Phenomenal Space and the Unity of Conscious Experience
Douglas B. Meehan
Philosophy and Cognitive Science
City University of New York, Graduate Center
365 Fifth Avenue
New York, NY 10016-4309
U.S.A.
dbmeehan@yahoo.com
PSYCHE, 9(12), May 2003
Previously held: http://psyche.cs.monash.edu.au/v9/psyche-9-12-meehan.html
Copyright (c) Douglas B. Meehan 2003
KEYWORDS: unity of consciousness, spatial perception, phenomenal space, sensory qualities, higher-order thoughts, cross-modal integration.
COMMENTARY ON: Barry Dainton. (2000) Stream of Consciousness: Unity and Continuity in Conscious Experience. (International Library of
Philosophy Series). London, New York: Routledge. xvi + 254pp.
ISBN: 0415223822.
One's contemporaneous conscious mental states seem bound in a single, unified experience. Dainton argues, against what he calls the S-Thesis, that we cannot explain such co-consciousness in terms of states' being located in a single phenomenal space, a functional space posited to explain our ability to locate ourselves relative to perceived stimuli. But Dainton's argument rests on a conflation of egocentric and allocentric self-localizing, and thus fails to undermine the S-Thesis. Nevertheless, experiments on visual neglect (Bertelson et al., 2000) suggest one can have unconscious mental states that are located in the same phenomenal field, so the S-Thesis fails after all. I examine a modified version of the S-Thesis according to which mental states are co-conscious when one is aware of them via a higher-order sensation that represents them as located in the same phenomenal field. But among other problems, this view fails to explain the co-consciousness of intentional states, which aren't located in phenomenal fields. Finally, I argue that a higher-order-thought model of consciousness (e.g., Rosenthal, 1997, forthcoming) best explains the apparent unity of experience in terms of one's tacit assumption that all the first-person thoughts in virtue of which one is conscious of one's mental states refer to the same individual.
1. Introduction
At any given waking moment, one consciously experiences an amalgam of
diverse mental states. My current experience includes visual sensations
of my computer screen, tactile sensations of the keyboard, auditory
sensations of a truck, a feeling of tension in my back, and thoughts
about what I will write next. These simultaneous mental states seem
unified in one consciousness.
Barry Dainton (2000) dismisses a number of explanations of such
simultaneous co-consciousness, concluding that it is a basic, but
inexplicable relation. Among the accounts Dainton rejects is the view
that two states are co-conscious in virtue of being located in the same
phenomenal space. Though I will argue that Dainton's argument against
this view fails, I offer other reasons for denying it. I then examine a
modified version that attempts to explain co-consciousness in terms of a
higher-order-sense view of consciousness. Finally, I argue that
co-consciousness is better explained in terms of a higher-order-thought
model of conscious, such as David M. Rosenthal's (1997).
2. Projectivism, Phenomenal Space, and the S-Thesis
Among the explanations of simultaneous co-consciousness Dainton rejects
is the S-Thesis:
.. simultaneous experiences are co-conscious solely in virtue of
occurring at the same time within a single unified three-dimensional
phenomenal space; being thus spatially connected is both sufficient
and necessary for co-consciousness. (p. 61)
Understanding this thesis requires an understanding of Dainton's theory
of perception.
Dainton adopts a Lockean projectivist theory of perception, according to
which the properties we perceive objects as having are actually
properties of our perceptual states. We indirectly perceive the
nonmental causes of sensations in virtue of being directly aware of the
properties of the sensations (p. 18). For instance, when one sees a
Coke can, one perceives it as being red. But it is one's visual
sensation, not the can, that is red. It is in virtue of being aware of
the redness of one's visual sensation that one perceives the can.
Projectivism is recommended by two familiar arguments. First, we see
surfaces as uninterrupted expanses of color. The can's surface seems
saturated with redness. But, according to physics, between the
molecules composing the surface are pockets of empty space. Since the
surface is full of empty spaces, but expanses of color are not, colors
are not properties of these surfaces (p. 15).
Second, visual sensations of color occur in the brain at the end of
causal chains usually beginning at photon-emitting surfaces. But
sensations of color can occur in the absence of such stimuli, as
hallucinations show. Since one can have the same kind of sensation with
or without the normal stimulus, the color must be a property of the
sensation, not the stimulus. Likewise for sounds, smells, textures, and
the other so-called secondary qualities (p. 15-16).
