300 THE STRUCTURE OF EVOLUTIONARY THEORY
Talk is cheap, and romantic notions of abstract, overarching unity can easily
be verbalized. Lorenz Oken, the leader of German Naturphilosophie, wrote
wondrous aphorisms (1809-1811, English translation, 1847), and produced solid
empirical work early in his career (1806), but never established a methodological
program or built a factual foundation for his formalist philosophy. Meckel, Carus,
and other Naturphilosophen extended the empirical side, but we rightly honor
Geoffroy as the legitimate focus of this movement by our primary scientific
criterion of fruitful utility. Geoffroy won his fame as a formalist because he
managed to "cash out" the common ideas of transcendentalism in a workable
program of research. His program included the two elements demanded of any
good theory in natural history: a method for identifying the central phenomenon,
and a reasonable explanation for exceptions.
The paradox and pitfall of unity of type as a working research program lies in
the vast range of modifications that the archetype experiences under the widely
varied adaptive regimes of our planet. Elements of the archetype should, in
principle, be named and identified by their form, but the idealized archetype may
be modified into incompatibility and unrecognizability along the copious adaptive
pathways of concrete earthly biology—and we therefore face the dilemma that
archetypal elements cannot always be identified by their shapes, or even by their
discreteness (for elements fuse, or appear in embryos and then drop out during
ontogeny). Some other criterion must be developed.
Geoffroy's major productive insight (still a favored basis for recognizing
anatomical homologies—see Riedl, 1978) lay in his "principle of connections"—
the claim that homology must be identified by the relative positions and spatial
interrelationships of elements, rather than primarily by form. Parts may expand and
contract according to utility, but topology remains unaltered, and the archetype can
be traced by unvarying spatial order.
Yet, as so often happens amidst the exuberant diversity of natural history, the
criterion must be nuanced as "traced by unvarying spatial order ... except when you
can't." Just as Haeckel bolstered recapitulation by bounding and taxonomizing
exceptions (heterochronies and heterotopies in his terminology—see Gould,
1977b), and as Darwin specified forces other than, but clearly subsidiary to, natural
selection (1859, p. 6), Geoffroy recognized a key class of exceptions to the
principle of connections in his concept of metastasis (we use the word in a
different, medical sense today, but the general meaning of movement to anomalous
places has not altered). Connections can break and blocks of elements can move
(though topology within blocks does not alter). For example, the shoulder girdle
attaches to the rear of the head in fishes. But, in tetrapods, this connection breaks
and several vertebrae may be interposed between skull and forelimbs (see p. 320
for the central role of this metastasis in Owen's interpretation of the vertebral
archetype).
In addition, Geoffroy tried to codify rules for secondary adaptive modification
of archetypal form. Why do elements vary so much in size, and why can they fuse
or even disappear? Geoffroy relied primarily upon a loi de balancement, or
principle of compensation. Only so much general material