SUSANNA SIEGEL
INDISCRIMINABILITY
AND THE PHENOMENAL
In Philosophical Studies volume 120, 90-112
In “The Limits
of Self-Awareness” (this volume) , M.G.F. Martin characterizes disjunctivism about perceptual experience as follows:
[S]tatements about how things appear to a perceiver [are] equivalent to statements of a disjunction that either one is perceiving such and such or one is suffering an illusion (or hallucination) . . . such statements are not to be viewed as introducing a report of a distinctive mental event or state common to these various disjoint situations. (p. 1)
The disjunctive
theory stands opposed to the common-kind
theory, according to which there are pairs of genuine
perceptions
and mere seemings to
perceive that have some fundamental kind of
mental state in
common.
The issue
between disjunctivism and the common-kind theory concerns the status of being indiscriminable from a veridical perception. Suppose, for example, that I see a green
cube, and my experience is veridical – no illusion is involved. An example of
the sort of indiscriminability property whose status is at issue is the
property of being indiscriminable from my experience of seeing the green cube.1
As Martin sees
it, the common-kind theorist and
the disjunctivist
have three bones of contention. First, the common-kind theorist affirms, while
the disjunctivist denies, that for any event indiscriminable from a (specific)
veridical perception,
there is a robust
property in virtue of which that event is so
indiscriminable.
According to the common-kind theory, once we
fix on an
indiscriminability property – for instance, the property of
being
indiscriminable from my seeing the green cube – any event
that has this property
has it in virtue of having some
robust property
or other, so the
indiscriminability property isn’t fundamental.
This brings us
to the second bone of contention. As Martin
characterizes it,
the common-kind theory goes further than merely
denying that
indiscrminability from veridical perception is a
brute fact: in
addition, any two events with the same (specific)
indiscriminability
property have it in virtue of having the same
(specific) robust
property. This entails that there is a common
kind:
a kind that
hallucinations and perceptions share.
In contrast,
Martin denies that there is any fundamental kind
to which
hallucinations and perceptions both belong. According
to Martin’s version
of disjunctivism, some experiences – but only
some – are indiscriminable
from a (specific) veridical perception,
even when there is
no robust property they have in virtue of which
they are so
indiscriminable. This is Martin’s line on causally matching
hallucinations: hallucinations with the same proximate
causal antecedents
as veridical perceptions. Such hallucinations,
Martin holds, belong
to the fundamental kind: being indiscriminable
from a veridical
perception.
Other
experiences, Martin thinks, belong to this kind, but it
is not their
fundamental kind. Veridical perceptions are, naturally,
indiscriminable from
veridical perceptions; but they are so
indiscriminable,
Martin thinks, in virtue of having robust properties.
My veridical
perception of the green cube, for instance, Martin
thinks, is
indiscriminable from a veridical perception of the green
cube in virtue of
the perceptual relation that holds between the
perceiver (me), on
the one hand, and the cube and the properties of
it that appear to
me, on the other. Veridical perceptions belong to the
fundamental kind: being
veridical perceptions; whereas causally matching
hallucinations
belong to the fundamental kind: being
indiscriminable from a veridical perception.
So Martin
agrees with the common-kind theorist that there is
a common element
between causally-matching hallucinations and
the veridical
perceptions they causally match;2 but disagrees about
the depth and
significance of the commonality. For Martin, it goes
no deeper than the
indiscriminability property, and it does not
constitute the
fundamental kind to which both experiences belong.3
So far, I’ve
mentioned two of the three bones of contention
between the Martin’s
disjunctivism and the common-kind theory.
The first was that
the disjunctivist (of Martin’s stripe)4 allows, while
the common-kind
theorist denies, that being indiscriminable from a
(specific) veridical
perception can be a brute fact. The second is that
the common-kind
theorist allows, while the disjunctivist denies, that
for any paired
hallucination and perception that (in some intuitive
sense) seem the same
to the subject, there is a single mental state
had by both
experiences that constitutes their fundamental kind.
The third
disagreement concerns the concept of perceptual experience. Martin takes
indiscriminability from veridical perception to
be definitive of
perceptual experience: “being indiscriminable from
veridical
perception,” he writes, “is the most inclusive conception
we have of what
sensory experience is” (p. 22). Now, Martin takes
it that the
common-kind theorist will agree that for an event to so
much as count as a
perceptual experience, it has to be indiscriminable
from a veridical
perception. So they will agree, Martin thinks,
that it’s a
conceptual truth that sensory experiences are indiscriminable
from veridical
perceptions. The disagreement is supposed to
concern whether
anything else is conceptually true of sensory experience.
As Martin construes
his opponent, she says that something
else is: it is part
of the concept of perceptual experiences that they
instantiate mental
properties that realize, or underlie, indiscriminability
from veridical
perception. As to the metaphysical nature of
the common kind
property, there are the options made familiar by
the history of the
philosophy perception so far: candidates include
sense-data, being an
adverbial modification, having propositional
content of some
sort, and combinations thereof.
In what
follows, I will challenge both the assumption that
phenomenality and
indiscriminability from veridical perception are
as closely linked as
Martin thinks they are, and Martin’s defense of
disjunctivism.
Sections 1 and 2 criticize Martin’s claim that every
experience is
indiscriminable from a veridical perception: section 1
focuses on
veridicality; section 2 on indiscriminability. In section 3,
I turn to Martin’s
argument against the common-kind theory, which
is supposed to
motivate taking the disjunctivist conception of experience
as the default. In
section 4 I address a residual question related
to the apparent
intransitivity of looking the same with respect to hue.