According to Dainton, projectivism suggests that conscious experience
has a spatial character. One ordinarily mistakes colors, sounds, and
smells as properties of perceptible stimuli, but these immediate objects
of experience are actually phenomenal properties of sensations. So
though secondary qualities seem to us to be located in perceptible
space, they are actually located in a three-dimensional mental, or
phenomenal, space.
The plausibility of phenomenal space is strengthened by Dainton's
account of hallucination. When one hallucinates a vase one presumably
has the same kind of experience one has when one actually sees a vase;
the difference is that only seeing is veridical. One sees the vase as
being in a particular location and as having a particular shape and
size. Since the vase is absent in the case of hallucination, the
apparent shape, size, and location must be properties of the
hallucinated vase, a phenomenal object. If there are phenomenal spatial
properties of such phenomenal objects, it makes sense to speak of
phenomenal space.
Phenomenal space is, then, a three-dimensional mental replica of
perceptible space in which secondary qualities such as colors and sounds
are phenomenally located. Since all of one's sensations seem to be
phenomenally located, perhaps being so located is what makes them
co-conscious. Dainton calls this claim the S-Thesis.
But there are problems with projectivism that may seem to undercut the
S-Thesis. After raising these problems, I will show that, in fact, the
S-Thesis does not rest on projectivism.
The first problem is that projectivism seems to preclude unconscious
perception. According to projectivism, we indirectly perceive nonmental
causes of our experiences in virtue of being directly aware of our
experiences. So we cannot perceive stimuli without being aware of our
perceptions of them. This seems to preclude unconscious perception,
since arguably an experience is conscious when, and only when, one is
conscious of it.<1> But there is good evidence for the distinction
between conscious and unconscious perception. For example,
masked-priming experiments show that subjects see stimuli they are
unaware of seeing. Though the subjects deny seeing the masked prime, it
affects their subsequent behavior and reasoning. Studies on blindsight
(Weiskrantz 1986), visual-form agnosia (Milner & Goodale 1995), and
change detection (Fernandez-Duque & Thornton 2000; Simons et al. 2002)
provide similar examples.
Second, according to projectivism secondary qualities, such as colors,
are mental properties not stimulus properties. So our ascriptions of
color to stimuli are always false. But both perceptual experiences and
stimuli have spatial properties, though the former have phenomenal
spatial properties and the latter have perceptible spatial properties.
So our ascriptions of location, shape, and size to stimuli are not
systematically false. Given that we can explain the spatial character
of sensory experience without concluding that our ascriptions of spatial
properties to stimuli are systematically false, we should be able to do
the same for ascriptions of colors, sounds, and smells. This is
important in light of the highly counterintuitive nature of the claim
that ascriptions of colors to stimuli are always false. And, though
sometimes unavoidable, such counterintuitive claims should be avoided
when possible. A theory that avoids attributing systematic error will
be better off for it.
3. The S-Thesis and the Homomorphism View of Sensory Quality
Projectivism claims both that we are mistaken in our ascriptions of
colors, sounds, and textures to perceptible stimuli and that we perceive
stimuli in virtue of being directly aware of our sensations. The latter
claim seems false because we can perceive stimuli without being
conscious of perceiving them. And a theory that maintains the truth of
our everyday ascriptions of colors, sounds, smells, and textures to
stimuli is preferable on commonsense grounds.
But even if these projectivist claims are false, perception still
requires mental properties similar to those posited by projectivism.
One can only perceive the difference in color between red and green
stimuli if one's visual sensations differ in some way corresponding to
the difference between red and green. For every color discrimination
one makes, one's visual states must have a corresponding mental color.
But one need not be conscious of a state's mental color in order for it
to enable the perception of a color. It is in virtue of having these
properties that sensations perform their perceptual roles.
Likewise, that one can visually perceive differences in stimulus shape,
size, and location suggests that one's visual sensations differ in ways
corresponding to these perceptible spatial differences. So there must
be mental shapes, sizes, and locations. Such properties constitute a
mental visual field, a functional space that enables one to perceive
spatial differences among colored stimuli. As such, mental spatial
properties are higher-order properties of mental colors. And like
mental colors, one need not be conscious of mental spatial properties
for them to enable perception of stimulus shape, size, or location.
