The role of gender-marked information in second language processing

Two main theoretical approaches have been proposed to explain what causes gender errors in L2 processing. Deficit accounts argue that L2 learners who are NS of languages without grammatical gender cannot acquire gender due to impaired underlying syntactic representations. Accessibility accounts argue that morphological errors are due to production problems or even to gender ‘mapping problems’. Recently, the ‘lexical gender learning hypothesis’ has argued that L2 learners exhibit better performances while being taught the gender of nouns when they are provided with the noun suffix together with the determiner rather than when they learn the endings of each gender class seperately from the gender-marked articles. This latter hypothesis is in line with the Competition Model. This thesis investigates the role of gender-marked information in processing Greek as L2. The target population is adult L2 learners of Greek. For the purpose of this study two tasks were elaborated. In Task 1 a non-word (stimulus) was presented to the participants containing three possible types of gender-marked information: (1) morphological (noun suffix), (2) morphological syntactic (determiner + noun suffix) and (3) morphological extensive syntactic (determiner + adjective + noun suffix). The participants’ reaction (in Greek) to the stimulus was to indicate the gender class of the non-word. The presented stimuli in Task 2 were identical to the Task 1. However, a reply was required in English. The corresponding gender value of the stimulus was indicated by the appropriate personal pronoun in English (he, she, it). The findings indicate that L2 learners are capable of processing more accurately and faster the grammatical gender of novel nouns when they are provided with morphological and extensive syntactic gender-marked information than when they are provided with morphological and syntactic information or morphological information only. These findings are in line with the lexical gender learning hypothesis and the principles of the Competition Model demonstrating that the quantity of gender-marked information has a positive effect on L2 processing. Our analysis also reveals that beginners and advanced L2 learners do not differ in terms of the employed processing strategies; they both process and use effectively the maximum amount of available gender-marked information.