attractiveness of a person may lead to our active engagement in
some sexual fantasy concerning him or her, although we would rather like
not to follow this course of action (e.g., because we take it to be morally
inappropriate, because we very much dislike the personality of that person,
or simply because it is not the right moment or leads to an awkward situation). In both examples, we would like to get rid of the respective images
and thoughts. Although these imaginative episodes are still actively produced by us, we may not succeed or only with great effort in our attempt to make them disappear again.
394
(iii) Directness
The final set of qualifications is concerned with the status of imaginings as
direct actions. This time, two considerations are of particular importance.
First, the fact that imaginative activity is direct with respect to the specific determination of which states of affairs are represented is completely
independent of the issue of whether imaginative activity is direct or indirect in other respects, or whether it differs in this from other kinds of mental
5
It seems that Wittgenstein has cases like these in mind when he writes that imaginings often force themselves upon us and stay against our will, do not let themselves
be scared away. But the will is able to fight against them (Wittgenstein ((1984b):
vol. II, sec. 86). (The translation is mine. The original reads: da Vorstellungen oft
gegen unsern Willen sich uns aufdrngen und bleiben, sich nicht verscheuchen
lassen. Doch aber kann der Wille gegen sie ankmpfen.). Budd endorses the same
interpretation in his book on Wittgenstein (see Budd (1989): 105 and 109; see also
Section 4.7). Armstrong (1969): 298, and Scruton (1974): 94f., make very similar
statements about the relation between imagining and the will. Compare finally the
observation that obsessive or intrusive images and thoughts are actively produced by
us against our will though, in the latter case, not acknowledged by us as such (see
Sartre (1940): 148ff., and Roessler (2001)).
Imaginative Agency
395
I follow here the traditional understanding of actions as non-basic just in case they
are done by performing another action (see: Danto (1963); OShaughnessy (1980):
vol.1, xiii; Searle (1983): 98ff.; McCann (1998): 4; Audi (1993): 171; Lowe (2000):
234).
396
Imaginative Agency
397
398
ode itself.10 The relation between these elements may be understood in two
different ways. The resulting views the Causal and the Constitutive View
differ in whether they treat the two relata as distinct mental phenomena,
in the sense that they could occur independently of each other. They are
therefore concerned with the internal structure of the active and direct
formation of imaginative episodes. More specifically, the two views disagree about whether the active origin of the relevant imaginative episodes
is essential to them, given that that actively formed representations are essentially active in so far as they are inseparable from and form an episodic unit with the event of their active formation.
The Causal View assumes that the occurrence of the imaginative episode is merely a causal result or product of the imaginative activity. Accordingly, the two elements are treated as distinct entities: the activity could fail
to bring about any imaginative representation; and the imaginative episode
could occur in a different, purely passive manner. The two elements merely
happen to be part of a more complex entity, namely the active formation of
an imaginative episode. Given that the very same imaginative episode
could have occurred without the involvement of agency, passive imaginings are possible. It is in this sense that the active origin of imaginative
episodes is assumed to be inessential to them.
The Constitutive View, by contrast, maintains that the two elements are
constitutively dependent on each other. On the one hand, actively formed
imaginative episodes could not have occurred without imaginative activity:
the active origin of imaginative episodes is essential to them. On the other
hand, imaginative activity could not occur without actively formed imaginative episodes: the occurrence of an imaginative representation is essential
to imagining. It is therefore no accident that the two elements form a
whole. Moreover, active imaginings differ essentially from passively occurring images and thoughts. That is, the two belong to different mental
kinds. Assuming that the actively formed episodic representations are central instances of imagining, it follows that passive imaginings are impossible.
10
In what follows, I assume but do not always explicitly mention that the active
formation of the imaginative representations under consideration is direct: that is, involves an active determination of content, without any reliance on epistemic or
merely causal mechanisms.
Imaginative Agency
399
Of course, which of the two mental phenomena should be labelled imaginings is primarily a terminological issue. Perhaps there are good reasons
to use that term ambiguously and refer to both active and passive cases as
imaginings. But this does not undermine the claim that there is an essential difference between actively and passively produced mental episodes;
and that, hence, a unified account of imagining can hope to accommodate
at best one of them, given that it aims to identify the essence common to all
instances of imagining. No one denies that there are voluntary instances of
imagining; while the possibility of involuntary imaginings is controversial.
Hence, the second view should be read as entailing that there are no passive imaginings.
So far, the characterisation of the two views has been neutral on the ontological categories of the two elements concerned. But it is natural to assume that both the episodic representations and their active formation are
event-like aspects of the stream of consciousness. For instance, actively
visualising a tree seems to consist of an event of trying or willing to visualise a tree (what I have called in Section 3.5 an intention-in-action), and of
an event of visually experiencing a tree. The two views therefore differ in
whether they treat the two elements as two distinct events, or rather as two
dependent parts of a single event. One particular issue is whether the resulting imaginative episodes constitute an inseparable part of the mental actions concerned, or whether they are merely causal consequences of those
actions. The other main issue is whether imaginative activity can fail to
result in imaginative representations.11
The view that I would like to defend is, of course, the Constitutive View,
given that it includes the second premiss of the argument outlined above,
according to which imaginative activity is essential to actively formed imaginative episodes. But the Constitutive View assumes more than this con11
The disagreement between the two views therefore parallels the debate about whether bodily movements are part of, or merely the causal product of, bodily actions
and whether, therefore, bodily actions can be reduced to tryings or similar mental
events, or not (see, for instance, the respective discussions in OShaughnessy
(1980), Hornsby (1980) and Ginet (1990). It also parallels the debate about whether
disjunctivism about actions is true: that is, whether there is an essential difference
between acting, on the one hand, and merely trying but failing to act, on the other
(see, for instance, Dancy (2008) and Hornsby (2008), as well as Dorsch (2011a)).
400
Another motivation for this assumption is that it is unclear what kind of event-like
entity could depend on an event of trying (or willing) without itself being part of that
event. The picture just developed is, however, in tension with the idea that actions
should be identified simply with tryings or willings, and that none of the immediate or distant results of such tryings should count as part of the actions, let alone
necessary parts (see Hornsby (1980); Pietroski (2000)). But not only is this very restrictive view of what actions are problematic in its own right (see, for instance,
Audi ((1993): ch. 3) and Noordhof (2001)). Its proponents also permit that action
descriptions can apply to complex events consisting in tryings and their causal results if there is a suitable or non-deviant causal link between the two (Hornsby
(1980): 122f.). Hence, I continue to characterise the respective events of thinking,
imagining, moving ones body, and so on, as actions with two dependent parts,
namely a trying and a representation or bodily movement.
Imaginative Agency
401
402
the occurrence of such a thought or image, the action would not have been
performed. This still allows for partially successful direct mental actions
for example, when we manage to only partly think or visualise what we
have intended to, or maybe also when we manage to think or visualise only
something completely unintended. Solely in cases of complete failure does
this view assume that no action of the desired kind has been performed.
Third, only the Constitutive View can account for the specific phenomenal character of actively formed representations.15 From the inside, we
usually experience actively formed images and thoughts differently from
passively formed ones. A natural explanation of this fact is that our experience of actively formed representations involves an awareness of agency.
That is, it is part of the phenomenal character of these representations that
they are experienced as actively formed. This normally enables us to tell
whether we have brought about a certain visual image by means of visualising something, or whether it has occurred spontaneously and unwilled.
Similarly, we can usually distinguish thoughts which simply cross our
minds from those which we actively produce.
Now, we can become aware of the active formation of a mental representation in an immediate and non-inferential way. It is not that we infer the
presence of a causal link between the action and its result in response to
two independent instances of introspective awareness, one concerned with
the action of forming a representation, and the other with an occurrence of
a representation of that type. Instead, we are typically able to tell whether a
mental episode has been actively formed or not simply by introspecting its
phenomenal character. In other words, the phenomenal character of mental
episodes reflects their active or passive origin.
The Causal View cannot account for this fact about our self-knowledge
of direct agency. If representations were the result of agency solely due to
their contingent and subjectively inaccessible causal origin in some preceding activity, their phenomenal character would not involve a marker of
their active character. This is due to the fact that our awareness of agency
does not extend to mere causal results of actions. For instance, if we deliberately attempt to improve our mood by humming a tune in our mind, we
15
See Dorsch (2009b) for a more detailed discussion of the following phenomenological description of mental agency and the presented argument in favour of this description.
Imaginative Agency
403
experience only our imaginative humming, but not our resulting change in
mood, as an instance of agency. It thereby does not matter whether the
causal results concerned are of such a kind that, under different circumstances, we could experience them as active. When we use our right hand
to move one of the fingers of our left hand, say, we are not aware of the
movement of the latter as actively formed; while we are so aware if we
move the finger in the normal direct way.
In addition, the phenomenal character of mental episodes reflects only
aspects of their essences, but not also some of their contingent features
such as their causal origin, according to the first view. This is due to the
fact that the phenomenal character is itself part of the nature of the episodes concerned and, hence, cannot be determined by their non-essential aspects (see Dorsch (2012a) and Soldati & Dorsch (2011)). That we experience imaginative episodes as actively formed indicates that their phenomenologically salient activity is therefore essential to them. Hence, the
fact that we experience the immediate results of mental actions as actively
produced can be satisfactorily accounted for only by reference to the fact
that these results (i.e., the formed representations) are part of the respective
actions, and not merely their immediate causal results. The Constitutive
View on the relation between the active and the representational element in
the active and direct formation of imaginative representations should therefore be preferred over the Causal View: the two elements are dependent
constituents of the mental actions in question.