In addition to perceiving differences and similarities between stimuli,
one can perceive degrees of difference and similarity. For instance, we
see red as more similar to orange than it is to green. This suggests
that, not only are there mental counterparts to perceptible differences,
but those mental counterparts resemble and differ from one another in
ways parallel to the ways their perceptible counterparts resemble and
differ from one another. So mental red is more similar to mental orange
than it is to mental green. The similarities and differences among
mental colors are homomorphic to the similarities and differences among
perceptible colors. This view applies equally well to the other
so-called secondary qualities, such as sounds, smells, and textures.
It also applies to spatial sensory qualities. One can see spatial
similarities and differences between colored stimuli. For instance, a
red patch off to the far left and a red patch slightly off to the left
are more similar to one another than either is to a red patch off to the
right. Being off to the far left is more similar to being slightly off
to the left than either is to being off to the right. One perceives
these similarities and differences in virtue of having visual sensations
with properties that resemble and differ from one another in parallel
ways. Being mentally off to the far left is more similar to being
mentally slightly off to the left than either is to being mentally off
to the right.<2>
According to this view, there are both perceptible properties and
corresponding mental properties in virtue of which we perceive them.
Perceptible colors are properties of light-reflecting surfaces and
mental colors are properties of visual sensations. Likewise,
perceptible shapes are properties of physical stimuli and mental shapes
are pr operties of sensations. But we refer to both a perceptible
property and its mental counterpart with the same predicate. Stop signs
are red and sensations of them are mentally red. We use 'red' to refer
to the perceptible color when we utter the sentence 'Stop signs are
red'. And we use 'red' to refer to the mental color when we utter 'When
I see a stop sign I have a sensation of red'. Since there are
perceptible colors, ascriptions of colors to surfaces are not
systematically false. So we avoid the inevitable systematic error to
which Dainton's projectivism is committed.
This view has the added benefit of explaining the counterpart relation
between mental properties and their perceptible counterparts
independently of their resembling or being identical to them. A visual
sensation does not enable the perception of red in virtue of being red.
It merely resembles and differs from other mentally colored sensations
in ways parallel to the ways that red stimuli resemble and differ from
other colored stimuli. And visual sensations need not be next to one
another in order for one to see two stimuli as next to one another. They
merely need to relate to one another in a way that corresponds to the
way stimuli that are next to one another in the field of view relate to
one another.
Since spatial sensory qualities are posited to explain how we perceive
spatial properties of stimuli, we might posit a cross-modal family of
spatial qualities, or cross-modal sensory field, to explain cross-modal
perception, e.g., when one perceives a sound as coming from where one
sees a bird. Since all of our sensory experiences seem to correspond to
some location of a stimulus, suggesting their mental location in a
sensory field, maybe being mentally located in such a field is what
makes simultaneous mental states co-conscious. This is, in effect, the
S-Thesis.
4. Dainton's Objections to the S-Thesis
Dainton claims, however, that we can imagine co-conscious mental states
located in different phenomenal spaces. This, he argues, suggests that
the S-Thesis is false.
Dainton offers the following thought experiment. Imagine your brain is
removed from your body and placed in a vat. Though your brain is
separated from your body, it remains connected to it by radio
transmitters. In addition, it is connected via radio to artificial eyes
and ears that are separated from your body.
With your new eyes and ears turned off, your body is dropped in the
ocean. You feel the water around you and realize where you are but you
see and hear nothing. You locate yourself just by your bodily
sensations.
Your eyes and ears are then placed on top of a mountain and activated.
You see a bird and hear its call. Meanwhile, your body remains
underwater.
According to Dainton, it might seem to you that you are in two places at
once; your body feels like it is underwater and your eyes and ears make
it seem that you are on a mountain. Since you seem to be in two places,
your bodily and audio-visual sensations must be located in different
phenomenal spaces. Nonetheless, they are co-conscious. Since, Dainton
argues, the states are co-conscious without being present in the same
phenomenal field the S-Thesis is false.
But even if we can imagine Dainton's scenario, it does not establish
that the S-Thesis is false. It merely establishes that our conception
of co-consciousness, or what we take that conception to be, is different
from the S-Thesis. Our imagination is not a reliable enough gauge of
reality for it to establish anything more.
Also, it is unclear that in this situation you would seem to be in two
places at once. It is likely that you would seem to be on the mountain
having hallucinatory bodily sensations of water, or visa versa. One can
hallucinate without thinking one's sensations are veridical. For
instance, one can hallucinate an elephant in the refrigerator while
thinking that there is no elephant there. One thinks one is just
"seeing things."
