Academic journal article Southwest Journal of Linguistics

Disentangling the Necessarily Entangled: The Phonology and Phonetics of Spanish Spirantization

Academic journal article Southwest Journal of Linguistics

Disentangling the Necessarily Entangled: The Phonology and Phonetics of Spanish Spirantization

Article excerpt

ABSTRACT. Previous accounts have treated Spanish spirantization either as a purely phonological process or as a purely phonetic one. Each of these approaches, however, is inadequate on its own. In this paper a new analysis of Spanish spirantization is proposed that takes into account both phonology and phonetics. By considering both phonology and phonetics together, a cleaner and more explanatory analysis can be offered. A phonological spirantization process by which the value [+continuant] spreads from a preceding segment to an underlyingly [-continuant] voiced obstruent is proposed. This process applies more broadly than similar processes that have been proposed, affecting all voiced obstruents following a [+continuant] segment. But this process is overwritten by a phonetic process in one case where an intrusive stop burst occurs as the result of gestural timing during the production of a particular consonant sequence. *

INTRODUCTION. For more than a decade the field of laboratory phonology has prospered. One of the many contributions that have come from work in this field is the realization that phonology and phonetics are largely inseparable (Pierrehumbert, Beckman and Ladd 2000). Many phonological processes are dependent on phonetic factors, both articulatory and acoustic, and phonetic implementation often depends on phonological factors. There are cases, however, where the entangling of phonology and phonetics makes it difficult to explain a particular set of data, since a portion of the data may be the result of a phonological process while another is attributable to a phonetic process. In this paper I examine one such case, that of Spanish spirantization, where no adequate analysis has been offered because previous accounts have tried to explain it as either exclusively a phonological matter or exclusively a phonetic one.

Due to the fact that phonology and phonetics are largely inseparable, it is important that any attempt to sort out what is the result of phonology and what is the result of phonetics in a given set of linguistic data be preceded by a clear definition of what is meant by these two terms. That is, what determines whether a particular pattern is a phonetic pattern or a phonological pattern? I follow recent definitions that consider phonological patterns to be categorical and phonetic patterns to be gradient (Cohn 1993, Keating 1988, 1996, Myers 2000, Ohala 1990, 1997, Pierrehumbert, 1980, 1990, 1994). An example of a categorical, and therefore phonological, pattern is that of the distribution of palatal consonants in Spanish. Palatal consonants never occur at the end of a word in Spanish, producing pairs such as that in 1.

(1)   Realizations of the Spanish palatal nasal/n/
      a. Word-final: desde [n] 'disdain (noun)'
      b. Not word-final: desde [n] ar 'disdain (verb, infinitive)'

In this pair, the palatal articulation is maintained when the palatal nasal is not the final segment of the word, but it loses its palatal articulation word-finally. Since this pattern is categorical (i.e. palatal when not word-final; not palatal when word-final), this pattern must be considered phonological.

An example of a gradient, and therefore phonetic, pattern is that of vowel length in Spanish. (1) Several studies have shown that Spanish vowels have a longer average duration in stressed syllables than in unstressed syllables (e.g. Quilis 1971, Face 1999). Therefore in the pair in 1, where the second [e] (represented orthographically as ) is stressed in desden but not in desdenar (where stress is on the final syllable), the stressed [e] of desden will be longer on the average than the unstressed [e] of desdenar. But since this is a gradient rather than a categorical pattern, the lengthening of the [e] when it is stressed does not change the category of the vowel (i.e. it remains [e], nor does it mean that there are no cases of unstressed [e] that are longer than the stressed [e] of desden. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.