The difficulty with
concentrating solely on neural
mechanisms is not simply that
we still know so little about the
brain. Ultimately, we may
discover the neural
mechanisms underlying a
mental or behavioral fact, but
this understanding would be
incomplete without
intervening levels of
explanation.

    The need for several 
levels of explanation arises in
all science. Darwin's
principles, for example, can be
said to explain the evolution of
species, but they tell us
nothing about the underlying
mechanism of evolution.
Suppose we discover how to
describe the mutation of an
organism in biochemical terms.
Would this description have
any meaning to us without
Darwin's principles? The
evolutionary effect of the
mutation only makes sense
within the context of the
organism's adaptation to an
environment and its
competitive chances for
survival. Thus, while reduction
to more fundamental levels of
analysis is desirable, higher
levels of analysis retain their
usefulness and are often
required first.