Lightfoot (1980) claims that the mechanism for syntactic change, reinterpretation, lacks constraints. An in depth study of particle reinterpretation in Ancient Greek will be presented to demonstrate semantic constraints on the process. Particle reinterpretation involves several steps, whereby adverbial particles are associated with Kase-phrases (KP) and are eventually reinterpreted as the case-assigning head of the phrase. The constraints on the process are specific and easily identifiable, but they should not be attributed to the "the very broad constraints of UG", as Lightfoot claimed. Familiarity with Lehmann and Venneman's (1974) theory of word-order change and Baker's (1988) theory of incorporation is helpful for understanding the issue, but not absolutely necessary to follow the argument as presented here.