<<

. 11
( 62 .)



>>

ning just with sensory primitives and general psychological mechanisms,
acquires the variety of knowledge that it ends up possessing; it is how
environmental complexity leads to or produces organismic complexity.
The centrality of learning for behaviorists is re¬‚ected in their emphasis
on techniques of conditioning and in their attempted development of a
generalized learning theory that applies to any behavior that is cognitive
in a broad and intuitive sense.
If the twentieth-century nativist-empiricist dialectic begins with the
Chomskyan challenge to the empiricist paradigm of behaviorism, it has
been continued by a newer empiricist alternative, that provided by con-
nectionism. Connectionism was developed in the early 1980s as a compu-
tational modeling technique alternative to that embodied by “classical”
or “traditional” arti¬cial intelligence. Like behaviorism, its empiricism
was manifest in its emphasis on explaining complex capacities in terms
of the operation of learning strategies or rules on relatively simple inputs.
Rather than viewing the innate cognitive contributions of the organism
to be speci¬c, rich, and “high-level” in structure, connectionists place an
emphasis on environmental stimuli (the initial data set) and powerful,
general computational algorithms that are reapplied to an initial data set
to generate complex outputs.
Disciplining the Individual and the Mind
56

One computational model that illustrates the challenge that connec-
tionist models pose to nativist views of language is the connectionist
model for the formation of past tense developed by David Rumelhart
and James McClelland. What is signi¬cant about this model in the cur-
rent context is that it does not have a preexisting linguistic rule that
distinguishes regular from irregular cases, but eventually develops this
distinction from repeated applications of an algorithm. The model was
taken by its proponents to exemplify the idea that a sensitivity to linguis-
tic structure could be developed through repeated interactions with an
environment, rather than having to be “built in” to the language learner,
as nativists had claimed. As Pinker notes, the “model irrevocably changed
the study of human language. . . . it is no longer possible to treat the mere
existence of linguistic productivity as evidence for rules in the head.”
While nobody views this model as depicting how we learn the past tense
in natural language, the model has spawned over twenty descendants
that have led proponents of “rules and representations” views of lan-
guage acquisition to attenuate their views. These have also motivated the
development of hybrid models that integrate the insights of such views
with their connectionist competitors.11

4 the two-dimensional approach
The basic idea of the two-dimensional approach to the debate between
nativists and empiricists, recall, is that two simple, independent theses de-
termine a two-by-two space in this debate. The two-dimensional approach
does three things. First, it helps to identify strong forms of both nativist
and empiricist views. Second, it points to views that are less extreme and,
in some sense, intermediate between the strong versions of nativism and
empiricism. Third, it provides a common framework for understanding
debates over nativism elsewhere in the fragile sciences.
The ¬rst thesis, the internal richness thesis, concerns the nature of an
individual™s internal structure, while the second, the external minimalism
thesis, concerns the causal role of an individual™s external environment in
the acquisition and development of a given cognitive process, structure,
or phenomenon. Where X stands in for the relevant cognitive process,
structure, or phenomenon, these two theses can be stated as follows:

Internal Richness Thesis: Structures and processes internal to the individual that
are important to the acquisition and development of X are rich.
External Minimalism Thesis: Structures and processes external to an individual
play at best a secondary causal role in the acquisition and development of X.
Nativism on My Mind 57