The preceding considerations have been concerned with establishing the
claim that the two elements of the active formation of imaginative episodes
the imaginative representations themselves and the acts of forming them
are constitutively dependent on each other. As a consequence, actively
formed imaginative episodes are essentially instances of agency: they
could not have had a passive origin. On the assumption that at least some
imaginings are actively and directly formed, it follows that there are no
passive instances of imagining. Furthermore, given that their active character cannot be traced back to some more fundamental aspect of their nature
(see Section 14.1), imaginings are fundamentally instances of agency.
Hence, the Agency Account of imaginings should be endorsed.
404
Imaginative Agency
405
This view has been put forward by Casey (1976): 34f. and 63ff.; and perhaps also
Walton (1990): 13ff. and Collingwood (1938): chs. 9 and 10.
406
tual world). People differ in their abilities, and what is possible for some
may be impossible for others. Abilities are also open to evolutionary developments and to individual training, so that what is impossible for all or individual human beings today might be possible for them in the future. It
therefore needs to be asked which kind of possibility of voluntary and direct control should be taken to be distinctive of imagining. For instance,
does what is possible for other people matter for whether one of my mental
episodes is an instance of imagining? And should we consider what is possible for the relevant people relative to how they actually are, or instead
relative to how they would be after five years of intensive training or after
the next step in the evolution of the human species? Unless the Subjection
to the Will Account does not answer these and similar questions and specifies the kind of possibility concerned, its characterisation of what it
means to imagine something remains unacceptably incomplete and inapplicable as a unified account.17 But it is doubtful that these issues can be
satisfactorily resolved. For the Subjection to the Will Account appears to
mix up the question of what the nature of imagining is with the question of
when it is possible whether for human beings or for other creatures to
engage in imagining. The act of forming an image of a chiliagon (i.e., a
regular polygon with a thousand sides) in a voluntary and direct way, for
example, seems to be a clear instance of imagining, independently of
whether any one will or would ever be able to deliberately perform or influence it.
Finally, even if the kind of possibility of voluntary control invoked by
the Subjection to the Will Account could be further specified in such a way
that it might become relevant for the question of whether mental episodes
count as imaginative or not, it would presumably require the potential answers to that question to be empirical. Whether it is possible for us to visualise a chiliagon for instance, after suitable and sufficient training, or
after changes to our environment, improvements in our diet, increases in
our brain size, and so on is by and large an empirical (i.e., experimental)
question. By contrast, we can normally settle the issue of whether one of
our own mental episodes is imaginative from our first-personal perspect17
Note that the problem here is not simply one of vagueness. There might be no clear
answer to the issue of how much training or evolution may make a difference. But
whether training or evolution matter in the first place is not a question of degree.
Imaginative Agency
407
18
CHAPTER 14
Meeting the Desiderata
At the beginning of this book, I specified two desiderata for any unified account of imagining (see Section 1.1). Such a theory should be extensionally adequate with respect to all central cases of imaginings, as well as all
paradigm instances of non-imaginative phenomena. And it should provide
an explanatorily powerful account of the imaginativeness common to all
imaginings. In this final chapter, I illustrate how the proposed Agency Account of imagining can satisfy these two requirements on a unified theory
of imagining, thereby discussing them in reverse order. While the first section of the chapter highlights the explanatory power of (ACT), the remaining three sections discuss the scope of its applicability. Two of these three
sections are concerned with the discussion of apparent instances of imagining that appear to be problematic for the Agency Account, and which may
therefore constitute counterexamples to it. The particular focus is on the
possibility of non-representational imaginings and on the issue of whether
certain kinds of passive representation namely spontaneously occurring
images and thoughts should count as imaginative.
410
property or set of properties distinctive to imaginings. Since mental actions can consist both in simple mental episodes and in complexes thereof,
(ACT) applies equally well to imaginative episodes, to experiential imaginings and to imaginative projects. I therefore focus on the other two aspects
of the explanatory power, namely informativity and fundamentality.
(i) Informativity
There should be no doubt that the main thesis of the Agency Account that
all imaginings are essentially mental actions of a particular kind, as specified by (ACT) is a substantial and illuminating claim. It says something
very specific about the nature of imaginings and their shared imaginativeness: it describes what it means for something to be an instance of imagining. Moreover, in contrast to, say, negation claims, it is not in obvious need
of supplementary theses about the constitution of imaginings. Indeed, reference to the particular kind of activity distinctive of imaginings and, especially, the presence of a conscious and concurrent motivational state (e.g.,
an intention-in-action, volition or trying) suffices not only to distinguish
imaginative phenomena from non-imaginative ones (see the next section),
but also to explain why imaginings share certain features with cognitions,
but lack the cognitive features that are unnecessary for the provision of
knowledge (e.g., a cognitive attitude, the status as providers of prima facie
reasons, and the function or aim of cognising reality).19
That imaginings can refer to reality, for instance, is partly due to their
19
411
412
that we recognise imaginings and their contents as the result of our own
agency helps to explain why we normally do not take them to reliably represent the world. It is not surprising then that, in most cases, we do not rely
on them in belief (nor, presumably, in action); and, hence, that they normally do not play the same epistemic role as cognitions.20
Finally, the absence of a cognitive attitude can derived from the presence of agency in a closely related way. In the Sections 2.1f., I suggested
that the cognitive attitude of perceptual experiences, judgemental thoughts,
and so on, constitutes one of the aspects of their phenomenal character that
reflect their function to cognise reality. That we experience cognitive episodes as making a claim about how things are is part of what renders it phenomenologically salient that their aim is to constitute knowledge. Hence, it
is not surprising that imaginative episodes, which do not possess this aim,
also lack a cognitive attitude: the functional difference between cognitive
and imaginative episodes is mirrored by their difference in attitude. This is
also part of the explanation of why the experience of agency involved in
imagining and the experience of endorsement involved in judging are incompatible with each other: while the second reflects the presence of the
function to cognise reality, the first reflects its absence (see Dorsch (2009b)
for more on this point).
The Agency Account is therefore in the position to provide a satisfactory
and illuminating explanation of why imaginative episodes share some of
the features characteristic of cognitions (e.g., the ability to refer to reality
and the possession of success conditions) and are without others (e.g., the
epistemic function, the cognitive attitude, the actual role in the acquisition
of knowledge and the lack of an immunity to error through misidentification of what is represented). Imaginings possess or lack these features exactly because of the active and direct determination of their occurrence and
their content.
20
This normative role of phenomenal character (i.e., how we experience our mental
episodes) for rational motivation is further discussed under the heading of Experiential Rationalism in Dorsch (2009b), (2012a) and (2013).
413
(ii) Fundamentality
It is also reasonable to maintain that the thesis (ACT) assuming that it is
true identifies the most fundamental feature which imaginings have in
common and which is responsible for their imaginativeness. Indeed, the
basicness of the property of being a mental action with the purpose of directly producing specific representations is very difficult to deny, given that
there does not seem to be any other, more fundamental, property that all
the bearers of the complex feature in question have in common, and which
is responsible for their exemplification of direct agency with respect to the
specific determination of what is represented.
In particular, as I have argued in Chapter 8, the property of lacking the
capacity to provide knowledge or to show any of the cognitive features
essential to the provision of knowledge (e.g., reliability or veridicality)
cannot account for the common imaginativeness of imaginings, given that
some imaginings (such as the two examples introduced in Chapter 4) can
cognise reality, while others (affective imaginings and imaginative projects) do not differ from their non-imaginative counterparts in their noncognitivity. On the other hand, the cognitive features that imaginings lack,
but which are not necessarily linked to the provision of knowledge, are not
responsible for the distinctive active character of imaginings, either. There
is no good reason to maintain that imaginings involve the active and direct
determination of what they specifically represent because they do not possess a cognitive attitude, do not provide us with prima facie reasons for belief, or do not have the function or aim of providing knowledge.
Similarly, the fact that sensory and affective imaginings are perhaps best
understood as representations of cognitive episodes does not help to explain why (ACT) is true of all imaginings. First of all, it cannot account for
the active character of intellectual imaginings or imaginative projects, say,
because they do not represent their cognitive counterparts. But even if this
problem of scope is ignored, the representational link essential to some
forms of imaginings cannot account for their equally essential voluntariness. For being a representation of a cognitive episode is fully compatible
with being a passive episode. Sensory memories are arguably representations of perceptual experiences as well (see Section 9.4). But it is not up to
us how we recall people, situations or other entities as being. Whether I re-
414
415
416
417
Given that the only way of directly forming representations with specific contents is
to generate imaginative episodes, imaginative projects may also be individuated as
projects that aim at the production of imaginative episodes. But their description in
terms of the direct production of representations with specific contents is to be preferred, not only because it reflects the fact that imaginative episodes and imaginative
projects count as imaginative for the very same reason, but also because it is more
fundamental. Imaginative projects aim at the production of imaginative episodes because they aim at the production of representations with directly determined specific
contents; and not the other way round.
