Modeling contested categorization in linguistic databases
MetadataShow full item record
A fundamental problem in the design of linguistic databases is finding effective ways to encode content which is of a contested nature—that is, content which involves data for which there is no general consensus on how it should be best interpreted. A clear example of such content is the grouping of languages into genealogical units. Numerous proposals abound, but only a relatively limited subset of these proposals are not, in some way, contested. For example, while current wisdom accepts the existence of numerous low-level genealogical units (e.g., Germanic or Algonquian) and a number of high-level units (e.g., Indo-European or Niger-Congo), there is no general consensus on the grouping of the larger families together or, for the most part, on the arrangement of the subgroupings within even relatively small families. There are proposals, of course, some of which have more support than others, but there is no consensus.