Rethinking Pleijel’s (1995) characters under a hierarchical point of view

Abstract

In 1995, Pleijel raised an issue on the coding of complex (multistate) characters: to find a unique coding approach satisfying both informational and semantical criteria. Following Pleijel′s work, this study aims to propose a new answer to this problematic. We proposed here to use hierarchical characters instead of classical partitional ones as hierarchy allows to deal effectively with the different uncertainties met by phylogeneticists since the beginning of phylogenetic data matrices: missing data, inapplicable data and polymorphism. We translated all previous proposed approaches into hierarchical ones and add three new approaches, only coded using hierarchy. Using phylogenetic and show with different metric how one of these new approaches displaces the other ones. The results from this study then supports the idea than phylogenetic characters should be coded using only hierarchies, as they offer a better management of uncertainties and propose new approaches more informative.

Revealing Evolutionary Patterns Behind Homogeneity: the Case of the Palaeolithic Assemblages from Notarchirico (Southern Italy)

“…The optimal trees resulting from the analysis are those that minimise the number of rejected hypotheses. 3ia addresses all these criticisms in a relevant way: characters are hierarchical (Cao et al, 2007), clades cannot be supported by losses (De Laet & Smets, 1998;Rineau et al, 2015), specific procedures exist for each abovementioned problem (Faure-Brac et al, 2020;Zaragüeta i Bagils & Bourdon, 2007), and maximising congruence (retaining the maximum number of 3is) is equivalent to minimising incongruence (minimising the number of rejected 3is). Furthermore, several examples have been proposed in the literature to show the preferable behaviour of 3ia over parsimony (Nelson, 1996;Nelson & Platnick, 1991).…”