<<

. 22
( 62 .)



>>

corresponding mental properties, and so in effect gives up the constitu-
tivity thesis, it must be revised in fairly signi¬cant ways: We are no longer
talking of token physical states of the brain, or compositional states of
individuals.
Nonreductionist forms of physicalism are also often expressed in terms
of there being “higher” and “lower” levels of explanation, the latter of
which provide a metaphysical but not a reductive basis for the former.
Whether we can adequately conceptualize mental states in general as be-
ing (totally) realized by “lower level” states seems to me doubtful. Those
articulating this idea further using a constitution-based conception of re-
alization either will be hard pressed to maintain the view that realization
is determinative, or will, in effect, concede that lower levels do provide a
reductive base for higher levels. Neither option should be attractive to a
nonreductionist. To tackle the ¬rst horn of this dilemma would require
a fairly radical rethinking of the concept of realization, one that gives up
on the suf¬ciency thesis altogether. Tackling the second horn threatens to
locate the site of one™s nonreductionism solely within the realm of expla-
nation, a threat exploited by Jaegwon Kim in his attacks on nonreductive
physicalism.7
We can make this point in another way and more positively. I said in
Chapter 5 that a constitution-based conception of realization appears to
provide the metaphysical grounding for the explanatory strategy of ho-
muncular functionalism. In the language of higher and lower levels, this
is the idea that things and properties speci¬ed by lower-level homuncu-
lar descriptions physically constitute those speci¬ed by higher-level ho-
muncular descriptions. If we grant that at least these latter things and
properties are often relationally individuated, this relation of constitu-
tion can be determinative only if the former things and properties are
likewise relationally individuated. This is to say that the things and proper-
ties speci¬ed by lower-level homuncular descriptions may be relationally
individuated. And if the relevant relations for the higher-level properties
extend beyond the boundary of the individual, so too must those for the
lower-level properties. So while there may be some sense in which lower
levels “constitute” higher levels, neither need be exhausted by the sub-
ject or bearer™s intrinsic, physical properties, that is, by those properties
usually taken to physically constitute an individual.
In effect, a homuncularly decompositional view that takes relational
individuation seriously entails rejecting a premise crucial to a smallist,
individualist view of the mind, a variation on the suf¬ciency thesis, viz.,
Individualism and Externalism in Cognitive Sciences
130

that the physical constitution of an individual metaphysically determines
what mental properties that individual has. There are metaphysical and
not merely pragmatic grounds for construing homuncular functionalism
as an externalist view, and proponents of homuncular functionalism need
to transcend their implicit reliance on the standard view of realization.


5 the modi¬ed standard view: causation
and realization
One might reasonably doubt whether the context-sensitive view is needed
to accommodate the ubiquity of relational properties and kinds in the
fragile sciences. Perhaps a view of realization embracing both the suf¬-
ciency and constitutivity theses can be modi¬ed in ways that “take rela-
tional individuation” seriously. In this section I shall consider the idea that
a modi¬ed standard view can do so simply by acknowledging a greater
and more explicit role for background conditions than the standard view
itself has.
Central to the standard view is the intuition that total realizations them-
selves always satisfy the constitutivity thesis. Even if ascribing any prop-
erty to an entity, including mental properties, presupposes that certain
beyond-the-individual background conditions hold, realizations them-
selves do not extend into the world beyond the individual. That is, all
total realizations are in fact entity bounded, and we should not mistake
some of these background conditions for (parts of) the realization it-
self. Putative examples of wide and radically wide realizations should be
reinterpreted within the parameters of the standard view, modi¬ed to
acknowledge just the ¬rst two forms of context-sensitive realization that
I identi¬ed in Chapter 5: That of core realizations, which are realiza-
tions only insofar as they are part of a total realization, and the context
sensitivity of total realizations to background conditions. On this modi-
¬ed standard view, realizations are entity bounded and the background
conditions necessary for the realization of a given property may be more
extensive than initially envisaged.
We can support the modi¬ed standard view by drawing a parallel be-
tween how we should think about realization and how we often think
about causation. We talk of Sarah™s carelessly discarding an un¬nished
cigarette as the cause of the bush¬re even though we recognize that this
event is not itself suf¬cient to bring about that effect. It does so only
given that a variety of conditions are in place: that it not rain, that the
Context-Sensitive Realizations 131

