If true that development is a "process of construction and reconstruction in which heterogeneous resources are contingently but more or less reliably reassembled for each life cycle'' (Oyama et al. 2001 p 1), a situation might arise in which a developmental biologist would want to invoke a concept of partial homology. The concept of modular development (Raff 1996; West-Eberhard 2003; for a review and references see Rieppel 2004c) builds on connectivity yet dissociability of developmental modules. If developmental modules are dissociable, they can potentially "be reorganized somewhat independently of each other'' (WestEberhard 2003 p 60), thus providing the resources for reassembly of traits derived from heterogenous resources in descendant life cycles. Such "mosaic development'' results in heterochrony, heterotopy, and other effects (Raff 1996) that led West-Eberhard (2003 p 60) to speak of homology as "not an all-or-nothing phenomenon but rather a continuum of more or less similarity due to common ancestry.'' This conclusion, startling from a systematist's point of view, would seem to be supported by Raff's (1996) review of appendage development in Drosophila and tetrapods. Why not speak of partial homology of appendages in those widely separated lineages in which "appendages have a deep underlying and (prior to the findings of developmental genetics) completely unexpected homology'' (Raff 1996 p 353; see also Shubin et al. 1997; Gilbert and Bolker 2001)? Very basal building blocks are used to develop vastly different structures. But is it true that through such mosaic development, "evolution makes a mess of homology'' (West-Eberhard 2003 p 497)? Such a conclusion seems subject to a naturalistic fallacy. Homology is a theoretical concept. It may, indeed, be the case that our theoretical concepts are somewhat messy, but this is not because ofevolution. The evolutionary process just unfolds the way it does, and this may be a way that is difficult to capture conceptually.
The call for partial homology can have two sources: the heterogeneous combination of ancestral resources in descendant life cycles as outlined previously, or gradual and continuous morphological change as in Sattler's (1990, 1996, and references therein) "continuum morphology,'' which essentially reduces structure(s) to process(es). Although at odds with cladistics, Weston (2000 p 141) considered it an avenue toward holism if partial homology was fitted as a distance measure on a tree. Why this appeal to holism?
The corollary of "continuum morphology'' is that it "acknowledges gradations between typical structures From this point of view, homology is a matter of degree'' (Sattler 1996 p 577). The tension identified by Sattler (1990, 1996) between "typical structures'' and the "continuity of developmental processes'' broadly corresponds to the tension between "evolution and stability of structure'' that was used by West-Eberhard (2003 p 488) in support of her claim for complementarity. Indeed, the contrast between pattern and process can quite generally be conceptualized in terms of complementarity (Rieppel 1988), although this may in turn result in some tension with scientific realism that adheres to ontological monism. Ontological monism holds that there is one, and only one, way to carve nature at its joints, i.e., that particles and waves are not two different kinds of light but rather a reflection of our imperfect understanding of light. Accordingly, McMullin (2002 p 252) argued that complementarity invoked in physics does not also invoke an ontological dualism, but rather points to the issue of partial reference of the terms deployed in complementary theories. The same seems true for biology.
The discussion of partial homology in terms of morphological stasis versus continuity of developmental change mirrors the paradoxical discussion of "species'' as spatio-temporally delimited chunks of a genealogical continuum. With respect to species, Hey (2001 p 168) found that "systematists and evolutionary biologists have both placed incommensurable demands on a word," in which the word "species" is supposed to function in the context of pattern as well as process analysis (Hull 1997), and the same is true of homology. But again: "a tendency for different perspectives by different biologists [must not] be confused with different kinds of reality'' (Hey 2001 p 150). Indeed, if species transformation is a gradual process, any species concept will be subject to the "Sorites Paradox'' (Rieppel 2003b): take a heap of sand and remove grains one by one—when would one stop calling it a "heap"? Under the paradigm of gradualism, species boundaries will be unsharp, fuzzy, no matter which concept is deployed in their delineation (Hey 2001). The same is true of homology in cases in which morphological transformation (developmental and phylogenetic) is a gradual process. There is, however, a conceptualization of species that specifically allows for such fuzziness, which is that of homeostatic property cluster natural kinds (Boyd 1999; see also Griffiths 1999; Wilson 1999; Keller et al. 2003). The same concept can be applied to homologs, as was discussed previously (Rieppel 2004c). The conclusion is not that evolution makes a mess of homology, but that it is the names associated with the natural kind terms (predicates) we deploy in statements of homology that may have imprecise, or even only partial, reference (the terms apply only imprecisely, or partially, to the organisms under analysis).
A second issue to consider in relation to the concept of partial homology is the fact that natural kinds, such as taxa, or homologs, come in hierarchies and thus raise the qwa-problem for natural kind terms as discussed previously (Devitt and Sterelny 1999). The statement "bird wings and bat wings are homologous'' makes little sense unless it is tied to a conditional phrase. The same is true of the statement "arthropod and tetrapod appendages are partially homologous,'' unless the question is raised in which sense they are homologous, partially or not. The concept of partial homology may reflect a tendency of developmental biologists to use broad-sense homology independent of a hierarchical conception of development. If the sense of partial homology in this example relates to certain gene expression patterns (a "similar genetic cascade'') (Shubin et al. 1997 p 646), then the homology relation exists not at the level of appendages, but at the level of genes, which ties the homology relation to its proper hierarchical level. Gilbert and Bolker (2001 p 447) speak of such gene-level homology as "deep homology,'' for insect "legs'' and vertebrate "legs'' obviously cannot be homologous as appendages, because "phylogenetically intermediate taxa (particularly basal chor-dates) do not possess comparable structures'' (Shubin et al. 1997 p 645). However, Gilbert and Bolker (2001 p 449) added a note of caution to this tale of wings and legs. Applying the "small toolbox argument'' to developmental processes in the name of parsimony, they draw attention to the fact that genetic resources that regulate basal (early) patterning processes in development are limited, for which reason convergent gene expression patterns are difficult to rule out. For example, much of the "toolbox'' that is involved in early limb development is also involved in the early development of vertebrate teeth.
Was this article helpful?