Germanic: an innovation cluster

In the conventional model of language relationships, as in the family trees in the previous chapter, the growth of linguistic diversity, the origin of ‘new’ languages, is imaged as a process of branching. The tree has one ultimate ‘mother’ node, and the rest of the languages in the family arise by successive splits. Genealogical trees of this kind are familiar from biology and other fields; in linguistic history however the parent-offspring relations are most often parthenogenetic (only one parent per child!). Multiple parentage (except in the case of pidgins and Creoles) is supposed to be relatively rare; though some ‘normal’ languages, like Dutch, may well be examples of something of the sort. There are problems in an oversimple interpretation of genealogical trees for languages (e.g. straight-line developments may be interrupted by diffusion of features from one dialect to another, etc.); but the metaphor is useful, is usually reasonably consistent with the facts, both linguistic and historical, and for most families is a useful organizing device.

Linguistically, ‘branching’ can be defined more or less as it is in biology: we propose a split in a lineage when one subgroup becomes different enough to merit being assigned to a new class. In other words, branchings are dialect splits; they represent the emergence of one or more structural innovations that are striking enough to make us give a new name to the innovating group.