web viewthis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis...

81
Imaginative Content Paul Noordhof Our imaginings are a disparate bunch of states that can differ from each other in, at least, three quite general ways. One point of difference is that between cognitive and sensuous imaginings. Cognitive imaginings have the same type of content as beliefs, judgements, guessings, and other propositional attitudes. Sensuous imaginings, by contrast, have a content, or character, which is phenomenally similar to the corresponding perceptions, for example, visual imaginings are phenomenally similar to visual perceptions. Cognitive imaginings often differ from the other propositional attitudes listed by not having the truth of their contents as a constraint upon the conditions under which it is appropriate to engage in them, or as placing a constraint upon their satisfaction (e.g. in the case of desire). Sensuous imaginings likewise can be unconstrained explorations of scenarios. However, there are tasks to which imagination can be put, for example, recreative imaginings geared at understanding another person by producing corresponding states in the imagination, which are subject to constraints. If I am imagining believing that elephants are evil creatures to predict how a pachydermophobe will react to a nearby trumpeting, then the content that elephants are evil creatures must be a content of our imagining 1

Upload: vulien

Post on 01-Feb-2018

219 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

Imaginative Content

Paul Noordhof

Our imaginings are a disparate bunch of states that can differ from each other in, at least, three

quite general ways. One point of difference is that between cognitive and sensuous imaginings.

Cognitive imaginings have the same type of content as beliefs, judgements, guessings, and other

propositional attitudes. Sensuous imaginings, by contrast, have a content, or character, which is

phenomenally similar to the corresponding perceptions, for example, visual imaginings are

phenomenally similar to visual perceptions.

Cognitive imaginings often differ from the other propositional attitudes listed by not having

the truth of their contents as a constraint upon the conditions under which it is appropriate to

engage in them, or as placing a constraint upon their satisfaction (e.g. in the case of desire).

Sensuous imaginings likewise can be unconstrained explorations of scenarios. However, there are

tasks to which imagination can be put, for example, recreative imaginings geared at understanding

another person by producing corresponding states in the imagination, which are subject to

constraints. If I am imagining believing that elephants are evil creatures to predict how a

pachydermophobe will react to a nearby trumpeting, then the content that elephants are evil

creatures must be a content of our imagining of a pachydermophobe’s beliefs. Recreative cognitive

imaginings are constrained by whether or not it is true that the subject, whose behaviour is to be

explained, has the relevant state in question, not by whether the content of the subject’s state is

true. By contrast, recreative sensuous imaginings must involve a phenomenally similar content to the

corresponding perceptions. For example, if I’m trying to predict how a subject will respond if they

see a ball thrown towards their face, I will visually imagine a ball coming at me. Visually imagining is

not phenomenally identical to visual perception. Thus, the constraint is weaker.

1

Page 2: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

A third point of difference between various kinds of imaginings lies in whether they are

active or passive: produced by the subject or triggered in the subject. Recognising this distinction

has implications for the proper explanation of borderline cases of imagination or related

phenomena, as well as paradigm cases of imagination (about which, more below). Both Brian

O’Shaughnessy and Colin McGinn take hallucinations to be internally triggered but will-impervious

imaginings (O’Shaughnessy (2000), p. 341; McGinn (2004), p. 15). McGinn suggests that dreams are

unconsciously willed imaginings (McGinn (2004), p. 90). Lying between these, as interesting

phenomena to compare with imagining, are what are generally called experiential or episodic

memories but which, to emphasise the important dimension of similarity and avoid confusion, I shall

call sensuous memories and, perhaps, also imaginative perceptions. Sensuous memories seem to

involve imagery guided by retained acquaintance with objects and properties previously perceptual

experienced. Imaginative perceptions involve such phenomena as seeing a landscape in a picture.

Here the organisation of paint on a canvas might be thought to control the development of a

particular piece of imagining. My intention, in remarking upon these potentially related phenomena,

is not to endorse a particular account of their character but rather to have them available as a

potentially illuminating comparison with the central cases of imagining.

In what follows, my focus is going to be on a certain kind of central case of imaginings: active

sensuous imaginings and their imaginative content. By ‘imaginative content’ I mean the phenomenal

content of these imaginings. The more specific question I am going to examine is whether the

phenomenal content, or character, of active sensuous imaginings provides considerations in favour

of one or other account of phenomenal properties. Before the discussion can proceed, I need to fix

the meanings of a few terms including that of phenomenal content itself.

First, let us understand the phenomenal content – others use the term phenomenal

character – of our mental lives to be what it is like to have the mental states or events which

constitute our mental lives. Standard examples are to see an elephant in front of us, to hear a

2

Page 3: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

baby’s cry, to feel somebody stroking your hand, to have a pain, to feel a very strong conviction, to

yearn for an ice cream, to feel blinding rage, to feel sexually excited, to remember the smile on your

face at last night’s dinner, to visually imagine ending the presentation prematurely and walking from

the room. Talk of content at this point is not meant to convey some notion of propositional content.

It is simply to capture the thought that our mental lives seem made up from various elements which

we cite to characterise what it is like to be undergoing them at a certain point in time. Some of these

elements may be no more easy to specify than as a feeling of yearning.

One reason why I use the term ‘phenomenal content’ rather than ‘phenomenal character’ is

that, often, the latter is used to stand for some complex property of our mental lives characterised in

terms of what it is like to be a subject with the mental life in question. Generally speaking, one

should avoid using weasel words like ‘character’, ‘feature’, ‘aspect’ or the like, rather than

‘property’, because it can assist to avoid the hard questions about properties that would otherwise

be asked. In brief, if you mean ‘property’, then say it. But this would only be an interesting scruple if

it helped to avoid certain confusions and I think it does. Talk of phenomenal character can help to

conflate two very different kinds of roles, corresponding to a difference between phenomenal

properties on the one hand and manifest objects and properties on the other, because it seems as if

it can appropriately cover either.

Let phenomenal properties be those properties of mental states or events which determine

their phenomenal contents (here properties may include relations). For those sceptical about the

existence of mental events or states because, for example, they take experience to be a relation to

the world, and not either an event or state, the definition of phenomenal properties may be

extended to include those (relational) properties of the brain, or the person, which determine the

conditions under which it is appropriate to characterise their mental lives as in such and such a way.

Manifest objects and properties are the constituents of phenomenal contents rather than those

properties which determine whether or not a mental state or event has a certain phenomenal

3

Page 4: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

content. Thus, if I see a certain grey elephant before me, then (plausibly) a manifest object is the

elephant, a manifest property is greyness, and the phenomenal property is that which determines

that the event of seeing the elephant has this manifest object and property as a constituent.

Broadly speaking, there are three types of account of phenomenal properties. According to

the first, representationalism, phenomenal properties are certain kinds of representational

properties, or supervene upon (in the sense of being metaphysically necessitated by) arrangements

of them. The second, non-representational relationism, holds that phenomenal properties are brute

relations of awareness to manifest objects and properties. The third approach takes them to

characterise modes of (not necessarily brute) awareness of objects and properties, non-

representationally. The last approach takes there to be phenomenal differences not to be accounted

for in terms of differences in the manifest objects and properties of a particular mental event or

state and inexplicable in terms of appeal to further representational properties.

Since the alternatives to representationalism are partly characterised by reference to

representationalism, grasp on the issues is advanced by arriving at a better understanding of the

nature of representational properties. To get a fix, representational properties are those properties,

of mental states or events, in virtue of which the mental states or events have contents that, in the

case of beliefs, thoughts, perceptions etc. yield truth or veridicality conditions, in the case of desires,

satisfaction conditions, and so on. This does not mean that phenomenal properties are, by

definition, representational properties. Phenomenal contents may not be the kind of content that

can be captured in terms of truth or veridicality conditions. That is an additional claim. Equally, such

a general characterisation of representational properties does not rule out the possibility of a non-

representational relationism. Suppose that a perceiver stands in the relation of (putatively) non-

representational awareness to Edward (the elephant) being grey – that’s what he or she is seeing. It

is open to the proponent of non-representational awareness to deny that this is a state with a

veridicality condition. There is not some condition of the state – its content - which may be either

4

Page 5: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

veridical or non-veridical. Rather, the state only exists if the perception is what their opponents

characterise as a veridical one.

I noted above that representationalism holds that phenomenal properties are certain kinds

of representational properties, or arrangements thereof. This reflects the obvious fact that, say,

unconscious beliefs, or early stages in the perceptual processing, may have no phenomenal content

at all and yet involve states and events with representational properties (hereafter ‘events’ will

cover both). So there are further conditions events with representational properties must meet for

what is represented to be an appropriate characterisation of the phenomenal content. Michael Tye

suggests that the events in question must be poised as an input for the practical deliberation system.

Here, the kind of representational properties is: poised representational properties (Tye (1995), p.

138). Others, for example Fred Dretske, focus, instead, on a certain way in which representation is

achieved, by natural systemic functional properties rather than acquired ones (Dretske (1995), p.

162). The key point, though, is that the kind of representational property in question does not add to

the proper characterisation of the phenomenal content itself as opposed to constitute what must

hold for the subject to be aware of the content of the representational properties in question.

Within this framework, then, the target question becomes this. Can the phenomenal content

of active sensuous imaginings be accounted for in terms of these events’ possession of

representational properties of a certain kind, together with, perhaps, the relation of these events to

other events with representational properties, without throwing into question representationalist

treatments of perception, or sensuous memory (as we shall see later)? The question will not be

approached from the perspective of a certain account of the nature of representational properties –

e.g. Tye’s, Dretske’s – but rather exploring options on this front which may help in the defence of

representationalism. So the emphasis will be on whether there are general considerations against

representationalism or whether recognising that representational properties may have a certain,

indeed variety of, nature(s) can help us to alleviate concerns.

5

Page 6: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

Active sensuous imaginings have two elements which are potentially determinative of their

overall imaginative content. The first is what is generally termed the mental image. By calling it that,

the idea is not that there is a mental picture of some kind. Instead, it is to recognise that these

imaginings have a certain kind of sensuous content that make them like our perceptual experiences.

The second element I shall dub the suppositional element, following Christopher Peacocke

(Peacocke (1985)). It is the latter that characterises the imaginative project which the image serves.

Supposition covers a variety of different ways in which this takes place. It can be an initiating

supposition which generates the image to begin with. Alternatively it can be a more integrated

interpretative response to images, the latter coming more or less spontaneously to mind. The

familiar thought is that what we are imagining is partly determined - some would say, entirely fixed -

by the supposition because the mental image may leave things under or overspecified.

