. 19
( 62 .)


While total realizations are in some sense complete states of S, they
are incomplete in two important respects. First, the distinctness of S and
the subject or bearer of P entails that total realizations do not include all
states of those subjects or bearers, for not all states a subject or bearer
is in form part of the system speci¬ed. For example, a person™s having
a toenail of two centimeters, while a property of that person, is not a
property of that person™s digestive or respiratory systems; “x has a toenail
of two centimeters” expresses a property of persons, not of digestive or
respiratory systems. Second, the total realization of P excludes the back-
ground conditions that are necessary for there to be the appropriate,
functioning system. While these may themselves be necessary for a given
entity to have P, since they are not states of S, they are no part of the total
realization of P. Thus, total realizations should be distinguished from the
broader circumstances in which they occur.
To illustrate these points, consider the mammalian circulatory system,
which is made up of various parts “ such as the heart, arteries, capil-
laries, arterioles, venules, and blood. Various states of these parts, con-
sidered together, determine what circulatory properties one has at any
given time. Related common sense and medical theories about circula-
tion specify what the circulatory system includes and excludes, but it is
clearly a (proper) part of an organism. For a given circulatory property “
say blood pressure “ not all parts are of equal causal importance. From an
intuitive point of view, one™s blood pressure is most saliently determined
by the condition of one™s heart and arteries. Thus, the core realization
of, say, having blood pressure of 120/80 would be identi¬ed with a state
of these parts of the circulatory system “ for example, having clogged
arteries and a strong heart. Yet such states do not by themselves and in-
dependent of the state of the rest of the circulatory system guarantee
blood pressure of 120/80 in a person. Rather, they need to be located
Individualism and Externalism in Cognitive Sciences

in a certain way within the rest of the person™s circulatory system. A total
realization of having blood pressure of 120/80 is a state of the circula-
tory system, including the states of having clogged arteries and a strong
heart, that determines the presence of that property. Excluded from to-
tal realizations are both properties instantiated by the individual that are
not properties of the circulatory system at all (such as her having brown
hair, or being six feet tall), as well as broader features of the individual™s
environment that are necessary for her to have a functioning circulatory
system (such as there being oxygen in the environment and the world™s
persistence through time). Such background conditions are no part of
the total realization of the corresponding property since they are not
properties of the circulatory system at all.
Since the distinctness between a total realization and background con-
ditions is important for the general alternative to the standard view of re-
alization that I want to present, consider one other way of coming to this
distinction. Consider our common sense view of circulation and how we
would expect it to be modi¬ed by the ¬ndings of circulatory physiologists.
While we would expect physiologists to offer a more precise speci¬cation
of both the core and total realizations of the properties of this system, we
wouldn™t expect them to contribute much to our understanding of the
nature of the background conditions of these realizations. (For this, they
defer to other scientists, or to common sense.) In investigating a given
biological system, scientists examine both the core and noncore parts
of a total realization, but the boundaries of the system to a large extent
delineate the boundaries of their inquiry.
Strictly speaking then, it is only the physical states constituting a total
realization together with the appropriate background conditions that metaphys-
ically suf¬ce for P. Our paradigms for the relevant systems are function-
ing, integrated physical systems, and without the appropriate background
conditions in each case there would be no such systems. This might be
taken to suggest that even total realizations, considered simply as complex
con¬gurations of physical matter and energy, are metaphysically context
sensitive in much the way that partial realizations are.
Let us catch our breath. In this section I have identi¬ed two types
of realization that violate the suf¬ciency thesis and pointed to ways in
which realizations are context sensitive. First, core realizations in and
of themselves are not metaphysically suf¬cient for the properties they
realize, but must be part of some larger functional system. This point is
of some signi¬cance because even if no one really believes the suf¬ciency
thesis to be true of core realizations, it is physical structures that in fact
Metaphysics, Mind, and Science: Two Views 111

are core realizations “ such as C-¬ber ¬rings “ that are typically invoked
in discussions of reductionism, realization, and functionalism. This is
especially true in discussions of mental states. By the end of this chapter
I hope to have made a prima facie case that this is also true of the sorts
of example, those of genes in developmental biology and collectives in
the social sciences, that I mentioned in the previous section. Second,
since total realizations are physical states of larger functional systems, and
there are background conditions necessary for their functioning, strictly
speaking even total realizations themselves do not satisfy the suf¬ciency
thesis. The signi¬cance of these points will unfold over the next few