In fact, there is further reason to think that even if you did seem to
be in two places, your so-seeming would be due to something other than
your bodily and audio-visual experiences' being located in separate
phenomenal fields. You would, no doubt, maintain cross-modal
integration, such as hand-eye coordination. When asked to point at the
bird, you would point even though your finger would feel like it was
underwater. Your kinesthetic sensations would still be calibrated to
your visual sensations even though you would not successfully point at
the bird. This suggests your bodily sensations would be located in the
same phenomenal field as your audio-visual sensations. So, if you did
seem to be in two places at once, this must be due to some other factor
involved in your locating yourself, not the phenomenal locations of your
sensations.
Dainton anticipates this objection. He claims this maintained
cross-modal integration would indicate only that the sensations are
located in the same imagined phenomenal space, not the actual one (p.
83).
But the only factor that could be cited to distinguish imagined from
actual phenomenal space is the connection between phenomenal and
perceptible space; imagined space inaccurately represents perceptible
space, as indicated by your pointing at the seabed when trying to point
at the bird. But phenomenal space often inaccurately represents
perceptible space, as cases of illusion show. So, again, your seeming
to be in two places at once does not entail that your sensations would
be co-conscious without being located in the same phenomenal field.
In addition, the thought experiment rests on the assumption that since
it would seem to you that you are in two different places, your bodily
and audio-visual sensations would be located in two distinct phenomenal
fields. And this rests upon the assumption that it is one's sensations
alone that enable one to locate oneself.
But we locate ourselves in two different ways: relative to perceived
objects and relative to unperceived landmarks. Only the first is
enabled by sensations alone.
One can locate oneself relative to the entities one currently perceives.
One is in front of a blue sign, on a hard sidewalk, below the sound of a
jet. Such localization involves perceiving these stimuli. And one does
this in virtue of having sensations with mental colors, sounds, and
pressures located in a phenomenal field. One has a sensation of blue
mentally ahead, a sensation of a hard sidewalk mentally below, and a
sensation of a jet's noise mentally above. To locate oneself in this
way, one just needs to have sensations with these mental locations.
But sensations are insufficient for locating oneself allocentrically. If
one is lost in a library, one can locate oneself relative to the books
one sees, but one cannot find the exit. Locating oneself relative to
perceived entities isn't enough to locate oneself relative to
unperceived ones, such as an unseen door. Likewise, standing in a movie
set of Times Square one might think one is in New York, when one is
actually in Hollywood. One correctly determines that one is in front of
the Sony sign and beside a subway entrance, but wrongly localizes
oneself in New York. One cannot allocentrically localize oneself simply
based on what one perceives in one's immediate surroundings because two
differently located environments can be perceptibly identical.
That one cannot distinguish between a replica of Times Square and Times
Square itself based on one's sensations alone indicates that locating
oneself relative to entities one is not currently perceiving requires
more than just one's sensations.
Nonetheless, to someone who makes this mistake it seems like they are in
Times Square. One will accurately localize oneself relative to
perceived landmarks, such as the Sony display, but not to unperceived
landmarks, such as the Mississippi River. So where one seems to be
allocentrically depends on more than just the mental locations of one's
sensations in virtue of which one perceives the locations of colors and
sounds in one's immediate surroundings.
So, in Dainton's scenario, that you seem to be in two places does not
entail that your bodily and audio-visual sensations are located in
different phenomenal fields. It would entail this only if it were in
virtue of one's sensations alone that one located oneself
allocentrically. Since this is not the case, something other than the
mental locations of your sensations makes it seem to you that you are in
two places at once. Locating oneself allocentrically may involve one's
beliefs, not just one's sensations. Perhaps you know that scientists
can put your body and eyes and ears in different places and, since you
also know that sensations of water and sensations of bird-calls never
coincide, you infer that they have done this to you.
Nonetheless, the S-Thesis is false. Sensations can be spatially
integrated without being co-conscious. Paul Bertelson et al. (2000)
show that visual stimuli presented in neglected fields of subjects with
unilateral visual neglect induce visual biasing of auditory perception.
Though the subjects are unaware of seeing visual stimuli presented in
their neglected fields, when they are asked to point to auditory
stimuli, they do so inaccurately. Moreover, the inaccuracy indicates a
strong influence from the visual stimuli.
This explanation relies on subjects' actually seeing the visual stimuli
and on the spatial integration of visual and auditory sensations. Since
the subjects see the stimuli without consciously seeing them, and since
seeing them affects their pointing at auditory stimuli, their visual and
auditory sensations must be co-present in the same phenomenal field
without being co-conscious. So co-presence in a phenomenal space is
insufficient for making two states co-conscious. The S-Thesis is false.