I conclude the
discussion in section 5.5
1. VERIDICALITY
Both Martin’s
disjunctivism and the common-kind theory, as Martin understands it, are
supposed to endorse a strong link between the
notions of
phenomenality and indiscriminability from a veridical
perception. More
exactly, both positions are supposed to agree that
any event that
counts as a perceptual experience is indiscriminable
from a veridical
perception.
One might
question the claim that the phenomenal is quite
as closely linked to
indiscriminability from veridical
perception
as Martin suggests.
Consider, for example, a virtual-reality scene
made to look just
like what’s depicted in Escher’s drawing of
the impossible
staircase. There seems to be no possible veridical
perception from
which an experience of seeing such a scene is
indiscriminable. Yet
it seems to be a perfectly good specimen of
a perceptual
experience.6
A different
sort of example casts additional doubt on the claim
that any event that
counts as a perceptual experience is indiscriminable
from a veridical
perception. Suppose the following sort of
error theory of
color is correct: visual experience represents color
properties, but
nothing external is, or could in principle be colored.7
On this view, when
earlier I described my experience by saying that
I saw a green cube,
strictly speaking this was a misdescription: what
I saw was a cube,
and though it appeared green, there was no color
property had by the
cube that my experience even so much as falsely
represented.
If such an
error theory were correct, then there would not be any
veridical perception from which my experience of seeing
the green appearing cube is indiscriminable. Many philosophers would find
such a theory
implausible.8 But presumably its implausibility has
nothing to do with
miscategorizing the event of seeing the greenappearing
cube as an
experience. Like the events of seeing Escher
drawing, this event
seems a fine specimen of an experience; and,
crucially, its
status as fine specimen seems independent of whether
the error theory
described is correct.
I’ve been
questioning the claim that any event that counts
as a perceptual
experience is indiscriminable from a veridical
perception. As we’ve
seen, Martin’s disjunctivism includes an
even stronger claim
connecting phenomenality to indiscriminability, a claim concerning the very
concept of perceptual experiences:
In
fixing on our concept of perceptual experience, we seem to have no more
resources
than we need to pick out something indiscriminable from a veridical
perception.
(p. 11)9
Martin considers it
a conceptual truth about perceptual experiences
that they are
indiscriminable from veridical perception. If this were
right, then error
theorists of color would be conceptually
confused
about what
experience is. But whatever errors such theorists may
be making, they do
not seem to include conceptual confusion about
what to count as an
experience.
2. INDISCRIMINABILITY
So far, I’ve
criticized the idea that phenomenality is linked
to
indiscriminability from veridical
perception. I now want to
challenge the idea
that it is linked to any notion of indiscriminability.
As Martin thinks of
it, indiscriminability is a notion defined
in terms of
judgment. “To discriminate two things,” Martin writes,
“is judge them
non-identical” (p. 26). This suggests that when A
and B are indiscriminable
for a subject, the subject cannot tell them
apart in judgment. Saying no more than this leaves much unsettled
about what
indiscriminability is, and shortly we will consider two
ways to precisify
the notion. For the moment, what’s notable is that
Martin’s notions of discriminability and indiscriminability are cognitive notions.
As we saw
earlier, Martin takes the common-kind theorist to be
committed to the
claim that any event that counts as a perceptual
experience is
indiscriminable from a veridical perception.10 Let S
be a subject, and
let I* be a robust property of the sort that, by
the lights of the
common-kind theory, is supposed to bestow on
any event that has
it the property of being indiscriminable from a
veridical
perception. So fix on an indiscriminability property, such
as the property of
my seeing the green cube, and by the lights of
the common-kind
theory (as Martin construes it) there is a property
I* that
characterizes what it is like to have an experience with that
indiscriminability
property. In the hands of Martin’s common-kind
theorist, then, I*
is supposed to play two roles: first, it is supposed to
make any event that
has it indiscriminable from a veridical perception;
second, it is
supposed to characterize what the experience is like
for the subject.
Martin, then,
takes his opponent to accept the following:
Sufficiency claim:
If S’s experience has I*, then S’s
experience is
indiscriminable from a veridical perception.
What the Sufficiency
claim comes to depends on how the
notion of
indiscriminability is understood. I will now consider two
notions of
indiscriminability, and argue that on each way, someone
sympathetic to the
main thrust of the common-kind theory could
reasonably deny the
Sufficiency claim – though for different reasons
each time. Both
doubts come into focus by considering creatures
who have perceptual
experiences, yet lack the cognitive resources
to make judgments
about them. In addition, as we will see, the
possibility of this
sort of creature also threatens Martin’s positive
view that two
events’ being indiscriminable from the same veridical
perception suffices
for their being phenomenally the same.
I’ll call the
first notion of indiscriminability the positive
notion:11
Positive: X is
indiscriminable from Y by a subject S at
time t iff S is
disposed to judge on basis b that X = Y.
This notion of
indiscriminability has a parameter for the basis of
S’s disposition to
judge. The motivation for having such a parameter
is as follows.
Suppose that X and Y look totally different to S, but S
is disposed to judge
on the basis of consulting an unreliable oracle
that X = Y. Without
the parameter for the basis, by the positive
notion, X and Y
would count as indiscriminable for S. This seems
like the wrong
result. In any case, it seems clear that the basis Martin
has in mind is
“introspection and reflection”,12 so let
us put that in
for b.
Now, suppose
there were a creature who had I*, but who was
not equipped to form
any judgments at all, ergo was not disposed
to judge that she
was veridically perceiving. On the positive
account of
indiscriminability, the sufficiency claim would predict
that if such a
creature had a perceptual experience, then S would
be disposed to judge
on the basis of introspection and reflection
that S’s experience
is a veridical perception. For the sort of creature
imagined, this
prediction would be false.
