Portal de Periódicos da CAPES

Symplesiomorphies and explanation

2012; Wiley; Volume: 29; Issue: 1 Linguagem: Inglês

10.1111/j.1096-0031.2012.00434.x

1096-0031

James S. Farris,

Coleoptera Taxonomy and Distribution

Sir, Platnick (2012, p. 2) declared himself unrepentant: “I remain an unrepentant Hennigian, rather than post-Hennigian, systematist, who groups by putative synapomorphy, who does not regard shared plesiomorphic absences as homologs, or as evidence that requires explanation, and who therefore views unordered multistate characters as, at best, impoverished sources of relevant phylogenetic evidence.” He was right about being unrepentant, but wrong about the rest. There is nothing Hennigian about Platnick’s supposition that symplesiomorphies are not to be explained, and indeed that supposition would lead to unsatisfactory conclusions, as will shortly be seen. Platnick’s aim was to find fault with Nixon and Carpenter’s (2012) treatment of homology, and while he did not mention it, they employed the usual phylogenetic concept of explanation, that applied in parsimony. As is well known, parsimony selects trees to minimize extra origins of features, and this maximizes explained similarities, as each additional origin of a feature reduces a tree’s ability to explain similarities in that feature by inheritance and common ancestry (see Farris, 1983, 2008). Suppose, as in Platnick’s first example, that the data comprise spiders with spinnerets and non-spiders without spinnerets (the plesiomorphic condition). Then maximizing explained similarities would lead to uniting the spiders into a monophyletic group, as this would require just one origin of spinnerets (in the stem species of spiders) and one origin of their lack (at or before the root of the tree). Support for spider monophyly would certainly be the Hennigian interpretation of spinnerets, and it is now seen that this reasonable conclusion is reached by explaining symplesiomorphies as well as synapomorphies. In contrast, if (as Platnick supposed) symplesiomorphies should not be explained, extra origins of lack-of-spinnerets would not need to be avoided, in which case the same data would no longer support monophyly of spiders, as now non-spiders could just as well be derived from spiders by secondary loss of spinnerets. Platnick’s supposition, then, has nothing to do with how phylogenetic explanation actually functions, let alone with Hennig. One would think that Platnick was simply unaware of the relationship between explanation and minimizing origins of features, but such is hardly the case (Platnick, 1993, p. 271): “Farris (1983) showed that those cladograms which imply fewer steps are to be preferred on grounds of their ability to summarize, describe, and explain the data.” Platnick (2012) never mentioned that, however, apparently having decided to ignore that result in order to adopt his present view. Ignoring relevant information also seems to have underlain Platnick et al.’s (1996, p. 249) similar position:1 “The notion that plesiomorphies require explanation is simply not a cladistic idea ... the fact that both lice and ticks lack spinnerets is no more in need of explanation than the fact that they lack mammary glands, or the fact that they lack Cadillac engines in their stomachs.” Farris et al. (1995) had criticized Nelson and Platnick’s (1991) three-taxon analysis (3ta),2 and Platnick et al. wanted to defend 3ta. They chose exactly the wrong argument for that purpose, however, as 3ta actually treats symplesiomorphies (as well as synapomorphies) as requiring explanation (Farris et al., 1995, p. 215, emphasis added): “That both apomorphies and plesiomorphies must be accounted for at once is seen in [Platnick’s (1993) fig. 1] Tree 4, which (as Platnick [(1993, p. 270)] pointed out) discards all 135 statements. For r(ab), for example, Tree 4 is consistent with a single origin for either state, but not with single origins for both.” Platnick et al. can hardly have been unaware of that, yet they never mentioned it. Farris (1997), moreover, called attention to the same passage again, but Platnick has never addressed it. Having decided, apparently as an article of faith, that symplesiomorphies should not be explained, Platnick seems to have become unwilling to discuss any subject that would undermine his belief. A similar disdain for reality seemed evident when Platnick (2012, p. 2) took his ideas on symplesiomorphy to imply a preference for additive characters (emphasis added): “[An unrepentant Platnick] does not regard shared plesiomorphic absences as homologs, or as evidence that requires explanation, and ... therefore views unordered [non-additive] multistate characters as, at best, impoverished sources of relevant phylogenetic evidence.” Even if it were not specious in itself, Platnick’s view of symplesiomorphy could hardly provide a reason for avoiding non-additive characters—a category that includes nucleotides. As additive characters also have symplesiomorphies, there is no logical connection between explaining symplesiomorphies (or not) and non-additive characters in particular—nor did Platnick attempt to describe one. Further, while Platnick wrote as if his second example (involving tarsal organs) somehow addressed non-additivity, it did not actually do so, but included only results from an additive coding of his data. Those arguments, then, can only have been pretexts for Platnick’s hostility to non-additive characters. Probably his real motive was simply that 3ta cannot analyse such data. As I have noted before (Farris, 2011), advocates of that ill-considered method generally resort to contrived excuses for disregarding non-additive characters. Finally, not only did Platnick want to leave symplesiomorphies unexplained, he also opposed recognizing them as homologies. Despite the plesiomorphic absence in his spinneret example, this last complaint did not concern only absences—he was also critical of Nixon and Carpenter’s (2012, p. 225) observation that “symplesiomorphous parts are also homologs”. Thus Platnick’s view was like Nelson’s (1994, p. 127):3 “Characters seen as part of the same transformation series were claimed by early cladists to be homologous, and it is now evident that the claim is defective, for it construes the transformation series as “fins”—“arms”—wings, that is as including symplesiomorphy.” As Nelson avoided doing so, it is informative to identify the author of what he called the “defective claim”. It was Hennig (1966, p. 93) who stated: “Different characters that are to be regarded as transformation stages of the same original character are generally called homologous. “Transformation” naturally refers to real historical processes of evolution.” Platnick’s purportedly “Hennigian” conception of homology, then, was actually based on rejecting Hennig’s. As Platnick said, he was unrepentant. Platnick thanked Andy Brower, Mario de Pinna, Gary Nelson, and Toby Schuh for helpful discussions of his manuscript.