Volume 48, Issue 2-4

Do the phonological predictions derived from phonetic models impose restrictions on the definition of’possible phonological system’? This question is addressed by examining the spectral integration model of vowel perception. The output of the model is investigated under a series of phonetic manipulations and translated into a set of phonological predictions, which are then compared against cross-linguistic patterns in vowel processes and inventories. It is argued that, despite the serious constraints on predictive power imposed by language-specific aspects of phonetic and phonological structure, phonetic models such as the spectral integration model enable us to define the set of phonetically motivated phonological phenomena.

It is the goal of this paper to consider how phonetics may interface with phonology. The role phonetics plays in phonological analysis is considered with respect to features, segments, phonological processes, and the abstractness of phonological solutions. Phonetic evidence is reviewed concerning the feature [strident] that questions the defining properties of this feature, its place in the distinctive feature framework, and its role in the analysis of natural classes for fricatives. The extent to which a sound segment is analyzed as a single unit or a cluster is considered with reference to prenasalized stops in Moru. The phonetic data are neutral with respect to the unitary or complex status of such segments, suggesting that the decision is more properly a phonological one. Phonetic evidence for the phonological process of neutralization is reviewed. Acoustic analysis of word-final devoicing in German, Polish and Catalan provides support for the reality of the process of neutralization in phonological solutions. Finally, phonetics helps provide a set of constraints or limiting conditions on the abstractness of phonological solutions. While phonetics cannot inform phonology about the appropriateness of a particular underlying form, there should be some phonetic transparency between the underlying form and its ultimate phonetic instantiation.

Phonologists have often held that phonetic ‘substance’ is more or less unrelated to phonological ‘form’. This view rests on assumptions about the phonetic domain that are highly questionable on empirical grounds. Evidence is reviewed suggesting that very few phonetic properties of vowels fail to serve the linguistic function of preserving and enhancing distinctiveness. Accordingly, much of what has been considered to be purely phonetic is also phonological in character; that is to say, the domains of phonetics and phonology overlap significantly. Finally, several well-known criticisms of the program of phonetic explanation in phonology are discussed and rejected.

Our aim is to present a sociopragmatic framework for phonological behavior and to comprehend phonology as part of a social action (in the sense of Schütz). This approach enables us to connect sociology, psychology, phonology and phonetics. Phonological variation is no longer seen as a secondary parameter, but can be analyzed as an integral part of language behavior. First we present Natural Phonology, then results of our investigations: phonological and phonetic variation due to psychological parameters on the one hand, steps of analysis in predominantly sociophonological work on the other hand.

In vowels contrasting for height, a large number of articulations covary with tongue height, which is supposed to be the principal bearer of the contrast. However, attempts to link these covarying articulations to tongue movement physiologically have been largely unsuccessful, and the particular pattern of covariation appears to make more sense as a concerted effort to influence the perceived height of F₁. The experiments reported here used the Garner interference paradigm, modified to assess the perceptual primacy of stimulus dimensions, to show that the acoustic effects of two of these covarying articulations, velum height (nasalization) and rate of vocal fold vibration (pitch) are integrated perceptually with the acoustic effects of varying tongue height. This perceptual integration suggests that the different articulations are not independently perceived, contrary to the predictions of direct realist theories of speech perception, that articulatory events covary so as to enhance each other’s perceptual effects, and that the surface phonological or initial phonetic representation of vowels might be quite richly specified (contrary to claims of phonetic underspecification).

After a brief review of important landmarks in the development of the phonetics/phonology dichotomy, the need for the integration of the two approaches is stressed. It is argued that, on the one hand, instrumental speech signal processing has to be subordinated to the language point of view, but that, on the other hand, the linguistic categories, such as phoneme and distinctive feature, should be critically assessed in respect of their adequacy as language categories, relevant for speakers and hearers. The phoneme is associated with word phonology, and problems are pointed out that arise when it is applied to the phonetics above the word. With reference to articulatory reduction processes in German, especially in function words, the interplay between the principles of economy of effort, on the part of the speaker, and of the need for distinctiveness for the listener are discussed. Empirical data are presented showing that reductions can be arranged along a continuous scale in spite of phonemic switches, if the physiological and articulatory dynamics are taken into account, providing deeper insight into human speech and language than the phonemic perspective. The concept of a ‘reduction coefficient’ is introduced and the call for a new ‘paradigm’ in the science of the spoken medium put forward.

Linguistic universals are prevalent patterns attributed to natural factors, either arising directly from human biology or from the ecology of language use. Phonological universals relating to segments can be studied using the UCLA Phonological Segment Inventory Database (UPSID). The nature of the data in this resource is described and some issues in interpreting cross-language uniformities and areal disparities are discussed. As an illustration, the modal number of vowel qualities in language inventories is shown to differ for the set of African languages in UPSID from those of other continents. This difference might be explicable as resulting from different ways of balancing various properties of phonological systems. But the correlation of this distribution with likely patterns of dispersion of ‘modern’ humans leaves open the possibility that the modal number of vowels in non-African languages reflects an inherited property of the parent of these languages.

Modern Hindi words such as [dãt] ‘tooth’ and [t ∫ãnd] ‘moon’ had similar phonetic structure in Middle Indo-Aryan, with conventional transliterations of danta and č anda , respectively. The development of the long nasal vowel is usually correlated with loss of the nasal consonant. If so, why does one form still contain a nasal consonant? We argue that a sequence of nasalized vowel + voiced stop (but not voiceless stop) can, for phonetic reasons, engender an epenthetic nasal, and we demonstrate that the same process can be found (nondistinctively) in present-day Hindi and French in the junction between a word-final nasal vowel and a following word-initital voiced stop. A nondistinctive epenthetic nasal can become a ‘full’ or ‘lexical’ nasal when listeners reinterpret this transitional event as purposeful or intended.