Martin’s
construal of the common-kind theory, then, would be
too restrictive,
given the positive notion of indiscriminability. The
existence of
creatures who have perceptual experiences without
capacities for
judgment is compatible with there being pairs of
veridical perception
and causally-matching hallucination that share
a common mental
kind. So rejecting the Sufficiency claim does
not seem tantamount
to giving up on the central claim of the
common-kind theory
itself.
I’ve complained
that Martin shouldn’t attribute the Sufficiency
claim to the
common-kind theorist. In my complaint, I’ve assumed
that if some sort of
first-person access to what experiences are like
is required, this
access can take a form other than judgment. It is
a difficult question
how to understand the nature of such access,
once it’s stipulated
not to involve judgment. Perhaps there is a more
primitive form of
introspective access of some sort.
The same sort
of creature presents a different reason for a
common-kind theorist
to deny the Sufficiency claim, when that
claim is taken to
involve indiscriminability understood differently.
I’ll call the second
notion of indiscriminability, which is proposed
by
TimothyWilliamson, the double-negative
notion of indiscriminability:
Double-negative: X
is indiscriminable from Y by a
subject S at time t
iff S is not able at t to activate
knowledge that X =/=
Y.13
Using this notion of
indiscriminability, the Sufficiency claim
comes to this:
Sufficiency
claim-TW: If S’s experience has I*, then S
cannot activate
knowledge that having-I* is distinct from
having a veridical
perception.
With respect to the
sort of creature lacking cognitive equipment
of the sort needed
to form judgments, Sufficiency claim-TW
is trivial. For such
a creature, there will be no pair of perceptual
experiences such
that the creature can activate knowledge that they
are distinct from
one another. So all perceptual experiences of the
creature will count
as indiscriminable from one another.14 This
gives
the common-kind
theorist reason to reject the Sufficiency claim, on
the grounds that it
doesn’t capture anything important in their view.
The case of
creatures with perceptual experiences who lack
capacities for
judgment also suggests an objection to Martin’s
positive view that
being indiscriminable from the same
veridical
perception is
sufficient for two events’ being phenomenally the
same. Assuming
either notion of indiscriminability, all of the experiences
of such creatures
will count as the same, if indiscriminability
suffices for
sameness of experience.
Martin
considers and responds to this very worry. His response is
that the relevant
notion of discriminability is impersonal:
when
we turn to the experiences of sentient but unselfconscious creatures, to the
extent
that we do have a positive grip on the kinds of experience that they can
have,
and which can differ one from another, we also have a grip on how such
experience
would be discriminable through reflection or not . . . a dog might fail
to
discriminate one experience from another, making no judgment about them as
identical
or distinct at all, [but] that is not to say that we cannot judge, in ascribing
them
such experience, that there is an event which would or would not be judgably
different
from another experience. (p. 28)
If the claim here is
that two of the dog’s experiences are discriminable
by someone other than
the dog, that seems correct. But it does
not seem correct to
say that they are discriminable by
reflection, if
reflection is
supposed to be on the part of the subject whose states
are in question.
After all, by hypothesis it is not the dog doing the
reflecting, and it
is not clear what it would be for us to reflect on the
dog’s experiences,
without doing some empirical investigating of a
sort that the dog
would be incapable of carrying out. If the relevant
sort of
(in)discriminability is (in)discriminability for a subject on the
basis of that
subject’s reflection and introspection, then the appeal
to the impersonal
notion won’t work in this case.15
On another
reading, Martin’s response to the worry is that there
is a sense in which
some of the dog’s experiences are discriminable
from one another, to
the dog. But the notion of discriminability that
would make this
claim true could not be a cognitive notion. And as
we’ve seen, it is a
cognitive notion that is at work in Martin’s central
claims.
3. MARTIN’S OBJECTION TO THE
COMMON-KIND THEORY
So far, I’ve
been criticizing Martin’s views of the relation between
indiscriminability
and the phenomenal. These views form the background
to his argument for
disjunctivism. I now want to turn to that
argument itself.
Martin’s
defense of disjunctivism aims to show that the disjunctivist conception of
perceptual experience should be the default
conception. It
should be the default conception, Martin thinks,
because otherwise
one’s epistemological assumptions about the
mind will be very
weighty.16 The fact (as Martin sees it) that
the common-kind
theory is committed to such weighty epistemic
assumptions is the
main objection he raises against that theory.
This fact, in turn,
is the main reason given for why the disjunctivist
conception should be
the default. I will now examine this objection.
Martin gives
the objection in a passage that contrasts the
supposed “modesty”
of disjunctivism with the supposed “immodesty”
of the common-kind
theory. Disjunctivism is supposed to
be modest, because
it takes indiscriminability from a veridical
perception as
necessary and sufficient for an event to count as an
experience. The
common-kind theory, in contrast, is supposed to be
immodest, because it
takes as necessary and sufficient for an event
to be an experience
that it instantiate a robust property that realizes
the
indiscriminability from a veridical perception. In the first part of
the objection,
Martin considers how the common-kind view would
classify a situation
in which a subject was unable to discriminate
her situation from
one in which she was seeing a street scene, and
yet had no robust
property of the sort that the common-kind theory
takes to
characterize perceptual experience:
For
the immodest view in question this could not be a case of visual experience as
of
a street scene, while by modest lights that would be exactly what it is. . . . Now
surely
this result would surely be unfortunate for any immodest view, given our
initial
assumptions. For we supposed that reflection on experience offers support
to
a naďve realist construal of sensory experience. When one reflects on one’s
experience
it seems to one as if one is thereby presented with some experience-independent
elements of the scene before one as constituents of one’s experience
and
not merely as represented to one as in imagination. (p. 10)
There seem to be two
steps here that Martin thinks the common
kind theorist is
forced to take. The first step is that an event is
indiscriminable from
a veridical perception, just in case it seems
to the subject as if
she is “presented with experience-independent
elements of the
scene before her as constituents of her experience”.
