Consonantal Structures in Phonetics and Phonology

Abstract

The purpose of this study is to present and compare two different approaches (a phonetic approach and a phonological one) for the speech sound systems of natural languages. To this end, this study investigates natural speech sound systems with the consonantal systems of four Slavic languages, Russian, Polish, Czech and Serbian and Croatian, on the basis of phonetic and phonological approaches. In the phonetic approach, the consonant inventories of the four Slavic languages are analyzed with the theory of maximal and sufficient dispersion and the size principle, together with a frequency-based statistical approach. Segmental universals are discussed regarding sound types such as obstruents and sonorants. From the phonetic approach, it is shown that Slavic consonant systems are very unusual in terms of natural languages. Palatalized sounds in Russian and affricates and fricatives in Russian and Polish support that the Slavic consonantal system is far removed from the general aspect of human languages. On the other hand, with the phonological approach, four of the five feature-based principles proposed by Clements are employed to reveal the universals of the languages. They are Feature Economy, Marked Feature Avoidance, Robustness and Phonological enhancement. What we have seen is that some unsolved problems from the phonetic approach are explained by phonological accounts. The fact that Russian has plenty of segments represented by [+palatal] may not be unusual with respect to a feature-based approach. In addition, while the phonetic approach claims that Slavic languages (in particular, Russian and Polish) have different consonantal systems from the general aspect of natural languages because of the marked segments, the phonological approach accounts for the universals of these languages in the light of Robustness and Feature Economy. In short, what we get from phonetic accounts are language universals, found by frequency-based statistical approach while what we get from phonological accounts, using a feature-based approach, are linguistic universals.