Περίληψη σε άλλη γλώσσα
This dissertation aims at investigating the acquisition of grammatical gender in Modern Greek (MG), as a second language (L2), through three instances of gender manifestation, assignment, predictability and syntactic agreement, in the light of second language acquisition theories which argue for or against access to the Universal Grammar (UG) (Chomsky 1957, 1981) in the L2. More specifically, regarding gender assignment, it was investigated whether the proficiency level affects NNS’ performance and reaction times, and whether the linguistic behavior of advanced NNS is native-like. At the same time, the role of animacy in the performance and reaction times of NNS was also tested.Gender predictability was examined by means of pseudo-nouns which provide morpho-phonological cues (Αναστασιάδη-Συμεωνίδη & Χειλά-Μαρκοπούλου 2003, Ράλλη 2003) but not semantic or syntactic cues. Regarding syntactic gender agreement, it was explored whether NS and NNS are able to recognize the (un)grammaticality ...
This dissertation aims at investigating the acquisition of grammatical gender in Modern Greek (MG), as a second language (L2), through three instances of gender manifestation, assignment, predictability and syntactic agreement, in the light of second language acquisition theories which argue for or against access to the Universal Grammar (UG) (Chomsky 1957, 1981) in the L2. More specifically, regarding gender assignment, it was investigated whether the proficiency level affects NNS’ performance and reaction times, and whether the linguistic behavior of advanced NNS is native-like. At the same time, the role of animacy in the performance and reaction times of NNS was also tested.Gender predictability was examined by means of pseudo-nouns which provide morpho-phonological cues (Αναστασιάδη-Συμεωνίδη & Χειλά-Μαρκοπούλου 2003, Ράλλη 2003) but not semantic or syntactic cues. Regarding syntactic gender agreement, it was explored whether NS and NNS are able to recognize the (un)grammaticality in nominal agreement structures as well as how much time the participants take to perform grammaticality judgments. Finally, in all three facets of gender acquisition, i.e., assignment, predictability and agreement, the grammaticalization of gender in the participants’ L1 was examined as an additional factor affecting accuracy and reaction times.Eighty-eight (88) non-native adult speakers (NNS) of MG (25 male), aged from 18 to 50 years, and twenty-five (25) native speakers (NS), matched to the NNS on age, took part voluntarily in the study. The NNS were recruited from the Modern Greek Teaching Center of the University of Athens and were divided into three subgroups according to their language proficiency level: (i) NNS-A: basic users (Breakthrough/A1 & Waystage/A2), (ii) NNS-B: independent users (Threshold/B1 & Vantage Level/B2), (iii) NNS-C: proficient users (Advanced/C1 & Master of Proficiency Level/C2). The NNS were further divided into two groups, of fifteen (15) individuals each, according to whether gender was grammaticalized in their L1. For this comparison fifteen (15) NS were used as a control group.Three on-line experimental tasks were conducted. The first on-line task aimed at investigating gender assignment through eighty (80) real nouns from basic inflectional classes of MG (masculine: -ος, -ης, feminine: -α, -η, and neuter: -o, -ι, -μα, -ος). The scope of the second on-line task was to yield/predict gender value according to the target-gender through forty (40) pseudonouns with a clear gender suffix (unambiguous suffixes: -is, -a, -o, -ma) and the prediction of gender through twenty (20) pseudonouns with an obscure gender suffix (ambiguous suffixes: -os and -i). The third on-line task was a self-paced reading task with a grammatical judgement integrated, aiming at exploring the acquisition of the syntactic agreement between the noun (gender-controller) and the adjective (gender-target) through three different types of structures (96 sentences in total): (i) agreement within the Nominal Phrase (NP) in structures Article + Adjective + Noun, (ii) agreement within the Verbal Phrase (VP) in predicative structures with a (subordinate) relative clause inserted, (iii) agreement within the VP in predicative structures. All three types of structures contained an equal number of grammatical and non-grammatical structures.The results indicate that the gender is acquired relatively early on from the initial language proficiency levels; some difficulties are observed, however, with certain inflectional suffixes (especially for neutral nouns ending in -ος). Animacy, and hence semantic cues, does not seem to play a significant role in gender performance for the first two proficiency levels, while, at the most advanced level, it seems to have a positive role in yielding the gender of [+animate] nouns. Additionally, the morpho-phonological cue of the suffix has a prominent role in the predictability of gender specifically to unambiguous suffixes (-is, -a, -o, -ma) for which it acts as a compensatory strategy in the absence of semantic information. For the ambiguous suffixes (-os, -i), a further investigation of the phonological variable (the syllable number and the stress position) is required, although the suffix -os seems to only trigger male gender values for the NNS. Therefore, regarding the assignment and predictability of gender, the (full access to UG) Missing Surface Inflection Hypothesis (Lardiere 1998, 2000, Haznedar & Scwartz 1997, Prévost & White 2000) seems to be confirmed, which predicts a native-like behavior for NNS-C and particular difficulties with complex morphological structures. The syntactic agreement of the gender (Adjective + Noun) is still to some extent problematic even for the NNS-C. This finding verifies the (non-access in UG) Fundamental Difference Hypothesis (Bley-Vroman 1989, 1990) and the Shallow Structure Hypothesis (Clahsen & Felser 2006), which anticipate difficulties for the NNS due to their limited, less detailed and shallow syntactic representations during the sentence processing. Furthermore, the distance between the two agreeing elements, i.e., the adjacency of the elements involved, does not seem to affect the performance and the reaction times of NNS. However, developmental differences were revealed even in the syntactic gender agreement task. Additionally, the grammaticality of a sentence seems to facilitate the grammaticality judgment of the participants in particular of NNS-B, but not those of the NNS-C and NS. Finally, the grammaticalization of gender in the participants’ L1 does not seem to affect gender acquisition via its three forms, namely assignment, predictability and agreement.
περισσότερα