To answer the challenge to representationalism, we need to understand how both these

elements work together to produce the overall imaginative content of active sensuous imaginings. In

the first section, I shall identify certain features of the mental image that, taken together, seem to

suggest that phenomenal content cannot be accounted for in terms of a single type of theory:

representationalism, relationism etc. Relationists typically address this problem by endorsing what

has become known as the Dependency Thesis, which, roughly, holds that sensuously imagining

something is sensuously imagining an experience of that thing. The Dependency Thesis is also of

some help to representationalists in accounting collectively for the features of the mental image

identified in the next section. Nevertheless, I argue in section 2, the Dependency Thesis should be

rejected. The result is that, unless relationists postulate mental objects and properties of which

subjects are brutely aware in the case of imagination, they have no account of the phenomenal

content of imaginative experiences. Representationalists too must look elsewhere for a defence of

their approach. An important component of this defence, I shall argue, concerns the way in which

the suppositional element determines the overall imaginative content, partly by what it supplies in

addition but partly, also, by what it reveals about what is subtracted from imaginative content when

6

Page 7: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

compared with perceptual experiences and sensuous (or experiential) memories, Thus the second

half of the paper – sections 3 and 4 – focuses on the role that this other component plays, and the

implications of it.

1. The Mental Image

The first three features often attributed to mental images all underline the ways in which they seem

to resemble the corresponding sensory modality: thus visual imaginings/visual perceptions; auditory

imaginings/auditory perceptions and so on. The fourth identifies a point of disanalogy. Together they

create the problem of the mental image for mono-theoretic approaches to phenomenal content.

The first feature of mental images is their perspectival character. A mental image of a/the F

takes over from sensory experience, the imagining subject, from the inside, having a certain

perspective on F (where F might be dog, or table or the sound of the sea). The perspective is partly

characterised in terms of monadic directional properties - such as to the left, to the right -where the

relations which ground such relations are dyadic: to the right of me, the subject stands in such and

such a relation to an object (Campbell (1994), pp. 119-120; Martin (2002), pp. 409-410).

The second feature is that the mental image is sensuous where this adverts to the fact that

different kinds of sensory imagination – visual, auditory, and so forth – are characterised by the

same, and limited to the same, bundle of properties to which the corresponding perceptions are

characterised and limited.

The third, and final feature, is characterised by what I have called the Multiple Use thesis:

the same sensuous type of image will serve more than one imaginative project. For instance visually

imagining a F and visually imagining a perceptual experience of a F involve the same mental image in

this sense. As does imagining a suitcase and imagining a suitcase behind which, totally obscured in

the imagined scene, is a cat.

7

Page 8: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

As we shall see when considering an objection to the Multiple Use Thesis later, the exact

character of the thesis needs careful specification. For now, the important point is that it is a

consequence of constraints upon how things are presented in a particular sense modality, and the

corresponding sensuous imagining. Consider the case of the cat hidden behind the suitcase. Our use

of the same mental image to imagine a suitcase or a cat wholly hidden behind a suitcase seems

grounded in the fact that visual experience of the suitcase and visual experience of the suitcase

totally obscuring a cat would involve no differences in what is presented because, on the assumption

that the subject has no beliefs or suspicion that a cat is hiding behind a suitcase, the presentation of

some parts of the scene occludes the presentation of other parts. There are complexities to this – to

which I will return – concerning how rich the content of a visual experience may be but the

straightforward case just described is untouched by them.

The fourth feature attributed to mental images is the point of difference with perceptual

experiences. It is said that mental images do not involve the presentation of sensory qualities. Our

visually imagining something red does not involve a presentation of redness, imagined pains and

itches don’t involve presentations of itches or pains. Here is Mike Martin on the subject

If we treat imagining an itch as a representing of an experience of an itch, then we can both

accept that the relevant quality is before the mind, as it is in experience itself, while

denying that there has to be an actual instance of it, in contrast to the case of experience

(Martin (2002), pp. 406-407, my emphasis).

The surrounding material constitutes Martin’s proposed solution to the difficulty. The emphasised

bits contain the key claim. There is certainly a phenomenon here that we need to explain even if we

do not, in the end, agree with Martin’s formulation of the issue.

On the one side, we have three features that emphasise the phenomenal similarity between

m-perceptions and m-imaginations (here ‘m’ stands for visual, auditory and the like). This suggests

8

Page 9: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

that the natural way to explain the connection is that they have phenomenally similar contents

plausibly thought of as involving, at least, the same manifest objects and properties. The various

accounts of phenomenal properties would, then, indicate how this explanation should be deepened.

Different theories would have different degrees of difficulty in explaining this similarity and that

would support favouring one theory over another. For example, relational theories of phenomenal

properties would face a problem because the entities to which we stand in a relation in the case of

perception are arguably absent in the case of imagination. This might tempt some relationists to

postulate the existence of common objects and properties for both cases. The relations of brute

awareness become those of the sense datum theorist: relations to mental objects and properties.

On the other side, the claim that m-perception involves presentation, where m-imagination

does not, seems to favour relationists’ views of phenomenal properties regarding perception in

contrast to imagination. This kind of relationism – often dubbed naïve realism – takes the objects

and properties to be items in the world, which is, of course, Martin’s position. The question then

becomes what explains the three observations that favoured the claim that there is a phenomenal

similarity between m-perception and m-imagination. The Dependency Thesis is alleged to provide a

satisfying answer. Because, according to the Dependency Thesis, imagining always involves

imagining an experience, the nature of the latter constrains the character of the former. However,

since imagination involves a representation of a perception of O being F rather than, as experiences

are taken to involve, a relation of awareness to O being F, we don’t have the presentation of objects

and properties.

This combination is the probably the most challenging, and popular, bi-theoretic approach to

phenomenal content. However, the Dependency Thesis does not just provide the beginnings of an

explanation of the phenomenal similarity for relationists. It can also form the basis of a

representationalist attempt to account for the alleged difference between m-perceptions and m-

imaginings since they can point to a difference in representational properties to accommodate the

9

Page 10: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

difference. But does the Dependency Thesis really help? In the next section, we shall look at its

character in more detail and find its explanatory promise wanting. It is neither a successful

component is a sophisticated bi-theoretic approach nor something that may be of assistance to the

representationalist.

2. The Dependency Thesis versus the Similar Content Hypothesis Thesis

Put a little more precisely, the Dependency Thesis is the claim that, in m-imagining O being F say, a

subject imagines a m-perceptual experience, from the inside, of O being F (where, once more, ‘m’

picks out a mode of perceptual experience like visual, auditory and so forth). For example, visually

imagining O being F is imagining a visual perceptual experience of O being F. The italicised phrase

indicates that we are not imagining looking at a subject perceiving O being F, for example, some

spectator of a dog being happy. Instead, we are imagining ourselves – or, at least, the view point of a

subject – undergoing an experience of O being F. Perhaps found in Berkeley, the thesis was given the

title of the Experiential Thesis by Christopher Peacocke and has its current name from Martin who

extended it to experiential memory (Berkeley (1710, 1734), Part 1, s. 23, pp. 83-84; Peacocke (1985);

Martin (2002), p. 404). In that context, the relevant thesis is that to m-experientially remember O

being F is to recall a particular m-experience of O being F from the inside (Martin (2001), pp. 279-

280).

It is important to be clear about exactly what is envisaged to do the explanatory work in the

Dependency Thesis. Discussion can very easily work at the level of something feels right about it and,

just as vaguely, that it this that is doing the explanatory work. Equally, the proper formulation of the

alternative – what I have dubbed the Similar Content Hypothesis – requires some attention.

The Dependency Thesis does not claim merely that, in imagining O being F, a subject

imagines something which is the content of a possible perceptual experience and, thus, a possible

10

Page 11: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

perceptual experience from the inside. The claim is the stronger one that what is imagined is a

perceptual experience which is part of the content of the scene imagined. So if you are imagining

yourself steering a yacht, then what you are imagining is that there is a perceptual experience of

yacht steering. Although the perceptual experience you are imagining is possible rather than actual,

within the content of the imagining, what you are imagining is an experience and not a possible

experience. Imagination might reveal the possible but it does not have the possible as part of its

content according to the Dependency Thesis.

A familiar example underlines the point. Consider Berkeley’s claim that we can’t imagine

something which is not experienced. If my imagining of a perceptual experience of a grey elephant

were of only a possible experience of a grey elephant, then what I am imagining – the grey elephant

- can be imagined unperceived. A possible experience of a grey elephant rather than an experience

of a grey elephant is compatible with the elephant not being experienced in the imagined scene.

Whereas, the Dependency Thesis insists that, in the imagined scene, what is being imagined is an

experience of a grey elephant.

With this in mind, there are two distinct theses that it is important to distinguish from the

Dependency Thesis. They are weaker than it and do not play the explanatory role envisaged for it.

The first is the Simulation of Perception Thesis: to m-imagine O being F is to simulate a m-perception

of O being F. Evan Thompson sometimes writes in these terms. Thus he says that to imagine O being

F is to represent mentally O being F as given to a neutralized perceptual experience of O being F. He

takes mental representation in these circumstances to involve subjectively simulating a neutralized

perceptual experience of O being F. Simulating neutralised perceptual experiences involve no

implication that a subject has had or will have the perceptual experience in question. Nevertheless,

these claims, he continues, involve an acceptance of the Dependency Thesis (Thompson (2008), p.

408).

11

Page 12: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

It is clear that, in fact, the Simulation of Perception Thesis is distinct from the Dependency

Thesis. One can hold that imagination involves the simulation of perception and yet deny that that

involves a perceptual state being part of the content of what is imagined. Brian O’Shaughnessy

nicely expresses the difference

If I stands here for imagining, if O is some physical object, and Ф is the mental type of the

‘filler’ phenomenon, then while we can perfectly well characterise the imagining as (say) I

(ФO), it is more perspicuously to be expressed as (ФI)(O). The latter formula is structurally

perspicuous, as the former is structurally misleading. For to repeat, there is no such thing as

merely imagining an object (O’Shaughnessy (2000), p. 364).

The simulation of a certain kind of perception figures as part of the characterisation of the mode of

what one is doing when one is imagining. One doesn’t just sensuously imagine, one visually sensually

imagines or aurally sensuously imagines and so forth. It is in this sense that we are simulating a

certain kind of perception. Taking perception to figure as part of the content is an additional step.

Interestingly, this is a point which Thompson himself recognises almost a page later in his

discussion. He writes

Although visualizing an object requires imagining visually experiencing the object, the visual

experience is not the object of the imagining; the intentional object is the visualized object.

As visualized, however, that object must be given visually in some way or other, and this

mode of visual givenness on the part of the object entails a correlative mode of visual

experience on the part of the subject. The visual experience co-imagined in visualizing an

object is thus simply the intentional correlate of the imaged object’s mode of visual

appearance in the visualization. The intentional object of the transitive imaging

consciousness is the imaged object; the correlative and co-imagined visual experience is

experienced only intransitively and prereflectively. In other words, this experience is “lived

12

Page 13: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

through” without usually being noticed, attended to, or reflected upon (if it is taken notice of

in these ways, then the imaging experience becomes a reflective one) (Thompson (2008), pp.

408-409).

The Dependency Thesis claims that if a subject imagines O being F, then the object of the imagining

is a perception of O being F. In the passage, Thompson explicitly denies that a perceptual state is

part of the content of the imagining.

