Commentary on: Labels, cognomes, and cyclic computation: an ethological perspective

A commentary on Labels, cognomes and cyclic computation: an ethological perspective by Murphy, E. Murphy (2015) seeks to gain insight into the phylogeny of the human language faculty by adopting a computational perspective. After surveying the ethology literature in an attempt to isolate what is lacking in other species at the computational level, Murphy claims that the process of Labeling, well studied in theoretical linguistics, is what is specific to human language. We applaud Murphy's effort to address the issue of language phylogeny in such a computationally explicit fashion, but we would like to take this opportunity to highlight a few challenges to building adequate cognitive phylogenies that Murphy does not seem to appreciate enough. While Murphy's critical remarks toward the " Chomsky hierarchy " as a tool to build biologically-sound cognitive phylogenies (Fitch, 2015) echoes some of our own concerns (Boeckx, 2013; Benítez-Burraco and Boeckx, 2014), we find his appeal to computational principles like Labeling to be equally inadequate. Quite apart from our skepticism coming from narrowly linguistic considerations surrounding Labeling (Boeckx, 2014), our main reason for not siding with Murphy is that issues of cognitive phylogenies must be firmly grounded in comparative studies. But it is well-known that linguistic principles of the sort Murphy trades in resist meaningful comparison across cognitive domains and species. As Newport (2010: p. 282) has correctly observed, " the generative tradition in language has given us an elegant and detailed articulation of how these principles work themselves out in language; whether the same principles apply in detail to any other domain remains to be seen, since few comparably sophisticated analyses have ever been done of other complex cognitive domains. " The generative tradition took some 50 years to arrive at the sophisticated level of computational characterization that Murphy seeks to exploit. Nothing like it exists in other domains of human cognition, let alone in other species. Accordingly, how are we to determine if, say, the " cognome " of baboons contains the Labeling operation when they conceptualize dominance hierarchies? It is not that questions of this type are meaningless, it is just that they are inapplicable in practice, casting doubt on the current feasibility of what Murphy calls computational ethology. We also find it difficult to believe that operations like Labeling are formulated at the right level of granularity to enable the formulation of linking hypotheses between the cognome and the dynome, connectome, …