418
419
day at the beach may actually trigger some unexpected, ambivalent feelings or thoughts about the prospect of spending so much time with the other person. Given that these feelings or thoughts may run contrary to the intention to imagine a happy day, it is not clear to what extent they still belong to the daydream, or to what extent they actually terminate it or turn it
into a different daydream. Imaginative projects are by no means unique in
this respect. Which feelings, memories or associated thoughts actually contribute to and not merely accompany ones deliberation about what to
believe or do in a particular situation may sometimes be difficult to decide
too. Equally, the question of which particular acts of thinking are part of
the project of solving a certain problem in ones head may resist conclusive
settlement as well.
Many experiential imaginings are imaginative projects (see Hopkins
(1998): ch. 7, and Noordhof (2002)) and therefore fit the proposed Agency
Account as well. The experiential imaginings in question are those consisting of at least two distinct episodes: an imagining of a mental episode together with some higher-order representation about the imagined nature of
that episode. For instance, I may imagine a visual perception as of some
land appearing on the horizon and simultaneously imagine that my imagined representation is the perception of Columbus when he thought he
had reached India. We actively engage in such experiential imaginings, and
we do so for reasons related to the formation of specific representations.
My experiential imagining may be motivated by, say, my intention to imagine being Columbus when he first saw land.
But not all experiential imaginings of mental episodes are complex in
the way described: there are also episodic instances of experiential imagining. For instance, visualising something is best analysed as imagining seeing it, while the latter is understood as an instance of object awareness concerned with imagining an instance of a certain phenomenal type of perceptual experience (see Chapter 10). Something similar is true of emotional
imagining (see Chapter 11). Contrary to what Peacocke seems to think (see
his 1985, and also Section 3.6), these cases of experiential imagining do
not have to involve any thought-like element specifying the nature of the
imagined experience, given that this nature is determined by default. Imagining a perceptual experience as of a tree, for instance, amounts, by de-
420
421
422
423
real pains).22
Given that there is no reason to doubt that the situation is the same with
respect to other examples of affective imaginings, that correspond to nonrepresentational feelings or moods, it should be accepted that there are no
non-representational imaginings. Accordingly, instances of imagining pain
or similar affective imaginings can be accommodated by (ACT): they are
imaginings formed with the purpose of directly producing a representation
of how it feels, say, to be in pain. This may also explain why discussions of
imagining typically do not address the issue of the actuality or possibility
of non-representational imaginings.
424
In the remainder of this chapter, I aim to do this with respect to spontaneous images and thoughts. In line with the specification provided in Section 5.6, spontaneous representations can be characterised as non-cognitive
sensory or intellectual representations that occur entirely unbidden or unwilled in our minds. This means, minimally, that they lack a cognitive attitude and the epistemic function characteristic of cognitions; and that they
arise in an unsolicited manner, that is, do not involve any agency or effort
on behalf of the subject (see Section 5.6; see also Casey (1976): 34f.; 68f.).
They are thus passive through and through, and not determined by any of
our motivational states.
Examples are thoughts that simply cross our minds, or images that appear unbidden when we close our eyes (such as hypnagogic representations). In this respect, they resemble perceptual experiences and strictly
differ from compulsive or seemingly involuntary imaginings, which we
actively bring about, though possibly sub-intentionally and contrary to
what we really want or take to be the best course of action (see Section
13.3). As Casey has observed, the fact that spontaneous episodes arise unwilled and without the involvement of any effort has also the consequence
that they have the capacity to surprise us. For we do not know in advance
what they will represent and of what type of representation they will be; or
even that they will appear in the first place (Casey (1976): 34; 63).
Spontaneous representations may still originate in some of our motivational states, as long as these merely cause their occurrence, that is, determine them only by some non-rational and possibly unconscious process. More often, however, the occurrence of spontaneous representations
is due to associative links, mental dispositions, and similar phenomena.
Non-mental causes, such as certain drugs or neural processes, may bring
them about as well.
Another important feature of spontaneous images and thoughts is that
they are by their very nature fleeting and elusive in character or, as Casey
puts it, they occur instantaneously [...] and without any sense of drawn-out
duration (Casey (1976): 34). In particular, they are typically transitory and
do not involve any of the development for instance, with respect to which
entities they represent, and how that active and deliberate imagining may
include (see Casey (1976): 34f.; 70f.). As a consequence, they usually dis-
425
See OShaughnessy (2000: 346ff., 479), who also mentions Macbeths vision of
Banquo which, however may instead be taken to be a perfect perceptual hallucination with a cognitive attitude (or a full sense of reality in OShaughnessys
terminology; see Dorsch (2010a) and (2012a)). It may not always be easy to decide
whether cases like Macbeths visions indeed lack any agency on behalf of the subject. Some seemingly spontaneous representations may in fact involve such activity
which is, however, not acknowledged by the subject as agency, or not as his own
agency (see Roessler (2001) on inserted thoughts). Cases in which the agency is acknowledged as ones own, but is in conflict with some of ones further desires or intentions, are different. Someone may be obsessed by a certain thought and not be
able to banish it from his mind, but still recognise himself as actively thinking it, or
even wanting to think it (see the involuntary imaginings discussed in Section 13.3).
426
427
428
cognise reality, at least not in the same ways as perceptions and beliefs (see
OShaughnessy (2000): 344f. and 349ff.); or that they do not involve a
cognitive attitude (see Sartre (1940): 12f.; 181f. who distinguishes there
sensory imaginings from both perceptions and sensory memories).25 The
respective negation claims about spontaneous representations may very
well be true. As a consequence, spontaneous images and thoughts may
resemble imaginings more than cognitions, not only in their easy
subjection to the will, but also in their phenomenal character and their role
in cognition.
But, as I have argued at length in Part Two, the non-cognitivity of imaginings is neither unique to them, nor responsible for their imaginative
status. Hence, the similarity between spontaneous and imaginative representations in this respect does not suffice to justify the classification of the
former as a subclass of the latter. Moreover, spontaneous representations
do not possess any other distinctive features that might plausibly be said to
be sufficient for imaginativeness. In particular, they share their passive occurrence and their origin in mental or neural causes with many cognitive
representations, such as perceptual hallucinations and associatively
triggered episodic memories. The claim that spontaneous representations
are imaginative remains, therefore, completely unfounded. Instead, they
should just be taken to be very similar to imaginings in important respects
a fact which the Agency Account of imagining can satisfactorily explain.
It is important to note that this line of reasoning applies not only to normal and short-lived instances of spontaneous representation, but also to
spontaneous images and thoughts that are the result of mental disorders,
subconscious forces or hallucinogenic substances and which may stay in
existence for longer periods of time, and usually against our will. The latter
are equally non-cognitive in virtue of their lack of a cognitive attitude and
their incapacity to cognise reality. But again, and for the same reasons, this
25
Casey does not provide any specific support for his position. Sartre notes also similarities between spontaneous images and cases of active visualising (e.g., their quasiobservationality, or their negating attitude towards their objects; see Sartre (1940):
ch. 1). But these similarities hold between visual imaginings and visual memories,
too (as Sartre accepts at Sartre (1940): 10; 181), and hence are not relevant for the
imaginativeness of episodes.
429
is not enough to render them imaginative. 26 The Agency Account is therefore able to accommodate, not only the imaginative counterparts of nonrepresentational and non-imaginative phenomena, but also cases of what
might wrongly seem to be passive imaginings notably spontaneous images and thoughts.
Conclusion
26
Wittgenstein also seems to embrace the view that these kinds of states are non-imaginative when he writes: Vorstellungen sind nicht Halluzinationen, auch nicht Einbildungen [Imaginings are not hallucinations, nor figments the author] (Wittgenstein (1984d): sec. 621, and (1984b): vol. II, sec. 63). See Dorsch (2010a) for more
on the nature of, and distinction between, short-lived and persistent spontaneous images and thoughts, as well as their relationship to imaginings and cognitions.
Conclusion
432
Conclusion
our voluntary control over what is represented is not mediated by the reliance on epistemic or merely causal processes as means. The resulting account can distinguish imaginative episodes from their cognitive counterparts and other non-imaginative episodes, given that the latter allow at best
for an indirect voluntary determination of their content. It also can distinguish imaginative projects from other kinds of mental projects, notably
from cognitive ones, which aim instead at the formation of cognising representations.
More generally, while the explanatory power and fundamentality of the
Agency Account with respect to the unity of imagining should be fairly uncontroversial, its extensional adequacy needs to be further defended against
several potential counterexamples. On the assumption that cognitive and
other non-imaginative representations do not allow for the voluntary and
direct determination of their content, my discussion has been primarily
concerned with cases that appear to be either non-representational, or passively or indirectly determined instances of imagining.
The best candidates for non-representational imaginings are imaginings
with non-representational counterparts. Plausible examples are imagined
pains or imaginatively felt anxiety assuming, for the sake of argument,
that genuine feelings of pain or anxiety do not show any form of
representationality. However, such cases are most fittingly understood as
instances of experiential imagining: namely as representations of their nonimaginative counterparts. Imagining a pain thus amounts to imaginatively
representing the occurrence of a real pain.
