The linguistic problem of morphology: structure versus homology and the standardization of morphological data

The present article discusses the need for standardization in morphology in order to increase comparability and communicability of morphological data. We analyse why only morphological descriptions and not character matrices represent morphological data and why morphological terminology must be free of homology assumptions. We discuss why images only support and substantiate data but are not data themselves. By comparing morphological traits and DNA sequence data we reveal fundamental conceptual shortcomings of the former that result from their high average degree of individuality. We argue that the delimitation of morphological units, of datum units, and of evidence units must be distinguished, each of which involves its own specific problems. We conclude that morphology suffers from the linguistic problem of morphology that results from the lack of (i) a commonly accepted standardized morphological terminology, (ii) a commonly accepted standardized and formalized method of description, and (iii) a rationale for the delimitation of morphological traits. Although this is not problematic for standardizing metadata, it hinders standardizing morphological data. We provide the foundation for a solution to the linguistic problem of morphology, which is based on a morphological structure concept. We argue that this structure concept can be represented with knowledge representation languages such as the resource description framework (RDF) and that it can be applied for morphological descriptions. We conclude with a discussion of how online databases can improve morphological data documentation and how a controlled and formalized morphological vocabulary, i.e. a morphological RDF ontology, if it is based on a structure concept, can provide a possible solution to the linguistic problem of morphology. © The Willi Hennig Society 2009.