Nevertheless, Thompson seems to have in mind an intermediate thesis to the Dependency

Thesis and the Simulation of Perception Thesis. Visually imagining O being F, Thompson starts by

assuming, involves a specifically visual mode of presentation of O being F. We might put it like this.

Visually imagining O being F involves imagining (O being F) visually. This is something that

proponents of the Simulation of Perception Thesis may endorse. From this, though, Thompson

concludes that the subject imagines a (correlative) visual experience, specifically, he takes this to be

an entailment of the visual mode of presentation. We might put the resulting position as follows: in

visually imagining O being F, a subject imagines visually experiencing O being F. This seems very

close to O’Shaughnessy’s view except that Thompson talks about the ‘visually experiencing’ in

‘imagining visually experiencing’ being experienced intransitively and, thereby, lived through. Call

this the Correlative Content Thesis.

Here are two related considerations against the Correlative Content Thesis. First, it is

undeniable that there are phenomenal differences between visual perception and visual

imagination. Even if it is accepted that there are sufficient phenomenal similarities between them to

make it appropriate to group both as involving a visual mode of presentation, the differences are

sufficient to make it illegitimate to conclude that a visual experience is present. Indeed, one might

suppose that the differences entail that there isn’t a visual experience there. Second, as Thompson

remarks, imagining visually experiencing O being F involves not having the visual experience as

object of the visual imagining but rather as intransitively experienced along with it. There are

13

Page 14: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

significant differences between intransitively experiencing visual perceptions, when we are visually

perceiving objects, and intransitively experiencing the corresponding visual imaginings. That implies

that the presence of visually experiencing to characterise a certain kind of imagining doesn’t even

get the character of what we are intransitively experiencing, when we are imagining, right.

Moreover, as we shall see in more detail below, m-imaginings vary in their phenomenal

similarity with m-perceivings depending upon the manifest objects and properties they concern. For

example, visually imagining a red object is phenomenally much closer to visually experiencing a red

object then imagining the pain of a cut is to having a pain experience as a result of a cut. This

indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon

with their own particular characteristics.

Although the Simulation of Perception Thesis does not imply either of the stronger theses

just identified, it is compatible with them. For example, although the Dependency Thesis asserts

that experiences invariably show up in the content of the corresponding imaginations, it does not

immediately follow that when we m-imagine we are, in part, theorising about our experiences using

background knowledge applied to the contents in question. A theorist might still insist that m-

experiences are being simulated rather than being the subject of theory because the content of the

imaginings involves the instantiation of no theoretical concepts in its realisation, for example, that of

a perceptual state. Talk of sensuous imagination being the simulation of perception also suggests a

certain task for such imagination – that it involves recreating the perceptual experience of others for

explanatory ends - that need not be mandatory.

For this reason, the more appropriate contrast with the Dependency Thesis is a view of

imagination that takes it simply to involve similar phenomenal contents to their corresponding

perceptions. Visual perceptions and visual imaginings, like auditory perceptions and auditory

imaginings, have a certain kind of phenomenal similarity of content that makes it appropriate to

characterise both by the term visual or auditory. One way of putting this is

14

Page 15: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

(SC) m-imaginings and m-perceptions are such that the manifest objects and properties of the

former are, perhaps a proper, subset of the manifest objects and properties of the latter,

manifested m-ly,

‘perhaps a proper’ because, intuitively, there is some loss of richness and other material when we

imagine, in a sense I shall seek to clarify below. In contrast to the Dependency Thesis, though, there

is no introduction of material into the content like, for example, it being an m-perception of O being

F rather than (O being F) visually. Equally, it is not implied that imagination is a dependent

phenomenon on perception in the sense described by the Dependency Thesis. Imagination’s

simulation of perception takes place simply because they involve similar phenomenal content.

Dominic Gregory has sought to characterise what seems very like the Similar Content

Hypothesis for visual content in the following terms.

(C) Given a visual image whose corresponding sensation-type is V, this image may serve to show

how things look from an imaginary viewpoint e. In so doing, it characterises e as being such

that the contents of the visual appearances which accompany V are accurate relative to e

(Gregory (2010), p. 748).

Gregory’s sensation types are characterised in terms of how things visually look to a subject

undergoing a particular visual sensation at a certain spatiotemporal position. V is taken to have a

certain content of appearances, which accompanies it, that may be accurate or inaccurate

depending upon whether the scene around the subject’s spatiotemporal position is as the

corresponding content of appearances says it is. Visual images are stipulated to have the content of

visual appearances of their corresponding sensation type (Gregory (2010), p. 737).

There two problems with this characterisation of the Similar Content Hypothesis. The first

relates to whether visual images having the same content of visual appearances as perceptual

experiences of a particular sensation type V implies that visual images fall under V too. Proponents

15

Page 16: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

of relational accounts of perception are apt to deny that the mental image and the visual sensation

could share a sensation type. This is compatible with supposing that differences of sensation type

have corresponding differences in the content of visual appearances and, indeed, recognising that

there are correspondences between visual images and visual perceptions, albeit falling short of

sharing the same sensation type. Of course, it might be stipulated that sensation types are to be

individuated by whatever similarities of content there are between visual perceptions and the

corresponding visual imaginings (to take one case). But then (C) is not a further articulation of the

Similar Content Hypothesis but rather relies upon it.

The more substantial difficulty is that those who are relationist about perception deny that

perceptual experiences have accuracy conditions and, hence, will resist the claim that there are

contents of visual appearances which may be accurate or otherwise depending upon how the scene

is around the experiencing subject. They see this as of a piece with denying that perceptual

experiences have contents (e.g. Travis (2004)).

By contrast, the characterisation of (SC) in terms of manifest objects and properties allows

that different accounts of phenomenal properties – as brute relations of awareness in the case of

perception, as representational properties in the case of imagination – may be grouped together

because they determine the same manifest objects and properties manifested m-ly. That means that

we can capture the idea, favoured by relationists, that perceptions don’t have contents but rather, in

the good case, just have manifest objects and properties as constituent parts of their nature. Talk of

the manifest objects and properties being manifested m-ly rather than presented m-ly allows for the

possibility that, while m-imaginings have the same manifest objects and properties as m-

perceptions, they are not presented m-ly, because m-imagination doesn’t involve the presentation of

objects and properties. As we shall shortly see, though, the matter is complex.

Purely relational approaches to phenomenal properties cannot accept (SC) unless they take

perceptions and imaginings to have mental manifest objects and properties since, typically, there

16

Page 17: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

will be no manifest objects and properties in the world to correspond to our imaginings. Without this

move, they would be committed to the manifest objects and properties of imaginings failing to be a

subset of those of the corresponding perceptions. Only by taking perceptions to involve mental

objects and properties could imaginings still be brute awareness of a subset of them. Those who

limit their relational approach, to phenomenal properties, to perception may also endorse (SC). They

can claim that the m-imaginings corresponding to m-perceptions have a subset of the latter’s

manifest objects and properties in virtue of the m-imagining’s representational properties.

Relationists about perceptual phenomenal properties have not tended to pursue this option

probably because some of the features of perception that led them to contrast it with belief and

judgement are present in sensuous imagination. So they look for a distinctive story to be able to tell

and the Dependency Thesis seems to fit the bill.

They are not the only ones who may be so motivated. As we have already noted, the Similar

Content Hypothesis constitutes a premise in favour of an argument against representationalism.

Unlike the Dependency Thesis, the Similar Content Hypothesis does not entail a difference in what is

represented by m-imagining and m-perception. So the phenomenal difference between these two

cases seems to lie with two distinct modes of awareness of the content that representationalists

cannot accommodate (see e.g. Thompson (2008), p. 398)).

The apparent advantage of the Dependency Thesis in this respect is only telling if the

explanatory claims that have been made for the Dependency Thesis with regard to the other

features of the mental image identified can be, in fact, made good over those of the Simliar Content

Hypothesis. We shall find them wanting. I will go through them in turn.

The first, upon which a lot of weight has been placed, is the perspectival character of

imaginative content. It is urged that if the Dependency Thesis is true, then imaginative experience is

perspectival because it has the perspective of visual perceptual experience built into it and, indeed,

that the Dependency Thesis is uniquely well-placed to provide the required explanation. I shall argue

17

Page 18: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

that both the Dependency Thesis and the Similar Content Hypothesis have to appeal to an additional

explanatory element to capture the perspectival character of imaginative experience and, in

addition, the Dependency Thesis cannot explain the perspectival character of imaginative experience

without an independent appeal to the special character of imaginative experience. In which case, it

seems to involve an additional, otiose, claim.

The standard argument for the Dependency Thesis appeals to the fact that m-imaginative

content shares with perceptual content characterisation in terms of monadic directional properties.

Suppose that you imagine a red light to the left and a green light to the right. Then, Martin claims,

you are imagining a distinct scene from a scene in which you imagine a red light to the right and a

green light to the left. Distinct scenes are presented only if the view point is fixed. Otherwise, we

could have the same scene presented from two distinct viewpoints. Thus imaginative experience

brings with it a viewpoint. Moreover, it is argued, this viewpoint must imply the presence of a

perceptual experience in the imagined scene because, unlike the case of perceptual experience, the

lights imagined are not in an environment in which the subject is actually located, with one light to

the left, the other to the right. So to capture the orientation of the lights in the imaginative

experience one has to appeal to an imagined perceptual experience (Martin (2002), pp. 408-411).

Appeal to the perspectival character of perceptual experience is insufficient, however.

Unless you, in addition, suppose that the viewpoint is fixed in both cases of imagining, you don’t get

the result that two distinct scenes are imagined as opposed to one scene from two distinct

viewpoints. The proponent of the Similar Content Hypothesis then can resist appeal to the

Dependency Thesis in the following way.

If the content of perceptual experience and the content of imaginative experience is similar

in the way identified, then each will convey perspective in a similar fashion, in terms of orientation

to a non-represented viewpoint. In the case of perceptual experience, the actual existence of a

relation between the two lights and the subject’s location fixes the other relata. Thus whether the

18

Page 19: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

same scene or a distinct scene is presented will turn on whether the subject has moved. However, in

the imaginative case, neither an imagined perceptual experience, nor simply an imaginative content

similar to that of a perceptual experience, will settle this issue. In both cases, it will have to be

supposed that the viewpoint is the same, or different, as part of what one is imagining. Once one has

done this, then there is nothing further for the Dependency Thesis’s appeal to an imagined

perceptual experience to do. Emphasising, as the Dependency Thesis does, that the viewpoint in

question is occupied by a subject undergoing an experience is unmotivated.

In a development of Martin’s discussion, Fabian Dorsch claims that monadic directional

properties are experience-dependent subjective properties (Dorsch (2010), pp. 179-185). This is

certainly questionable as my characterisation of them relative to a, not necessarily occupied,

viewpoint reveals. Nevertheless, if Dorsch is right, the considerations I adduce later to account for

other differences between perceptual experience and imaginative experience – for example, the

difference between being in pain and imagining one is in pain – would also have relevance to the

present case. No further issues of principle seem to arise.