8 physical constitutivity and wide realizations
Thus far, I have assumed the constitutivity thesis and thus individualis-
tic realizations, having used that assumption to challenge the suf¬ciency
thesis. That thesis was challenged as a global view of realizations in two
ways by realizations that are context sensitive: Core realizations are both
metaphysically and epistemically context sensitive, and in presupposing
background conditions necessary for the existence and functioning of
the corresponding system, total realizations are metaphysically context
sensitive. I turn now to the constitutivity thesis and how it is undermined
by a more far-reaching type of context sensitivity. Here the suf¬ciency the-
sis will be my ally, and I shall return to focus initially on mental properties
in particular.
I begin by elaborating on my claim, made in section 5, that homuncu-
lar functionalism is often construed as a methodological counterpart to
the constitutivity thesis. The idea of the prevalent strategy of homuncu-
lar decomposition in cognitive science is to explain complex, intelligent,
representational capacities by functionally analyzing them into simpler
(but typically more numerous) capacities, and then reapplying this ¬rst
step recursively until we have simple abilities that require neither repre-
sentation nor intelligence. If each homuncular level of analysis provides
a realization of the level above it, and realizations satisfy the constitutivity
thesis, then any view of homuncular functionalism that purports to be a
physicalist view should proceed via physical decomposition.8
The constitutivity thesis itself implies that realizations of mental prop-
erties are individualistic, in that two molecularly identical individuals
must also share the same realizations of mental properties. And if realiza-
tions are determinative of the properties they realize, mental properties
Individualism and Externalism in Cognitive Sciences

must be individualistic, too. Indeed, re¬‚ection on the relationship be-
tween the above bodily systems and the individuals to whom they belong
supports this as a general view of realizations: Because bodily systems are
parts of individuals, there is no way for molecularly identical individuals
to differ in the bodily systems that each has.
This general view overlooks, however, that there are two species of
total realizations, only one of which can be understood in terms of the
notion of constitution above. While it is often the case that S is a part
of the individual that has P, there are a variety of examples in which the
converse is true, examples in which the individual that has P is a part of S.
These are cases in which S extends beyond the boundary of the individual,
and I shall call the type of total realization that exists in such cases a wide
Let IB be the subject or individual bearer of P. In constitutive de-
composition, of which homuncular functionalism is often construed as
a paradigm, S is a part of IB. By contrast, in cases of integrative synthesis
IB is a part of S; in these cases, P has total realizations that are wide. We
can summarize the distinction between wide and entity-bounded realiza-
tions in terms of the location of the noncore part of a total realization as

(c) entity-bounded realization: a total realization of P whose noncore
part is located entirely within IB, the individual who has P
(d) wide realization: a total realization of P whose noncore part is not
located entirely within IB, the individual who has P

Figure 5.1 provides a simpli¬ed depiction of the metaphysical parallels
between these two forms of realization, as well as of the differences be-
tween the corresponding strategies of constitutive decomposition and




¬gure 5.1. (a) Constitutive Decomposition. Involving Entity-Bounded Realiza-
tion (b) Integrative Synthesis Involving Wide Realization. S = system B = bearer
Metaphysics, Mind, and Science: Two Views 113