This is supposed to
be an upshot of the initial assumption that reflection
on experience
supports a naďve realist construal of experience.
The second step is
that it seems to one as if one is presented with
such elements, just
in case one is having a perceptual experience
(perhaps a
hallucinatory one). And this seems plausible. Putting
these steps
together, being indiscriminable from veridical perception
suffices for being
an experience. As we’ve seen, Martin also thinks
there are grounds
(acceptable to disjunctivist and non-disjunctivist
alike) for the
converse – that all experiences are indiscriminable
from a veridical
perception. I raised some worries about that earlier,
but let us set them
aside here. Combining these commitments gives
us
(1)
I iff E,
where ‘I’ is for
indiscriminability from a veridical perception, and
‘E’ is for being a
perceptual experience, and (1) abbreviates “all
and only the events
indiscriminable from veridical perceptions are
experiences.”
This brings us
to the second part of the objection:
A
proponent of the immodest view can only hope to offer necessary as well as
sufficient
conditions for having an experience – and hence to explain the having
of
an experience in terms of its favored conditions – if it can ensure that
themodest
approach
and its favored form of immodesty coincide in the extension they give
the
concept of experience. (p. 11)
Here the relevant
part of the sentence is the first part, with its
assumption that the
common-kind theorist aims to give necessary as
well as sufficient
conditions for having an experience. The common kind
theory, recall,
takes it to be a conceptual truth about perceptual
experience that such
experiences have a certain robust property (the
exact metaphysical
nature of the property is left open – it could the
property of having
sense-data, or of having propositional contents of
sort, some
combination, etc.)Where this property is R (for ‘robust’),
the assumption comes
to this:
(2) E iff R,
or more exactly, all
and only the experiences have robust property
R.
In the passage
quoted, Martin suggests that the common-kind
theorist can accept
(2) only if she accepts (3):
(3) I iff R,
that is, all and
only the events indiscriminable from veridical perceptions
have the property R.
In embracing (1)–(3), the common-kind
theorist ensures
that it classifies as experiences all the same events
as the
disjunctivism-a-la-Martin does: as Martin puts it, she ensures
that the modest and
immodest approaches “coincide in the extension
they give the
concept of experience.”
The final part
of the objection connects (1)–(3) to a substantive
epistemic principle:
A
proponent of the immodest view can only hope to offer necessary as well as
sufficient
conditions for having an experience – and hence to explain the having
of
an experience in terms of its favored conditions – if it can ensure that the
modest
approach and its favored form of immodesty coincide in the extension
they
give the concept of experience. In turn, this coincidence of extension can by
guaranteed
only if the proponent of the immodest account embraces a substantive
epistemic
principle . . . one must assume that a subject
couldn’t but be in a position
to
discriminate a situation which lacked E1 . . . EN from one which possessed
them.
. . . A responsible subject who
wishes to determine how things are with him
or
herself through reflection must be, on this view, infallible in the answers
they
come
up with. They must not only correctly identify phenomenal properties of a
specific
sort when they are present, but also they cannot be misled into judging
them
present when they are not. (p. 11)
It is the
common-kind theorist’s commitment to (1)–(3) that
supposedly forces
her to accept the substantive epistemic principle,
which from now on I
will call Hefty.
Hefty:
A responsible subject who wishes determine how things are with him or
herself
through reflection must not only correctly identify phenomenal properties of a
specific sort when they are present, but also they cannot be misled into
judging them present when they are not.
The success of
Martin’s objection to the common-kind theory
(and, given the
dialectical context, his success in motivating
disjunctivism) hangs
on two things: first, whether a commitment to
Hefty really does
follow from (1)–(3); second, on how plausible or
implausible Hefty
is. I will consider the second issue first, and then
return to how
exactly the commitment to Hefty on the part of the
common-kind theorist
is supposed to arise.
In assessing
the plausibility of Hefty, a crucial interpretive question is what sorts of
properties E1 . . . EN, the “phenomenal properties of a specific
sort”, are meant to be. Either they are robust
properties that
characterize a specific type of veridical experience,
such as my seeing
the green cube; or else they are the robust general
properties, shared
by all such specific ones, such as the property of
having propositional
content, or being sense-data. Call these strong
and weak robust
properties, respectively.
Up until the
statement of Hefty, it seems clear that it is weak
robust properties
are at issue: notably, premises (2) and (3) concern
the property that
(by the lights of the common-kind theory) is
supposed to be necessary for having an experience at all. And it
is surely not
necessary for an event to count as a perceptual experience
that it have the
strong robust property (if such there be) that
characterizes my veridically
seeing the green cube. As pleasant as it
is for me to see the
green cube, it is thankfully not the only type of
experience one can
have. It seems undeniable that the properties at
issue are the weak
ones, rather than the strong ones. Martin’s case
against Hefty goes
like this:
[T]he
doctrine of infallibilism about the mental is particularly problematic in
relation
to
sensory states once we are forced to admit that appearances. systematically
appear
to us other than they are. For if we can be misled with respect to some
properties
of sensory experiences, there is a question as to what can motivate
the
claim that we are infallible in other judgments about them . . . part of the
motivation
of disjunctivism is precisely the thought that introspection of our sense
experience
supports Naďve Realism, and hence forces us to see both sense-datum
and
intentional theories as forms of error theory. (p. 12)
The idea here seems
to be that Hefty does not sit well with the
common-kind theory,
because the common-kind theory already
accepts some sort
of fallibility to introspection in rejecting Naďve
Realism. In
rejecting Naďve Realism, the common-kind theorist
denies that
veridical experiences consist in part in the objects
perceived. Combining
this with the claim that Naďve Realism introspectively seems to be true – a claim that, as we’ve seen, Martin
thinks the
common-kind theorist should accept (see discussion of
(1) above) – the
result is that introspection is fallible about the metaphysical nature of
perceptual experience. If the common-kind theory
is already committed
to introspection delivering fallible results
about the
metaphysical nature of experience, the thought seems to
be, then the idea
that it would be infallible about some other aspect
of experience seems
to be undermined.17 This is Martin’s reason for
thinking that Hefty
is implausible.