Simple appeal to the fact that imaginative content is similar to perceptual content is likely to

look explanatory trivial for two reasons. First, since ‘similar’ does not mean ‘phenomenally identical’,

the perspectival character of perceptual experience might be one of the ways in which imaginative

experience is not similar. Second, the explanation offered by proponents of the Similar Content

Hypothesis seems to amount to no more than saying that perspectivality is one of the dimensions of

similarity both types of experience share.

The first concern can be avoided almost by stipulation. The Similar Content Hypothesis holds

that m-perceptions and m-imaginings share manifest objects and properties, manifested m-ly. If

visual experience – say – has a perspectival character, then that will be part of the way in which the

manifest objects and properties are manifested m-ly, in this case, visually. While this might address

the first concern, it is likely to be felt that it exacerbates the second. However, this is to mistake the

19

Page 20: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

explanatory ambitions of the Similar Content Hypothesis. It is offering the same kind of explanation

as one might offer to the following question: why can’t we have desires about spanners without

beliefs about spanners and vice versa? Answer, because they involve the same kind of contents,

conceptual contents, so that inability to have one sort of state would suggest inability to have the

other kind of state and vice versa. It is perfectly appropriate to note that two kinds of states share a

certain kind of content, so an ability to have one implies an ability to have another; the character of

one resembles the character of the other. In noting this, we are, in effect, taking the deeper

explanatory question to be that of how phenomenal contents of this character are possible at all.

Everybody has to answer this question at some point: relationists about perception with regard to

perception; and the kind of representationalism I seek to defend here with regard to both

perceptual and sensuous imagination. Proponents of the Similar Content Hypothesis just deny a

certain way of dealing with this problem, namely deriving the character of sensuous imaginative

content from the character of the phenomenal content of perception. There is no problem reduction

because each can take the answer to the general question – how perspectival phenomenal content

as possible to deal with both cases.

While the Similar Content Hypothesis has limited explanatory credentials, the Dependency

Thesis might be thought to gain in plausibility by having more. This impression is quite overstated. It

is tempting to suppose, for example, that if a subject is imagining a perceptual experience, then

simply because he or she is doing that, we have some account of the similarity in phenomenal

content between imagination and perception and, in particular, their perspectivality. This is not so. If

a subject thinks (in the sense of entertains the thought) that he or she is perceptually experiencing O

being F, then there is no way in which the thought takes on a perspectival character. If having the

perception of O being F as part of the content of state gave it perspectival character, then we would

expect such thoughts to be, just in that way, perspectival. They are not. The fact that our imagining a

perceptual experience of O being F has a perspectival character shows something about the nature

of sensuous imagination and not something about the impact of a certain content being embedded

20

Page 21: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

in the imaginative act. Thus, the Dependency Thesis cannot claim that it captures the perspectival

character of imagination.

There are two differences between sensuous imagination of a perceptual experience of O

being F and entertaining the thought that one is having a perceptual experience of O being F. If the

perspectival difference could be traced to either of these, then the Dependency Thesis would still

provide a candidate explanation. However, neither is a plausible source of the difference.

The first is that the relevant thought is propositional – that he or she is having a perceptual

experience of O being F - whereas imagining takes an object: perceiving O being F. But it is debatable

whether perceptual, or imaginative experience, takes an object rather than a state of affairs or fact.

It is also implausible that there is a straight consideration in favour of the former to be drawn from

noting that perception and sensuous imagination are perspectival. There are other attitudes that,

plausibly, take an object e.g. love, hate, and so on, that don’t seem perspectival thereby.

The second is that the difference between thought and the relevant kind of imagination is

that the latter is specifically sensuous imagination and it is because of this that, if perceptual

experience is part of the content of the imagination, then the imagination is perspectival. This

difference may be relevant but it appeals to the very feature that needs explaining. Taking

perceptual experience as the content of a certain kind of mental state was supposed to illuminate

the sensuous character of the state. Now it seems that the explanation goes the other way around.

It is because of the sensuous character of this kind of imagination that imagining perceptual

experiences of O being F involve perspective. Nor does appeal to perceptual experience in this way

provide anything in addition. As we saw above, it then becomes dependent upon the same materials

as we found to be required to characterise the perspective of sensuous imagination if the similar

content hypothesis were true.

21

Page 22: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

The Dependency Thesis does not do any better in explaining the Multiple Use Thesis. The

thought that I am perceiving a suitcase and the thought that I am perceiving a suitcase which totally

obscures a cat hiding behind it are phenomenally distinct in virtue of the difference in content

between these two thoughts. However, the Multiple Use Thesis claims that a mental image with the

same sensuous content would subserve each imagining. We cannot capture this similarity of

sensuous content by seeking to appeal to the fact that perceptual experience in the content of an

attitude – a thought, or an imagining – is the basis of it.

The most compelling consideration in favour of the Dependency Thesis remains its ability to

explain the non-presentational character of imaginative experience. However, even here, it is in two

respects explanatorily clumsy. First, it is hard to accommodate the fact that the Dependency Thesis

seems more plausible in some senses than it does for others. For instance, if you sensuously imagine

the taste of bacon or the touch of somebody’s hand upon your skin, it is much more plausible that

what you are doing is imagining an experience of them. That seems to be the means by which one

generates the content that one wants. By contrast, although it is true that there are phenomenal

similarities between visually imagining an orange and visually perceiving it, to visually imagine it, you

don’t need to imagine you are experiencing it. If the Dependency Thesis were generally true, we

would need a distinct explanation of why it is more obvious in some cases than others. Of course, if

there were independent reason to believe in the Dependency Thesis (something I have sought to

undermine above), then the search for such an explanation would be of significance. However, in the

absence of such a reason, the difference in plausibility is sign that the thesis does not hold in

general.

The second respect in which it is explanatorily clumsy is that the non-presentational aspect

of m-imagining, when compared with m-perceiving, alters depending upon the kind of perception in

question. A visual perception of a red object is phenomenally quite similar to the visual imagination

of a red object. By contrast, our perception of an itch or a pain, is phenomenally significantly

22

Page 23: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

different from imagining an itch or a pain. I take it that this difference is, in part, to be explained by

the fact that there is a way in which visually imagining a red object seems to involve the sensuous

nature of redness but imagining a pain or itch does not involve the sensuous nature of hurt or

itching. The putatively ‘non-presentational’ character of imagination is complex and qualified in this

respect. If the Dependency Thesis were the explanation of the non-presentational nature of

imagination, we would not expect this variation because the same kind of explanation would be

offered of both cases.

In spite of its promise, then, the Dependency Thesis is no help in resolving the issue we

identified at the end of the last section. Bi-theoretic accounts do well in capturing the putative non-

presentational phenomenal difference between perception and sensuous imagination. However,

they do badly in capturing the first three features of sensuous imagination, that they are

perspectival, sensuous, and the image is subject to multiple use. Removing this motivation for

combining relationism about perception with representationalism about sensuous imagination, it is

worth considering whether a representationalist approach can be developed that can make sense of

the apparent non-presentationality of sensuous imagination or whether, as Thompson suggests, we

have to account for the phenomenal difference between perception and sensuous imagination in

terms of a difference of mode (Thompson (2008), p. 398). However, before I can present my answer

on behalf of the representationalist, I need to set out the second element in the determination of

imaginative content, the first being the mental image itself. Many of the differences between

perceptual and imaginative phenomenal content are a consequence of recognizing the implications

of this feature of imaginative experience.

3. The Suppositional Element

23

Page 24: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

Visually imagining a suitcase or a suitcase totally obscured behind which is a cat involves the same

mental image. Likewise, there is no difference in mental image between visually imagining being at

the helm of a yacht/visually imagining having an experience of being at the helm of a yacht/visually

imagining being stimulated by brain scientists to experience from the inside being at the helm of a

yacht or between visually imagining a red tomato/visually imagining a wax tomato/visually imagining

the outside surface of a hollow tomato, making standard assumptions about the perfection of the

wax model of the tomato. Nevertheless, it seems that a subject imagines something very different in

all of these cases.

The cases reveal the contribution of the, second, suppositional element in the determination

of imaginative content. Recognition of this contribution is behind much of the plausibility of the

multiple use thesis, for obvious reasons. The suppositional element can take different forms.

Sometimes it is the intention behind a piece of active sensuous imagining, sometimes it is an

interpretation that naturally arises along with it. I have discussed the different possibilities in greater

detail in Noordhof (2002).

Kathleen Stock has challenged the orthodoxy that I expressed in the previous two

paragraphs (Stock, Manuscript).1 In effect, she identifies three related problems. First, what is the

role of the suppositional element – does it supplement or de-emphasize? The latter – de-emphasis –

involves elements of the mental image being discounted from being part of the imaginative content

of the imagining (see below for discussion of examples). Second, what is the common content

neutral between the various imaginative projects differentiated above? Third, can conservatism

about the mental image be defended? The third problem falls out of her answer to the first two

although she does not put it in quite the way I am putting it.

1 I am very grateful to Kathleen Stock for letting me see her very interesting and provocative paper and prompting me to defend my earlier dogmatism. I would recommend it as an addition to anybody’s reading list on the issue.

24

Page 25: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

It is pretty clear when we examine the various cases detailed above that there can be no

single answer to what the suppositional element does. Sometimes it will supplement, other times it

will de-emphasise. Contrary to Stock, I don’t think that this presents a problem. In the case of the

suitcase and the cat, the mental image is clearly supplemented by supposing that there is a cat

behind it. In the case of imagining being stimulated by brain scientists to experience oneself at the

helm of a yacht, both supplementation and de-emphasis looks to be in play. You sensuously imagine

being at the helm of a yacht – which if the Dependency Thesis is not true is distinct from imagining

experiencing being at the helm of a yacht – supplement the mental image by taking it to be an

experience and de-emphasise that there is a yacht there, supplementing instead with the idea that

the experience is produced by stimulation by brain scientists.

In the yacht case, it may be tempting to argue that you can just imagine the experience

(without committing oneself to there being a yacht there) and supplement this with the idea that the

experience is the result of brain stimulation. However, I note that this would not be available to one

of the proponents of the Dependency Thesis, Martin, because imagining a perceptual experience is

supposed to involve a commitment to it being a good case of perception – that is, involving a

relation to the helm of a yacht – in order to get the epistemology of imaginative experience right.