But the S-Thesis is not the only way to explain co-consciousness in
terms of phenomenal space. I will examine an inner-sense view of
consciousness that provides a modified version of the S-Thesis.
5. The Higher-Order S-Thesis
The S-Thesis is false because mental states can be co-present in a
phenomenal space without being co-conscious. However, a different
phenomenal space might be responsible for their being co-conscious. Such
a space is suggested by inner-sense, or higher-order-sense, theories of
consciousness.
According to higher-order-sense theories of consciousness (e.g.,
Armstrong 1980; Locke 1975/1700; Lycan 1996), mental states are
conscious in virtue of being detected by an internal sense. Sensing
one's mental states would make one conscious of them just as seeing
visual stimuli makes one conscious of the stimuli.
Such a view could explain co-consciousness in terms of the higher-order
sensory field needed to explain detection of first-order mental states.
Two visual stimuli are spatially unified in virtue of being located in
one perceptible field of view, the space in front of one's open,
functioning eyes. The stimuli are seen simultaneously because they are
both located in that field. And one sees the stimuli in virtue of one's
visual sensations' having mental locations corresponding to the
perceived locations of the stimuli. So one is conscious of the stimuli
as unified in virtue of the mental spatial relations between one's
sensations of them.
To explain the co-consciousness of mental states in a similar way, those
states must be unified in a mental space detectable by the higher-order
sense. And to enable perception of the first-order states as spatially
unified, the higher-order sensations of those states must have their own
higher-order space corresponding to the first-order space of those
target mental states.
Perhaps the first-order states one senses are mentally located in the
cross-modal space needed to explain cross-modal perception. Then
higher-order sensations will have mental locations corresponding to the
mental locations in this cross-modal space. And these higher-order
mental locations correspond to the cross-modal locations of first-order
states in virtue of homomorphisms, like those that explain first-order
sensing.
But Dainton objects to higher-order-sensing views of consciousness (p.
45). He claims that if one is conscious of one's mental states in
virtue of having a higher-order sensation of them, then the first-order
states will be absent from consciousness. Only the higher-order
sensation will be conscious. Since this is absurd, the
higher-order-sensation view of consciousness fails.
But this reductio rests on the assumption that a higher-order
sensation would have to be conscious itself in order to make a
first-order sensation conscious. If it were in virtue of the
higher-order sensation's being conscious that one were conscious of
one's first-order state, then one would be conscious of the first-order
state only when one were conscious of the higher-order sensation. So,
it would seem, one is only indirectly conscious of the first-order state
in virtue of being directly conscious of the higher-order one; only the
higher-order sensation is present to consciousness.
But one need not be conscious of one's mental states for them to make
one conscious of stimuli. So one's higher-order sensation need not be
conscious to make one conscious of a first-order state. The first-order
state itself is conscious in virtue of one's having a higher-order
sensation of it.
However, there are two problems for the higher-order-sensation view of
co-consciousness. First, the cross-modal sensory field needed to
explain cross-modal perception is different, in important ways, from the
sensory fields needed to explain modality-specific perception. The
mental visual field is posited to explain how we can discriminate
between colored surfaces based on their spatial properties. Two
identically colored stimuli can differ with respect to location, shape,
and size. Visual spatial qualities are the mental analogs of these
properties. But we posit them only because we can perceive colored
surfaces. Visual spatial qualities are the mental boundaries of mental
colors corresponding to the perceptible boundaries of perceptible
colors. Those perceptible boundaries, however, are fixed by the
physical limits of our visual system, e.g., the eyes can only detect the
colors of stimuli in front of them.
Since each sensory modality differs in its physical limits, the
perceptible boundaries of modality-specific properties and, thus, their
mental counterparts, are distinct. No mental visual locations are
identical to any mental auditory locations, nor are they identical to
any mental tactile locations.<3> So the cross-modal mental space must
be constructed by calibrating modality-specific spaces; it is not,
itself, a family of mental properties homomorphic to the spatial
properties of cross-modally perceptible stimuli.<4>
Without a cross-modal family of mental spatial properties, we have
nothing for the similarities and differences among higher-order spatial
sensory qualities to correspond to, so no reason to posit them. So we
have no reason to think that simultaneous mental states are co-conscious
in virtue of being perceived by an inner-sense.
