The Phonological Typology of Coronal Palatalization and Affrication

Abstract

This study explores cross-linguistic patterns of palatalization and affrication of coronal consonants. Although Humes (1992) Vowel-Place model predicts that coronal palatalization and affrication are distinct phonological phenomena, the results of the previous typological studies (e.g., Bateman 2007

Bhat 1978

Hall & Hamann 2006

Kochetov 2011) have suggested that these two processes have many similar properties. However, given that no previous study has systematically compared coronal palatalization and coronal affrication, it is difficult to determine whether these two processes are different phonological phenomena or not. In order to answer this question, this thesis investigates the two processes separately, in terms of their typology and phonetics. The results of this in-depth survey strongly suggest that they are separate phonological processes which have distinct typological patterns and phonetic backgrounds. A survey of 69 languages produced the following four implicational universals, two for palatalization and two for affrication. (i) Non-high front vocoids do not trigger palatalization unless high front vocoids do so (the trigger height asymmetry). (ii) Laterals do not undergo palatalization unless nasals do so. Nasals do not undergo palatalization unless obstruents do so (the target manner asymmetry). (iii) High back or central vocoids do not trigger affrication unless high front vocoids do so (the trigger frontness asymmetry). (iv) Voiced stops do not undergo affrication unless voiceless stops do so (the target voicing asymmetry). In addition, the results of the present survey show that coronal palatalization is exclusively triggered by front vocoids whereas coronal affrication is triggered only by high vocoids. Given that coronal palatalization and affrication have different restrictions on their triggers and targets, and that they are subject to the absolute process-specificity of the triggers, the separation of the two processes are typologically supported. The investigation of the phonetic backgrounds of palatalization and affrication of coronal consonants shows that the typological universals, presented above, correlate well with perceptual asymmetries. The key observations are that more likely target coronals have lower perceptibility of the relevant contrast than less likely ones and that coronal consonants occurring before more likely triggers have lower perceptibility than those occurring before less likely ones. Specifically, I investigated the perceptibility of the feature [anterior] for palatalization and that of the feature [strident] for affrication. First, the perceptibility of the feature [anterior] is mainly determined by the similarity in F2 transitions between alveolar and palatal consonants. They have more similar F2 transitions before high front vocoids than before non-high ones. In addition, alveolar and palatal obstruents (e.g., [s-ʃ]) have less distinct F2 transitions than their nasal counterparts (e.g., [n-ɲ]), which in turn have less distinct F2 transitions than lateral counterparts (e.g., [l-ʎ]). These phonetic facts suggest that the anteriority contrast is less salient before high front vocoids than before non-high ones, and that the anteriority contrast of obstruents is less salient than that of nasals, which in turn is less salient than that of laterals. These findings are consistent with the implicational universals of coronal palatalization (i-ii). Second, the perceptibility of the feature [strident] depends on the similarity in intensity and duration of friction between coronal stops and affricates. The intensity and duration of the stop release noise are higher and more elongated, respectively, and therefore more similar to those of the affricates, before front vocoids than before back or central vocoids. In addition, the intensity and duration of the friction noise of a voiceless stop are higher and longer and thus more like those of its affricate counterpart, than those of a voiced stop. Due to these acoustic properties, the stridency contrast between coronal stops and affricates is less salient before front vocoids than before back vocoids, and the stridency contrast is less salient for voiceless stops and affricates than for voiced ones. These findings are consistent with the implicational universals of coronal affrication (iii-iv). In order to capture the correlation between the implicational universals and perceptual asymmetries within the framework of Optimality Theory (Prince & Smolensky 2004), I follow the P-map hypothesis (Steriade 2001, 2009) that Faithfulness constraints for perceptually prominent segments and positions are ranked above those for perceptually less prominent ones. This analysis shows that the typological universals of the two processes can best be explained under the phonetically-based approach to phonology (Hayes, Kirchner & Steriade 2004), especially providing further support for the role of the phonetic perception in shaping phonological grammar (Boersma 1998

Jun 1995, 2004

Steriade 2001, 2009, inter alia).

본 논문은 설정 자음의 구개음화 및 파찰음화의 횡언어적 양상에 대해서 탐구한다. Hume(1992)의 모음-위치 모형은 설정구개음화와 파찰음화가 서로 다른 음운 현상임을 예측하고 있으나, 선행 유형론 연구(Bateman 2007

Kochetov 2011 등)의 결과는 이 두 음운과정이 서로 유사한 속성을 다수 공유함을 제시하고 있다. 그러나 설정구개음화와 설정파찰음화의 두 현상이 체계적으로 비교된 바가 없기 때문에, 이 둘이 서로 다른 음운현상인지 아닌지 판별하기가 쉽지 않다. 이러한 문제에 대답하기 위해서, 본 연구는 두 음운과정을 그 유형론 및 음성학적 측면에서 별도로 조사하였다. 그 결과는 두 과정이 서로 다른 유형론적 패턴과 음성학적 배경을 가진 별개의 음운과정임을 제시한다.