The word is not enough: on morphemes, characters and ontological concepts

Abstract

The field of morphology has recently seen the arrival of computer-aided ontologies, tools which permit the semantic organization of defined concepts and which therefore promise to be extremely useful in computer-mediated approaches involving morphological data, for example in cladistics. The theoretical relationship between ontologies and cladistics, however, has hardly been explored. Here we examine the ontological status of the main terms in morphological cladistics, i.e. morpheme, character, character state and ontological concept. Morphemes are units of the descriptional perspective, whereas character states are units of the evolutionary perspective and refer to identical stages of transformation. Both morphemes and character states represent things, i.e. real entities. However, character state and morpheme denote different perspectives on these entities (description vs. evolution). Characters (transformation series; Hennig's ideographic character concept) and ontological concepts are both classes, but not of the same nature. Ontological concepts which are used to classify morphemes are constructs, i.e. totally man-made classes that serve only as a way of classifying real entities for human recognition. Characters, however, are classes that encompass all character states of common descent and which therefore have objective, human-independent properties. Characters, then, are natural kinds, classes which exhibit a natural identity in the same way as monophyla do, for example, which also have common descent as an objective property. Against this background, ontology-based morphology can be a valuable addition to phylogenetic systematics. Formalized, machine-parsable descriptions permit the generalization of morphemes in large datasets and can facilitate the recognition of identical character states. However, the expertise of the researcher is indispensable, and full automation of the transfer from the descriptive level to the evolutionary level thus appears impossible.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 20 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.