The array of potential examples of passively or indirectly determined
imaginings, on the other hand, is more varied, ranging from spontaneous
images and thoughts via obsessive mental representations to pictorial experiences or non-purposive associations (although the last two are not
really discussed in this book). The already mentioned argument to the effect that the direct mental agency involved in voluntary imagining is essential to its instances entails, however, that there is a basic difference in
nature between active imaginings and all passive representations, indicating that the latter belong to a different and, presumably, non-imaginative
mental kind. Hence, none of the listed phenomenal poses a threat to the
Agency Account of imagining. Spontaneous representations, for instance,
Conclusion
433
See Dorsch (2009b) for further and independent arguments and observations in
support of the phenomenological salience of direct (but not indirect) mental agency.
434
Conclusion
Conclusion
435
damental.
The focus on the non-cognitivity of imaginings is problematic since
characterising imaginings solely in terms of, and in contrast to, cognitive
phenomena significantly limits the scope of the Epistemological and the
Dependency Account. For certain forms of imagining notably affective
imaginings and imaginative projects do not have cognitive counterparts
and, hence, resist specification in terms of them. As a result, the relevant
theories turn out to be explanatorily unilluminating if not extensionally
inadequate with respect to certain central cases of imagining. For instance, it is not a substantial truth about affective imaginings that they cannot constitute knowledge, because this applies to all kinds of affective
states, including non-imaginative ones. And that emotional imaginings are,
indeed, representations of non-imaginative experiences does not distinguish them from, say, episodic memories and, moreover, does not amount
to a characterisation of them in terms of cognitive phenomena.
However, accounting for imaginings instead by reference to both cognitive and non-cognitive phenomena threatens to lead to a disjunctive theory:
while the nature of some imaginings (e.g., sensory or intellectual ones)
would be elucidated in terms of how they differ from or depend on cognitions, the nature of others (e.g., affective imaginings) would be described
by mention of how they differ from or depend on cognitions. The pair of
intellectual and affective imaginings shows that this would result in some
form of disjunctivism, given that intellectual imaginings depend merely
generally and semantically on judgements or beliefs, whereas affective
imaginings are representationally dependent on specific emotional feelings.
Such a disjunctive treatment is, by itself, unproblematic but becomes untenable once it is presumed to provide a unified account of imagining.
In addition, the Epistemological Account and, in particular, OShaughnessys rich and sophisticated version of it faces objections that are more
specifically directed at its central claim that imaginings lack a cognitive
concern with reality, or a closely related cognitive feature. To begin with,
the Epistemological Account possesses only limited explanatory power.
For postulating such an absence of a cognitive property does not say anything positive about what it means to imagine something which is in part
why OShaughnessy, as the principal proponent of the Epistemological Ac-
436
Conclusion
count, supplements his theory with a representational version of the Dependency Account.
Another problem is that, depending on which cognitive feature imaginings are said to lack, the respective versions of the Epistemological Account are either false, or uninteresting, with respect to the ambition to
provide a unified account. Some are interestingly false because they cannot
allow for the already noted possibility that sensory imaginings may, under
suitable circumstances, constitute knowledge and possess the required cognitive features (see above and Chapter 4). Accordingly, imaginings are not
always deprived of the properties of being reliable, informative, grounded
in perception, rationally integrated with our picture of the world and, possibly, constitutively linked to the relevant aspects of reality. The other versions of the Epistemological Account, by contrast, are uninterestingly true
because their denial that imaginings do not show certain other cognitive
features does not distinguish imaginings from other non-cognitive phenomena. The observation that imaginings lack a cognitive attitude, have no epistemic function and possess no prima facie justificatory power with respect
to first-order beliefs does not reveal much about their nature, given that the
same may be discerned, say, in emotional or conative states and in nonimaginative mental projects.
On the other hand, much of the Dependency Account notably in its
representational version can be preserved, not the least its identification
of a core divide, on the level of sensory and affective episodes, between
original or genuine occurrences and mere representations or imitations of
such occurrences. This crucial distinction has already constituted a central
element of Humes approach to (non-intellectual) imaginings and is as illustrated in Part Three still present in many contemporary theories, the
history of which can often be traced back to Humes theory of the mind.
However, the Dependency Account does not apply to intellectual imaginings and imaginative projects and should therefore be rejected as a theory
of the unity of imagining.
Moreover, the main truths noted by the Dependency Account as well
as by the Epistemological Account can be accommodated and explained
by the Agency Account. For example, visualising a tree amounts to imagining seeing a tree in part because the only direct way of actively bringing
Conclusion
437
438
Conclusion
Besides, any ontological difference between imagined and cognised objects would
very likely to be linked to and perhaps even due to some co-extensional difference in how imaginings and cognitions relate to the world or the will (e.g., if imagined objects would turn out to be mind-dependent). The investigation of the latter
difference might thus render the investigation of the former at best supplementary,
and at worst superfluous.
Conclusion
439
Finally, the Subjection to the Will Account (which was discussed in Section 13.5 rather than Chapter 5) is a watered down cousin of the Agency
Account, given that it demands the possibility rather than the actuality
of voluntary and direct control over what imaginings represent. But, partly
because of this weakening of the requirement on the involvement of imaginative agency, it is to be rejected for three reasons. First, it cannot pay
justice to the fact that directly determined mental actions are essentially
active in this way. Second, it wrongly locates the difference between imaginative and non-imaginative phenomena in the ability to engage in imaginative activity, rather than in the nature of that activity. Third, it wrongly
turns the question of whether an episode is an instance of imagining into an
essentially empirical question, thus ignoring the phenomenological salience of the difference between imaginative and non-imaginative phenomena.
440
Conclusion
the Attitude Account may have much to say about the nature of imaginative
episodes (if not also of complexes thereof) and, in particular, about the distinctive commitment of imagining.
It would also be interesting to investigate the extent to which mental
phenomena, that are neither clearly imaginative nor clearly non-imaginative, involve or are at least closely related to imagining. I already briefly
listed examples of such phenomena in Section 1.3 and discussed there also
why their status is rather undecided. But the truth of the Agency Account
may have interesting consequences for them, which are worthwhile of further exploration. Pathological or psychologically unusual phenomena (e.g.,
intrusive or inserted thoughts, compulsive representations, delusions, hallucinations, etc.) may turn out to involve some form of suppressed, non-deliberate or unacknowledged imaginative activity. Hopes, wishes, expectations, anticipations or speculations may likewise involve imagining, perhaps to the extent to which they mark or recognise the concerned non-actual states of affairs as unrealistic or unlikely to occur (i.e., to the extent to
which they fail to take the represented states of affairs to be present in the
actual future). Finally, although experiences of grouping phenomena, or of
noticing aspects, allow for voluntary switches between experiencing one
thing and experiencing another (or, perhaps more precisely, between
experiencing something as one thing or experiencing it as another), it is unlikely that they involve some imaginative activity as part of, or in addition
to, active attention. For which groupings or aspects we can choose between
does not seem to be up to us. In particular, we cannot decide to see a horse
in the famous duck-rabbit picture, just as we cannot decide to see a horse
in an unambiguous picture of a duck (see Dorsch (2012c)). And this list of
applications of the Agency Account to borderline cases can surely be extended much further.
Besides, it is very appealing to bring the conclusions about the distinctive nature of imagining to bear on relevant issues in other areas of philosophy. As mentioned in the Introduction to this book (see especially Note
2), the imagination is a central ingredient in accounts of such diverse phenomena as thought experiments, modal knowledge and knowledge of
counterfactuals, our simulation of and empathy with other people, pretense
and acting, games of make-believe, the aesthetic appreciation of artworks
Conclusion
441
Bibliography
In this book, I discussed three specific applications of the imagination to other philosophical issues: to the acquisition of factual knowledge (see Chapter 4); to the acquisition of conditional and modal knowledge (see Sections 2.1 and 2.3, as well as
Williamson (2008)); and to our (emotional) engagement with art and fiction (see
Chapter 11, as well as Walton (1990) and Moran (1994)). I intend to elaborate on
these and other utilisations of the imagination in epistemology and aesthetics in a
future monograph.
Bibliography
444
Bibliography
Bibliography
445
446
Bibliography
Nida-Rmelin & F. Dorsch (eds.), Phenomenal Presence. Oxford: Oxford University Press.
Dretske, Fred (1979). Simple Seeing, in: F. Dretske (2000), Perception,
Knowledge, and Belief: Selected Essays. Cambridge: Cambridge University Press.
(1981). Knowledge and the Flow of Information. Cambridge (Mass.):
The MIT Press.
(1986). Misrepresentation, in: S. Stich & T. Warfield (eds.) (1994),
Mental Representation. Oxford: Blackwell.
(1995). Naturalizing the Mind. Cambridge (Mass.): The MIT Press.
Eco, Umberto (1994). Reflections on The Name of the Rose. London:
Minerva.
Evans, Gareth (1982). The Varieties of Reference. Oxford: Clarendon
Press.
Fish, William (2009). Perception, Hallucination, and Illusion. Oxford: Oxford University Press.
Fodor, Jerry (1990). A Theory of Content, in his: A Theory of Content
and Other Essays. Cambridge (Mass.): The MIT Press.
Frege, Gottlieb (1884/1980). The Foundations of Arithmetic. Evanston:
Northwestern University Press.
(1892). ber Sinn und Bedeutung, in: Zeitschrift fr Philosophie und
Philsophische Kritik, NF 100, pp. 25-50.
