| The work reported in this study is based on empirical data collected from three groups of speakers: native speakers of English, native speakers of Persian and Persian speakers of English as a second language.; The study addresses, among others, two of the most important aspects of second language speech: (1) it provides an extensive sociolinguistic description of L2 speech by conducting multivariate analyses on features of English spoken by Persians. In this way it throws light on the issue of the systematicity of second language speech by providing evidence for systematic variation in L2, (2) it explores the relationship between native language of the speakers and their L2, on one hand, and the relation between the L2 and the language L2 learners try to acquire, on the other. In other words, it examines the sources of the patterns of variability found in second language speech.; The focus of our investigation is the variable contraction (deletion) of auxiliary verbs and variable use of relative and resumptive pronouns by L2 learners. We first use the principles of variation theory to make a detailed assessment of the behavior of the L2 learners with respect to the contraction of auxiliary verbs in their natural L2 speech at both low and high proficiency levels. The comparative method, then, is applied to systematically compare patterns of contraction in low and high proficient L2 speech with: (a) patterns of contracting auxiliaries in our native English data and, (b) patterns found in the samples of Persian speech produced mostly by the same L2 informants. Then, we extend the same procedures to examine variable use of relative and resumptive pronouns in restrictive relative clauses in the three contexts of English, Persian and L2 (high proficiency only).; Our results produced conclusive evidence that the variability found in the features of L2 speech studied in this thesis was indeed systematically conditioned by at least one linguistic factor group. There was found to be no difference between high and less proficient learners in this regard. With respect to the sources of variability in L2 speech, we consider at least three possibilities: (1) that the variation can be explained by processes derived from English as the target language, (2) that the variation can be explained by the processes derived from Persian as the native language of the learners, and (3) that the variation and its conditioning system is unique to our learners' L2 speech, different from native or target languages. Parallel analysis and comparison of features of L2 and those of the informants' native and target languages demonstrated that variability in advanced second language speech was mostly conditioned by factor groups and factor weights that constrain native English speech. Informants' native language was found to play a trivial role in this respect. Less proficient learners, however, relied more on their native patterns of variability in cases where they have not acquired the target language patterns of variation yet. It was also demonstrated that patterns of variability change as learners advance in their L2 development. We did not observe any pattern unique to our informants' second language behavior, different from either English or Persian. |