If such cases were the only ones in which de-emphasis were required, it may be appropriate

to consider whether the complexity I have identified – and to which Martin would also be committed

– constitutes a flaw which, when treated, renders de-emphasis an otiose element in the

determination of imaginative content. However, it is clear that de-emphasis has a role to play in

determining the imaginative content of a piece of imagining even if there are differences of opinion

over the extent of the role it might play. Consider the case of somebody who seeks to imagine a

building A but produces a mental image which, in fact, only approximates it but which, by accident

since the subject has never come into contact with it, is an exact image of building B. We might allow

that the subject had a mental image of a building exactly like B but we would not conclude that they

25

Page 26: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

were imagining B rather than A. We would still claim that they were imagining building A. They had

just got that building a little wrong. It would be a mistake to insist that, in order to imagine building

A, the subject must get it exactly right. On the other hand, we would not want to say that they could

imagine whatever they liked, they would still be imagining building A. So the imagining subject must

get the building in the image right in certain respects and that means that there will be an emphasis

on certain respects, and a de-emphasis on other respects when the subject is imagining, or perhaps

just a de-emphasis of any (unspecified) minor departures.

This minimal role for de-emphasis is to be contrasted with the stronger position that we can

never be wrong about what we are imagining. Suppose that a subject experienced New College in

winter but was told, mistakenly, that it was Hertford College (which is, in fact, further down the

street). Later, they imagine, as they suppose, Hertford College in winter. What would we say that

they are imagining? Certainly their mental image corresponds to New College rather than Hertford

College – and the buildings are rather different – but says Peacocke, it is clear that what they are

imagining is Hertford College. He takes this to be an example of knowing what you are thinking

(Peacocke (1985), p. 27).

Knowing what you are imagining is not a straightforward case of knowing what you are

thinking. In the case of thinking, the conceptual capacities to which you appeal in thinking a certain

thought are amongst those you will use in attributing to yourself that thought. Thus, apparent

misattributions of conceptual ability don’t infect self-knowledge because the belief you have

concerning the thought you are currently thinking will appeal to the conceptual resources you have

rather than the conceptual resources you think you have. For example, suppose I think that Proteus

is a computer game rather than a urinary infection. Then when I think the thought ‘George has just

got Proteus’ and self-attribute the thinking of that thought, I know that I am thinking that George

has just got Proteus regardless of whether we suppose that, in fact, I am thinking a thought whose

content concerns a urinary infection, or putative computer game. In the case of sensuous

26

Page 27: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

imagination, self-attribution of a certain imaginative content uses different resources. In the former,

a mental image is being used, in the latter, appeal to our conceptual abilities. Of course, if the self-

attribution were re-using the sensuous imaginative capacity, no error would be introduced for the

same reason. The point only concerns the typical case in which we make a non-sensuous self-

attribution of a piece of sensuous imagining.2

Nevertheless, although there is this difference, it does not follow from that that any mistake

in the mental image will vitiate the imaginative project. De-emphasis rules out certain mistakes in

the image undermining the success of the imaginative project. For example, unless the imaginative

project is to imagine a building with the architectural details of Hertford set in snow, the subject is

successfully imagining Hertford College in the snow. Sometimes, as already remarked, this will be a

matter of the subject’s intentions but on other occasions, when the imagining just comes to him or

her with an interpretation along with it, the de-emphasis will be rooted in the concurrent

interpretation. Hence, talk of de-emphasis allows us to be neutral about the various ways in which it

may arise.

A radical case of de-emphasis is that described by Bernard Williams involving a Bonnard

corrupted man. Although the man tries to imagine an empty bath, repeated exposure to Bonnard’s

painting of a Woman in Bath means that he inadvertently entertains a mental image with a woman

in it. Williams supposes that, so long as the man is trying to imagine an empty bath, that is what he

has imagined. The project de-emphasises the presence of the woman (Williams (1966), pp. 33-34).

The Bonnard corrupted man is a borderline case. The verdict changes depending upon how

the case is filled out. If the man’s focus is very much on a bath and the woman occurs simply

because she was extremely salient in the man’s most recent experiences of baths, then it seems

plausible to allow that he is imagining an empty bath, albeit not by producing a mental image that

best serves this project. Obviously the mental image is not of an empty bath but as I have already

2 I am grateful to Peter Langland-Hassan for pressing me to be clear on this point and subsequent discussion.

27

Page 28: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

indicated earlier, this does not mean that we should conclude that what is being imagined is not an

empty bath. The combination of bath imagery together with the supposition that it is empty makes

this attribution plausible. The presence of a woman is a distraction rather than a defeater.

Nevertheless, it is undeniable that it is more plausible to suppose that the man successfully

imagines a bath rather than an empty bath. The former imaginative project does not conflict with

the mental image, the mental image just contains irrelevant elements, whereas the mental image is

flatly in conflict with the latter imaginative project. I suspect that this is because there is some

confusion, or difference of emphasis, between theorists about what sensuous imagination is. In

Williams’ discussion of the issue, he distinguishes between what is visualised, which is more closely

tied to the visual image, and what is (visually) imagined where the project behind the image holds

sway. Those who allow for radical cases of de-emphasis don’t have to deny that the subjects can

only visualise things more closely tied to the mental image. They just recognise an additional state

more firmly in the grip of the supposition about the image, either via the intention behind the image

or how it is interpreted when apprehended.

If both supplementation and de-emphasis can come into play, then, contrary to what Stock

claims, the proponent of the multiple use thesis does not have to give an account of the content of

the mental image that keeps open all interpretative options for an image with application of only

one kind of supposition: supplementation or de-emphasis. For example, they don’t have to suppose

that there is a minimally specified content of the image that is neutral between whether the image is

of a real tomato or a wax tomato, and whether this object is experienced or not, to be

supplemented in one interpretative way or another. Rather, the basic case could be that of being a

tomato, say, with its genuineness de-emphasised and it interpreted as wax. If there is a basic

characterisation of mental images, it must be motivated by other considerations.

A natural place to look for such a motivation will be how we tend to think about our

perceptual experiences and the corresponding imaginations. In the case of our perceptual

28

Page 29: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

experience, it is plausible that thinking about the perceptual experience draws, first, on a

characterisation of what the experience is of, and second upon our concept of a perceptual

experience (see, for example, Dretske (1995)). Thus, when I visually experience a grey elephant, I

first characterise what the experience is of, a grey elephant, and then, in virtue of my grasp of the

concept of a visual experience in general, talk about a particular experience I am undergoing as an

experience of a grey elephant. I do not, as it were, come at my perceptual experience first by some

further act of mental perception which reveals to me I am undergoing such and such a mental state.

If this is the correct structure for understanding how we think about individual experiences,

then it is plausible to take it across when we consider how a subject should go about imagining that

they are having a visual experience of a grey elephant. They start with visually imagining a grey

elephant and then, by supplementation and their grasp of visual experience in general, attribute to

themselves an imaginary perceptual experience of a grey elephant. So the basic content which the

subject may supplement or de-emphasise at will is the objects and properties in the world,

characterised in one way or another, depending upon whether the subject is visualizing or tactually

or auditorily imagining. It was this idea which was behind my characterisation of what is involved in

imagining yourself stimulated by evil neuroscientists to experience being at the helm of a yacht.

I have argued that there is no technical reason – in terms of requiring inconsistent acts of

supplementation or de-emphasis – why one should reject the multiple use thesis. Nevertheless, the

multiple use thesis does rely upon there being a further contribution by the suppositional element

beyond that which, as one might suppose, is there strictly speaking in the image. The idea that there

is an additional contribution to be made might derive part of its plausibility from relying upon a

certain view about the nature of perceptual content and how it shows up in imaginative content,

namely that perceptual content is to be understood austerely (or conservatively) rather than richly

(or liberally).

29

Page 30: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

According to proponents of the rich (or liberal) view, it is legitimate to suppose that proper

characterisations of perceptual contents may include kind terms – thus we perceptually experience

that something is a pine or a tiger – and related notions such as that of causation (Siegel (2010)).

Rich theorists may also insist that, for example, part of the characterisation of perceptual experience

should include the fact that objects appear to have back sides or occluded parts where, by definition,

there is a sense in which this not present in the subject’s perceptual experience. Austere (or

conservative) views are the denial of rich views. Between proponents of austere views, there can be

disagreement about exactly what may correctly be taken to be a characterisation of the content of

perceptual experience.

Let me illustrate how this difference touches upon the debate. Consider our case of the cat

hiding behind the suitcase. Proponents of the rich view will say that our experience of the suitcase if

we suppose that a cat is hiding behind it is phenomenally different from what it would be if we

didn’t suppose a cat were hiding behind it. Taking this point across to the case of imagination, such a

proponent might insist that the mental image is different too.

The rich vs austere debate seems to be centred around two very strong intuitions which are

not, in fact, in straight contradiction. Bringing them out will suggest a middle way. The first is that

there seems a phenomenal difference between two experiences one might have, one in which one

experiences something as tomato-like before one, the second in which one experiences something

as a tomato. Experiencing something as tomato-like is not meant to suggest there is some noticeable

difference between something experienced to be tomato-like and a tomato. Rather the thought is

that, even if something is a perfect replica, it will seem different if one experiences it simply as just

like a tomato but noncommittal over whether it is one. The coherence of this thought might be

contested but, to make it out, we can draw upon the materials used to express the second intuition.

The second intuition is that it is possible that two things – say a wax tomato and a tomato –

will be impossible to discriminate visually without reference to spatiotemporal position.

30

Page 31: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

The second intuition can be used as a basis for defining a certain kind of property: a sensory

property for a particular sense modality. Thus suppose we stipulate that

Two objects A and B are sensory duplicates for sensory modality S if and only if, for any

subject with sensory modality S, the subject cannot discriminate between them by that

modality without appeal to information about their distinct spatiotemporal location.

Then we can say that

Sensory properties are those that must be shared by sensory duplicates.

In putting matters in this way, I pick up some of the structure of Martin’s discussion (Martin (2010),

pp. 202-203). There are complexities but not ones which vitiate the basic contrast I am seeking to

make. Corresponding to this notion of property, we can introduce the notion of a sensory content, a

content which (only) attributes sensory properties.

If the sensory content of a perceptual experience exhausts its phenomenal content, then the

first intuition is to be rejected. If two kinds of thing cannot be distinguished, by a subject, on the

basis of sense experience in the way indicated, then there is no possible difference in phenomenal

content that they can have. Discriminability in experience settles the range of properties that can be

attributed as part of the phenomenal content of a perceptual experience. However, if it is allowed

that the phenomenal content of perceptual experience is not to be fixed by what is discriminable in

perceptual experience, then the following position is available. The sensory contents of the relevant

perceptual experiences, either correctly characterised partly in terms of a tomato or in terms neutral

over whether the experience is of a tomato or a wax tomato, are the same. It is just that the former

has, in addition, a non-sensory content that there is a tomato before one.

Non sensory perceptual content might seem like a contradiction in terms. Of course, it is

sensory if that is just taken to mean a content legitimately attributed to particular sensory states.