In order to
assess this reason to reject Hefty, one more interpretive
issue concerning
Hefty needs to be settled. Hefty, recall, was
the epistemic
principle that “a responsible subject who wishes to
determine how things
are for themselves sensorily must . .
. not only
correctly identify
phenomenal properties of a specific sort when they
are present, but
also they cannot be misled into judging them present
when they are not”
(p. 10). This principle can be interpreted in two
ways: extensionally
or intensionally.
Taken
extensionally, Hefty entails that a subject who satisfies it
can discriminate on
the basis of introspection between events that
have the weak robust
property and events that lack it. But it is not
required, on the
extensional interpretation of Hefty, that the subject
know what sort of
properties she is discriminating between. It is
enough simply that
she make the discriminations, on the basis of
introspection,
infallibly.
Taken
intensionally, Hefty is much more demanding. A subject
who satisfies
intensional Hefty will be disposed to judge correctly
that she has weak
robust properties of sort R, where R specifies
the metaphysical
nature of the robust properties: e.g., sense-data,
or representational
properties. Intensional Hefty says, in effect, that
introspection can
reveal the basic metaphysical nature of the properties
that characterize
what it is like to have an experience. If intensional
Hefty is true, then introspection
can reveal that perceptual
experience is the
having of representational properties, or sense-data,
or some combination
– whichever (if any) of these properties
perceptual
experiences turn out to have.
In an earlier
paper, Martin gave convincing grounds for doubting
intensional Hefty.
If intensional Hefty were correct, Martin has
noted, it would
difficult to explain how there can so much as be
philosophical
disagreement about the nature of experience. The
disagreements
philosophers have about metaphysics of experience
don’t seem plausibly
to result from variation in their inner lives.18
It does not seem to
be introspection alone that decides between,
say, the sense-datum
theory and the theory that experience consists
entirely in
representational properties. Other sorts reasoning are
needed. So
introspection doesn’t seem to provide substantive knowledge
of weak robust
properties.
The reason
Martin gives in the text for rejecting Hefty also counts
against intensional
Hefty. If the common-kind theorist embraced
intensional Hefty,
she would be saying, in effect, “introspection
wrongly tells me that experiences partly consist in
external objects
that I perceive; yet
it also tells me infallibly that my experience
consists in weak
robust property R.” There is a tension here, given
that R is by
definition the sort of property that can be had by
hallucinations, and
so cannot be object-involving.
But neither of
the arguments Martin gives against Hefty work
so decisively
against the extensional version of Hefty. Introspection
may not infallibly
reveal the nature of weak robust properties (if
there are such), but
that’s compatible with it providing infallible
grounds for
discriminating between events that are experiences and
events that aren’t.
And this is also compatible with introspection
being wrong – even
systematically so – about the nature of weak
robust properties.
So for all Martin says, extensional Hefty has not
been shown to be
implausible, or even in tension with the commonkind
theory. Being
committed to extensional Hefty does not seem
to be a reason to
reject the common-kind theory, or to regard the
disjunctivist
conception of experience as the default.
But what
exactly is the common-kind theorist committed to? As
wes saw earlier, one
of the claims Martin thinks the common-kind
theorist is
committed to is that an event is an experience just in case
it has a robust
property of type R:
(2) E iff R.
From here, it is a
simple step to Hefty, with the following assumption:
subjects can
infallibly discriminate between events that are
experiences and
events that aren’t. For if a subject “couldn’t but be
in a position to
discriminate” experiences from non-experiences,
then she also
“couldn’t but be in a position to discriminate an situation
which lacked E1 . . . EN
from a situation which possessed
them”, where E1 . . . EN
is a weak robust property. This is one way
in which the
common-kind theorist’s commitment to Hefty could
arise.
It seems
plausible to suppose that any common-kind theorist is
indeed committed to
(2), and from (2) plus the assumption about
infallible
discrimination of experiences from non-experiences, some
version of Hefty
follows. The version that follows, however, is the
extensional version.
That version isn’t obviously implausible, and
in any case Martin’s
arguments against Hefty, as we’ve seen, work
only against the
intensional version.
Another
argument that the common-kind theorist is committed to
some version of
Hefty proceeds from two premises: I iff R (claim
(3) above), and the
transitivity of indiscriminability. The main idea
is the same as
above: if I and R are co-extensive, then infallible
discrimination of
events with I from events without I just is infallible
discrimination of
events with R from events without R. And the
latter just is
Hefty.
Why think that
subjects can infallibly discriminate events with
I from events
without I? Let x be an event that is indiscriminable
from a veridical
perception. Let e be an event that is discriminable
from a veridical
perception. Now suppose that subjects can’t
infallibly
discriminate events with I from events without I. In particular,
suppose that e and x
are themselves indiscriminable. Then e is
indiscriminable from
x, and x is (by definition) indiscriminable from
a veridical
perception. If indiscriminability is transitive, then e will
be indiscriminable
from a veridical perception. But by hypothesis, e
was discriminable from a veridical perception. So
we get a contradiction.