(1918), Der Gedanke. Eine logische Untersuchung, in: Beitrge zum
deutschen Idealismus 2, pp. 58-77.
Furlong, E. J. (1961). Imagination. London: George Allen & Unwin.
Gendler, Tamar (2000a). Thought Experiment: On the Powers and Limits
of Imaginary Cases, New York: Garland Press.
(2000b). The Puzzle of Imaginative Resistance, in: Journal of Philosophy, vol. 97, pp. 55-81.
(2006). Imaginative Resistance Revisited, in: S. Nichols (ed.), The Architecture of the Imagination: New Essays on Pretence, Possibility, and
Fiction. Oxford: Oxford University Press.
(2011). Imagination, in: E. Zalta (ed.), The Stanford Encyclopedia of
Philosophy (Fall 2011 Edition), http://plato.stanford.edu/archives/
fall2011/entries/imagination/.
Gendler, Tamar & Hawthorne, John (2002). Conceivability and Possibility.
Oxford: Oxford University Press.
Giaquinto, Marcus (1992). Visualizing as a Means of Geometrical Discovery, in: Mind and Language, vol. 7, pp. 382-401.
Bibliography
447
448
Bibliography
Bibliography
449
450
Bibliography
Bibliography
451
452
Bibliography
Siegel, Susanna (2010). The Contents of Visual Experience. Oxford: Oxford University Press.
Siewert, Charles (2000). The Significance of Consciousness. Princeton:
Princeton University Press.
Soames, Scott (1999). Understanding Truth. New York: Oxford University
Press.
Soldati Gianfranco (1994). Bedeutung und psychischer Gehalt. Zur sprachanalytischen Kritik von Husserls frher Phnomenologie. Paderborn:
Schningh.
(2006). Review of Currie & Ravenscrofts The Recreative Mind, in:
European Journal of Philosophy, vol. 14, pp. 448-452.
Soldati, Gianfranco & Dorsch, Fabian (2005) Conceptual Qualia and
Communication, in: Interaction Design Institute Ivrea (ed.), The
Foundations of Interaction Design, pp. 1-14.
(2011). Intentionalism, Phenomenal Error and Experiential Error. Manuscript.
Sorensen, Roy (1992). Thought Experiments. Oxford: Oxford University
Press.
Stock, Kathleen (2008). The Role of Imagining in Seeing-In, in: Journal
of Aesthetics and Art Criticism, vol. 66, pp. 365-380.
Strawson, Galen (1994). Mental Reality. Cambridge (Mass.): MIT Press.
(2003). Mental Ballistics or The Involuntariness of Spontaneity, in:
Proceedings of the Aristotelian Society, vol. 103, pp. 227-256.
Strawson, Peter (1970). Imagination and Perception, in: L. Foster and J.
W. Swanson (eds.), Experience and Theory. London: Duckworth.
Stroud, Barry (1977). Hume. London: Routledge.
Thomas, Nigel J. T. (1999). Are Theories of Imagery Theories of Imagination?, in: Cognitive Science, vol. 23, pp. 207-245.
(2010). Mental Imagery, in: E. Zalta (ed.), The Stanford Encyclopedia
of Philosophy (Winter 2010 Edition), http://plato.stanford.edu/archives/
win2011/entries/mental-imagery/.
Travis, Charles (2004). The Silence of the Senses,in: Mind, vol. 113, pp.
57-94.
Tye, Michael (1991). The Imagery Debate. Cambridge (Mass.): MIT Press.
(2000). Consciousness, Color, and Content. Cambridge (Mass.): MIT
Press.
Velleman, David (2000). The Possibility of Practical Reason. Oxford: Oxford University Press.
(2006). Self to Self, in his: Self to Self: Selected Essays. Cambridge:
Bibliography
453
454
Bibliography
Index
456
Index
agency (contd.)
non-mental 179, 386-7
sub-intentional 93, 111, 143, 249-50, 392, 401, 424, 440
voluntary 2, 8, 10-1, 13-4, 63, 101-2, 112, 138-44, 146, 157, 178-82,
184-7, 199-200, 210, 247, 357-8, 364, 372, 381-2, 385, 388-90,
392, 394, 397, 399, 404-6, 413, 431-3, 437-40
Agency Account 2, 8, 10-1, 13, 17, 19-20, 24, 30-1, 55, 83, 85, 104-6,
111-2, 139, 143, 146, 149-50, 176, 188, 192, 247, 262, 271-2, 279,
364, 370, 378-9, 381-429 passim, 431-4, 436-7, 439-41; see also:
(ACT); Constitutive View of voluntary imaginings
Alain 140
anaesthesia see: states of consciousness
anticipation 51, 55-7, 354, 374, 440
anxiety 260, 421, 432
Argument from Attention 214, 218, 221, 247
Argument from Origin 200-1, 214, 216, 219, 221-2, 242, 247, 251, 257,
268, 271, 378
Armstrong, D. 54, 394
art 2, 46, 53, 328, 339-41, 343, 345-7, 349-51, 355-6, 360, 363-4, 440-1;
see also: aesthetic appreciation
assertion 35, 63, 72, 375
association 11, 93, 104, 110-1, 140, 145, 179-80, 184, 227, 249, 266, 281,
393, 405, 419, 424, 428, 432, 439
assuming 42-3, 64, 68
attention 20, 39-40, 47, 52-3, 72, 91-3, 124-5, 138, 140-1, 174, 182, 195,
197, 214, 218, 221-5, 228, 247-51, 318, 323, 349, 376-7, 390, 392,
404, 425, 433, 439-40; see also: Argument from Attention
object-directed 92, 178, 222
attitude/commitment 12-3, 18, 23, 32-4, 59-68, 73, 75, 83-4, 87, 129, 143,
149-52, 154, 156-8, 161-78 passim, 184-5, 187, 206, 224, 243, 245,
254, 296-7, 303, 307, 333-4, 351, 368, 377, 407, 412, 416, 422,
437-8, 440
conative see: attitude, practical
cognitive 13, 17, 54, 59, 62-4, 72-3, 80, 83, 128-9, 144, 163, 166,
169-70, 172-4, 238, 245, 258, 260, 267-8, 271, 290, 294, 297, 305,
Index
457
458
Index
real vs. imagined itches 4, 41, 50, 61, 306, 315, 318, 322-3, 361, 4201, 423
bodily sensation (contd.)
real vs. imagined pain 36, 41, 82, 85, 150, 292, 315, 322, 338, 361-2,
386, 421-3, 432
Bodrozic, D. 154, 156-157
Braddon-Mitchell, D. 71
brain see: neural/neurofunctional
Budd, M. 5, 10, 45, 48, 59, 73, 82, 143, 145, 180, 270, 277, 382, 394
Burge, T. 48, 69, 71, 294, 325-6, 329-30, 366
Campbell, J. 316, 324
Carroll, N. 392
Carruthers, P. 41, 47, 370
Casey, E. 1, 3-4, 12-13, 20, 37, 42, 46, 73, 80, 85-7, 139, 152, 155-6, 164,
172-3, 177-8, 180-3, 185-6, 188, 269, 405, 424, 426-8
Causal Account see: imagining, accounts of
Causal View of voluntary imaginings 398-403
chess 261, 387
children 44, 330-1, 374
cognising episodes/states 3, 12, 18, 24, 60, 71, 73, 81, 115-8, 122-5, 1279, 131-4, 136-7, 141, 144, 146, 154-5, 195, 199-200, 202, 205-6,
211-6, 221-2, 226, 228-32, 234, 236, 239-41, 244-5, 257-65, 267-8,
270-2, 293, 305, 311, 378, 381-2, 396, 410-4, 428, 432, 434-6, 438;
see also: cognition; knowledge; perception; recognition
cognition 8-9, 32, 56, 59, 62, 70-2, 84, 103, 115, 128-9, 149, 154, 187,
195, 199, 203, 207, 245, 259-62, 296, 298, 300-1, 428, 433, 437;
see also: cognising episodes/states; perception; knowledge; recognition
cognitions 8-10, 12-3, 18-20, 23-4, 27, 30, 34-5, 56, 59-64, 68-73, 80, 845, 115, 124, 128-9, 134, 143, 146, 149, 151-7, 159-63, 165-7, 16970, 172, 175, 177, 187-8, 192-3, 195, 199, 201, 208, 211, 213-5,
218-20, 222-3, 231, 241-5, 247, 254-5, 258-9, 262, 265-8, 272,
275-9, 288, 293-302, 306, 308, 310, 365-73, 375-7, 381, 389, 407,
410-3, 424, 426-9, 433-5, 437-8; see also: belief; judgement; per-
Index
459
460
Index
248-9, 264, 275, 277, 285, 287, 293, 295-300, 303-7, 309, 334-5,
339-40, 343, 347, 351, 353-4, 356-61, 365, 368, 370-1, 388-90,
393, 395, 397-8, 404-5, 411-2, 414-8, 420, 422, 425-7, 432, 439,
441; see also: mental representation
conventions 7, 328, 346-7, 349
copies see: dependency of imaginings on cognitions
counterfactuals 3, 66, 310, 369, 440
counterparts see: echo claims
Crane, T. 33, 62, 96, 168, 260, 421
creativity 3, 24, 32, 46-7, 64, 177-8, 184, 307, 338, 383; see also: imaginativeness
Crowther, T. 182
Cruz, J. 69-70, 240, 256, 305
Currie, G. 3-4, 9, 15, 20, 37, 41-4, 46, 59, 84, 276-7, 304, 326, 330, 3345, 337, 353, 366, 368-71, 374-5, 395
Dancy, J. 270, 399
Danto, A. 395
Davidson, D. 240
daydreaming see: imagining, forms of
delusions 15, 54-6, 216, 378, 440
Dennett, D. 47
dependency of imaginings on cognitions; see also: echo claims
asymmetric 9, 370-2
causal (EC) 284-5, 288, 292-4, 296, 298, 365, 368; see also: imagining, accounts of
conceptual 369-71, 373
constitutive 272, 301, 369-71, 373
copies/reproductions 80, 84, 87, 159, 206, 208, 279, 282, 285, 287-9,
291-3, 298, 300, 302, 410; see also: Humes Copy Principle; pictures
imitational (EI) 9, 19-20, 205, 275-7, 280, 365, 368-9, 373, 375-7,
436; see also: imagining, accounts of
representational (ER) 84, 193, 197, 201-2, 205-6, 208-10, 212-3, 216,
220, 247, 279-80, 288, 292, 294-311, 313-6, 322, 325-8, 330-1,
Index
461
334-5, 337, 362, 364-5, 370, 377; see also: imagining, accounts of
dependency of imaginings on cognitions (contd.)
