552 THE STRUCTURE OF EVOLUTIONARY THEORY
should only be invoked when the simpler explanation is clearly inadequate
(1966, pp. 123-124).(Note that Williams here introduces the ultimate causal reduction to genes as units
of selection. He speaks of adaptation at the organismic level—but only as the
consequence of genie selection. Thus Williams's book also becomes the manifesto
for the ultimate—and, I think, erroneous—Darwinian reductionism still popular
today as "selfish gene" thinking in such fields as socio-biology and evolutionary
psychology. See Chapter 8 for a critique.)
In closing, Williams waxes messianic in his pointed comparison of natural
selection with the teachings of Jesus (see John 8:12 and 14:6): "Perhaps today's
theory of natural selection, which is essentially that provided more than 30 years
ago by Fisher, Haldane, and Wright, is somewhat like Dalton's atomic theory. It
may not, in any absolute or permanent sense, represent the truth, but I am
convinced that it is the light and the way" (1966, p. 273).
I love Williams's book; his austere and incisive argument shaped my thinking
and that of all evolutionists in my generation. But Williams's central thesis
includes a disabling problem in logic, one that produced unfortunate effects in
evolutionary practice. Parsimony, or Occam's razor, embodies an important logical
principle when properly applied. William of Occam, a 14th century English
philosopher and divine (a Franciscan), strongly espoused nominalism against the
Platonic concept of ideal types as entities in a realm higher than material existence.
(For nominalists, our designations of general categories only have standing as
names [nomina] based on abstraction from objects in the material world, not as
ideal and "excess" archetypes in a non-material realm.) Occam devised his famous
motto, "non sunt multiplicanda entia praeter necessitatem" (entities are not to be
multiplied beyond necessity), as a weapon in this philosophical battle—an
argument against the existence of an ideal Platonic realm (for nominalists regard
names of categories only as mental abstractions from material objects, and not as
descriptions of higher realities, requiring an additional set of unobserved ideal
entities, or essences).
Occam's razor, in its legitimate application, therefore operates as a logical
principle about the complexity of argument, not as an empirical claim that nature
must be maximally simple. Williams's key invocation of parsimony— to reject
group selection when an explanation based on organismic selection can be devised
for the same results—fails as a general argument, and does not use Occam's razor
in a valid manner, on two grounds:
- Whereas Occam's razor holds that we should not impose complexities upon
nature from non-empirical sources of human argument, the factual phenomena of
nature need not be maximally simple—and the Razor does not address this
completely different issue at all. The Lamarckian one-step route to adaptation, for
example, operates more simply and directly than the Darwinian two-step process
of variation and selection. But nature happens to follow Darwin's path. Similarly,
the simultaneous operation of several hierarchical levels in selection may represent
a more complex system than the idea that selection