That is another
argument that the common-kind theorist is
committed to Hefty.19
There seem to
be two arguments suggested by Martin’s text that
the common-kind
theorist ends up committed to Hefty. Neither argument,
however, shows that
the common-kind theorist is committed
the intensional
version of Hefty. And it’s the intensional version that
Martin argues
against.
4. A RESIDUAL PROBLEM
As Martin
characterizes the common-kind theory, it ought to be
committed to the
view that for each indiscriminability property (e.g.,
the property of
being indiscriminable from my veridically seeing
the green cube),
there is a single robust mental property such that
any event with the
indiscriminability property has the robust mental
property. Some doubt
is cast on this claim by considerations related
to intransitivity of
some indiscriminability properties. Notably, if
being
indiscriminable with respect to hue is intransitive, then there
is reason to reject
the common-kind theory as Martin characterizes
it. In this section,
I discuss whether this counts in favor of
disjunctivism.
Consider three
red swatches that differ slightly in hue: they
are, say, red38,
red39 and red40. Many philosophers think that
indiscriminability
respect to color is an intransitive relation. Such
intransitivity is
supposed to be illustrated by cases of the following
sort: a subject
comparing the red38 swatch with the red39 swatch
(at a time) cannot
discriminate between their hues. The same subject
comparing the red39
swatch with the red40 swatch (with the red38
swatch out of view)
cannot discriminate those hues either. But the
same subject
comparing the red38 swatch with the red40 swatch
(with the red39
swatch out of view) can discriminate the hues of
those.
The case just
described involves only three swatches, but that’s
not essential to
what it is supposed to illustrate. What’s essential is
that there be
discriminable hues on the outer edges of a range of
indiscriminable
ones. How many indiscriminable ones occupy the
range is not
important. If there are cases of the sort just described
– and many
philosophers think there are, though some disagree20 –
they illustrate the
intransitivity of looking the same with respect to
hue.
For the sake of
argument, I am going to ignore the controversy,
and assume that
indiscriminability with respect to hue really is
intransitive. I will
also assume, for simplicity, that the intransitivity
shows up with just
three swatches. This is empirically dubious,
but makes it easier
to get to the main point.21
If it turns out that
indiscriminability
with respect to hue isn’t intransitive, then there
is one less reason
to reject the common-kind theory as Martin
characterizes it.
It will be
useful to spell out a (putative) case of intransitive
indiscriminability
in a bit more detail. Let I38 be the property of
being
indiscriminable from a veridical perception of red38, and let
E38 and E40 be,
respectively, an event of veridically seeing a swatch
of red38 and an
event of veridically seeing a swatch of red40. And
finally, make the
following assumptions about E38 and E40: E38
is indiscriminable
from a veridical perception of red39 and red40;
and E40 is
indiscriminable from a veridical perception of red41 and
red42. So E38 has
three indiscriminability properties: I38, I39 and
I40. Given these
assumptions, E38 and E40 share the indiscriminability
property I40. Both
are indiscriminable from a veridical
perception of red40.
What must
Martin’s common-kind theorist say about the robust
properties in virtue
of which E38 and E40 each have I40? One
version of the
common-kind theory would predict that these robust
properties have to
be different. For by the hypothesis of intransitivity,
red38 is
discriminable from red40. If these shades are
discriminable, then
at least some common-kind theorists will want
to say that the
specific robust properties are too. These will include
common-kind
theorists who take the specific robust properties to
be representational
ones, so that differences in what is represented
result in phenomenal
differences.
This last move
has a notable consequence: one and the same
indiscriminability
property could be had in virtue of having different
specific robust
properties. And this is to deny one of the claims that
Martin attributed to
the common-kind theory: that any two experiences
sharing an
indiscriminability property share the same specific
robust property.
Is the view
just sketched a disjunctivist view? There is a similarity
with disjunctivism:
both disjunctivism and the view sketched
deny that any two
experiences that are indiscriminable from the
same veridical
perception share the same robust mental property.
However, there are
also notable differences. Whereas the disjunctivist
holds that there are
pairs of hallucinations and veridical
perceptions that are
indiscriminable from the same veridical perception,
yet are of
fundamentally different kinds, the view just sketched
can allow that pairs
of experiences with different specific robust
properties
nevertheless are of the same fundamental kind. Both
experiences have, as
it might be, representational phenomenal
properties, and for all the view says, they have a general fundamental kind in common. The difference is only at the level of the specific robust properties.
So the view
sketched does not seem to capture the heart of
disjunctivism. It
does, however, seem to be incompatible with a
version of intensional
Hefty – a version that says that introspection
infallibly reveals
which specific robust properties a subject has.
But as we have seen,
there are independent reasons to think that
intensional versions
of Hefty are wrong anyway.
5. CONCLUSION
I’ve suggested
that someone could accept the main thrust of the
common-kind theory,
while rejecting two other claims that Martin
attributes to it.
The two claims are these: first, that the kind common
to perception and
hallucination is a property of an event in virtue
of which it is
indiscriminable from a veridical perception; second,
that it is a
property in virtue of which it is indiscriminable from
anything at all
(where indiscriminability is a cognitive notion). The
main thrust of the
common-kind theory, I suggested, is that there
are pairs of
hallucination and veridical perception sharing a fundamental
kind of mental
state, and this part of the theory survives the
denial of the two
claims just mentioned.
For all my
complaints about the link Martin sees between
indiscriminability
and the phenomenal, rejecting it leaves a
serious question
unanswered. Even to state the debate between
disjunctivism and
its opponents, one needs a way to characterize
the relevant pairs
of perceptions and hallucinations. Only some
such pairs raise the
question whether they share a fundamental
mental property.
