Strong linearity and the typology of templates
MetadataShow full item record
The term template is commonly employed in linguistic description and analysis when salient aspects of the linear arrangement of the subconstituents of some larger constituent appear to be specified independently from syntactic, semantic, or phonological concerns.1 However, there has been essentially no work on the issue as to whether or not the term is used in a coherent way across the diverse grammatical phenomena to which it has been applied.2 For example, are Semitic root-and-pattern morphology (McCarthy 1981) and Athapaskan position-class systems (Kari 1989) – both of which have been labeled “templatic” – on some level, the same basic kind of thing? If they are, then we will need to capture the nature of their relatedness in any system we develop for typologizing templates across different languages. If they are not, then we need to develop separate models for the two phenomena and treat the fact they have been given a common label as superfluous. In Good (2003) I attempted to lay the foundations of a framework for the categorization of different kinds of linear relations, with the goal of being able to use that framework to come to a better understanding of templatic constructions. This paper will highlight those aspects of Good (2003) which I believe to be of most relevance to typological study. As such, the paper will be largely programmatic in nature, though Good (2003) contains not only programmatic elements but also a detailed examination of “templatic” data from three case studies involving diverse morphosyntactic phenomena: Bantu verb suffixes, Chechen verb phrases, and Saramaccan serial verb constructions.