Also, the higher-order-sense view of co-consciousness does not explain
how an intentional state is co-conscious with a sensation. Even if
there was a cross-modal mental space, there is no reason to locate
thoughts there. The space would be posited to explain cross-modal
perception of the spatial properties of stimuli. But thoughts need not
even refer to spatial entities. The thought that Sundays are melancholy
isn't about anything spatial. So it has no properties corresponding to
the spatial properties of its referents. But that thought can be
co-conscious with a visual sensation of a square at the center of one's
visual field. The higher-order-sensation view will have to explain what
properties bind these two states, such that the higher-order sensation
will correctly represent them as bound.<5>
Indeed, some mental states involve both sensory and intentional aspects.
When one visually perceives a Coke can, one has a visual sensation of
the color, shape, and size of the can. But one also perceives that
there is a can there. Emotions also involve both sensory and
intentional aspects. For instance, when chased by a bear, one has a
sensation of horror and one is afraid that the bear will tear one to
shreds. If the mechanism that makes states co-conscious is what makes
states conscious, higher-order sensing fails to make perceptual and
emotional states conscious because intentional states are not located in
any mental field.
In fact, since one can have a thought about something completely void of
perceptible, sensible, or emotional qualities--e.g., the thought that
the law aims at justice--one might wonder what qualities, akin to mental
colors, shapes, or sounds an internal sense would detect. A
higher-order-sensing view of co-consciousness must explain the
properties in virtue of which one is conscious of one's conscious
thoughts. Without such an explanation, the higher-order-sense
modification of the S-Thesis is untenable.
6. An Alternative to the Modified S-Thesis: Higher-Order Thoughts
Nonetheless, we need not conclude that co-consciousness is basic and
inexplicable.
Though one cannot see two entities at the same time unless they are both
present in the same space, i.e., the field of view, one can think about
two things regardless of their spatial relations. For instance, one can
think that Sundays are melancholy and Toledo is in Spain. Likewise, one
can think about both another thought and a perceptual experience at the
same time even if they are not both located in a mental space. Perhaps
if one were conscious of one's mental states in virtue of having
higher-order thoughts about them, those conscious mental states would be
co-conscious in virtue of being targets of the same higher-order
thought.
Rosenthal (1997) has argued for a higher-order-thought view of
consciousness according to which a mental state is conscious when one
has a suitable first-person thought that one is in the state in
question. Two mental states could be co-conscious in virtue of one's
ascribing them to oneself in the same higher-order thought. My thought
that Sundays are melancholy and my sensation of blue are co-conscious
because I have the higher-order thought that I both think that Sundays
are melancholy and have a sensation of blue.
But Dainton raises an objection that may seem to apply to this view.
According to Dainton (p. 49-50), co-consciousness cannot result from an
awareness that simply reveals, without adding anything to, the nature of
one's phenomenal states. Those states would already have phenomenal
qualities; they are experiences independent of any act of awareness
directed upon them.
This would apply to the higher-order-thought view if the states Dainton
refers to as phenomenal states were conscious in virtue of having
phenomenal properties. For instance, if phenomenal properties were
intrinsically conscious, as many people think, then a revealing act of
awareness would be superfluous. And Dainton's projectivism, as I've
suggested, may commit him to the intrinsicness of consciousness to
sensations in virtue of holding that one is indirectly aware of stimuli
in virtue of being directly aware of one's sensations. But phenomena
such as masked priming and blindsight suggest that one can have visual
sensations without being conscious of them. And arguably a mental state
is conscious only when one is conscious of it. So phenomenal properties
are either conferred by higher-order thoughts or they do not, by
themselves, make states conscious.
But the homomorphism view of perception avoids this problem. According
to this view, mental colors and sounds are simply the properties that
enable the perception of stimulus colors and sounds. These are the
properties that resemble and differ from one another in ways parallel to
the ways their perceptible counterparts resemble and differ from one
another. These states can have these properties without one's being
conscious of them.
When one has a higher-order thought to the effect that one is seeing
red, one is conscious of one's visual sensation in virtue of its being
mentally red; having that property makes the state a token of that
particular type. The higher-order thought does not alter the state's
intrinsic character to make it conscious; it merely represents to one
the kind of state one is in.
One might object that, except in cases of introspection, there simply is
no such higher-order thought about one's conscious experiences. We have
conscious experiences without thinking about them.