semantic (ES) 9, 146, 275-6, 280, 366-71, 393, 411; see also: imagining, accounts of
types vs. tokens 277-8, 286-7, 298-301
Dependency Account 9-11, 18-9, 30, 84, 149, 159, 188, 192-3, 197, 2002, 207, 213, 220, 255, 272-3, 275-80, 282, 295, 377-8, 381, 431,
433-7; see also: dependency of imaginings on cognitions; echo
claims
depiction see: pictures
Descartes, R. 240
desire 5-6, 9-10, 33-35, 41, 44, 52, 56, 67, 71, 73-4, 133, 146, 163, 209,
227, 260-1, 282, 293-294, 297, 303-4, 307-308, 337, 359, 370-372,
389, 405, 411, 415-6, 420, 425; see also: intention; mental representation; volition
Devil 168
direct agency see: agency
direct object/reference see: mental representation
disjunctivism 24, 33, 47-9, 89-90, 153, 258, 263, 305, 314, 334-5, 399,
435
disorders, mental see: states of consciousness
dispassionate imagining 266, 339, 351, 353, 356, 358-9, 361
dramatic imagining see: imagining, forms of
dreams see: states of consciousness
Dretske, F. 33, 71, 82, 95, 120-121, 128-129, 134, 153, 223, 233, 248
drugs see: intoxication
(EC) see: dependency of imaginings on cognitions
echo claims 34, 77, 84, 159, 171, 192-3, 197-8, 202, 205, 208-9, 212-3,
216-20, 255, 275-80, 288, 292-5, 298-9, 301-4, 307-10, 365-6, 3689, 371, 377, 422; see also: Dependency Account; dependency of
imaginings on cognitions
idea of counterparts 4, 9, 12, 19, 27, 37, 41, 59, 73, 80-1, 133, 149,
160, 173, 193, 195, 200, 204, 216, 243-4, 257, 260-1, 265, 272,
275-9, 288, 291, 293, 295, 300, 337, 365, 367, 369-71, 373, 377,
462
Index
Index
Evans, G. 134, 146, 258
evolution 71-2, 120, 129, 406, 411
examples
calculating a sum in ones head 130, 261, 386
chiliagon 406
Columbus 419
counting windows 3, 124, 393
Devils tuning fork 168
elephant in the city 64-6
field of daffodils 247-8
frogs and their lips 3, 125, 137, 141-2
Hamlet 341
hearing voices 54
Hercules 328
Hertford College 94-9, 103-4, 106-7, 411
imposter 95, 374
Jean of Arc 328
Macbeth 6, 155, 216, 244-5, 425
missing shade of blue 290, 299, 366
Mller-Lyer illusion 110
Napoleon 44, 46-7, 50, 287, 332, 417
Odysseus 154, 162
old friend 116-33 passim, 144-5, 267, 269
Pantheon 140-1, 249-50
Pegasus 176
Perkys experiments 151
Pierre 102, 179, 186
pink elephant 224-5, 232
Rambutan 367
Rome in snow 35-6, 41-2, 44, 46, 141, 391
Saint John 328
Sherlock Holmes 52-3, 68, 156
sofa in the shop 116-33 passim, 144-5, 267, 269, 414
suitcase with a cat behind it 97, 99, 328
Twin Earth 366
463
464
Index
examples (contd.)
Ulysses 107
unicorns 154, 299, 386, 417
Waterfall illusion 168
Wellington 395
wax replicas/fakes 90-1, 108-11, 254, 277
Wozzeck in the Opera Bastille 73-5, 229
expectations 6, 42, 55-7, 163-4, 173, 440
experiential imagining 4, 17, 37, 44-7, 49-50, 84, 97-100, 207, 295, 298,
308, 313-364 passim, 419-20, 422, 432; see also: imagining seeing;
S-imagining
Experiential Rationalism 412
experiments see: empirical evidence
explanatory power 17, 27-9, 106, 108, 210, 253-6, 259, 261, 264-5, 268,
334, 396, 409-10, 432-3, 435, 437
fundamentality 2, 11-3, 15-6, 29-30, 33, 48, 59, 77, 101, 133, 149,
152-4, 156-7, 161, 165, 169, 176, 187, 192-4, 197-8, 200, 202, 204,
207, 210-3, 220, 254-7, 264-5, 268, 271, 275, 293, 295, 302, 304,
345, 365-6, 372, 378, 382, 390-1, 394, 396, 403, 405, 409-10, 4134, 417, 432, 437-8, 441
informativity 20, 28, 254-7, 371, 409-12
extensional adequacy 2, 17, 27-28, 33, 51, 210, 213, 224, 253, 261, 2667, 310, 396, 409, 414-29 passim, 432-3, 435
external world 3, 13, 18, 69, 80, 115-116, 134-5, 141-2, 144-6, 158, 169,
184, 194, 219, 224, 233-5, 237-9, 241, 248, 256, 258-9, 267, 416,
433-4; see also: actual world; reality
factivity see: mental representation
faculty of imagination 31-2, 281
faith 338
fancy/fantasy see: imagining, forms of
fear 33, 42, 55, 164, 342-6, 349-50, 355, 357, 361, 363-4, 374
ficta/fictional objects 41, 154, 340, 342, 344-345, 349-352, 356, 363-364,
437
fiction 6, 15, 36, 41-3, 46, 68, 71, 97, 146, 154, 156-7, 170, 283-4, 338-
Index
465
52, 355-6, 363-4, 410, 416, 437, 441; see also: paradox of fiction
fictional worlds/situations 97, 338, 340-1, 343, 345-9, 352, 356
filling-in 140-3
first-order thought see: thinking
first-personal see: perspectivalness
Fodor, J. 71
foundationalism 240
Frege, G. 162, 416
function, biological/evolutionary 71-2, 120, 411
Furlong, J. 42
games of make-believe see: making-believe
Gendler, T. 3, 15, 42-3, 162, 168, 392
generic memories 93, 137, 141, 145, 393; see also: memory
Gettier, P. 232, 241
Giaquinto, M. 3
Gill, A. 77
Ginet, C. 104, 399
God 416
Goldie, P. 2, 46, 163, 260, 345, 354
Goldman, A. 15, 368
Gordon, R. 309
Goya, F. 395
grounding knowledge see: cognising episodes/states
grouping phenomena see: noticing an aspect
guessing 11, 187, 327, 386, 388
hallucinations see: perceptual experience
Harris, P. 15
Hawley, K. 67
Hawthorne, J. 3, 162
Heal, J. 15, 368
hearing see: sense modalities
higher-order thought see: thinking; introspection
Hjort, M. 42, 343
466
Index
Hobbes, T. 110
hope 33, 51, 55-6, 163-4, 340, 343, 371, 440
Hopkins, R. 1, 3, 5, 12, 37, 39-40, 46-7, 65, 84, 92, 101, 111, 151, 158,
175-6, 201, 250, 269, 277, 287, 305, 307, 309, 320, 348-9, 382,
419, 423
Hornsby, J. 104, 399-400
Hume, D. 4, 8-9, 12, 18, 20, 32-3, 62, 79-82, 150-1, 159, 184, 194, 201,
208, 269, 275, 279, 281-94, 296, 298-302, 333, 436
Humes conception of imagining see: imagining, accounts of
Humes copy principle 281-94 passim, 299, 333
Husserl, E. 1, 13, 20, 33, 47, 62, 64, 79-81, 83-4, 152, 158-9, 169-71,
173-5, 281
hypnosis 231, 234, 236, 240, 389, 397, 415, 417
hypothetical imagining 353; see also: intellectual imagining
illusions see: perceptual experiences
images 3, 6, 11, 19, 34, 38, 44, 54, 64, 74, 77, 79, 82, 86, 88-9, 92, 94, 98102, 112-3, 117-9, 121-5, 128, 131-2, 134, 139-41, 145-6, 172,
175-6, 178-9, 181, 183, 186, 224, 232, 249, 282, 285, 287, 296,
301, 309, 320, 329, 333, 357, 367, 369, 376, 378, 386, 393-4, 3978, 401-2, 405-6, 409, 423-9, 432-3, 437-8
images and thoughts 423-9; see also: images; imagining; mental episodes/states; thinking
difficult to banish 55, 393, 425
fleeting/transient 113, 140, 182, 357, 424-6
hypnagogic 140, 249, 378, 424
instantaneous 35, 140, 181-2, 424
passive/involuntary 13, 82, 102-3, 138-43, 177-82, 184-6, 199-200,
244, 247, 272, 383, 394, 397, 399, 402, 415, 424-5, 427, 431, 438
persisting 140, 175, 243, 246, 429
spontaneous 423-9; see also: spontaneity
surprising/unexpected 113, 181-2, 249, 419, 424
sustainment 103, 113, 177, 244, 357, 425-7, 433
transformation 113, 425-7, 433
imagination as faculty see: faculty of imagination
Index
467
468
Index
Index
469
imagining (contd.)
