Utterance structure in the initial stages of Polish L2 acquisition: from semantics to case morphology

This thesis is devoted to the processing of morphosyntax in the earliest stages of the acquisition of Polish L2, with special attention to the role of input and to elicitation techniques. The target structure of our work is the morphosyntactic opposition between the nominative and accusative case, respectively corresponding to the subject and object function. Within the VILLA project (Dimroth et al. 2013), 90 adult learners evenly distributed among five L1 groups (Dutch, English, German, French and Italian) took part in a first-exposure 14-hour Polish course taught by a specially trained native speaker. As participants had never been exposed to Polish or other Slavic languages, the experiment portrays the very first contact with a completely new target language. The experiment was carried out under strictly controlled input conditions: input was planned, recorded and transcribed, in order to thoroughly correlate it to learner output. The processing of inflectional morphology is first investigated through a repetition and a comprehension test. In the former, learners hear a target sentence and are asked to repeat it: the literature generally agrees that this task does not require test takers to merely repeat a string of sounds, but rather to de-code and re-encode the meaning of the target sentence according to the present state of the interlanguage grammar. In the comprehension test, participants hear a target sentence and are asked to indicate which picture out of a set of two best describes it. The methodological heart of both tasks is the simple intuition that the processing of inflectional morphology can be studied by manipulating word order. In order to process SO targets, two strategies whould prove equally effective. First, a morphosyntactic principle can be adopted, whereby case endings are associated to the corresponding meaning. Alternatively, learners can rely on a positional principle, in which the subject is simply instantiated by the noun which occurs first in the utterance. As OS targets can only be processed correctly using a morphosyntactic principle, above-chance accuracy in their processing constitutes evidence that the learner has established a solid form-function association between case endings and the corresponding syntactic function. The two tests are first analysed separately. A powerful, though not unexpected role for word order and the L1 emerges. In both tests, SO targets are processed with far greater accuracy that their OS equivalents; moreover, German speakers exhibit overall better performance than the other L1 groups, which is attributed to the greater complexity of nominal morphology in their native language. In the repetition test, however, the Italian group slightly outperforms them, in spite of the greater typological distance from Polish. It is hypothesised that questions of perceptual prominence and phonology may be involved. If confirmed in further work, these tendencies would suggest a role for these factors in the processing of morphosyntax and at the same time raise challenging questions as to the nature of the repetition task. Next, the results of the two tests are combined, highlighting a hierarchy of contexts defined as the interaction between task and word order: OS repetition ⊃ OS comprehension ⊃ SO repetition ⊃ SO comprehension. While the hierarchy confirms the role of word order, it also suggests that the repetition test may be harder for learners than the comprehension test, not unexpectedly as it can be argued that the former entirely encompasses the latter. The case of learners performing above chance in comprehension, but not repetition with the same type of targets witnesses to the slower development of productive as opposed to receptive abilities. The target structure is then investigated in the context of semi-spontaneous interaction as produced by a subset of learners (n=17), with the aim to verify whether or not morphosyntactic competence could be influenced by the communicative situation. Indeed, accuracy rates are markedly lower in this new context than in the structured tests. Communication efficacy does not seem to suffer from this lack of target-like case, though, as semantic and phrasal principles such as animacy contrasts and unmarked SO word order make it possible to encode and decode meaning with no apparent break-down, at least in the simple task at hand. Interestingly, the same principles are in operation even if the learner demonstrates to be able to process case marking in a consistent and productive manner. A preliminary analysis of the input, finally, shows that even the utterances produced by the native speaker can be easily decoded on the basis of semantics and default word order alone. Morphosyntax would prove indispensable only in a negligible proportion of input transitive sentences. This last observation prompts us to morphologically tag the entire input in order to analyse in greater detail the transitive sentences comprised in it, paying particular attention to the interaction of morphosyntax and semantics. Thanks to a specially designed software we perform several computations on this enriched data-set, including form-function association between case endings and syntactic functions, type frequency for subjects, objects and transitive verbs, as well as token frequency for a variety of transitive structures, defined by the interaction of word order, grammatical category, gender and animacy. Although strong tendencies are identified, learners appear to be able to go beyond the models contained in the input, applying principles of utterance organisation derived from their L1 or from semantics and universal cognitive constraints, thus effectively dealing with infrequent or even absent patterns. To summarise, the study concludes that while several learners proved able to process inflectional morphology in a structured test after only a few hours of exposure to the input, much fewer could do the same in the context of spontaneous interaction, in which they relied on semantic and phrasal principles. While this conclusion may be seen in itself as a significant contribution to the debate regarding the initial stages of L2 acquisition, we believe this work highlighted sensitive methodological points regarding the role of semantics in determining morphosyntax, on the one hand, and the effect of elicitation technique on the observable strategies of morphosyntactic processing. These insights seem precious to better describe the complex operations by which adults speakers come to approach a completely unfamiliar language.