The issue, though, is whether there is perceptual content not governed by the particular constraint

31

Page 32: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

concerning discriminability on the basis of experience identified above. Putting the two intuitions

together we have the thought that, although experiences of sensory duplicates should be attributed

the same perceptual content – hence capturing the fact that they are sensorily indiscriminable –

which content that is may go beyond what the senses can discriminate. Thus, for subjects who are

able to recognise tomatoes, both real tomatoes and wax tomatoes will be presented in exactly the

same way: as tomatoes. Prior to their ability to recognise tomatoes, real tomatoes and wax

tomatoes will be presented as roundish red objects (say). Subjects’ inability to differentiate between

them is the explanation for why they are deceived, when they are, by wax tomatoes. The deception

at hand is a perceptual one (for further defence of non-sensory perceptual content, see Noordhof

(forthcoming)).

In the case of perception, its non sensory phenomenal content seems to be the result of

manifested perceptual recognitional abilities. It is because we have the perceptual ability to

recognise tomatoes that a certain object is presented as a tomato in experience. The corresponding

imaginative content seems to reflect this, not by the same recognitional abilities being exercised on

the mental image, but rather those abilities informing our mental images. It would be a mistake to

suppose that a necessary condition of being able to perceptually recognise something is that we can

sensuously imagine that thing. There are many things about which we can feel: I can’t bring an

image to mind but I know that I’d be able to recognise it if presented to me. Nevertheless, being able

to produce the relevant kind of mental image can be one way of developing our perceptual

recognitional abilities further.

Consider now, the case of imagining a wax tomato. It might be thought that the mental

image serving imagining a wax tomato or a genuine one would be the same. The difference between

imagining a wax tomato and imagining a real one would lie in the suppositional element. However,

once we recognise the existence of nonsensory imaginative content, the matter is not so

32

Page 33: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

straightforward. The mental image content itself may reflect the nature of the suppositional

element.

Even though this is the case, the multiple use thesis has plausibility. It is unlikely that the cat

wholly hidden by the suitcase forms part of the image content. Nevertheless, I do not have to

commit myself to this and, indeed, I would misrepresent the plausibility of the extent of the multiple

use thesis, by relying upon that kind of case. Instead, we should recognise that the multiple use

thesis concerns the sensory content of the mental image and reformulate it thus.

Images with the same sensory content may figure in sensuous imaginings with different

imaginative contents in virtue of the role of different suppositional elements.

This formulation leaves open the possibility that, although the images have the same sensory

content, they are different types of images (as far as their overall content is concerned) because the

suppositional element partly determines the image content – and not just the overall content of the

imagining.

This version of the multiple use thesis, thus, would accommodate Sartre’s view that

In a word, the object of perception constantly overflows consciousness; the object of an

image is never anything more than the consciousness one has of it; it is defined by that

consciousness: one can never learn from an image what one does not know already (Sartre

(1940), p. 10).

If we take the suppositional element as providing the content of which we are conscious, then

Sartre’s view seems to be that a contrast between perception and imagination is that there is

nothing more to the mental image than how we take it. Even if this is so, we might take it that

diverse images can have the same sensory contents, appropriately interpreted, and the additional

content provided, by the suppositional element.

33

Page 34: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

In fact, I think Sartre’s position is overstated for two reasons. First, sensuously imagining

something can involve a performance that can be assessed for its success in producing an image

which expresses the supposition behind it. For example, suppose that I set myself the task of

imagining something which is on the borderline between green and blue but just on the blue side.

The first image I might produce could fail in this task, on reflection it seeming more green to me.

That’s a simple task. Suppose I set myself the task of imagining myself doing some complex activity

in detail. Then I might fail in various ways to produce a sequence of images that seems like a

successful imagining of this activity (see, for example, Noordhof (2002), p. 432). The activity of

supplementation and de-emphasis can help to fix the imaginative content but it cannot overcome all

the inadequacies of the image.

Second, there is the case of deriving knowledge from experientially informed imagining.

Suppose I want to work out how many windows are in the front of my parent’s house. Then while

there might be no particular case of experiencing the front of my parent’s house that I am

remembering, my prior experiences enable me to form an image of it. With the image in mind, I can

count the windows. It is not the case that, in imagining my parent’s house, I supposed it had such

and such a number of windows – certainly not consciously. I discover the number of windows at the

front of my parent’s house by counting them in the image. So, in a certain sense, I am finding more

in the image than I put in it explicitly in the supposition behind the image.

The case just described might be called one of ‘general experiential recollection’ rather than

imagining. A certain kind of experience is recollected even if no particular experience is. However, it

would be a mistake to suppose that the existence of some cases which are more plausibly thought of

as general experiential recollections – I reserve judgement over whether the case offered is one –

should vitiate the general point. Imaginings draw upon, and are constrained by, materials that we

have previously come across. These constraints are not necessarily things that we put into our

34

Page 35: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

imaginings by supposition. Thus, there will be cases in which we have stuff to learn from our

imaginings regarding these constraints.

If Sartre merely meant to insist that experientially informed imagining must encode stuff we

already know about the world in some sense, even if, by imagining, we then are enabled to assert

propositions we could not assert before – for example, regarding the number of windows in the

front of my parent’s house – then nothing I have observed runs counter to this. Nevertheless, it does

not secure the total dependence of the content of the image on the act of interpretation or

intention behind its production.

The two cases described do throw into question, though, a particular diagnosis of the reason

why we never learn anything from an image. Ludwig Wittgenstein, and those influenced by him,

claims that our knowledge of the content of our imaginings is non-observational. We do not discover

features of the imaginative content of our imaginings, rather imaginings are things in which we

actively, even if sometimes nonintentionally, engage. Some seek to explain the fact that we cannot

visually perceive that O is F and imagine that O is F at the same time as a result of this (e.g. Budd

(1989), p. 110). However, that doesn’t explain why we couldn’t perceive that O is F and, with regard

to a quite different token content, also sensuously imagine that O is F. The constraint seems to be

one which derives from attention and not because two attitudes are in conflict over the same token

content. In any event, taking the active nature of imaginings to ground their nonobservationality

ignores the fact that what is active may still be observed even if we have non-observational

knowledge of it too. Thus, it is possible that there are details that we might arrive at via observation

which are not available non-observationally. Appeal to the non-observational character of some of

our knowledge of an activity cannot establish that what we take ourselves to be up to, or how we

interpret a mental image, entirely constitutes its content.

Physical activities allow for the possibility that there might be another observer to arrive at

further truths about their nature. The actor is not the only one who has access to the activity in

35

Page 36: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

question. However, even if one denied that this could be the case for imagination, this point would

not vitiate what I have argued (see e.g. Wittgenstein (1967, 1981), s. 641). Our non-observational

knowledge of the intention behind an image, say, still leaves open the possibility that we may have

observational knowledge of the image concerning whether or not we have been successful in

imagining what we have intended. Getting better or worse at imagining a friend’s face, a certain

situation, and so on, are cases where the gap is revealed. We can also pick up detail from the image

that was not present in the supposition behind it. Imaginers may be the final arbiter (unlike the case

of physical activities) nevertheless we have clear examples of the role of something which looks very

like their observation of the image content. That is not to say that they are introspecting a mental

picture. It is rather that in having a state with imaginative content, everything is not explicit prior to,

or by initiating, the state in question.

4. The Suppositional Element and the Defence of Representationalism

In the first two sections, I rejected an account of the putative non-presentational character of

sensuous imagination by the Dependency Thesis. In the third section, I explained the role of the

suppositional element. These set the parameters for the representationalist account of imaginative

content I am now going to provide.

The difference between the phenomenal content of visually perceiving that O is F and

visually imagining that O is F and, more generally, the difference between perception and the

corresponding imagination, is due to a number of interacting factors. The first, and most familiar, is

that a reduced amount of representations are at work in the case of the sensuous imagination when

compared with that of perception.

It might seem as if, when we imagine a red object, we are really getting nothing like the

presentation of red, even at a reduced level. The two experiences are just very different. This

36

Page 37: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

concern depends upon an unfavourable comparison. If you compare our experience when foveating

on a red object with that we generate as a result of imagining it, then there is an admitted

difference. The phenomenal difference is far less, though, if we compare the nature of an

imaginative content with our experience of objects towards the periphery of our vision. There it can

seem like we are not being presented with them in the same way either. A significant difference

between the foveated and the nonfoveated is that we have far more sensory transducers in the

fovea with, correspondingly more representations of the scene produced as a result. Also the

balance of receptors alters. Towards the periphery there are many fewer cones and more rods, the

former being implicated in the perception of colours (e.g. Wolfe, Kluender and Levi (2009), pp. 36-

40). This suggests that it is, at least, plausible that part of the phenomenal difference can be

accounted for in this way. An appeal to reduced number of representations is, of course, something

which a representationalist can happily make.

Nevertheless, although this is part of the story, it cannot explain why, for example, when we

imagine pains or itches, we don’t feel that we are undergoing something awful or have the desire to

scratch. In these cases, the phenomenal content of what we experience is closely related to how we

respond to what is presented. This makes the following kind of treatment available to

representationalists. They can say that the dispositions to respond are, in fact, the means by which

certain properties are represented to be present. For example, in the case of pain, the awfulness is

represented by the fact that we are strongly disposed to rid ourselves of the pain and are upset by

its presence. The itchiness of an itch is represented by the disposition we have to respond to the

itching part in the way that we do, and so on. The metaphysically necessary connection between

something being presented as awful or itchy, and our dispositions to respond to this, is explained by

the fact that the dispositions are the representation. ‘Does it hurt because we are disposed to be rid

of it or are we disposed to be rid of it because it hurts?’ presents a false choice if the representation

of the latter is provided by the former. The implausibility of functional or dispositional accounts of

37

Page 38: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

consciousness derives, in part, from failure to recognise this connection (for further discussion, see

Noordhof (forthcoming)).

A detailed defence is needed of such a proposal one part of which will be to challenge the

idea that there is just one way in which representational properties are instantiated, for example, as

a result of causal covariances in optimal circumstances, or natural systemic functions (Tye (1995), p.

101; Dretske (1995), p. 162). The more important observation in the present context concerns how

the connections between a certain kind of sensory content and the representational dispositions, to

which it gives rise, change in the case of imagination. The role of supposition, in generating the

mental image, or interpreting it so adjusting its content, shows how sensuous imagining is a mental

activity for which we are responsible, rather than something which is imposed on us as passive

recipients. This alters the way it will strike us. The responses that we make to things which are not

the result of our own activity are turned off, or reduced, when we are.

Two illustrations of this phenomenon are attempts to tickle yourself and causing oneself

pain. In the first case, even if you can bring about the tingling in certain cases, it is hard to generate a

full giggling response. This seems to be because, to some extent, we are able to predict what

sensations are coming by being the source of them and so our response to novelty is much reduced

(see e.g. Blakemore, Wolpert and Frith (2000)). In the case of pain, we often find that, if something

painful is going to be done to us, and we can bring ourselves to do it, then the response is much

reduced if we do it than if somebody else is doing it to us. Consider the lengths we are prepared to

go to get a splinter out when compared to our response to other people doing it.

