ABSTRACT
Languages have been assigned to distinct types on the basis of two main sets of criteria: morphological and syntactic. The classification of languages according to their morphological structure has a long history, extending back to the early nineteenth century. Typology based on word order (or ‘the order of meaningful elements’, in Greenberg’s phrase (1963)) is much more recent.1 But both kinds of classification, if they provide criteria for identifying the characteristics of a language which are inconsistent with its dominant type, can generate expectations about morphological change. Whether these changes take place or not therefore contributes to deciding whether the typology in question is sensible or useful. Phonology is relevant here too, since a language’s morphology can be affected by phonological change in far-reaching ways. A standard example of this is the neutralisation and loss of many unstressed vowels in Early Middle English, which is traditionally seen as a major factor in destroying the English case system and radically simplifying English verb morphology. Such changes can in turn impinge on syntax, as when loss of case leads to more rigid word order. These syntactic changes can then generate new expectations about morphology, and so on. Theories of linguistic types ought, therefore, in principle to have plenty to contribute to at least some kinds of theory about morphology, and vice versa, with language change as the main testing ground for the mutually relevant predictions.
