An outline of grammatical evolution
Among the questions that we raised there were the following: Which is older—the lexicon or grammar? What was the structure of human language like when it first evolved? How did language change from its genesis to now? Was language evolution abrupt or gradual? It is these questions that will be our concern. After having provided a general introduction to grammaticalization theory, we will now use this theory to reconstruct some major lines in the development of functional categories. The ultimate goal is to trace grammar back to its beginnings in early language.
Introduction
The findings presented, are based on a wide range of data from over 500 languages across the world. In Heine and Kuteva (2002a), more than 400 common pathways of grammaticalization were identified; we will narrow down this range to a smaller number of more general grammatical developments. Most of the processes to be discussed have been documented to some extent in previous works on this issue (e.g. Lehmann 1982; Heine and Reh 1984; Heine, Claudi, and Hünnemeyer 1991; Hopper and Traugott 1993, 2003; Bybee, Perkins, and Pagliuca 1994; Heine and Kuteva 2002a; see also Dahl 2004), and the reader is referred to these and other works cited below for more details. But there are also a few pathways that have not been identified so far, and in such cases we will provide more detailed evidence to substantiate the hypotheses concerned. Since grammaticalization appears to be regular across different sensori-motor modalities, we will not be confined to spoken and written languages but will also include findings made on signed languages (e.g. Sexton 1999; Pfau and Steinbach 2005, 2006).
In our presentation we will begin with categories that previous research has established to be the least grammaticalized, that is, categories that cannot be derived historically from any other categories. Subsequently, we will proceed to reconstructing increasingly more strongly grammaticalized categories. Note that, in accordance with our methodology, we will not be able to analyze grammar as a whole but rather will be restricted to a range of morphosyntactic exponents of grammar.
A review of the grammatical descriptions available on the languages of the world yields a bewildering diversity of grammatical taxonomies, and reducing the taxa figuring in these descriptions to an uncontroversial and crosslinguistically stable set of categories is near to impossible. The categories figuring in our presentation therefore have to be treated with some caution. Selection was determined on one hand by what are particularly common patterns in the languages of the world—irrespective of whether these patterns are defined in terms of syntactic, morphological, or semantic criteria, or any combination of these. On the other hand, we are aiming to select the most inclusive categories available; to this end, we have chosen, for example, a more comprehensive category ‘‘pronoun’’ instead of less inclusive categories such as ‘‘personal pronoun’’, ‘‘indefinite pronoun’’.
