Historical reconstruction of languages relies on the comparative method, which itself depends on the notion of the regularity of change. The regularity of sound change is the famous Neogrammarian Hypothesis: "sound change takes place according to laws that admit no exception." The comparative method, however, is not restricted to the consideration of sound change, and neither is the assumption of regularity. Syntactic, morphological, and semantic change are all amenable in varying degrees to comparative reconstruction, and each type of change is constrained in ways that enable the researcher to distinguish between regular and more irregular changes. This volume draws together studies by scholars engaged in historical reconstruction, all focussing on the subject of regularity and irregularity in the comparative method. A wide range of languages is represented, with detailed discussion of data from Australia, Papua New Guinea, Austronesia, North and Central America, East Asia, and Europe.