simplest instances 47, 49-50, 99, 400
simultaneity with perception 38-41, 320-1
specificity in content see: imagining, concreteness
sub-intentional see: agency
surprising 131-2, 249-50
transparent see: transparency
unity 1, 5, 7, 11, 14, 16, 20, 57, 59, 105, 149-50, 152, 154, 167, 174,
176, 185, 294, 310, 359, 382, 431-3, 436, 441; see also: unified accounts of imagining
vagueness 81-2, 86, 406
veridical/correct see: veridicality
imagining, accounts of
Agency Account see: Agency Account
Attitude Account 13, 164, 167-70, 173, 175, 438, 440
Causal Account 18, 279; see also: dependency of imaginings on cognitions
Dependency Account see: Dependency Account; dependency of imaginings on cognitions
Epistemological Account see: Epistemological Account
Humes conception of imagining 18, 81, 279, 281-92, 294
Imitation Account 275-276, 365, 368-77; see also: dependency of
imaginings on cognitions
Modal Account 12, 24, 161-70, 176, 372, 438-9
Ontological Account 12, 24, 76, 154-6, 158-61, 167, 437-8
Origin Account 14
Phenomenal Account 12, 152-5, 169, 437
pictorial model 158-9, 170-1
Quantitative Account 11, 24, 150-1, 437
Representational Account 18, 30, 76, 159-60, 208, 275-313 passim,
370, 436; see also: dependency of imaginings on cognitions
Semantic Account 275, 366-8; see also: dependency of imaginings on
cognitions
Spontaneity Account 13, 24, 176-88, 438
Subjection to the Will Account 13, 18-9, 24, 143, 149, 383, 404-7,
470
Index
439
imagining, accounts of (contd.)
Subpersonal Account 14-8
unified accounts see: unified accounts
imagining, forms of
affective imagining see: affective imagining; emotional imagining
daydreaming 1-2, 4-6, 11, 30, 46, 53, 141, 269, 310, 337, 353, 357,
377, 385, 417-9
dramatic imagining 352-356
experiential imagining see: experiential imagining
fancy 43, 51, 55-6, 283
fantasy 4, 46, 175, 393
imaginative projects see: imaginative projects
imagined pain see: bodily sensation
imagined itches see: bodily sensation
intellectual imagining see: intellectual imagining
reverie 46, 72
sensory imagining see: sensory imagining; visualising
imagining seeing 5, 37, 45, 47, 49-50, 64, 97-8, 100, 207, 266, 298, 300,
313-35 passim, 348, 419-20, 436
imagining that see: intellectual imagining
imitation see: dependency of imaginings on cognitions
Imitation Account see: imagining, accounts of
immediacy see: imagining; mental representation
immunity to error through misidentification 79-81, 88, 94-6, 103, 106-8,
112-3, 131-3, 136, 139, 306, 396, 411-2
impossibilities 68, 168, 170
impressions vs. ideas see: Hume's Copy Principle
inclination see: motivation
indefeasibility see: justification
indeterminacy see: imagining
inference 15, 43, 66, 89, 103, 120, 123, 130, 185, 187, 246, 263, 270-1,
305, 316, 321, 353, 368-71, 376-7, 402, 407, 416, 418; see also: attitude, practical/theoretical
informativity see: explanatory power; new information
Index
471
472
Index
Index
473
240, 245, 248, 250, 254, 256, 258, 267, 276-8, 288, 297-9, 302,
306, 313-6, 318-23, 325-7, 333-4, 355, 361, 423
McCann, H. 104, 395, 401
McDowell, J. 90, 126, 146, 180, 184, 240, 258, 267
McGinn, C. 1, 3-4, 8, 10, 46, 52, 54, 155, 194, 269, 337, 357-8, 382
Meinong, A. 154
memory 2-3, 8, 10, 12, 24, 30, 32-5, 39-40, 48, 50, 54-5, 61, 63, 69, 73-4,
76-7, 79, 81-4, 87-8, 93-5, 98, 107, 111, 113, 117-9, 121-30, 136-7,
140-1, 144-5, 151, 153, 158-9, 164-5, 174-5, 185, 194-5, 199-201,
207, 214-5, 234, 241-2, 249, 257-8, 261, 267-9, 278, 280, 282-3,
285, 289-94, 296-9, 301, 309, 319, 321, 334, 353, 361-4, 367, 371,
383, 385-6, 388-90, 393, 395, 413-5, 417-9, 421-2, 427-428, 434-5,
437-9; see also: generic memory
mental agency see: agency
mental episodes 47-49; see also: imagining; mental representation
as indivisible 50, 400
as self-contained 49, 99, 176, 400, 416
complexes thereof 46-51, 68, 74, 100, 105, 262, 357-8, 396-403, 410,
417-420, 433, 440; see also: mental projects
difference in attitude 23, 59-68, 87, 152, 157-8, 162-76, 407
difference in degree/quantity 4, 11, 80, 143, 150, 290-1, 296, 418,
422, 437
difference in epistemic function 59-60, 68-73, 152, 169; see also: epistemic function/role
difference in kind/quality 4, 11, 59, 80, 108-9, 127, 143, 150-2, 251,
290-2, 296, 362, 422, 437
difference in nature 143, 177, 181, 188, 218-9, 222, 277, 399, 432
mental/natural kinds 5, 30-1, 150, 181, 201, 211, 340-1, 391, 398, 432
non-representational 19, 62, 82, 85, 260, 265-6, 340, 383, 388, 409,
418, 420-3, 429, 432
phenomenal difference 12, 48, 63, 73, 150-5, 157-9, 169, 177-9, 1845, 334-5, 437; see also: mental episodes, difference in attitude
introspective error 143, 151, 334-5
mental projects 46-50, 105, 115-118, 133, 180, 261-5, 355, 382, 415-6,
432-6
474
Index
Index
475
222, 224, 229, 276, 302-3, 307, 337, 339, 352-4, 356, 358-64, 4156; see also: propositions
mental representation (contd.)
relational/non-intentional 19, 49, 178, 187, 202, 206-9, 216-7, 258,
267, 276, 282, 284, 286, 298, 300-2, 361-3
representational element 34, 39, 93, 125, 159, 165, 171, 175, 293,
359-60, 377, 403, 416, 426-7; see also: images and thoughts; propositions
sensory vs. intellectual 32-3
teleological 71, 262
truth-evaluable 144-6, 195
vehicle 65
vivacity see: vivacity
mental rotation 66, 80, 88-9, 117-8, 130-2, 138, 145, 414
Meskin, A. 15, 168
Millar, A. 67, 90
Millikan, R. 71
Modal Account see: imagining, accounts of
modality 12-3, 61, 66, 161-4, 167-9, 175, 204-5, 327, 367, 391, 440-1
Mondrian, P. 270
moods 35, 255, 260, 266, 385, 388, 390, 402-3, 415, 421, 423
morality 2, 43-4, 338-9, 356, 392-3, 441
Moran, R. 20, 42, 337-339, 341, 345, 350-363, 441
motivation 9-10, 35, 42, 50, 59, 69-71, 100, 103, 142, 145-6, 175, 180-1,
185, 232, 238, 283-4, 318, 337, 339, 344, 346, 359, 363, 391, 3934, 410-1, 417, 424
Mulligan, K. 39
multiple use thesis 37, 98-100
negation claims 133, 193-5, 197-201, 205, 209-14, 216-20, 247, 251, 25369, 271, 279, 306, 308, 310, 410, 428; see also: Epistemological
Account
(NC) 194-5, 197, 200-2, 206, 212, 221, 253, 264-66, 268-71, 305
(NC*) 194-195, 197-202, 206, 210-2, 221, 247, 251, 253, 256-9, 2678, 271, 304, 306, 308, 378
476
Index
Index
477
478
Index
perspectivalness (contd.)
