Homology In Evolutionary Morphology

The issue of homology testing cannot be divorced from any discussion of adaptation and phylogeny. So these remarks are very relevant here. It was only in the 1980s that many primatologists and other students of fossil mammals increasingly accepted the notion that the determination of levels of related-ness between lineages was not tied to any one kind of evidence, such as teeth or skulls only, but that the whole skeleton (along with other attributes, of course) was at least as important. What matters in phylogenetic estimation is the nature of complexity of properties that are being utilized, as well as the relevance of these to adaptive solutions. The latter assessments aid in the recognition of heritage features, and the particular stage of evolution designated of a character complex (its polarity), not necessarily in that order.

But what renders discourse sometimes nearly impossible, however, is the assumptions (both implicit and explicit) of some workers about homology. Some have stated recently that phylogenetic or Darwinian homology (as opposed to Owen's views) is "logically" flawed. Such remarks overlook the fact that a theoretical definition of homology requires specific hypothetical statements regarding properties in different species, and that these hypotheses are to be operationally and independently tested against specific criteria relevant to the proposition. Much more cannot be asked of any other science (contra Cartmill, 1990).

So, impediments to the practice of testing phylogenetic homology are views that relate to the credo of parsimony cladists, whose assumptions were explicitly espoused by Cartmill (1994). The roots of such a change are difficult to trace in anyone's contributions, but the issue of morphological homology was undoubtedly troubling for Cartmill. In spite of the long and erudite introduction and his selective use of the literature that led up to his changed views, what remains is Cartmill's acceptance of algorithmic analysis as the ultimate arbiter of homology testing. The tone of the bottom line has the customary declarative "truth component" of theorizing by parsimony cladists. "The concept can be made intelligible in an evolutionary context only by giving it a cladistic interpretation that makes homology judgments dependent on the outcome of a phylogenetic analysis. It follows that such judgments cannot play a role in evaluating conflicting phylogenetic hypotheses" (p. 115). Clearly, for Cartmill, they cannot, but they certainly did and continue to do so for the assessment of a large and growing body of phylogenetic hypotheses, even if many feel the necessity for an algorithmic, a posteriori cloak to legitimize their efforts within a Kuhnian community.

Similarly, Lieberman's (1999) generally peculiar stance on the "relative goodness" of homology hypotheses, but particularly Lieberman's (2000: 152)

opinion, misses the theoretical versus operational empirics of homology evaluations. In his deceptively authoritative sounding essay on homology, he overlooks the fundamental requirement for any (Darwinian, hence phylo-genetic) homology hypothesis, namely, its phylogenetic (and level specific) context and a rigorous delineation of either the phenotypic or genotypic condition about which a hypothesis is proposed (gene trees, character transformations, taxograms, and phylogenetic trees express different things). Generally the same may be stated regarding the confusion of levels for the equivocating perspective of Lockwood and Fleagle (1999), who analyzed the meaning of homoplasy. Hypothesis and operational testing are (or rather should be) independent from one another. Lieberman (2000: 152), when he states that he agrees that phylogenetic homology concepts are fine "...but it remains true that the concept is logically problematic in the absence of a priori knowledge of the phylogeny in question" (italics supplied), adds an unwanted level of confusion to the already enormous literature. I note here that Lieberman, like Cartmill, obviously does not believe in the independence of homology testing, and therefore, neither can they logically consider testing phylogenies against independently tested and corroborated homologies. So for both Lieberman- and Cartmill-proposed phylogeny hypotheses of taxa should remain just that, vacuous proposals, as they cannot test these against independently corroborated homologies. Lieberman, or anyone else, who holds forth in detail about homology (or homoplasy) without some experience in the procedures of phylogenetic estimations in systematics, and who vaguely cites EvoDevo studies and equivocates on the level-specific meaning of these concepts to somehow support their taxic perspective has a serious problem. These workers have to grapple with the fact that the key conceptual contribution of evolutionary developmental genetics (that character complexes are modular in spite of the phenomenon of epistasis) obviously means that phylogeny estimation of these modules are likely to be independent from those of others in the same species, and therefore, in higher taxa as well. Mosaic evolution is back under the cloak of modularity (contra the opinion expressed by Tattersall, 2000, that it is a "hoary old concept"), showing us that the logical positivism of cladism is incapable of setting the ontological foundations for the theory of descent. Consequently both the choices of characters for analysis and the taxic approach to phylogenetic estimation may have to be seriously reconsidered in the near future.

0 0

Post a comment