Structural Constraints, Spandrels, and Exaptation 1247
remain adaptive throughout: they originate for one function (presumably by natural
selection), and then undergo quirky shift to a different utility.
However, the principle of functional shift, combined with Nietzsche's argument
about the invalidity of inferring historical origin from current utility, implies a
disarmingly simple and logical extension that does challenge the rule of Darwinian
mechanics and functionalist control over evolutionary change. Ironically, the very
simplicity of the argument has often led to its dismissal as too obvious to hold any
theoretical importance—a "feeling" that I shall try to refute in this section, and whose
disproof represents an important step in the central logic of this book.
The deeper challenge posed to orthodox Darwinism by the principle of
functional shifting flows from the implication that, if current utility does not reveal
reasons for historical origin, then these initial reasons need not be adaptational or
functional at all—for features with current adaptive status may have originated for
nonadaptive reasons in an ancestral form. In other words, and in the terminology of
Table 11-1, when current aptations rank as exaptations rather than adaptations, their
coopted source will be identifiable as an ancestral structure with either adaptive
origins (for a different function) or nonadaptive origins (for no function at all). (I do
accept the standard view that strongly wadaptive features hold little prospect for an
evolutionary legacy because natural selection must soon eliminate them. But
raoraadaptive— that is, effectively or nearly neutral—features may persist for several
reasons, including the "invisibility" of true neutrality to pressures of selection, and the
status of many nonaptations as automatic architectural byproducts, as in Darwin's
"correlations of growth" or Gould and Lewontin's "spandrels.")
The logical validity and evident application of this simple argument cannot be
gainsaid. Indeed, several examples, mentioned inter alia in the preceding section on
exaptation, fall into this category of features with current utility exapted from a
nonadaptive ancestral status. The optical property of transparency, shared by all the
diverse proteins and enzymes that have been exapted as lens crystallins, may
represent a trivial and automatic consequence of physical and chemical structures
evolved for other reasons. But this purely derivative and nonadaptive feature still
stands as a gatekeeper and prerequisite for exaptation to vision. We may regard
Darwin's example of non-fusion of skull sutures in mammalian neonates as a far
richer and less obvious case. For we need, in this example, to unravel enough specific
history of mammalian descent to know that this property arose in ancestors born from
eggs, and therefore cannot be a direct adaptation, initially evolved to compress the
head and permit passage through the narrow mammalian birth canal. (We should also
remember that Darwin explicitly declined to call non-fusion an "adaptation" for this
reason, even while he acknowledged the functional necessity for such a property in
the evolution of mammalian live birth.)
The general conclusion may be stated in a simple manner, but I believe that the
resulting implications for evolutionary theory are both profound and curiously
underappreciated: If many features that operate as adaptations under