Acceso usuarios
, Jon Franco Elorza (dir. tes.) , Jon Ortiz de Urbina (secret.) , Salvador Hernán Urrutia Cárdenas (voc.) , Francesc Roca Urgell (voc.) , Rosa Miren Pagola Petrirena (voc.) This dissertation investigates hiatus resolution of non-high vowels in Spanish, under the effect of five phonemic contexts: vowel position and quality, word boundaries, word type, syllable type and coda type. The main purpose is to gain knowledge about the gradient nature of hiatus resolution in this language, which resolves via different strategies at the phonetic output, namely acoustic diphthongization, gliding, coalescence or deletion. Hiatus resolution is a phonetic process in Spanish, totally optional and speaker dependent. In order to resolve hiatus, a durational reduction of the vowel sequence is required. This is claimed to be motivated by the five phonemic factors at study, some which are revealed to be more relevant than others. Previous studies dealing with vowel sequences in Spanish focus on the phonemic hiatus/diphthong alternation of such vowel sequences (Harris 1983, Roca 1997). But more recent studies based on phonetic data show intermediate stages of realizations -allophonic variations-between phonemic hiatus and diphthong (Aguilar 1999, 2003; Hualde 2004, Hualde & Chitoran 2003), suggesting a continuum of processes, which has led to a number of unresolved questions on the subject, some of which are addressed in this work.
The origin of the present study lies, first, in the analyses on rhythm (Pike, 1945; Abercrombie, 1971), which categorize languages based on the concept of isochrony into stress-timed or syllable-timed languages (Spanish). Posterior studies defend a gradual rather than a categorical division of the languages based on isochrony. This gradual division is motivated by factors such as the syllabic structure, the position of stress and vowel reduction, among others. Second, studies based on the phonemic division between hiatus and diphthong in Spanish observe vocalic sequences that respond to both categories within the same contexts (exceptional hiatus, Hualde 2004) and for which there is not a clear phonological explanation. Third, certain varieties of Spanish show a lexicalization of words containing vowel sequences that distinguish two phonetic outputs: the cohete cases (cohete à [¿kwe.te] versus [ko. ¿e.te]). These three point of departure guide the present work, which focuses on the reduction of vowel sequences underlyingly in hiatus and which questions the traditional categorization of diphthong and hiatus, defending a gradual phenomenon of vowel reduction.
The core of this dissertation consists of a double experiment involving a production task and a lexical decision task. The double experiment offers two major contributions to the literature on hiatus resolution in Spanish. First, it provides us with clear instances of the different strategies used by Spanish speakers to resolve hiatus. A description is proposed of four strategies that conform to the gradient nature of hiatus resolution in Spanish through the analysis of acoustic data: acoustic diphthongization, gliding, coalescence and deletion. Durational variation of the vowel sequence at the phonetic output is the main criterion used for the classification of those strategies. Linked to that, a binary division of the strategies is also proposed, which is determined by the existence of two vocalic segments (in hiatus, acoustic diphthongization and gliding) versus a single vocalic segment (in coalescence and deletion). Secondly, the results of the experiment help us understand in depth the relevance of the segmental and prosodic factors that contribute to the phenomenon of hiatus resolution. The findings show that the interaction of vowel position and quality has an effect on hiatus resolution, since durational reduction of the vowel sequences has a direct influence on the quality of the vocalic segments involved, in terms of perception of the vowels; and, added to that, the choice of the strategy of hiatus resolution is dependent on the context in which both vowels appear in the sequence (V1 or V2 position). Word boundaries and word type also affect hiatus resolution. Indeed, hiatus resolution strategies are often applied in vowel sequences across word boundaries, but within the word limits, the speaker prefers to maintain the canonical distribution of the vowels. Furthermore, in contexts across word boundaries, hiatus resolution seems to affect vowel sequences in function+content word combinations the least. In turn, vowel sequences in content+content and content+function word combinations allow hiatus resolution more often. Finally, the results demonstrated that codaless sequences resolve hiatus less often than sequences with a coda. Among the various coda groups in the language, obstruent codas, with the exception of the apico-alveolar sibilant, block hiatus resolution more often than sonorant codas. The apico-alveolar sibilant /s/ coda shows a radically different behavior, since it does not block hiatus resolution. Moreover, it allows deletion more often than any other coda groups. This fact confirms the special status of this consonant within the phonological consonantal system of the language.
The analysis on vowel sequences in contact within the research framework of this dissertation has a theoretical import, since hiatus resolution involves discrete phonological categories that appear in constant relation with gradient phonetic parameters, shared among speakers in real time. Phonology is primarily the study of a categorical system with a specific structure and based on lexical contrasts, shared among speakers. Phonetics, in turn, describes gradient phenomena of production and perception in real time which are speaker dependant. However, it is arresting to note that hiatus resolution shows an interaction of both disciplines. Crucially, this dissertation provides data that support an integrated approach to Phonetics and Phonology.
