296 THE STRUCTURE OF EVOLUTIONARY THEORY
of the old bipartite division of vertebrate and invertebrate (and the equation of
vertebrate classes with invertebrate phyla), for a system of four equal
embranchements based on necessarily separate and untransformable anatomical
plans: Radiata, Articulata, Mollusca, and Vertebrata. This appeal to limited and
untransformable anatomical designs as a basis for taxonomic order smacks of
structuralism, but Cuvier, true to his guiding philosophy, presents a purely
functional interpretation. Appel (1987, p. 45) explains: "The unity within an
embranchement came not from a comprehensive unity of plan, but from a common
arrangement of the nervous system, functionally the most important system of the
animal. The forms of the other major systems remain constant throughout an
embranchement because the other systems—respiration, circulation, etc.—were
functionally subordinate to the nervous system and determined by the requirement
of the nervous system. Animals within an embranchement could vary almost
arbitrarily in their accessory parts, precisely because accessory parts were not
necessitated by the choice of the nervous system." Both unity and diversity
therefore achieve a functional interpretation—unity by operational design,
diversity by local adaptation. Conditions of existence set both major aspects of
taxonomy.
I emphasize a primary intellectual correlation throughout this chapter—
formalism with commitment to internal constraint (in the positive sense of
channeling change, not only the negative definition of restriction). To render this
connection meaningful, the converse must also hold: functionalism must correlate
with denial of constraint. Cuvier's arguments test and affirm this implication.
In an overly broad (and therefore operationally meaningless) construction of
constraint, all biologists acknowledge some restriction on organic form, if only
because all conceivable shapes and sizes have not been realized. But we usually do
not apply this term to nature's avoidance of obviously unworkable creatures (flying
elephants or large dinosaurs with pencil-thin legs in Galileo's world of laws
regulating the ratio of surface to volume), for no one disputes the underlying
physical basis for their nonexistence. (For historically contingent reasons of
modern professional life within a Darwinian functionalist paradigm, we currently
apply the term "constraint" primarily to internal channels and limitations not set by
adaptation—see my full argument for this usage in Chapter 10, pp. 1027-1037.
That is, we apply the concept of "constraint" to sources of influence outside a
favored explanation—see Gould, 1989a.)
Thus, Cuvier cheerfully acknowledged limits set by function, but did not view
such boundaries as constraining because aborted, unworkable creatures offend the
very notion of a rational creating force. Instead, and thereby affirming the link of
functionalism to a denial of constraint, Cuvier clearly cherished his general theory
as a principle for maximizing God's liberty to create (translated as "adaptation to
alter" in the modern evolutionary version of functionalism). Cuvier wrote in an
1825 essay on "Nature": "If we look back to the Author of all things, what other
law could actuate Him but the necessity of providing to each being whose
existence is to be continued the means