Imagining a pain or an itch is something which, almost invariably, we are aware of our

mental activity in so producing. In contrast to tickling oneself, or causing pain in oneself, the control

over these imaginings is far greater and the expectations reduced even further. As a result, the

responses we are usually inclined to make are much reduced. That’s why these cases seem

particularly problematic for the representationalist. However, once one recognises that our

38

Page 39: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

dispositions to respond in various ways are themselves representation of certain response

dependent properties, we have a ready representationalist explanation of the phenomenal

difference between imagining oneself having a pain in one’s foot and perceptually experiencing that

one has one.

If dispositions to certain responses have a representational role, can the subject being active

in imagining in the ways indicated also have representational consequences? Consideration of the

contrast between sensuous memory and sensory imagination suggests this possibility. Certain kinds

of sensuous memory seem not to present its objects and properties but imply their reality in the

past. Compare the following two kinds of state: imagining that A occurred in the past and sensuously

recollecting that A occurred in the past. There is nothing in these ascriptions of contents that suggest

that the second is presented as really occurring. Nor do we achieve the impression of reality in the

past by imagining that A really occurred in the past. The characterisation of this imaginative project

does not make it into a sensuous memory, though it might be an imagination of a sensuous memory.

Nevertheless, the phenomenal content of sensuous memory implies the reality where that of

sensory imagination does not.

John Searle seeks to capture the character of sensuous memory by giving it a certain kind of

self-reflexive character. To give his example, a visual memory of a particular flower, f, has the

content: that f is present, f is the cause of a particular visual experience, which is a cause of this

memory that f is present (Searle (1983), pp. 95-97). The immediate problem with Searle’s proposal is

that one conjunct of the content provided is that f is present which would seem to entitle the

subject to judge that the flower is present rather than was present.

Appeal to a variant of the Dependency Thesis is of no assistance in accounting for the sense

that we are having a sensuous experience of something past in sensuous memory either. Suppose

the content of sensuous memory is that I experienced a particular flower, or I experienced that a

particular flower was thus and so. One problem is that it makes our sensuous memories primarily

39

Page 40: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

concern our experiences and not the events which occurred in the past (Evans (1982), p. 240,

Recanati (2009), p. 140). But the more fundamental difficulty is that we would have to conclude that

imagining that I experienced a particular flower, or experienced that a particular flower was thus and

so, would give us the same phenomenal content: having a sensuous experience of something in the

past. This is not how those sensuous memories deriving from particular experiences we had in the

past – what is commonly called episodic or experiential memory – strikes us (hereafter, I’ll dub them

‘particular sensuous memories). These sensuous memories are generally phenomenally distinct in

the relevant respect from any imaginings concerning past experiences. The exception may be

unrecognised memories discussed below.

Martin appeals to the Dependency thesis together with the claim that the representation of

the experience in recall links the subject to an actual episode which was experienced, rather than a

type of episode (Martin (2001), p. 280). However, the link to a particular episode experienced cannot

remedy the deficiency identified in appealing to the Dependency Thesis to account for the sense of

past. Consider Martin and Deutscher’s case of unrecognised remembering. A painter imagines a

scene which, in fact, is something they experienced in childhood. This satisfies the particularity

condition without invoking a sense of pastness or the presentation of something in the past (Martin

and Deutscher (1966), p. 168).

It might be thought that the difference between imagining and sensuous memory can be

captured by the fact that, in the case of the latter, a subject is disposed to judge upon the basis of its

content that certain events happened in the past. However, suppose we try sensuously to imagine

how things were in the past and, by working at it, are increasingly inclined to think we are doing a

particularly good job. Over time, we might be disposed to form beliefs in virtue of what we imagine

about the past. Or suppose that we are intellectually arrogant historians who have vivid sensuous

imaginations. Even then, we won’t be inclined to take our sensuous imaginings to imply the reality of

the manifest objects and properties of the sensuous content in the past.

40

Page 41: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

François Recanati has suggested that the phenomenal difference between sensuous memory

and imagination stems from our awareness of the mode of respective mental states in question.

Unlike sensuous imaginings, the sensuous memory that O is F involves the content that O is F being

presented as true for a past occasion (Recanati (2009), pp. 141-142). In the case of perception, the

content that O is F is presented as true for the time of the perception. Sensuous imagination has its

context of evaluation settled by the imaginative project. For example if I sensuously imagine that it is

raining in London then the content of my mental image is true if it happens to be raining in London

as I am imagining it (Recanati (2009), pp. 199-202).

The problem with this proposal as an account of the phenomenal content of our sensuous

memories (and how these contrast with perceptions and imaginings) is that it involves the rather

sophisticated idea of something – a content – that is not a means of representation, and not what is

represented, O being F – being presented as true. O being F cannot be presented as true because O

being F is an object having a property. These are not the kind of things to be presented as true or

false. Nor is the means of representation presented as true either because it is not even clear what

the means of representation are in the case of sensuous memory. In all three cases, we seem to go

right through to what is represented. We don’t take something else – the means of representation –

to be presenting something as true for a past occasion.

This is not to reject Recanati’s framework for understanding the two determinants of truth

conditions for a content-bearing mental state, viz what is represented by the state and a context of

evaluation which identifies the situation in which the content of the state should be evaluated. The

point is simply that this does not capture the phenomenal content of the state any more than the

positions Recanati has just rejected.

The implication of reality seems, rather, to stem from the way in which imaginings and

sensuous memories are produced. In the case of imaginings, we are the organisers of the sensory

content, in the case of experiential memory, while we might intentionally seek to remember

41

Page 42: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

something, we are not the organisers of the sensory material in question. Rather, they come up, to a

substantial degree, pre-organised by our prior experiences. Although the recall of particular

sensuous memories are often initiated by us, the organisation of what they involve is not - if we

have the kind of state people generally have in mind by the notion of a particular sensuous memory.

Recognising that particular sensuous memories arise as a result of a substantial degree of

pre-organisation by our prior experiences is consistent with psychological results that note that all

memories involve a substantial amount of construction and addition, for example, generation of

sensory content from a general sense of what happened or a gist (see e.g. Schachter and Addis

(2007), p. 778). This only qualifies the extent to which what is remembered is a recreation of the

content of a previous experience. It is still an organisation of sensory content that is relatively

independent of us, for example, as something over which we are capable of control in the way that

we can control what we sensuously imagine.3 The difference between sensory imagination and

particular sensuous memory to which I am adverting has some support in empirical work. In

particular, the initial constructive phase of the sensory content involves different patterns of

activation in sensuous imagination to particular sensuous memory (see Okuda et al (2003); Addis et

al (2007), and Schachter and Addis (2007), p. 781).

Obviously there are complexities here that merit further discussion. The basic idea, though,

might be developed in one of, at least, two ways. The first way would be to take the implication of

reality, or otherwise, of the manifest properties and objects to stem from whether or not the subject

is consciously aware that he or she is the organiser of the content. Thus, in the case of dreams, we

take them to be real, when we are undergoing them, because we are unaware of the fact that they

are the workings of our imagination whereas when we are imagining something we are aware of

what we are up to.

3 I am grateful to Peter Langland-Hassan for pressing me to consider how the basic idea relates to constructive theories of memory.

42

Page 43: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

According to the alternative way of developing it, certain properties of objects are

represented by the presence, or absence, of our organising the representations of these objects

regardless of whether or not we are aware that we have organised these representations. For

example, unreality is represented by our organising process or, reality, by the absence of

organisation. In the latter case, the features of the processes that give rise to perception or

particular sensuous memory which rule out these processes being the result of our organisation are

responsible for further representational properties. The apparent reality of our dreams will be the

result of them not being organised by us rather than our failing to be aware of the process of

organisation.

How might we decide between these two positions? And, within these two positions, which

should be taken to be fundamental: the representation of reality or unreality? Sufferers of de-

realisation dissociative disorder may provide an answer as to which is fundamental. At least some of

them have the feeling that what is sensorily presented to them is unreal. At the same time, they also

experience a loss of agency (Ackner (1954), pp. 842-843, quoting previous authors, Sierra and

Berrios (2000)). If the awareness of one’s agency, or the involvement of agency of a certain kind, is

required for the representation of something as unreal, then we would have no explanation of what

was going on in this case.4 One would expect loss of agency to make everything seem real not

unreal.

This suggests that objects and properties are represented to be real if they are unorganised,

or are experienced to be unorganised, by the subject in the context of a subject’s appreciation of his

or her own agency. Reality is primary but it is not simply the features that make the process

unorganised that makes something represented as real. Rather, this must be present when one has

a sense of one’s own agency. Thus, when subjects sense of their own agency declines, everything

begins to seem unreal. Sensory imagination presents things as unreal because it is just one obvious

way in which the condition for their representation as real is unmet.4 I am grateful to Anya Farennikova for making me thing more clearly about this phenomenon.

43

Page 44: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

If reality is primary, then dissociative disorder also presents a consideration in favour of the

second way of developing the representation of reality/unreality. According to the first way, a

subject would have the sense of the reality of what is presented by being aware of the fact that

sensory material was unorganised by themselves. However, with loss of agency, presumably they

have that awareness and yet they experience things to be unreal.

Nevertheless, there might appear to be a consideration on the other side. In cases of

unrecognised remembering, a subject takes themselves to be imagining something but, in fact, they

are recalling something which occurred in the past. If the features of the process which generated

the particular unrecognised experiential memory are responsible for the representation of reality,

then the unrecognised remembering shouldn’t (it might be thought) present the objects and

properties remembered as unreal (that is, as something imagined). Yet, that is precisely what

happens.

The first way of developing the position would struggle too, though. Unrecognised

remembering would have to involve non-factive awareness that one was organising the sensory

material that constituted the memory. After all, in the case of unrecognised remembering, it is

assumed to be precisely not the case that the subject is, in fact, organising the material in question.

But how could the subject have such an illusion of mental activity? It is more plausible to think that

what is, in fact, going on is that a subject is organising the sensory material in a way which, although

they do not realise this, is guided by information stored as a result of their experience of a past

event.

If that’s right, then we must recognise two relevant kinds of sensuous memory. Standard

particular sensuous memory is often characterised in terms of the idea of retained acquaintance

with what was previously experienced. It is a nice question how this is characterised but, at the

minimum, it might be thought of as an activation routine which enables one to recreate an

experience of the same type as the prior experience, with whatever variation or supplementation

44

Page 45: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

there is due to the element of construction within a gist mentioned above. Because it is an activation

routine, the idea is not that the process in question is one over which we, in principle, have some

kind of control. Rather, it is a process which is relatively automatically triggered by the attempt to

recollect. Sensory imagination, on the other hand, involves the possibility of control. If our

experiences in the past guide our sensory imaginings, then we have a feature of standard memory,

relevant structure of the activation routine, this time being run through a process over which we

have control: sensory imagination. That’s what the second kind of sensuous memory is, those we

don’t recognise as such: past-guided sensuous imaginings.

If that is what is going on in the case of unrecognised remembering, then there is an obvious

response on behalf of the second way of developing the position. Unrecognised memories present

things, without being committed to their existence, because they are the result of organisation by

the subject, albeit guided by past information. They are not the result of an unmediated process of

organisation – the activation routine - which makes the representations present the manifest objects

and properties as real, specifically, as having existed in the past.

