What counts as evidence in historical linguistics?

The main aim of this paper is to establish the position of historical linguistics in the wider field of linguistics. Section 1 centres on the immediate and long term goals of historical linguistics. Section 2 discusses the type of data that play a role and looks at tools to be used for the analysis of the data. It also addresses the question whether the explanation of the data should be in terms of grammar change (as advocated by formalist linguists) or language change. This latter point automatically leads to the question as to what type of grammatical model or theory the historical linguist should work with, and more particularly, in how far the innate, syntacto-centric generative model is adequate for studying grammar change (Section 3). This is followed by a brief conclusion in which a semi-independent position for the historical linguist is advocated.