Richness here is a technical term, but it is intended to draw on the ordinary
sense that the word has. Internal structures are rich, recall, just if they
are antecedently specialized, localized, internally complex, and causally
powerful. Proponents of the “modularity of mind” are paradigms of those
who accept the internal richness thesis about cognition, while those who
hold that cognitive processes are “domain general” or close to nonexis-
tent (for example, behaviorists) are paradigms of those who reject that
thesis. Likewise, those who conceptualize the environment as a “trigger”
for cognition accept the external minimalism thesis, while those who
view learning as the central mechanism for the acquisition of cognitive
structure reject that thesis.
As these examples might suggest, strong nativist positions “ exempli-
¬ed by Chomsky and Fodor “ accept both the internal richness and the
external minimalism thesis. Although Chomsky and Fodor are strong
nativists with respect to a range of mental processes and structures,
Chomsky™s view of generative grammar and Fodor™s views of concepts
exemplify their commitment to strong nativism as I have characterized
it. Strong empiricist positions “ exempli¬ed by behaviorists and at least
early connectionists “ reject both the internal richness and the exter-
nal minimalism thesis. For both behaviorists and early connectionists,
an organism™s cognitive structures are acquired through the process of
learning. This process is not internally rich in the sense speci¬ed above,
and ascribes a primary role to the structure of the environment in the
process of acquisition. That both of these theses capture standard nativist
and empiricist positions is re¬‚ected in Steve Laurence and Eric Margolis™s
recent characterization of these views:

Though there is clearly a continuum of positions, empiricist models are on the
side that attributes few innate ideas, principles, and mechanisms, and generally
considers the innate material of the mind to be domain-neutral and relatively
simple; empiricist models also tend to give special weight to the role of sensory
systems, in the most extreme cases maintaining that all concepts are constructed
from sensory constituents. By contrast, nativist models are on the side that views
the mind as highly differentiated, composed of far more innate elements, includ-
ing domain-speci¬c systems of knowledge or principles of inference, and innate
concepts of arbitrary levels of abstraction.12

Although these two theses are independent, strong nativists view them
as mutually reinforcing. Here™s why. Suppose that we accept the internal
richness thesis. If our cognitive architecture is internally rich, there is a
complexity to what we bring to any cognitive task that leaves less role for
an appeal to the complexity of the environments in which we operate.
Disciplining the Individual and the Mind
58

In the extreme case, environments are just triggers for the operation
of innately complex structures. Conversely, suppose that we accept the
external minimalism thesis. Since external structures play a minimal role
in the acquisition and development of cognition, any complexity that
cognition generates must be due to something else, and what lies within
the cognizer is a strong candidate for that something else.
To the extent that the two-dimensional analysis accurately classi¬es
paradigmatic nativist and empiricist positions, it passes one relatively un-
demanding test for the adequacy of an analysis of the debate. Yet the
two-dimensional view is well suited to capture more of this logical geog-
raphy, including positions less extreme than either of the paradigmatic
nativist and empiricist views, and I consider this one of its chief virtues.
There are various ways to reject paradigmatic nativist and empiricist views
of cognition, and the two-dimensional view is not only “empirically ade-
quate” but informative about this aspect to the logical geography of the
debate over nativism and cognition.
Consider the option of accepting the external minimalism thesis but
rejecting the internal richness thesis about cognition. Classic ethologists
and proponents of general intelligence as a built-in, global cognitive abil-
ity are both advocates of such a position. In fact, we might locate much of
cognitive psychology here, insofar as it gives little role to the environment
in structuring cognition yet is cast largely in terms of domain-general ca-
pacities, such as memory and reasoning, emotions and imagination.
The converse alternative “ of holding the internal richness thesis while
rejecting the external minimalism thesis about cognition “ has also been
endorsed. For example, there are modularity theorists, such as Annette
Karmiloff-Smith, who view modules as being assembled through environ-
mental interaction, rather than endogenously prespeci¬ed or governed.
For forms of cognitive activity in which there can be specialized, acquired
expertise, such as chess playing or mathematical problem solving, such
a view is particularly plausible. Proponents of socially distributed cogni-
tion, such as Edwin Hutchins, seem also to endorse this sort of view. On
Hutchins™ view of cognition, the structure of the social environment of
the individual plays a primary role in generating the structure of both
individual and group cognition. Yet individuals themselves are, in effect,
internally rich devices for the performance of cognitive functions.13
Figure 3.1 provides a summary of how the two-dimensional view applies
to this range of positions.
The two-dimensional view also sheds some light on how each of two
neurally inspired views depart from strong nativist views of cognitive
Nativism on My Mind 59