cigarette make contact with combustible material, that the cigarette is not
extinguished before it ignites that material, and so on. This distinction
between causes and conditions is commonplace in theories of causation,
but it provides little reason, one might argue, for viewing causes as either
“wide” or “radically wide.” It is Sarah™s action itself that is the cause, this
event here and now. Likewise, one might think, for realizers.
I shall return to the analogy to causation in a moment, but let us con-
sider the modi¬ed standard view itself. Of critical importance is how well
that view allows us to make sense of the full range of properties and kinds
posited across the various sciences, as well as those found in our common-
sense discourse. Both social actions and mental states pose problems here,
and suggest that utilizing only an entity-bounded notion of realization,
together with background conditions, requires abandoning some central
intuitions about realization.
Most problematic are social actions, which not only have political, eco-
nomic, and legal background conditions but also literally extend into the
world beyond the individual who enacts them. For these actions, such as
signing a check, even what I have been calling their core realizations do
not stop at the skin. On the modi¬ed standard view, realizations must
be entity-bounded, and so the realization of the action of (say) signing
a check can appeal to holding a pen and writing on paper only as back-
ground conditions for the realization of the action. Yet these are things
that the agent herself does, part of the action itself, not (like other back-
ground conditions) merely general features of the social and institutional
environment in which she acts, and this distinction is ignored within the
modi¬ed standard view. The real problem here is that viewing something
like the bodily movements of the agent as the total realization of her
signing a check is to collapse the distinction between total and core real-
izations. In effect, it is to accept that the realization of the action is not
determinative of the action, and so give up the suf¬ciency thesis. Thus,
this is not a view that modi¬es the standard view, but abandons it.
There is a fundamental, general problem for the modi¬ed standard
view here. Treating total realizations as core realizations makes them
not simply metaphysically context sensitive but also epistemically context
sensitive. That is, what counts as the realization for any given property
or kind depends on us. On the modi¬ed standard view, any part of the
world that is a physical constituent of the bearer of the property or kind
can be the realization of that property. All that needs to be done is sim-
ply to assign the state of the remainder of that individual to the set of
Individualism and Externalism in Cognitive Sciences
132

background conditions that, together with that physical constituent, are
metaphysically suf¬cient for the property or kind. In the case of signing
the check, we could push further back into the body (in fact, all the way
to the brain) in identifying the realization of that action, provided that
we make a corresponding addition to our set of background conditions.
Thus, on the modi¬ed standard view there is no mind-independent way
to specify the realization for any given property or kind.
Consider another example that illustrates the general problem here.
Any psychological state, such as being in pain, involves the ¬ring of thou-
sands if not millions of neurons. In a given case, which of these is the
realization of the pain? On the modi¬ed standard view, a single neuron
¬ring in my head at a given time could realize this state, provided that we
specify the remaining neural ¬rings as part of the background conditions
for that single neuron™s ¬ring to metaphysically suf¬ce for the individual
to be in pain. To put it the other way around, on the modi¬ed standard
view there is no answer to the question of what the realization of a given
psychological state is that is independent of our decision of how exten-
sive to make the corresponding background conditions. This constitutes
a reductio of the modi¬ed standard view.
We can return to the analogy to causation to diagnose what has gone
wrong here. Causes have the effects they do only given the presence of
certain conditions. If one has only the distinction between causes and
the conditions that necessitate their effects, then the arbitrariness of the
distinction between cause and condition lead one to consider the indi-
viduation of some events as causes and others as conditions as dependent
on us. Essentially, this is the situation that the modi¬ed standard view of
realization is faced with.8
On the context-sensitive view that I have introduced, background con-
ditions are necessary for there to be a functioning system that (totally)
realizes an individual™s properties. To be a realist about properties, and so
about mental properties, is to be a realist about at least their total realiza-
tions, and thus about the systems with respect to which total realizations
are de¬ned. Thus, the distinction between background conditions and
noncore parts of total realizations is required by the realism implicit
in the view I have defended (see also section 7). We simply don™t get
to decide where to draw the line between realization and background
conditions, and in particular it is not simply up to us to decide that
the realizations must be entity bounded, or where within an individ-
ual they begin and end. Given that the systems in terms of which re-
alizations are characterized are robust entities either that form parts of
Context-Sensitive Realizations 133

individuals or that individuals form a part of, background conditions have
a more restricted role to play than this modi¬cation of the standard view
suggests.
In terms of the analogy to causation, this is something like introducing
the notion of a causal chain between that of cause and condition. Causes
form causal chains, and these causal chains have their effects only given
the existence of certain other conditions. When we single out some event
as the cause of another, we imply that it is the most salient part of that
causal chain, and that without the remainder of that causal chain linking
cause and effect, “the cause” would be no cause at all. But it is also true
that without further supporting conditions, even the whole causal chain
would not produce that effect. What distinguishes causal chains from
the conditions against which they bring about their effects is a further
issue. But the parallel between two trichotomies “ between cause, causal
chain, and conditions; and core realization, system (total realization),
and background conditions “ is useful in understanding why one needs
to go beyond the resources of the modi¬ed standard view of realization.


