Unitary cardinals are a common source for indefinite markers. This thesis is a quantitative diachronic study of the development of Spanish un, from its cardinal value to its use as an indefi nite article. Based on a corpus comprising texts from the thirteenth to the seventeenth century, I present an analysis and chronology of the main changes undergone by un throughout this period, notably its increasing use as a marker of non-speci c indefinites, and its further incorporation in generic noun phrases and predicates. Additionally, I demonstrate that the development of the plural indefinite determiner unos is, with a few restrictions, parallel to that of its singular counterpart, not only in its increasing frequency, but also in its introduction into new contexts. Furthermore, I present a comparison between un and alg un in terms of speci city and conclude that although there are evident links between them, both being inde nite determiners derived from Latin unus, they have always had di erent functional domains. Finally, I show that one of the consequences of the incorporation of un into generic contexts is the rise of the so-called impersonal uno, and explain that this event is crucial to explain the disappearance of another generic pronoun, omne, whose last examples are found in the sixteenth century, that is, precisely the moment where the first instances of impersonal uno occur.

This study offers an assessment of the non-native speech intelligibility of a group of English learners of Spanish at word level and in connected speech. Specifically, we aimed at analysing the impact of certain categories of phonemic errors, as well as three temporal variables of L2 speech (speech rate, pause frequency and pause duration) on intelligibility scores. In addition, the possible correlation between degree of intelligibility and certain individual factors (gender, level of proficiency, motivation, aptitude and L1) was also studied. Sixty evaluators, native speakers of Peninsular Spanish, transcribed different speech samples belonging to a group of 20 Key Stage 4 English learners of Spanish. The transcription of the different speech samples served to assess intelligibility at word level and in connected speech (sentence, passage and semi-spontaneous production). Results revealed an intelligibility loss at all levels of analysis, as well as a high correlation between intelligibility scores in the single word test and those obtained in connected speech. At a segmental level, deviations affecting vowels, especially unstressed vowels, seemed to play a more important role than inaccuracies affecting consonants. Moreover, correlation analyses underscored the importance of speech rate, pause frequency and pause duration for intelligibility loss. The predictability of our multiple-regression models was high for speech samples obtained at sentence and passage levels. However, multiple-regression models for speech samples obtained through the semi-spontaneous production task exhibited a more limited capability in predicting variation in students’ intelligibility scores. Results suggest the existence of additional variables affecting intelligibility at this level of analysis. All individual differences under study, with the exception of gender, were highly correlated with speech intelligibility. From a pedagogical perspective, it is argued here that any successful instructional treatment of speech intelligibility will depend on an appropriate integration of temporal aspects of speech within the time devoted to pronunciation instruction in the foreign language classroom.

Zugl.: Münster (Westfalen), Univ., Diss.

Thesis (doctoral)--Universität des Saarlandes, Saarbrücken, 2009. / Includes bibliographical references.