136 Paul Sheldon Davies
inward and synthesize laterally. This expectation can be expressed as a directive for
inquiry:
(E) As we make progress understanding natural systems—as we analyze inward
and synthesize laterally—expect that the concepts in terms of which we understand
high-level systemic capacities will be altered or eliminated.
The history of human knowledge exhibits a very general pattern. For natural systems
we understand poorly, the discovery of low-level mechanisms implementing high-level
capacities often forces us to revise or discard the concepts in terms of which we understand
the high-level capacities. The evolution of our concept “natural purpose” is illustrative.
We should, in light of (E), calibrate our hunches, our intuitions, and our expectations so
that the potential for conceptual alteration becomes our default orientation toward
inquiry.
8.2.3 Human Psychology
Though lessons from the history of science are invaluable for understanding ourselves, I
also want to focus on what we are learning about our psychological constitution, for some
of these lessons bear directly on the felt importance of the concepts “design” and “purpose.”
Consider, for example, the “theory of mind.” The psychologist Alan Leslie hypothesizes
that humans develop at a very early age a set of capacities that cause us to “see” certain
objects as endowed with mental states. One capacity is a selective-attention mechanism
attuned to objects that exhibit characteristic features of persons or, more generally, of
minded agents. Another is a set of conceptual categories akin to “belief,” “desire,” “inten-
tion,” and more. The basic idea is that the attention mechanism causes us to pay prefer-
ential attention to things that might be agents and also triggers the application of mental
concepts, causing us to “see” those objects as endowed with beliefs, desires, and more
(Leslie 2000). Evidence for this hypothesis draws on work in several areas, including the
study of autism.
Autistic children suffer defi cits, some quite severe, in social intelligence. They appear
far less capable than nonautistic children of seeing persons as mental agents. And
as Leslie describes, autistic children as old as twelve regularly fail the false belief
test—a test designed to detect the ability to attribute to another agent a mental state clearly
distinct from one’s own. In contrast, nonautistic children as young as four years old
regularly pass the test, as do Down syndrome children as young as ten years old. And
Leslie is careful to point out that the defi cits in social intelligence associated with autism
do not appear to be the effects of general cognitive defi cits, as might be caused, for
example, by mental retardation. The fact that 25 percent of autistic children are not
mentally retarded but still suffer impairments in social skills and language is surely
relevant. So too are experiments suggesting that autistic children, though unable to con-
ceptualize persons as minded agents, nevertheless possess the cognitive sophistication