Portal de Periódicos da CAPES

Confusing homologs as homologies: a reply to “On homology”

2012; Wiley; Volume: 28; Issue: 3 Linguagem: Inglês

10.1111/j.1096-0031.2011.00387.x

1096-0031

David M. Williams, Malte C. Ebach,

Biomedical Text Mining and Ontologies

Sir, “To believe that the present day novel [scientific paper] will be read in a hundred years is not to praise the novel [scientific paper] but to condemn the world”, Greil Marcus, Lipstick Traces, (2009, p. 230). Like the species concept, it seems that systematists/taxonomists/phylogeneticists/historians and philosophers of biology, perhaps all biologists and palaeontologists, never tire of discussing, expanding and redefining the homology concept (see Remane, 1952; Patterson, 1982; De Pinna, 1991; Hoßfeld and Olsson, 2005). Prior to the general acceptance of cladistics as a theory of taxonomy, and after the Darwinian paradigm in biology, homology has been discussed mainly in terms of its evolutionary implications rather than in its original context as “the relations of the parts of animal bodies” (Owen, 1849, p. 2). Seen as a product of an emergent property of some process or processes, homology gained an explanation generally acceptable to most biologists: homology is caused by common ancestry. After the cladistic paradigm of the 1970s and 1980s, discussions of homology mainly focused on its discovery, its meaning having achieved a level of acceptance. Trouble started with the term cladistic paradigm, as it was exclusively associated with the Wagner parsimony algorithm, itself rooted in the numerical taxonomy of the 1960s and 1970s: “At present, Cladistics has reached the status of standard method for systematic research, and can be considered as the current paradigm in systematics”(De Pinna, 1991, p. 369). That “method” is simply an implementation of the Wagner parsimony algorithm, therefore it would be much easier to say “Numerical Cladistics = Wagner Parsimony” and leave it at that. We feel, however, that this is exactly what is wrong with numerical cladistics. Promoting a particular kind of implementation (Wagner Parsimony), Nixon and Carpenter avoid summarising much of the early cladistic literature (and misinterpret much pre-cladistic literature) on the topic of homology and evidence. The key to their views is found in the definition of synapomorphy, which “is one kind of homology” (Nixon and Carpenter, in press, p. 000). Synapomorphy is better thought of as one or more homologs—a homolog is a part. Examination of their term “operational homology” illustrates yet another reference to homolog, rather than homology, of which even a cursory look at Owen’s definition, via “Darwin11 We refer to Darwin in quotes because the definition is not Darwin’s. The text cited by Nixon and Carpenter comes from the glossary of the 6th edition of On the Origin of Species (in 1872), which they note, not the first (in 1859). The glossary was written by W.S. Dallas. Note too that the 6th edition was published after Lankester’s (1870a,b) papers on homoplasy, which Darwin discusses in the text of that edition. We wrote: “This definition [of homology] is in complete contrast to the passages Darwin rewrote in the Origin to accommodate Lankester’s homogeny—causing Gavin De Beer, some years later, to exclaim, that “... is just what homology is not” (de Beer, 1971; p. 14)” (Williams and Ebach, 2008, p. 172). ” (Nixon and Carpenter, in press, p. 000), would have revealed: “The parts which stand in such a relation to each other are said to be homologous, and one such part or organ is called the homologue of the other”. This confusion—homolog as homology and vice versa—occurs in their text too many times to list. Nixon and Carpenter compound their confusion by contrasting yet more pairs of terms: “The use of the terminology “primary homology” and “secondary homology” is unfortunate, and we instead use “hypothesis of homology” and “homology” in their places” (Nixon and Carpenter, in press, p. 000). If “hypothesis of homology” (and primary homology, for that matter) is understood as we feel it should be, then it is better expressed as “hypothesis of homologs”, and placing homologs (the parts) on a tree becomes merely an exercise in distributing “observations” economically on any particular tree. In a similar fashion to the pheneticists, unit character becomes “hypothesis of homology” and they are arranged using a kind of overall similarity. That process excludes any kind of notion of relationship, even phylogenetic relationship, a cornerstone to Hennig’s thinking. Viewed this way, Owen’s homologs, and their discovery, are the very stuff of systematics and taxonomy, as is evident in the very many pages of taxon descriptions in Zootaxa and Phytotaxa, for example. Related to this, and equally bizarre, is their view that homology (“shared similarity due to common ancestry”, Nixon and Carpenter, in press, p. 000) now includes both symplesiomorphy and plesiomorphy. If so, then homology (as similarity) includes all known homologs, again, an example related in spirit, if not in fact, to Sneath and Sokal's (1973) overall similarity. A further interesting exercise might be to try reading Nixon and Carpenter’s following comment from the point of view of homolog discovery: “A concept of homology that seeks to eliminate common descent and shared ancestry as the basis for formulating character states and scoring taxa have been promoted by those self-identifying as pattern cladists” (Nixon and Carpenter, in press, p. 000). Never mind the latter remark about self-identification (mere poser politics), we simply request the reader to ponder what kind of homology concept (theirs included) has ever included “common descent and shared ancestry as the basis for formulating character states and scoring taxa”? Even Lankester saw “common descent and shared ancestry” as interpretation, not part of a method of discovery. Nixon and Carpenter unwittingly muddy the distinction between numerical cladistics and numerical taxonomy. If numerical cladistics seeks to find a measure of overall similarity, how does this differ theoretically and historically from numerical taxonomy? After all, it is the foundations of cladistics that are at stake, not its methods and implementations. Nixon and Carpenter cannot theoretically justify their claims based on their knowledge base (which, apparently is limited to two sources: Willi Hennig and James S. Farris). What we see here is a form of systematic malpractice—distorting history to favour a particular method or implementation with disregard to resolving the confusion between homolog and homology. As Julian Barnes (2011) has noted “History isn’t the lies of the victors. It’s more the memories of survivors, most of whom are neither victorious nor defeated”. Nor are they reliable.