No doubt, this is how things seem to us. But it would only seem to one
that one had a higher-order thought about one's current experience if
that higher-order thought was itself conscious. But, just as an
unconscious thought about an apple makes one conscious of an apple, an
unconscious higher-order thought about a current mental state can make
one conscious of that state. In this case, it would not seem to one
that one has a higher-order thought about the mental state because one
would not be conscious of having that higher-order thought.
But another problem arises for the higher-order-thought model of
co-consciousness I've proposed. One can have any number of thoughts at
a given time. So there is no reason to think that one cannot have more
than one higher-order thought at a time. According to the view I have
proposed, this would result in simultaneous conscious states that don't
seem co-conscious. And, even if this never happened, an account of
diachronic co-consciousness must explain how higher-order thoughts at
different times can unify mental states in one stream of consciousness.
We can explain this in terms of the subject to which a higher-order
thought ascribes mental states (Rosenthal, forthcoming). The mental
analog of 'I' functions like the first-person pronoun; it is an
essential indexical. When one says, "I am hungry," one ascribes hunger
to oneself in virtue of 'I' referring automatically to the speaker of
the sentence. Likewise for first-person thoughts. When one thinks, 'I
am in pain,' one simply ascribes pain to the thinker of that very
thought. The mental analog of 'I' automatically refers to the thinker
of that thought. A higher-order thought, inasmuch as it is a
first-person thought, thus attributes a mental state S to the thinker of
the thought 'I am in S'.
One might object that this does not guarantee that mental states
ascribed to oneself in different higher-order thoughts will be unified
in one conscious experience. To guarantee that, one's distinct
higher-order thoughts would have to identify the same thinker; 'I' would
have to refer to the same thing in all of one's higher-order thoughts.
One way of avoiding this problem is to claim that there is a self to
which all of one's higher-order thoughts refer. Perhaps then any mental
state attributed to this self will be co-conscious in virtue of that
self's being conscious of them all.
Nevertheless, we need not commit to such a Cartesian self to explain the
unity of conscious experience. If one implicitly assumed that all of
one's first-person thoughts referred to the same thinker, one would have
a sense of the co-consciousness of all states ascribed to that thinker
in distinct higher-order thoughts. It would seem to one that all of the
states attributed in those higher-order thoughts were unified though
perhaps there is no single self that is conscious of all of them.
Such implicit assumptions are not unique to first-person ascriptions.
For instance, when one thinks, 'My keys were here a minute ago,' it
seems to one that the place 'here' refers to is the same place it
referred to a minute ago when one thought, 'Here is a good place to
leave my keys.' Of course, the mental analog of 'here' need not
actually refer to the same place at two different times for it to seem
to do so. And this is usually the case when one cannot find one's keys.
It seems to one that 'here' referred to the same place at both times
simply because one implicitly assumes that it did.
The unity of conscious experience could be apparent in the same way that
the identity of the referent of 'here' is only apparent in the above
case. I f one assumes that the two places referred to at different
times by 'here' are identical it will seem to be the same place.
Likewise, if one implicitly assumed that 'I' refers to the same thinker
at different times, or at the same time, the mental states one
attributes to the thinker of each thought would seem to be attributed to
the same thinker. This sense of unity explains the apparent
co-consciousness of mental states ascribed to one in distinct
higher-order thoughts.<6>
Notes
<1>. Dainton seems to recognize this distinction. He claims that there
are phenomenal experiences one fails to notice. However, his account is
not clear on the distinction between being directly aware of and
noticing an experience.
<2>. This view was pioneered by Wilfrid Sellars (1956) and has been more
recently argued for by David Rosenthal (1999) and Douglas Meehan (2002).
<3>. One might object that though an auditory sensation can represent
locations that visual sensations cannot, e.g., a location behind one's
head, both sensations can represent the same locations as well, e.g.,
when one hears a sound coming from a bird one sees. Since a visual
sensation and an auditory sensation can represent the same location,
there must be mental visual locations that are identical to mental
auditory locations.
But even though auditory sensations and visual sensations can represent
the same perceptible locations, they do so in different ways. One can
locate a stimulus only relative to a frame of reference. Consider the
case of an object between two people facing each other. If one person
sees the object as off to the left, the other person will see it as off
to the right. This is because the object is off to the left in the
first person's perceptible visual field--the space visible at a
particular moment--whereas it is off to the right in the other's. And
since one sees objects as located in virtue of having sensations with
mental locations that correspond to the objects' locations, and the
objects' locations are relative to the perceptible visual field, the
mental locations must be relative to the mental visual field that
corresponds to that perceptible visual field.