External Minimalism Thesis

YES NO



Y
Socially Distributed/
Strong Nativism E
Acquired Cognition Internal
S
Richness

General Cognitive Thesis
Strong Antinativism N
Psychology
O
¬gure 3.1. Nativism and Cognition

development, and how they are related to one another. The “constructive
learning” view of neural and cognitive development that computational
neuroscientists Steve Quartz and Terry Sejnowski articulate exempli¬es a
strong rejection of nativism. It ascribes a central role to the structure of the
environment in the construction of representations (and so rejects the ex-
ternal minimalism thesis), while denying the need to posit antecedently
richly structured cognitive modules (and so rejects the internal richness
thesis). By contrast, neuronal selection theory, whereby neural develop-
ment is explained as a selective process responsive to endogenous events,
shares the internal richness thesis with strong nativist views, but appeals
to underlying neural features, rather than cognitive modules, to explain
the structure of cognition. This af¬nity between neural selection theory
and strong nativist views, as well as their shared rejection of the external
minimalism thesis, underpins, I believe, Quartz and Sejnowski™s rejection
of both.14
Another criterion for evaluating the two-dimensional view is to see how
adequately it captures more speci¬c debates over nativism amongst cogni-
tive scientists. For example, we might turn to debates over the acquisition
of concepts in general, of word meanings or other aspects of language,
such as syntax, or those over speci¬c postulated cognitive domains, such
as the theory of mind.15
Consider ¬rst the last of these examples, that of the theory of mind
within developmental psychology. The two-dimensional account classes
the views of Alan Leslie and Simon Baron-Cohen, who hold that children™s
Disciplining the Individual and the Mind
60

psychological attributions and understanding of the mind are governed
by a domain-speci¬c module that develops maturationally, as strong na-
tivist positions. Strong antinativist positions, by contrast, are exempli¬ed
by Peter Hobson and Michael Tomasello. They view these abilities as de-
riving not from dedicated modules but from more general symbolizing
capacities acquired through social experience.16
But again there are those who have maintained views less extreme
vis-` -vis nativism. For example, “theory theorists,” such as Alison Gopnik
a
and Joseph Perner, take children to develop a theory of mind through the
same mechanisms of theory construction that children apply elsewhere
in acquiring knowledge of the world (and so reject internal richness).
Yet Gopnik and Perner ascribe only a limited role to the environment in
the generate-and-test process that governs their conceptual change, and
so share the external minimalism thesis with strong nativists. By contrast,
theorists such as Jay Gar¬eld and Candida Peterson maintain a view that
appropriates the mediational views of Vygotsky “ and so reject the external
minimalism thesis “ but recognize a modular component to children™s
acquisition of knowledge of the mind, even though they do not think that
there is a domain for the theory of mind per se, and so accept a form of
the internal richness thesis.17
Figure 3.2 summarizes the two-dimensional view of the nativist aspect
to the theory of mind debate. As one might expect, there are further
aspects to the debate over the theory of mind that the two-dimensional
view itself does not capture: For example, that between theory theorists
and simulationists, that over the relationship between language and the
theory of mind, and that concerning the signi¬cance of false belief task
performance. This should remind us that although nativism is one issue
in play in work on the theory of mind, it is not the only one.


5 making do with less?
Traditional philosophical analyses, cast in terms of the conditions puta-
tively necessary and suf¬cient for the concept of innateness, have typically
tried to understand nativism in one dimension, even if that dimension
has some internal complexity to it. Several recent accounts of nativism
attempt to make do with just one dimension, and I want to concentrate
in this section on problems that such views face. The two views on which I
shall concentrate, those of Muhammad Ali Khalidi and Richard Samuels,
have been developed as alternatives to existing one-dimensional views,
such as those that view innate structures as those present at birth, or
Nativism on My Mind 61


External Minimalism Thesis

YES NO



Y
Modularists
Gar¬eld, Peterson, and Perry E
Leslie
2001
Baron-Cohen
S Internal

Richness

Thesis
Enculturists
Gopnik Tomasello 1999 N
Perner Carpendale and Lewis
(in press) O



¬gure 3.2. Nativism and the Theory of Mind


<<

. 11
( 62 .)



>>