104 THE STRUCTURE OF EVOLUTIONARY THEORY
previously regarded as unrelated and independent, into stages of a single historical
process. Here we cannot observe the changes between configurations directly and
we must therefore work by recognizing them as temporally ordered products of a
single underlying process of change. Call this, if you will, the coral reef principle
to honor Darwin's first book (1842) on a scientific subject. His successful theory
proposes a single historical process for the formation of coral atolls by recognizing
three configurations of reefs—fringing reefs, barrier reefs, and atolls—as
sequential stages in the foundering of oceanic islands.
CONSILIENCE (CONCORDANCE OF SEVERAL). We now reach a break in types of
information. Methods 1 and 2 permit the reconstruction of historical sequences,
either by extrapolating up from the most palpable and testable of daily changes
(method 1), or by ordering a series of configurations as temporal stages (method
2). In many cases, however, we cannot reconstruct sequences, and must infer
history from the configuration of a single object or circumstance. Of the two major
methods for inferring history from single configurations, consilience calls upon a
greater range of evidence. This word, coined by William Whewell in 1840, means
"jumping together." By this term, Whewell referred to proof by coordination of so
many otherwise unrelated consequences under a single causal explanation that no
other organization of data seems conceivable. In a sense, consilience defines the
larger method underlying all Darwin's inference from historical records. In a more
specific context, I use consilience (see Gould, 1986) for Darwin's principal tactic
of bringing so many different points of evidence to bear on a single subject, that
history wins assent as an explanation by overwhelming confirmation and unique
coordination. Call this, if you will, the different flowers principle to honor the
extraordinary range of evidence that Darwin gathered (1877) to forge a historical
explanation for why some taxa bear different forms of flowers on the same plant.
DISCORDANCE (DISSONANCE OF ONE). Here we reach a rock bottom of
minimalism—unfortunately all too common in a world of limited information. We
observe a single object, but not enough relevant items to forge consilience about its
status as the product of history. How can we work from unique objects? How shall
we infer history from a giraffe? Darwin tells us to search for a particular form of
discordance—some imperfection or failure of coordination between an organism
and its current circumstances. If such a quirk, oddity, or imperfection—making no
sense as an optimal and immutable design in a current context—wins explanation
as a holdover or vestige from a past state in different circumstances, then historical
change may be inferred. Call this, if you will, the orchid principle (though I have
also designated it as the panda principle for my own favorite example, perforce
unknown to Darwin, of the panda's false thumb, Gould, 1980d), to honor Darwin's
argument (1862) for orchids as products of history. Their intricate adaptations to
attract insects for fertilization cannot be read as wonders of optimal design,
specially created for current utilities, for they represent contraptions, jury-rigged
from the available parts of ordinary flowers.