first-personal 48, 63, 113, 151, 160, 178, 186-7, 283, 314, 323, 348,
361, 406-7
hidden aspects of objects 80, 89, 99, 131-2, 138, 340
point of view 16, 40-1, 44-5, 80, 88-91, 131-2, 136, 317-8, 320, 325,
333, 342, 348, 354-5, 392
quasi-observational see: quasi-observationality
egocentric orientation 92, 316-8, 320, 322-7, 329-30, 332, 388
third-personal 47, 314, 348
visual 45, 88-9, 93, 108, 131-2, 136, 277, 306, 314-22, 325-7, 330,
332-3, 347-8, 355, 361
Phenomenal Account see: imagining, accounts of
phenomenal character/aspects 45, 47-9, 60, 62-3, 73, 81-2, 96, 152-4,
157, 169, 177-8, 183, 186-7, 201, 254, 283, 287, 289, 303, 310,
314-20, 322, 327, 329-30, 332, 337, 339, 361-3, 402-3, 412, 437
of judgements/beliefs 48, 303, 307
phenomenological salience 47, 59-61, 63, 65, 73, 81, 83-5, 129, 151, 157,
169, 175, 177, 179-80, 183-5, 188, 238, 245, 254, 272, 289-90,
297, 306, 339, 403, 407, 412, 433, 437-9
phenomenology 13, 49, 79, 143, 152-3, 157, 159, 169-70, 176, 186, 437
physics 15-7
physiology 249, 342-4, 425
pictorial experience 5, 11, 39, 45, 51, 64-5, 76, 101, 151, 159, 171, 175-6,
205, 210, 266-7, 309, 347-50, 358, 378, 423, 432, 438-9; see also:
pictures
imagination-based accounts 39, 347-50
as imaginative perception 205, 349
twofoldness 349-50
pictures 5, 34, 39, 64-5, 67, 69, 74, 76-7, 80, 91, 102, 111, 140, 158-60,
168, 171, 175-6, 210, 266-7, 270, 286-7, 298, 309, 318, 321, 328,
333, 335, 338, 347-50, 358, 395, 423, 440; see also: pictorial experience
ambiguous 358, 440
content 110-1, 175, 327-8, 441
internal/mental 34, 76, 88, 158-60
Index
479
pictures (contd.)
paintings 39, 111, 159, 176, 270, 287, 298, 328, 333-5, 347-8, 389
photocopies/reproductions reproductions 287-9, 333-5
photographs 102, 121, 130, 287-8, 333
portraits 158, 298, 328, 338
representing types vs. tokens 286-7
Pietroski, P. 400
Pink, T. 104, 139, 388, 415
pleasure 36, 41, 50, 63, 283, 342, 344, 385
Pollock, J. 69-70, 240, 256, 305
positing see: attitude
possibilities 12, 56, 66, 108, 161-4, 309, 352
possible worlds/situations 12, 67-8, 73, 155, 161-70, 319, 321, 324, 366,
405, 438
practical attitude/reasoning see: attitude
pretense 4, 9, 15, 89, 280, 365, 373-7, 440
propositions 34, 43, 45, 68, 162-3, 166-7, 205, 207, 270-1, 276, 285, 300,
303, 305, 307-9, 343, 347, 350, 353-4, 359, 370, 376-7, 388, 415-6,
425
Priest, G. 167
priming 110
proprioception see: sense modalities
propositional imagining 207, 212, 215, 229, 352-4, 359-61, 364; see also:
intellectual imagining
Pryor, J. 240
psychology 14
Putnam, H. 48, 294, 366
Pylyshyn, Z. 3, 15, 124
Quantitative Account see: imagining, accounts of
quasi-emotion 342-4, 346, 349-52, 356-8, 360, 363-4
quasi-observationality 18, 79-81, 86-94, 106, 108-9, 111-2, 131-3, 136-9,
156, 396, 410, 431
rational integration 18, 126-9, 172, 183, 187, 231-42, 246, 256-8, 260-1,
480
Index
Index
481
Routley, R. 154
Russell, B. 157
Ryle, G. 1, 4, 9, 20, 34, 194, 276, 281, 373-6
sadness 125, 338, 354, 356, 389
Salmon, N. 154, 156
Sartre, J.-P. 1, 3, 8, 12-3, 20, 24, 34, 38, 62, 64, 73, 79-80, 86-92, 94, 96,
100-3, 116, 134-5, 138-40, 142-4, 151-2, 154-6, 158-9, 170-5, 17781, 183-6, 188, 193-5, 253, 258, 269, 272, 281, 320, 382-3, 394,
427-8
science 14-7, 33, 137, 248-9, 338; see also: empirical evidence; experiments
Scruton, R. 1, 3-4, 6, 10, 42, 46, 51-2, 54, 61, 151, 180, 304, 337, 357-8,
367, 369, 381-2, 394, 423
sculpture 85-6, 171
Searle, J. 33, 68, 103, 395
Sellars, W. 32
Semantic Account 287
seeing an aspect see: noticing an aspect
seeing-as 52
sense-data 49, 65, 160, 198, 206-7, 223-7, 259, 302
sense modalities 37-8; see also: bodily sensation
hearing 33, 35, 84, 207-8, 307, 348, 353-4, 359, 391
kinaesthesia 38
proprioception 38
smell 367
taste 71, 270, 286, 289, 293, 327, 367
touch 38, 88, 176, 353, 359, 367, 391
vision see: perception; perceptual experience; perspectivalness; visualising
sense of reality 243-4, 286, 289, 425
sensory imagining 4, 9-10, 12, 15, 17-8, 23, 27, 36-9, 41, 50, 54, 62, 65,
67, 74, 76-7, 79-98, 100-8, 111-3, 115-6, 118-9, 121-2, 127-31,
133-46, 151-2, 155-6, 158-9, 161, 165-6, 170-5, 177-80, 184-5,
199-202, 205, 210, 214, 217-8, 221-2, 246-51, 256-7, 263, 267,
482
Index
276-8, 292, 297, 299, 301-2, 305, 307, 313-4, 316, 325-7, 348, 3524, 357-8, 366-7, 369, 376-7, 382, 391, 393, 396, 410, 428, 431,
436; see also: affective imagining; visualising
Shakespeare, W. 341
Shorter, J. 373-374
Siegel, S. 144
Siewert, C. 47, 303
S-imagining 96-8, 100, 103, 105
simulation 4, 9, 19, 271, 275, 280, 365, 368-73, 377, 440; see also: imitation; Imitation Account
simulationalism 15, 371
sleep see: states of consciousness
smell see: sense modalities
Soames, S. 162
Socrates 287
Soldati, G. 41, 48-9, 73, 75, 85, 96, 169, 178, 187, 245, 259, 303, 307,
318, 370, 403
somnambulism see: states of consciousness
Sorensen, R. 3
Soteriou, M. 180
space 11-2, 39-40, 91-2, 99, 117, 145, 154-6, 173, 235, 316-7, 319-24,
330; see also: perspectivalness
spontaneity 6, 11, 13-4, 19, 24, 51, 100-4, 138-9, 175-88, 296, 301, 309,
357, 382-3, 393-4, 397, 402, 409, 420, 423-9, 432-3, 437-8
Spontaneity Account see: imagining, accounts of
Stich, S. 15, 374
Stock, K. 51, 358, 423
Strawson, G. 303
Strawson, P. 32
Stroud, B. 150
subjection to the will 82, 142-3, 146, 181, 243, 357-8, 404-5, 426, 428
Subjection to the Will Account see: imagining, accounts of
subjective access/indistinguishability 12, 16, 47-9, 54, 63, 87, 89, 127,
151-2, 183, 216, 245, 254, 283, 289-91, 297, 299, 339-40, 361-3,
378, 402; see also: introspection
Index
483
subjectivity 13, 60, 139-40, 159, 177, 179-80, 183, 185-6, 222, 242, 254,
283-4, 289, 315-6, 320-7, 329-30, 332, 335, 338, 342, 348, 350,
361, 438; see also: perspectivalness; phenomenal character; subjective access
subpersonal phenomena 14-8, 93, 104, 110, 152, 160, 179, 185, 428; see
also: neuronal/neurofunctional processes
Subpersonal Account see: imagining, accounts of
supposing 1-6, 12, 15, 30, 34, 42-3, 48, 61, 63, 69, 73, 96, 137, 150, 155,
164, 229, 270-1, 292-3, 295, 298, 307, 328, 351, 353, 359, 370,
416-7, 420, 433
states of consciousness/mind
anaesthesia/coma 230, 234
disorders/pathologies 27, 51, 54-5, 187, 231, 425-6, 440
dreams 51, 53-4, 216, 230-1, 234, 236, 238, 240, 245-6, 378
insanity/madness 282, 378
sleep 198, 230, 237, 241, 249, 282
somnambulism 236, 240
trance 230, 236
unconsciousness 52, 179, 230, 234, 238, 241, 424-5
wakefulness 54, 197-8, 230, 232-7, 239-42
stream of consciousness 47, 49, 140, 198, 233, 238, 284, 399, 425; see
also: consciousness; phenomenal character
taste see: sense modalities
theology 338
theoretical attitude/reasoning see: attitude
thinking 4, 38-9, 64, 93, 104-5, 110, 125, 150, 164-6, 251, 282, 295-6,
331, 347, 385, 400-1, 416, 419, 421, 425; see also: mental episodes; images and thoughts
first-order 68-71, 73, 149, 151, 207, 260, 262, 411, 422, 436
higher-order 209, 216, 260, 270, 295, 303, 307, 419
Thomas, N. 15, 367
thought experiments 3, 43, 310, 338, 353, 416, 440
thought insertion 54, 425, 440
touch see: sense modalities
484
Index
Index
485