ABSTRACT
Two experiments investigated morphological decomposition in ambiguous novel compounds such as BUSHEATER, which can be parsed as either BUS-HEATER or BUSH-EATER. In first experiment, the predictions of the competing hypotheses were tested by simply asking subjects to read ambiguous novel compounds aloud. Second experiment proceeds from an account of word recognition that is predicted by the view that orthographic constraints affect the parsing of a stimulus prior to lexical access. The results of this experiment indicate that both constrained and unconstrained ambiguous novel compounds are classified more slowly than unambiguous ones. The Automatic Parsing and Lexical Excitation (APPLE) model proposed in Libben is a model of the morphological parsing component of the word recognition process. The APPLE model is a serial scanning procedure that isolates the constituent morphemes in a left-right fashion. The parsing module is distinct from the lexical excitation module and it is assumed that the operation of both these modules is independent and obligatory.