integrative synthesis. As a species of total realization, wide realizations
satisfy the suf¬ciency thesis. Yet since they extend beyond the physical
boundary of the individual, they are not exhaustively constituted by the
intrinsic, physical properties of the individual subject, and so do not sat-
isfy the constitutivity thesis.
The concept of a wide realization allows us to make metaphysical sense
of Putnam-Burge externalism introduced in the last chapter. On this view
the propositional attitudes are not individualistic, or they at least have
a nonindividualistic aspect. The propositional attitudes have a physical
total realization, albeit one that is not entity bounded. The realization
of particular folk psychological states is wide, and given the framework
I am proposing that entails that those states should be understood by
using integrative synthesis to locate their bearers in some broader system,
presumably one that involves social relations between individuals. I shall
call this our folk psychological system.
The width of our folk psychological system is not anomalous in
psychology; in fact, the strategy of integrative synthesis also applies readily
to computational psychology. Many computational systems that govern
cognition are themselves wide, where the computational system S extends
beyond the boundary of IB, the individual who bears the psychological
properties, and the appropriate type of total realization is a wide realiza-
tion. I have previously argued that we should expect wide computational
systems of cognitive states just when there has been sustained mind-world
constancy over evolutionary time of the type that one ¬nds in the case
of many perceptual and behavioral systems. Such systems include our
mechanisms for form perception and the navigational systems that ants
and bees deploy.9
This general view “ that world-mind constancy creates the opportunity
for cognitive loads to be shifted from inside the head to beyond it “ has
been advocated by a number of others. The philosopher Andy Clark™s 007
Principle “ “know only as much as you need to know to get the job done” “
was postulated to explain informational off loading and exploitation by
organisms in an evolutionary context, and his more recent endorsement
of “the extended mind” has taken this perspective as implying that the
mind itself may include parts of the world beyond the individual. Likewise,
the cognitive scientist Edwin Hutchins views his study of navigation as a
socially distributed, embedded computational process to imply that the
computations performed during navigational tasks extend beyond the
boundary of the individual. I shall discuss wide computationalism and
these sorts of views of computation and cognition more generally and
Individualism and Externalism in Cognitive Sciences

in more detail in Chapter 7. My chief point here is that the invocation
of wide realizations is not an ad hoc ploy introduced simply in order
to add a supporting metaphysics to our externalist intuitions about folk

9 wide realizations in the biological
and social sciences
The further point that I do want to develop in some detail is that there
is nothing strange about mental states having wide realizations. In fact,
I think that wide realizations are typical across the fragile sciences,
at least if we assume a notion of realization that satis¬es the suf¬-
ciency thesis. The claim that wide realizations are ubiquitous in such
sciences should sound, at least prima facie, counterintuitive. But once
we make the connection between the metaphysics of realization and the
methodologies of constitutive decomposition and integrative synthesis,
this point becomes easier to understand. Let me begin with the biological
A variety of evolutionary and ecological properties themselves have
wide realizations and are pro¬tably understood through the strategy of
integrative synthesis, that is, by locating their bearers in the correspond-
ing wide systems. Such properties include ¬tness, being highly special-
ized, and being a predator, properties of individual organisms or even
species; and properties of phenotypic traits or behaviors, such as being
an adaptation, a homology, or a spandrel.
Consider ¬rst the example perhaps closest to common sense, the prop-
erty of being a predator. Predators play a certain role in an ecological sys-
tem, occupy a particular ecological niche, that of preying on other living
things, typically (but not solely) for nourishment. This property is rela-
tional in at least two ways: An organism is a predator for certain other
living things (their prey), and in certain types of environments (their
ecosystem). Here the relevant system is the predator-prey system, whose
dynamics are captured, in part, by the Lotka-Volterra equations and is
explored in population ecology. While the part of this system most read-
ily identi¬able as playing a crucial causal role in producing or sustaining
the property of being a predator, that is, that property™s core realization,
might be thought to be contained within the organism that is a predator,
the total realization for this property is clearly wide. Since being a preda-
tor involves a relation between an organism and something beyond its or-
ganismic envelope, what is metaphysically suf¬cient for particular bodily
Metaphysics, Mind, and Science: Two Views 115

or behavioral phenotypes to be those of a predator extends beyond the
boundary of the individual.
The same is true of being highly specialized, the ability to occupy a rela-
tively small number of the ecological niches available. Although we can
and do speak of certain organisms as realizing this property by virtue of
having speci¬c intrinsic properties, at best these states or properties of
individual organisms are core realizations for that property, and the cor-
responding total realizations include physical con¬gurations in a system
that extends beyond the individuals who are highly specialized, that is,
the total realizations are wide. In the case of being highly specialized, the
relevant system is the ecosystem, this being the system in terms of which
ecological niches are de¬ned.
Finally, an organism™s ¬tness is its propensity to survive and reproduce
in its environment. We can represent the former as a probability between
0 and 1 (the organism™s viability), and the latter as a number greater
than or equal to 0 (the organism™s fertility) where this number represents
the organism™s expected number of offspring. In either case, although
¬tness is a dispositional property of individual organisms (or even whole
species), this disposition is not individualistic, since physically identical
organisms may differ in ¬tness because they have been or are located in


. 19
( 62 .)