Which pairs are these? The claim that they are
the pairs that are
indiscriminable from the same veridical perception
provides a simple
answer. If this answer is rejected, it’s not
clear what to
replace it with. Replacing it with some other notion
of
indiscriminability, or with some notion of phenomenal sameness,
brings in weighty
theoretical commitments at the outset – just
as Martin’s
cognitive notion of indiscriminability does. The moral
seems to be that in
this debate, it is difficult to escape making
theoretical
commitments from the very start about the kind of access
to experience that
introspection provides.22
NOTES
1 So “being indiscriminable from
a veridical perception” should be taken to
mean
something stronger than simply being indiscriminable from some veridical
perception
or other. How much stronger? If I had two successive veridical experiences
of
numerically different but qualitatively identical green cubes, would
the
property of being indiscriminable from each be different indiscriminability
properties?
For the purposes of this discussion, they could just as well count as
instances
of the same property.
2 Martin grants the conclusion
of the Causal Argument from Hallucination (see
section
4), which he formulates as follows: “whatever kind of experience does
occur
in situations like h
[hallucination],
it is possible that such a kind of experience
occurs
when one is veridically perceiving” (p. 12). Of course, “whatever
kind
. . .” is restricted to exclude the
kind: hallucination.
3 “. . . while the perceptual event is
of a fundamental kind which could not
occur
when hallucinating, nonetheless this very same event is also of some other
psychological
kind or kinds which a causally matching hallucinatory event . . .
belongs
to” (p. 17).
4 Martin’s stripe of
disjunctivism grants the conclusion of the Causal Argument
from
Hallucination (see previous note). There could be versions of disjunctivism
that
don’t, however.
5 Before beginning the critical
part of the discussion, a final expository note,
and
a warning. Martin’s disjunctivism is officially undecided on two matters:
first,
it leaves unsettled whether it applies to each sensory modalities, or only
to
some, or only to perceptual experience in general. Second, it is unsettled on
the
status
of illusions (where these are distinct from hallucinations) – whether they
always
instantiate robust properties in virtue of which they are indiscriminable
from
veridical perceptions or not, and if so, what sort of robust properties do
this.
I
believe that Martin has views on both of these matters, but that they don’t
bear
on
the issues raised in his paper.
The
warning: my discussion, like Martin’s discussion, takes as understood
the
idea that a particular, unrepeatable experience could be indiscriminable from
another
(perhaps repeatable) event. It’s not entirely clear what this means. The
relevant
notion of indiscriminability can’t be a statistical notion, since the event
said
to be indiscriminable is unrepeatable. Nor can the notion be reasonably
understood
by considering what would happen if the subject had two simultaneous
experiences,
compared them, and found that they were the same in the relevant
respect
(as one might be able to do with two physical objects). Perhaps one
could
think of how the subject would regard the pair of experiences, if she had
them
successively; but that introduces complications about memory that seem
extraneous.
Nevertheless, there is some intuitive sense in which certain pairs of
experiences
seem the same to the subject. Such a notion is needed even to state
the
debate between the disjunctivist and the common-kind theorist, and that is one
role
that is played by Martin’s notion of indiscriminability. I return to this point
in
section 5.
6 Another example: consider an
experience of seeing the Müller-Lyer lines.
Which
veridical perception is such that having an experience of these lines is
indiscriminable
from it? Presumably, if two lines really did differ in length,
but
each had the characteristic arrows drawn around them, they would look
different
than the lines look when they are the same length, at least given how
our
perceptual systems work.
In
the very latest version of Martin’s paper (written after these comments),
Martin
grants that Escher-experiences as a whole are not indiscriminable from a
veridical
perception, but suggests that they have constituent parts, each of which
is
indiscriminable from a veridical perception (p. 31). This strikes me as an ad
hoc
move tailored to specific example, and a retraction of a central claim.
7 Such an error theory might
hold that color is a property of mental, internal
objects
such as sense-data, that it couldn’t even in principle be a property of things
external
to the mind, and that color experience nevertheless represents external
things
as being colored.
8 One sort of worry that a
materialist about the mental might press is how experience
could
ever get to represent color properties, if such properties were never
instantiated.
9 The indefinite description “a
visual perception” should be taken to mean the
same
as “some veridical perception or other.” Cf. p. 20: “[t]he concept of
perceptual
experience
in general is that of situations indiscriminable from veridical
perception”;
and the passage cited earlier from p. 22: “I argued above that being
indiscriminable
from veridical perception is the most inclusive conception we
have
of what sensory experience is.”
10 In section 3 I discuss some
passages from pp. 10–11 of Martin’s paper that
show
that he construes the common-kind theory is this way.
11 In the very latest version of
Martin’s paper (written after these comments),
Martin
makes explicit that he accepts the other notion of indiscriminability, not
the
positive notion (fn. 11, p. 10). To state the issue between disjunctivism and
its
opponents,
some way is needed to characterize the similarity between the relevant
experiences,
and various notions of indiscriminability are candidate ways to do
this.
So it seems worth having the positive notion on the table.
12 Cf. “some event is an
experience of a street scene just in case it couldn’t be
told
apart through
introspection from
a veridical perception of the street as the
street”
(p. 9); “If the condition of indiscriminability is to be met, then a situation
of
experience must not lack any property necessary for veridical perception
the
absence of which is recognisable simply through reflection” (p. 20); “if any
property
of a veridical perception is introspectible – i.e. is recognisagbly present
in
perception through reflection . . .” (p. 22). Although in section 9 of his paper
Martin
also appeals to impersonal
notions of
indiscriminability, which seems in
tension
with this; I discuss this notion briefly at the end of this section.
