102 THE STRUCTURE OF EVOLUTIONARY THEORY
place observation that sufficient variation does exist, and is inherited often enough:
Whatever the cause may be of each slight difference in the offspring from
their parents—and a cause for each must exist—it is the steady accumu-
lation, through natural selection, of such differences, when beneficial to the
individual, that gives rise to all the more important modifications of
structure, by which the innumerable beings on the face of this earth are
enabled to struggle with each other, and the best adapted to survive (p.
170 —see also p. 131 for Darwin's argument that when we ascribe variation
to "chance," we only mean to express our ignorance of causes).Having established a domain of testability by exclusion, Darwin laid out his
methodology for history—never explicitly to be sure, but with such accumulating
force by example that the entire book becomes "one long argument" for the
tractability of his new science. Those of us who practice the sciences of
reconstructing specific events and unravelling temporal sequences have always
fought a battle for appropriate status and respect, no less so today than in Darwin's
time (see Gould, 1986), against those who would view such work as a "lesser"
activity, or not part of science at all. History presents two special problems: (1)
frequent absence of evidence, given imperfections of preservation; and (2)
uniqueness of sequences, unrepeatable in their contingent complexity, and thereby
distancing the data of history from such standard concepts as prediction, and
experimentation.
We may epitomize the dilemma in the following way: many people define
science as the study of causal processes. Past processes are, in principle,
unobservable. We must therefore work by inference from results preserved in the
historical record. We must study modern results produced by processes that can be
directly observed and even manipulated by experiment—and we must then infer
the causes of past results by their "sufficient similarity" (Steno's principle—see
Gould, 1981c) with present results. This procedure requires, as Mill (1881) and
other philosophers recognized long ago, a methodological assumption of temporal
invariance for laws of nature. Historical study manifests its special character by
placing primary emphasis upon comparison and degrees of similarity, rather than
the canonical methods of simplification, manipulation, controlled experiment, and
prediction.
Darwin had done some paleontological work, particularly in his treatises on
barnacles (1851-1858), and his important discoveries of South American fossil
vertebrates (formally named and described by Owen, at Darwin's invitation). But
Darwin was not primarily a paleontologist, and he did not intend to base his
argument for evolution on the evidence of fossils—especially since he viewed the
stratigraphic record, with its vast preponderance of gaps over evidence, as more a
hindrance than an aid to his theory (see chapters 9 and 10 of the Origin). Thus, of
the two major sources for historical reconstruction— direct but imperfect
information from fossils, and indirect but copious data from modern organisms—
Darwin preferred the second as his wellspring of