Seminari del CLT
Miquel Simonet (The University of Arizona)
Two Spanish phonological processes that may not be (fully) phonological after all
Seminari del CLT
Divendres, 23 de maig de 2025
Horari: 15:30h
Aula 202
Abstract
In this presentation, I report on two experimental research projects concerning two phonological processes in Spanish. The first project was concerned with Spanish external vowel sandhi. In Spanish, words may begin or end in a vowel, which may create sequences of vowels (or vowel clusters) across word boundaries, as in mono enano. It has been said that external vowel clusters such as these are affected by a syllable-contraction process that turns them into diphthongs. Traditionally, this process is known as synalepha. Based on intuitive introspection, scholars have postulated that syllable contraction in these situations turns one of the two vowels into a nonsyllabic vowel (or glide) while the other retains its syllabic status. The discussion has centered around identifying the vowel in each cluster combination that becomes nonsyllabic. A phonetic production experiment with Castilian Spanish data was conducted to corroborate whether prior descriptions were in line with acoustic findings. The clusters examined included /ea ae oe eo/. An examination of formant tracks showed that, for the most part, vowel clusters across a word boundary are resolved with a sort of blended vowel that preserves some of the linearity (or recoverability) of the underlying vowels. The second project was concerned with Spanish resyllabification. Phonologists have postulated that word-final consonants followed by a vowel-initial word are resyllabified so that they become the onset of a syllable whose nucleus is found in the following word: las alas [la.sa.las]. As a result of this process, sequences such as dices oso and dice soso are postulated to be homophonous. The results of four phonetic experiments we conducted recently challenge this view. Firstly, a production study showed that the /s/ indices eco is slightly shorter than the /s/ in dice seco. Secondly, a perceptual identification experiment showed that listeners make active use of duration to determine whether an intervocalic consonant is word initial or word final. A third and a fourth experiment were concerned with segmentation strategies. In a word-monitoring study, listeners were played sequences such as dice salgo and dices algo and asked to press a button as soon as they recognized the word <salgo>. We hypothesized that participants would respond to both dice salgo and dices algo when monitoring for <salgo>, and they did. Interestingly, however, it took them slightly longer to respond to mismatching trials (monitor: <salgo>; stimulus: dices algo) than to matching trials (monitor: <salgo>; stimulus: dice salgo). A fourth experiment revisited this issue with a syllable monitoring (rather than word monitoring) experiment. We conclude that dice soso and dices oso may not homophonous. Taken together, the results of our studies suggest that the facts pertaining to two Spanish phonological processes may be more complex than previously thought. While there is no doubt (in my mind) that many phonological processes are indeed phonological, I argue that we do not need phonological explanations for processes that may not be phonological after all.