13 TimothyWilliamson, Identity and Discrimination (Oxford: Blackwell, 1990),
p.
8. Two notable differences between the two notions are these. First, the
positive
notion makes use of the notion of disposition, whereas the double negative
notion
makes use of the notion of an ability. Second, where the positive
notion
defines indiscriminability in terms of judgment, Williamson’s
notion
defines it in terms of knowledge. Interestingly, only the first of these
differences
seems to survive the attempt to define corresponding notions of
discriminability.
For suppose X and Y were discriminable by S at t just in case S
was
disposed to judge X distinct from Y. Then X and Y could count as discriminable
for
S at t, even in the case that X = Y. And this seems to stretch the notion
of
discriminability unacceptably far beyond the ordinary notion. If this
infelicity
were
fixed by adding that the S at t be disposed to make a correct judgment, then
it
starts to look like it involves an appeal to knowledge (or something very much
like
it, given the link between knowledge and reliable judgment) after all. (Given
the
two dimensions of variation in the pair of notions of indiscriminability we’ve
been
considering, there are clearly other notions of indiscriminability that could
also
be considered.)
14 There may be an additional
reason for the common-kind theorist to reject
the
link between phenomenality and indiscriminability, at least on Williamson’s
notion
of indiscriminability. Consider once more what the sufficiency claim
comes
to, givenWilliamson’s notion:
If
S has I*, then S cannot activate knowledge that having-I* is distinct
from
having a veridical perception.
It
is not clear why a common-kind theorist should accept this. Suppose the
epistemic
basis in question is supposed to be reflection, where this includes not
only
introspection but philosophical theorizing as well. Let us grant that if S has
I*,
then on the basis of reflection, S cannot activate knowledge that she’s not
having
a veridical perception. Nonetheless, if S is a common kind theorist, S will
take
herself to know by philosophical theorizing that having I* is distinct from
having
a veridical perception, because S will take herself to know that she could
have
I* even if she weren’t veridically perceiving. So it seems that, by her lights,
at
least,
she can activate knowledge that having I* is distinct from having a veridical
perception.
Of
course, Martin thinks that the philosophical theorizing that typically leads
common-kind
theorists to this conclusion is mistaken, and perhaps he’s right. But
if
the claim is that a subject cannot activate the relevant knowledge on the basis
of
philosophical reflection, then it seems unmotivated to attribute the
sufficiency
claim
to the common-kind theorist in the first place.
15 In considering this worry in
the text, Martin also discusses the inattentive,
hasty
subject John, who treats samples of scarlet and vermillion indifferently.
Intutively,
his experiences of each should count as distinct; yet, as Martin points
out,
it does not seem inappropriate to say that John can’t discriminate scarlet
from
vermillion. Martin takes the moral to be that the disjunctivist should adopt
an
‘impersonal’ notion of indiscriminability, where this impersonal notion applies
equally
to the dog. I think the moral is rather that it’s the notion of an ability to
discriminate,
rather than a disposition to discriminate, that should figure in the
definitions
of indiscriminability.
16 Cf. last paragraph of Martin’s
section 2.
17 The same consideration that
makes Hefty implausible for a common-kind
theorist
should make it implausible for anyone who accepts Martin’s assumption
that
introspection supports Naďve Realism. If, as per Martin’s assumption,
introspection
on experience tells you that it consists in part of an object that
you
are veridically perceiving, it will be incorrect in cases of hallucination, no
matter
whether disjunctivism or the common-kind theory is true. By Martin’s
lights,
the common-kind theorist is forced to admit that appearances systematically
appear
to us other than they are, because no experiences are the way
introspection
(supposedly) says they are; the disjunctivist, in contrast, allows that
some
experiences are that way. Whether the disjunctivist also has to admit that
appearances
appear other than they are systematically depends on the extent to
which
experiences are not veridical. For discussion of the idea that most experiences
are
not veridical, see A.D. Smith, The Problem of Perception (Harvard
University
Press, 2002), chapter 1.
18 Cf. ‘Beyond Dispute’, in T.
Crane and S. Patterson (eds.), The History of the
Mind-Body
Problem, p.
197.
19 The defense just mentioned
itself has a controversial premise – the premise
that
indiscriminability is transitive. This has been defended by Graff (2001), but
is
widely thought to be shown false by cases of the sort discussed in the next
section.
20 For assumptions that there are
cases structured like the one described, see N.
Goodman,
The Structure
of Appearance,
1st edn., Cambridge: Harvard University
Press,
1951 and D. Armstrong, A Materialist Theory of Mind, London:
Routledge,
1968. For dissent, see D. Graff, ‘Phenomenal Continuaa and the
Sorites’,
Mind (2001); D. Raffman, ‘Is
Indistinguishability Non-transitive?’,
Philosophical
Topics 28(1)
(Spring 2001), pp. 153–175.
21 I take it that the heart of
the controversy being ignored concerns whether the
relation
in question is transitive or not; simplifying assumptions about what form
it
takes won’t change anything.
22 For helpful comments on
earlier drafts, thanks to Alex Byrne and Bernard
Nickel.
For extensive discussion of every issue discussed here, thanks to Maja Spener,
Scott Sturgeon, and David Chalmers. Additional thanks to the last two for
criticizing later drafts. Finally, many thanks are due to Mike Martin, for
writing such a rich and rewarding paper, and for so many nice discussions of it
before and during the 2002 Oberlin Colloquium.
Department of
Philosophy
208 Emerson Hall
Harvard
University
Cambridge MA
02138
USA
E-mail: ssiegel@fas.harvard.edu