The same applies to auditory perception. A sound is heard as being at a
particular location within a perceiver's auditory field. And one hears
it there in virtue of having an auditory sensation with a particular
mental location that is, itself, relative to the mental auditory field.
Mental visual locations correspond to visible locations, which are
relative to one's perceptible visual field. And mental auditory
locations correspond to audible locations, which are relative to one's
perceptible auditory field. And one's perceptible auditory field and
one's perceptible visual field have distinct boundaries. So one's
mental visual field and one's mental auditory field will be distinct.
Therefore, mental visual locations are distinct from mental auditory
locations.
Thanks to Barry Dainton for pressing me on this point.
<4>. George Berkeley (1975/1732) thought that modality-specific mental
spatial qualities were calibrated by inferences. According to this
view, one perceives that a visual stimulus is located at the same place
as an auditory stimulus in virtue of having learned that visual
sensations of that type and auditory sensations of that type represent
the same locations. This view, however, may pose a problem for simple
organisms that exhibit cross-modal behavior but lack thought.
<5>. The higher-order-sense theorist could invoke an error theory here.
It could be that the first-order states are not really bound. But if
the higher-order sensation is modeled on first-order sensations, this is
not an option. Though we misperceive things, e.g., a car as being red
when it's actually orange, we only perceive them as having perceptible
properties that other objects of their kind do have. This is because
the sensory counterparts of sensations are posited to explain our
relations to the properties we can, in fact, sense. But for a
higher-order sensation to represent first-order states as bound would be
for it to represent them as having a relation that diverse mental states
simply do not bear to one another. This is, in effect, another reason
for adopting a higher-order thought model of co-consciousness, as I do
below.
<6>. Thanks to Barry Dainton and David Rosenthal for comments on an
earlier draft of this paper, and to the editors of Psyche, especially
Tim Bayne.
References
Armstrong, D. M. (1980). What is Consciousness? In The Nature of Mind
and Other Essays. Ithaca: Cornell.
Berkeley, George (1975/1732). An Essay Towards a New Theory of Vision.
In Ayers, M. R. (Ed.), George Berkeley: Philosophical Works (pp.
3-59). Rutland, VT: Charles E. Tuttle Co., Inc.
Bertelson, P., Francesco, P., Ladavas, E., Vroomen, J., and de Gelder,
B. (2000). Ventriloquism in Patients With Unilateral Visual
Neglect. Neuropsychologia, 38, 1634-1642.
Dainton, Barry (2000). Stream of Consciousness: Unity and Continuity
in Conscious Experience. London, New York: Routledge.
Fernandez-Duque, Diego and Thornton, Ian M. (2000). Change Detection
without Awareness: Do Explicit Reports Underestimate the
Representation of Change in the Visual System? Visual Cognition,
7, 323-344.
Locke, John (1975/1700). An Essay Concerning Human Understanding.
Peter H. Nidditch (Ed.). Oxford: OUP.
Lycan, William G. (1996). Consciousness and Experience. Cambridge:
MIT.
Meehan, Douglas B. (2002). Qualitative Character and Sensory
Representation. Consciousness and Cognition, 11, 630-641.
Milner, A. David and Goodale, Melvyn A. (1995). The Visual Brain in
Action. New York: OUP.
Rosenthal, David M. (1997). A Theory of Consciousness. In Ned Block,
Owen Flanagan, and Guven Guzeldere (Eds.), The Nature of
Consciousness (pp. 729-753). Cambridge: MIT Press.
Rosenthal, David M. (1999). The Colors and Shapes of Visual Experience.
In Denis Fisette (Ed.), Consciousness and Intentionality: Models
and Modalities of Attribution. Dordrecht: Kluwer Academic
Publishers.
Rosenthal, David M. (forthcoming). Unity of Consciousness and the Self.
Proceedings of the Aristotelian Society, 103, 3, (April 2003).
Sellars, Wilfrid (1963). Empiricism and the Philosophy of Mind. In
Science, Perception and Reality (pp. 127-196). Atascadero:
Ridgeview.
Simons, Daniel J., Chabris, Christopher F., and Schnur, Tatiana (2002).
Evidence for Preserved Representations in Change Blindness.
Consciousness and Cognition, 11, 78-97.
Weiskrantz, Lawrence (1986). Blindsight: A Case Study and its
Implications. Oxford: OUP.