6 context sensitivity within the standard view
Several physicalist proposals that can be viewed as sympathetic to the
standard view have acknowledged the role of context in the metaphysics
of mind, and in this section I shall consider whether they adequately ac-
commodate the sort of context sensitivity introduced in Chapter 5. These
proposals are Terry Horgan™s idea of regional supervenience, and Denis
Walsh™s more recent defense of a view he calls wide content individualism
or alternative individualism.
While neither of these proposals focus on the notion of realization per
se, both views can be construed as attempts to provide a role for context
that maintain some version of the constitutivity thesis about realizations.
Their shortfalls qua modi¬cations of the standard view are my concern
here.
Terry Horgan introduced a thesis of supervenience that he later chris-
tened regional supervenience:

There are no two P-regions [spatio-temporal regions of a physically possible
world] that are exactly alike in all qualitative intrinsic physical features but dif-
ferent in some other qualitative intrinsic features.9

Regional supervenience was introduced to account for what Horgan
calls an individual™s context-dependent properties. Horgan™s examples
Individualism and Externalism in Cognitive Sciences
134

include being president of the United States, being a bank, and knowing
that Oscar Peterson is a jazz pianist. It does so by extending the subvenient
base to the spatio-temporal region that contains that individual so as to
include the relevant contextual factors in that base.
If we were to consider this as extending the realization base beyond the
individual, we would have something like a wide realization (though note
that I have de¬ned these in terms of entitylike systems, rather than spatio-
temporal regions). But this, of course, would be to give up the constitu-
tivity thesis. Horgan himself thinks that the realizers for such properties
are typically narrower than the corresponding subvenient base, suggest-
ing that he views realizations as satisfying the constitutivity thesis. Such a
view, however, can claim only that realizations, together with the larger
spatio-temporal region of which they are a part, determine the proper-
ties they realize, giving us realizers that by themselves do not satisfy the
suf¬ciency thesis. What satis¬es the suf¬ciency thesis is the regional su-
pervenience base, but clearly that does not satisfy the constitutivity thesis.
In short, I think that we see here something like the tension between the
two halves of the standard view of realization, with what metaphysically
determines the presence of a property in an individual being something
more than what is bounded by that individual.
This argument can be reexpressed as follows. Consider the constitu-
tivity thesis reformulated with respect to the entire region, rather than
the individual in that region who has the properties:

regional physical constitutivity thesis: Realizers of states and properties are ex-
haustively physically constituted by the intrinsic, physical states of the region
containing the individual whose states or properties they are.

The metaphysical suf¬ciency thesis and the regional physical constitutiv-
ity thesis are jointly satis¬able. But since the latter thesis explicitly requires
going beyond the boundary of the individual property bearer in order
to specify what determines that property, realization on this view is not
entity bounded.
Denis Walsh™s chief aim is to articulate a position that resolves what he
calls the antinomy of individuation: that combining the plausible claim
that thoughts of the same psychological kind have the same (wide) con-
tents with individualism entails the implausible conclusion that “thoughts
which are instances of the same physiological kind have the same wide
contents.” Walsh™s solution to the antinomy is to reformulate each of
these three claims so as to make explicit the way in which psychological
states are context sensitive. These three principles, which together he
Context-Sensitive Realizations 135

calls wide content individualism, are:
(1) Necessarily, if individuals have thoughts of the same psychological
kinds with respect to a context, then their thoughts have the same
(context-sensitive) content with respect to that context.
(2) Necessarily, states of the same physiological kind which share a
context realize states of the same psychological kind with respect
to that shared context.
(3) Necessarily, states of the same physiological kind which share a
context realize thoughts with the same content with respect to that
context.
As in the original antinomy of individuation, (1) and (2) entail (3), but
(3), Walsh suggests, is true.10
(2) implies that identical intrinsic, physical states of individuals in the
same context realize the same psychological states with respect to that
context, while (3) spells out this implication for the special case of in-
tentional psychological states. While Walsh suggests that this is a way of

<<

. 22
( 62 .)



>>