FUZZIFICATION OF FAILING DISCRETE PALEOTAXONOMY WITHIN LINEAGES

Ever since the dawn of Homo sapiens’ linguistic faculties, we classify our universe in discrete categories. Even when confronted with continuous variability, we are inclined towards compulsive discrete classification; a problem evolutionary biologist Dawkins dubbed “Tyranny of the Discontinuous Mind”. Accordingly, we have been categorizing our fellow biota for thousands of years. For larger organisms this works relatively well because the phenotypic disparity – utilized in the initially morphology-based categorization – is caused by genetic isolation, constituting truly discrete (i.e. non-interbreeding) genotypic bins. This is different in paleontology: if we were to know fossils of all individuals in an evolving lineage, we would generally observe quite gradually changing distributions of morphological variation over time. Yet, our Discontinuous Minds employ discrete taxonomic models, inadequately covering continuous anagenetic change. Incomplete fossil records mask the inadequacy: hiatuses facilitate the definition of well-separated chronospecies – i.e. morphologically distinct fossil remains. However, as more intermediate specimens are uncovered, discrete separation becomes untenable.

As an example we use Pika lineages (Ochotonidae, Mammalia) from the Miocene of Spain; abundantly represented by their fossilized teeth. Variation in teeth morphology encompasses local intraspecific variation, invading subpopulations, as well as anagenetic changes. Our view of the nature of the morphologic variability – both temporal and isochronous – changes as more data become available, thus affecting traditional paleotaxonomy of the group. We here equip the existing discrete labels that a specimen would get (e.g. chronospecies A or B), with fuzzy edges, using the original chronospecies’ morphologies as anchor points within the continuous ranges of morphologies. For each specimen we then determine its distance to the chronospecies archetypes in morphospace, expressed in units of morphological disparity between subsequent chronospecies in a lineage. Thus, we preserve long-established taxonomic labeling, while circumventing the inadequate nature of discrete paleotaxonomy in the inherently continuous morphologic variation within lineages.