If representation of something as existing is dependent upon something not being organised

by the subject, in the context of a subject’s sense of his or her own agency, then how can a subject

think that something is real or imagine that something is real? For these states, the lack of

organisation conditions don’t seem to be met. On the other hand, if it is argued that we can imagine

that something is real or think that it is, then there seems to be another way of representing that

something is real. In which case, we have not, the charge would go, supplied a pure

representationalist account of the representation of reality. The phenomenal difference between

these two cases – experience or recollecting something as real, and imagining or thinking it is –

cannot be captured in terms of their representational properties but rather the different ways in

which these representational properties are realised.

45

Page 46: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

To deal with this objection, it is important to distinguish between two ways in which

something may be represented as real. According to the first, that which shows up in perceptual

experience and sensuous memory, the representation of reality is internal to what is represented to

be real. It has roughly the following structure: Really (O being F). No further state or predication is

required. In the case of thinking that something is real, what is represented is that O being F is real.

Here, the addition of reality is as a predicate. In the case of sensuously imagining that something is

real, we sensuously imagine O being F and suppose O being F is real. There is an additional role for

supposition with the relevant content. The latter two cases involve the representation of reality as

external to O being F. These differences are representational differences even if the resulting states

can all be described as having the content that O being F is real. They each have a different

representational structure to generate this content. Only when the representation of reality is

internal to what is represented to be real can we take O being F as manifesting its existence in the

content in question.

As a premise in an argument for the claim that existence is not a property, Hume writes

‘There is no impression or idea of any kind, of which we have any consciousness or memory, that is

not conceived as existent’ (Hume (1739), 1. 2. 6). As an observation about the existence of a certain

mental state with the manifest objects and properties that we suppose them to have, this may be

unexceptionable. Even those who deny that there is a content in the case of hallucinations may

allow that we conceive that what is presented in a hallucination exists. However, as an observation

about the content of a mental state, the observation is false if what I have argued above is correct.

When I imagine O being F, I don’t represent O being F as existing. Indeed, the phenomenal contrast

between sensuous memory and sensory imagining illustrates exactly this point.

It may be true – I see no harm in it – that when I visually imagine that O is F, I am taking O

being F to exist in the world imagined. However, I don’t suppose that it follows from this that I am

taking O being F to exist in the world about us, viz to be a part of reality. Instead, it follows simply

46

Page 47: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

from what is required to constitute an imaginary world in much the same way as describing various

characters in a fiction is sufficient for them to exist in the fiction. Representing that really (O being F)

involves something more and imaginings and thoughts are ill equipped to do that internally. Instead,

they must add an existence claim to capture the same content.

The recognition that the case of imagination shows that there is something more present in

perceptual experience and experiential memory, than what is sensorily presented, does not

undermine the thought that there is nothing more to existence than an object and its properties.

Existence is not some further property to be added. That’s why the treatment of existence does not

require recognising, according to this view, a property of existence independent of the objects and

properties that exist. The important point is that, unlike reality as it were, sensuous states do allow

for the possibility of objects and properties being part of the sensory content of a state without it

being presented that they are existing items, namely when we are sensuously imagining them. The

gap falls on the side of representation and not reality.

The distinction between two ways in which the reality of something may be represented is a

contribution to the characterisation of the distinctive modes of perception and sensuous memory by

contrast with sensuous imagination. The case of the arrogant historian revealed potential problems

with alternative accounts appealing to differences in our dispositions to form beliefs. Moreover,

representationalists cannot afford to take the difference in modes as primitive, perhaps related to

different conditions of evaluation, and there is reason not to do so. As we saw in the discussion of

Recanati’s position, it is difficult to characterise the different commitments in which a sensory

content may stand without rooting this in a presentational difference of the content itself.

5. Concluding Remarks

47

Page 48: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

We have explained the phenomenal difference between m-imaginings and m-perceptions in three

ways compatible with the Similar Content Hypothesis. The first was that m-imaginings were

representationally more impoverished: less detail was represented. The second was that m-

imaginings failed to elicit the same responses and, hence, certain response-dependent properties

(represented by the disposition to have these responses) failed to be represented. The third was

that the conditions under which the internal representation of the reality of what was represented

were not met. It is in the light of these three that the manifest objects and properties of imagination

are a proper subset of the manifest objects and properties of perception. Taken together, they

account for why it seems that sensuous imaginings are non-presentational when, in fact, they are

but just in attenuated form.

A representationalist approach to imaginative content has not been defended because of an

unmotivated commitment to representationalism in the face of challenges from other approaches.

Rather, the defence of imaginative content has taken up what seem to be phenomenal differences in

what is presented in sensuous imaginings and sought to evaluate whether these might correspond

to representational differences. If the phenomenal differences were non-presentational, that would

be a different matter.

A distinctive component of the defence is that there are different ways in which things are

represented. By that I don’t mean that some phenomenal differences are accounted for in terms of

different kinds of representation of the same phenomenal content. Rather, the point is that, the only

way in which a certain phenomenal difference could correspond to a different representational

content is if there were a different kind of representational property in play e.g. not causal

covariance in optimal circumstances but rather whether a process is organised by a subject or not; or

whether there is a certain kind of response or not (for other illustrations, see Noordhof (2008) and

Noordhof (forthcoming)). The grounds for recognising that there are different kinds of

representational property derive from taking these presentational differences seriously.

48

Page 49: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

What unifies these different representational properties apart from the fact that they all

have contents, albeit different contents? The provisional answer is that they play, and in particular,

certain of their features play, the representational role. This is not a functional account of

representational properties but it is a functional account of what makes something a representation

and which of its properties are representational. Failure to respect this distinction is responsible for

some premature obituaries of representationalism although, of course, over use of it may give rise

to overexcited eulogies.5

References

Brian Ackner (1954), ‘Depersonalisation I. Aetiology and phenomenology’, British Journal of

Psychiatry, 100, pp. 838-853

D. R. Addis, A. T. Wong and D. L. Schachter (2007), ‘Remembering the past and imagining the future:

common and distinct neural substrates during event construction and elaboration’,

Neuropsychologia, 45, pp. 1363-1377.

Sarah Jayne Blakemore, Daniel Wolpert and Chris Frith (2000), ‘Why Can’t You Tickle Yourself’,

Neuroreport, 11, 11, 3, R11-16.

John Campbell (1994), Past, Space and Self (Cambridge, Massachusetts, The MIT Press).

Fabian Dorsch (2010), ‘Transparency and Imagining Seeing’, Philosophical Explorations, 13, no. 3, pp.

173-200.

Fred Dretske (1995), Naturalizing the Mind (Cambridge, Massachusetts, The MIT Press).

5 Many thanks to the audiences of the Perceptual Imagination and Perceptual Memory conference, an Imagination and Belief workshop in Gargnano organised by Cain Todd and others, and a workshop in Gargnano organised with Milan on Nonperceptual Sensuous States, for comments on previous versions of this material. I also wish to thank the Leverhulme Trust for giving me a Major Research Fellowship enabling me to undertake some of the research in this paper.

49

Page 50: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

Gareth Evans (1982), The Varieties of Reference (Oxford, Oxford University Press).

Dominic Gregory (2010), ‘Imagery, the Imagination and Experience’, The Philosophical Quarterly, 60,

no. 241, pp. 735-753.

David Hume (1739), A Treatise of Human Nature (Oxford, Oxford University Press).

M. G. F. Martin (2001), ‘Out of the Past: Episodic Recall as Retained Acquaintance’, Christoph Hoerl

and Teresa McCormack (eds.), Time and Memory (Oxford, Oxford University Press), pp. 257-

284.

M. G. F. Martin (2002), ‘The Transparency of Experience’, Mind and Language, 17, no. 4, pp. 376-

425.

M. G. F. Martin (2010), ‘What’s in a Look?’ Bence Nanay (ed.), Perceiving the World (Oxford, Oxford

University Press), pp. 160-225.

Paul Noordhof (2002), ‘Imagining Objects and Imagining Experiences,’ Mind and Language, 17, no. 4,

pp. 426-455.

Paul Noordhof (2008), ‘Expressive Perception as Projective Imagining’, Mind and Language, 23, no.

3, pp. 329-358.

Paul Noordhof (forthcoming), 'Evaluative Perception as Response Dependent Representation', Robert

Cowan and Anna Bergqvist (eds), Evaluative Perception (Oxford, Oxford University Press).

J Okuda et al (2003), ‘Thinking of the future and the past: the roles of the frontal pole and the medial

temporal lobes’, Neuroimage, 19, pp. 1369-1380.

Christopher Peacocke (1985), ‘Imagination, Experience and Possibility’, John Foster and Howard

Robinson (eds.), Essays on Berkeley (Oxford, Oxford University Press), pp. 19-35.

Brian O’Shaughnessy (2000), Consciousness and World (Oxford, Oxford University Press).

50

Page 51: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

Francois Recanati (2009), Perspectival Thought (Oxford, Oxford University Press).

Jean-Paul Sartre (1940, 1986), The Imaginary (translated by Jonathan Webber) (London and New

York, Routledge).

Daniel L. Schachter and Donna Rose Addis (2007), ‘The cognitive neuroscience of constructive

memory: remembering the past and imagining the future’, Philosophical Transactions of the

Royal Society, B, 362, pp. 773-786.

John R. Searle (1983), Intentionality (Cambridge, Cambridge University Press).

Mauricio Sierra and German E. Berrios (2000), ‘The Cambridge Depersonalisation Scale: a new

instrument for the measurement of depersonalisation’, Psychiatry Research, 93, pp. 153-

164.

Susanna Siegel (2010), The Contents of Visual Experience (Oxford, Oxford University Press).

Kathleen Stock, ‘Mental Images and the Multiple Use Thesis, Manuscript.

Evan Thompson (2008), ‘Representationalism and the phenomenology of mental imagery’, Synthese,

160, pp. 397-415.

Charles Travis (2004), ‘The Silence of the Senses’, Mind, 113, no. 449, pp. 57-94.

Michael Tye (1995), Ten Problems of Consciousness (Cambridge, Massachusetts, The MIT Press).

Bernard Williams (1966), ‘Imagination and the Self’, reprinted in his (1973), Problems of the Self

(Cambridge, Cambridge University Press), pp. 26-45.

Ludwig Wittgenstein (1967, 1981), Zettel (Oxford, Basil Blackwell).

Jeremy M. Wolfe, Keith R. Kluender and Dennis M. Levi (2009), Sensation and Perception

(Sunderland, Massachusetts, Sinhauer Associates, Inc.)

51

Page 52: Web viewThis indicates that, whatever the phenomenal similarities, m-imaginings are a sui generis phenomenon with their own particular characteristics. ... In a word, the

Paul Noordhof,

Department of Philosophy,

University of York

Heslington,

York YO10 5DD

[email protected]

52