3. Classification and subclassification techniques

Resemblances in morphological structure and basic vocabulary between specific languages sometimes are so striking that a common ancestry remains the most likely explanation. Once genetic relatedness has been posited as the most plausible hypothesis for the grammatical and lexical affinity observable between a set of languages, one may start determining the exact nature of the relationship. With three or more languages that are assumed to be genetically related, the question arises what the genetic relationship looks like. Of course, the higher the number of related languages involved, the more logical possibilities there are for representing their affinity. Ever since the birth of comparative linguistics some 150 years ago, a number of techniques have been developed to this end. The standard procedure is by way of family trees or arboreal schemata representing the closeness or distance of genetic relationships. Deciding on the most plausible and probable family-tree representation is what subclassification or subgrouping involves. The family tree, as an abstract representation of relationships and common origin, was probably first introduced into linguistics by the Indo-Europeanist Schleicher (1861, 1869, 1873). Similar classificatory techniques had of course been used in evolutionary biology by Darwin (1859), whose methods Schleicher used. When we view a language family as a kind of biological organism, the descendants of an ancestral language, or proto-language, may be referred to as daughter languages; and if we want to further the analogy with kinship terminology, we may call languages belonging to the same subgroup (or subfamily) sister languages. Next to the notion of family, higher level genetic units such as phylum are sometimes used. An even higher level grouping may be referred to as super-phylum or stock (corresponding to taxon in evolutionary biology). The positing of language families and subgroups in principle are based on so-called shared innovations, a technique explained below. Apart from this standard classification technique, lexicostatistics and mass comparison are other techniques that have been used. These are also discussed below.