ABSTRACT
This chapter focuses on the relationship between reanalysis and ambiguity. According to the classical view, reanalysis is made possible by the potentially ambiguous character of surface output. This view has been challenged on the grounds that the ambiguity of linguistic items is usually a result of (rather than a prerequisite for) reanalysis. As I will show, the traditional notion of ambiguity—generally defined as ambiguity of the linguistic code—fails to capture relevant aspects of the topic. Hence, my discussion will also take pragmatic underdetermination into account, which is defined in terms of implicature and inference. In the second part of the chapter, I will argue that reanalysis is rooted in language usage and brought about by the hearer. Moreover, all reanalyses are usually controlled by one of two principles: the Principle of Relevant Usage Effect and the Transparency Principle. On the basis of an overview of examples from lexical, morphological, and syntactic change, I will show that reanalysis can, but need not, be triggered by code ambiguity or pragmatic underdetermination. By the same token, code ambiguity can, but need not, be a consequence of reanalysis. Despite the apparent difficulties in pinning down a systematic one-to-one relationship between ambiguity and reanalysis, I will propose a unified, usage-based account for the various types of reanalyses discussed in this chapter.
