| The first stage of language production forms the bridge between non-linguistic thought and linguistic expression and is essential in understanding how and what concepts can be linguistically output. Despite its importance, it remains poorly understood. This project aims at investigating the product of the initial stage of language production the preverbal message (PVM). The PVM, although not yet linguistic in nature, should contain all the information necessary to guide subsequent grammatical encoding. Due to the difficulties of manipulating a speaker?s conceptual contents in order to get experimental evidence and of detecting the equivalent of slips-of-the-tongue at this level, there is a scarcity of data related to this stage of language production. There are, however, speakers for which ´conceptual slips-of-the-tongue´ are both predictable and detectable: second language (L2) speakers. I propose to elicit experimentally in proficient L2 speakers (by means of question and answer and recounting tests) specific errors´ errors that can be hypothesized to arise at a conceptual level. Since the preparation of the PVM depends in part on the specific grammatical requirements of a given language, some cross-linguistic differences between an L2 speaker´s two languages will sometimes give rise to conceptual errors which will result in sentences which, although otherwise grammatically correct, do not correspond with the intended concepts. These errors reflect a fault in the construction of the PVM derived from preparing it with respect to the linguistic requirements of one language while it should have been prepared with respect to the linguistic requirements of the other (target) language. In the same way monolingual speech errors opened a window into grammatical and phonological encoding, these predictable conceptual anomalies will open a window into the structure and composition of the PVM. |
