For the establishment of polarity of character transformation prior to phylogenetic analysis, various logical and biological criteria are discussed; some are rejected on grounds of liability to systematic error, circularity or unwarranted assumptions about evoluation. Parallel evoluation is used as a non-polar term, with forward and reverse to indicate polarity. A computer program for the detection of parallel evoluation is described which takes taxa in groups of four. The characters, with two derived: two primitive states or three derived: one primitive state, are listed according to the distribution of states over the four taxa. To each of the 15 phylogenies there corresponds a compatible pair of character patterns. Discordant 2: 2 patterns are unconditionally incompatible (Le Quesne test failure), discordant 3: 1 patterns are incompatible conditional upon correct scoring of polarity. For any putative phylogeny the concordant and discordant characters are identified. In cases of competing alternatives these character sets have to be weighed against one another. Character weighting is discussed; it is argued that it is the individual character transformation which should be weighted, in each direction separately.
No takes yet. Share an insight, caveat, or question.
Garth Underwood (1982) studied this question.
Synapse has enriched 5 related papers. Consider them for comparative context: