The Generative and the Structuralist Approach to the Syllable [1 ed.] 1443897027, 9781443897020

Citation preview

The Generative and the Structuralist Approach to the Syllable

The Generative and the Structuralist Approach to the Syllable: A comparative analysis of English and Slovak By

Renáta Gregová

The Generative and the Structuralist Approach to the Syllable: A comparative analysis of English and Slovak By Renáta Gregová This book first published 2016 Cambridge Scholars Publishing Lady Stephenson Library, Newcastle upon Tyne, NE6 2PA, UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Copyright © 2016 by Renáta Gregová All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-9702-7 ISBN (13): 978-1-4438-9702-0

TABLE OF CONTENTS

List of Figures.............................................................................................. x List of Tables ............................................................................................. xii List of Abbreviations ................................................................................ xiii Acknowledgements .................................................................................. xiv Introduction ............................................................................................... xv Chapter One ................................................................................................. 1 The Syllable in Generative Phonology 1.1 The Internal Structure of the Syllable .............................................. 2 1.1.1 Extrasyllabic Consonants ........................................................ 4 1.2 Syllabification and Morphology ...................................................... 5 1.3 The Underlying Representation and the Surface Representation ..... 6 1.4 Universality and Diversity ............................................................... 9 Chapter Two .............................................................................................. 12 The CVX Theory of the Syllable 2.1 An Outline of the Theory ............................................................... 12 2.1.1 The Affix Rule ...................................................................... 13 2.1.2 The Potential Vowel Rule ..................................................... 13 2.1.3 Anti-Allomorphy ................................................................... 14 2.1.4 Complex Sounds ................................................................... 14 2.2 The General Theory of Complex Segments ................................... 14 2.2.1 Feature Geometry .................................................................. 15 2.2.2 A Complex Segment in Duanmu’s Theory ........................... 18 2.2.3 A Consonant Cluster or a Complex Segment? ...................... 21 2.3 Word-initial and Word-final Clusters in English and in Slovak .... 22 2.3.1 What is a Consonant Cluster?................................................ 22 2.3.2 Consonant Clusters in English............................................... 22 2.3.3 Consonant Clusters in Slovak ................................................ 24 2.4 A Phonemic Complex-sound Analysis of the Word-initial CC Clusters...................................................... 27

vi

Table of Contents

2.4.1 A Phonemic Complex-sound Analysis of the English Word-initial CC Clusters .......................................................... 27 2.4.2 A Phonemic Complex-sound Analysis of the Slovak Word-initial CC Clusters .......................................................... 31 2.5 A Phonetic Complex-sound Analysis of the Word-initial CC Clusters .................................................................................... 35 2.5.1 A Phonetic Complex-sound Analysis of the English Word-initial CC Clusters .......................................................... 39 2.5.2 A Phonetic Complex-sound Analysis of the Slovak Word-initial CC Clusters .......................................................... 40 2.5.3 Do Complex Sounds Exist? ................................................... 41 2.6 The CVX Syllable Theory and the English Language ................... 43 2.6.1 A Single-slot Analysis of the Word-initial Consonant Clusters in English .................................................................... 44 2.6.2 The Analysis of the Word-final Rhymes in English .............. 45 2.6.3 The Syllable Boundaries in the English Word-medial Consonant Clusters ................................................................... 47 2.6.4 Summary ............................................................................... 48 2.7 The CVX Syllable Theory and the Slovak Language .................... 49 2.7.1 A Single-slot Analysis of the Word-initial Consonant Clusters in Slovak ..................................................................... 49 2.7.2 The Analysis of the Word-final Rhymes in Slovak ............... 52 2.7.3 The Syllable Boundaries in the Slovak Word-medial Consonant Clusters ................................................................... 56 2.7.4 Summary ............................................................................... 57 Chapter Three ............................................................................................ 60 The Syllable Structure in Lexical Phonology 3.1The Three-dimensional Phonological Representation .................... 60 3.2 The Syllable Structure Algorithm .................................................. 62 3.2.1 The Syllable Structure Algorithm in English ........................ 67 3.2.1.1 The Analysis ................................................................. 69 3.2.2 The Syllable Structure Algorithm in Slovak ......................... 74 3.2.2.1 Some Special Issues of Slovak Phonology ................... 78 3.2.2.1.1 Is Slovak ‘ j’ a glide? ........................................... 79 3.2.2.1.2 The Problem of the Slovak ‘v’ ............................. 80 3.2.2.2 The Analysis ................................................................. 82 3.2.3 The Sonority-based Analysis................................................. 86 3.2.3.1 The Sonority-based Analysis of the Word-initial CC and CCC Consonant Clusters in English............................. 87

The Generative and the Structuralist Approach to the Syllable

vii

3.2.3.2 The Sonority-based Analysis of the Word-initial Consonant Clusters in Slovak ............................................. 88 3.2.3.2.1 The Sonority-based Analysis of the initial CC Clusters in Slovak.......................................................... 88 3.2.3.2.2 The Sonority-based Analysis of the initial CCC Clusters in Slovak.......................................................... 89 3.2.3.2.3 The Sonority-based Analysis of the initial CCCC Clusters in Slovak.......................................................... 90 3.2.4 The Sonority and the Structure of the Syllable...................... 91 3.2.5 The SSA and Monosyllabic Words ....................................... 91 3.2.5.1 The Analysis of the English Monosyllabic Words ....... 91 3.2.5.2 The Analysis of the Slovak Monosyllabic Words ........ 92 3.3 Summary ........................................................................................ 94 Chapter Four .............................................................................................. 95 The Syllable in Generative Phonology Again 4.1 On the Universality of the CVX Syllable Theory .......................... 95 4.2 On the Universality of the SSA in Lexical Phonology .................. 97 Chapter Five .............................................................................................. 99 The Syllable in Structuralism 5.1 Ferdinand de Saussure and the Syllable ......................................... 99 5.2 The Syllable in the Prague School of Linguistics ........................ 100 Chapter Six .............................................................................................. 102 The Synthetic Phonological Theory 6.1 General Outline............................................................................ 102 6.2 The Syllable and the Synthetic Phonological Theory .................. 105 6.3 The Criteria for the Syllabification .............................................. 107 6.3.1 The Interface of the Semantic and the Phonic Part of a Linguistic Sign................................................................. 107 6.3.1.1 The Sample Analysis of the Slovak Language ........... 109 6.3.1.2 The Sample Analysis of the English Language .......... 110 6.3.2 The Transgressive Nature of Consonants ............................ 111 6.3.2.1 The Sample Analysis of the Slovak Language ........... 113 6.3.2.2 The Sample Analysis of the English Language .......... 114 6.3.3 The Criterion of Coarticulation ........................................... 115 6.3.3.1 The Sample Analysis of the Slovak Language ........... 117 6.3.3.2 The Sample Analysis of the English Language .......... 118 6.3.4 The Degree of Stricture ....................................................... 188 6.3.4.1 The Sample Analysis of the Slovak Language ........... 121

viii

Table of Contents

6.3.4.2 The Sample Analysis of the English Language .......... 124 6.3.5 The Phonotactics of the given Language and Kuryáowicz's Rule ............................................................ 127 6.3.5.1 The Sample Analysis of the Slovak Language ........... 128 6.3.5.2 The Sample Analysis of the English Language .......... 131 6.3.6 The Frequency Criterion ..................................................... 132 6.3.6.1 The Sample Analysis of the Slovak Language ........... 133 6.3.6.2 The Sample Analysis of the English Language .......... 134 6.3.7 The Power of Syllable Welds .............................................. 135 6.3.7.1 The Sample Analysis of the Slovak Language ........... 135 6.3.7.2 The Sample Analysis of the English Language .......... 138 6.4 The Complex (Synthetic) Approach to the Syllable Structure ..... 139 6.4.1 The Sample Complex Analysis ........................................... 142 6.4.1.1 The Analysis of the Slovak Word ............................... 142 6.4.1.2 The Analysis of the English Word .............................. 143 6.4.2 The Syllable and the Different Levels of Abstraction ......... 144 Chapter Seven.......................................................................................... 146 The Generative and the Structuralist Approach to the Syllable Structure 7.1 The Syllabification of the Sample Words .................................... 148 7.2 Different and the Same ................................................................ 150 7.3 Final Remarks .............................................................................. 152 Conclusions ............................................................................................. 155 Appendix 1 .............................................................................................. 161 Appendix 2 .............................................................................................. 163 Appendix 3 .............................................................................................. 165 Appendix 4 .............................................................................................. 175 Appendix 5 .............................................................................................. 195 Appendix 6 .............................................................................................. 198 Appendix 7 .............................................................................................. 203 Appendix 8 .............................................................................................. 205

The Generative and the Structuralist Approach to the Syllable

ix

Bibliography ............................................................................................ 213 Author Index............................................................................................ 222 Subject Index ........................................................................................... 225

LIST OF FIGURES

Figure 1-1 Figure 1-2 Figure 1-3 Figure 1-4 Figure 2-1 Figure 2-2 Figure 2-3 Figure 2-4 Figure 2-5 Figure 2-6 Figure 2-7 Figure 3-1 Figure 3-2 Figure 3-3 Figure 3-4 Figure 3-5 Figure 3-6 Figure 3-7 Figure 3-8 Figure 3-9 Figure 3-10 Figure 3-11 Figure 3-12 Figure 3-13 Figure 3-14 Figure 3-15

The Flat Syllable Structure.................................................... 3 The Hierarchical Syllable Structure–binary Branching with Body .............................................................................. 3 The Hierarchical Syllable Structure–binary Branching with Rhyme ........................................................................... 3 The Position of Phonology in Classical Phonemics and in the SPE ....................................................................... 9 The Structure of the Syllable in the CVX Theory ............... 12 A Labio-Coronal Complex Segment /pࢎ t/ ............................ 15 A General Model of Feature Organization .......................... 16 A Formal Model of Articulator-based Feature Geometry ... 18 The Oscillogram of the Phrase ‘is almost the same’, the Word ‘same’, and the Segmented Consonant [s] .......... 38 The Maximal Phonological Structure of the English Syllable ........................................................ 43 The Maximal Phonological Structure of the Slovak Syllable ......................................................... 49 Phonological Representations in Three-dimensional Phonology ................................................................................. 61 The N-Placement Rule ........................................................ 62 The Gliding Rules ............................................................... 63 The CV-Rule ....................................................................... 63 The Onset Rule ................................................................... 64 The Coda Rule .................................................................... 64 The Syllable Structure Algorithm ....................................... 65 The Complex Coda Rule ..................................................... 65 The Syllabification of the Word ant /ænt/ ........................... 66 The Sonorant Syllabification Rule ...................................... 67 The Syllabification of the Word extra /ekstrԥ/ .................... 68 The Syllabification of the Word yogurt /j‫ܥ‬gԥt/ ................... 68 The Syllabification of the Word bottle /b‫ܥ‬tl/ ...................... 69 The Syllabification of the Word kompromis 'compromise' ....................................................................... 75 The Syllabification of the Word with Prefix following the Basic SSA ..................................................... 77

The Generative and the Structuralist Approach to the Syllable

xi

Figure 3-16 The Syllabification of the Word with Prefix following the Basic SSA and the Prosodification Constraint .............. 78 Figure 3-17 The Syllabification of the Word klamstvo 'falsehood' ......... 78 Figure 3-18 The Syllabification of the Word jajþaĢ 'to whine' ............... 80 Figure 3-19 The Postcyclic Adjunction Rule.......................................... 86 Figure 3-20 The Syllabification of the Word nervstvo 'nervous system' .................................................................. 86 Figure 3-21 The Syllabification of the English Word-initial Three-consonant Cluster ..................................................... 91 Figure 3-22 The Syllabification of the English Word-final Three-consonant Cluster ..................................................... 92 Figure 3-23 The Syllabification of the English Word-final Four-consonant Cluster ....................................................... 92 Figure 3-24 The Syllabification of the Slovak Word-initial Three-consonant Cluster ..................................................... 93 Figure 3-25 The Syllabification of the Slovak Word-initial Four-consonant Cluster ....................................................... 93 Figure 3-26 The Syllabification of the Slovak Word-final Three-consonant Cluster ..................................................... 93 Figure 3-27 The Structure of the Syllable in Lexical Phonology ........... 94 Figure 6-1 The Relation between the Individual (I) and the Universal (U) ........................................................ 103 Figure 6-2 The Oscillogram and the Sonagram of the Word krajiny 'countries' .............................................................. 113 Figure 6-3 The Oscillogram and the Sonagram of the Cluster [mn] ... 114 Figure 6-4 The Sonority Graph of the Word pestrý ‘colourful’.......... 123 Figure 6-5 The Sonority Graph of the Word nervstvo ‘nervous system’ ............................................................... 124 Figure 6-6 The Sonority Graph of the Word sharpness [‫ܤݕ‬:pnԥs] ...... 125 Figure 6-7 The Sonority Graph of the Word esprit [espri:] ................ 126 Figure 6-8 The Sonority Graph of the Word Sanskrit [sænskrܼt]........ 127 Figure 6-9 The Hierarchy of the Criteria Used for the Syllabification in the SPT ................................................. 142 Figure 7-1 The Generative and the Structuralist Levels of Abstraction.................................................................... 152

LIST OF TABLES

Table 2-1 Table 2-2 Table 2-3 Table 2-4 Table 2-5 Table 2-6 Table 2-7 Table 2-8 Table 3-1 Table 3-2 Table 3-3 Table 3-4 Table 6-1 Table 6-2 Table 6-3 Table 7-1 Table 7-2

The Phonetic Classification of the English Consonant Sounds ................................................................................. 27 Feature Specification of the English Consonant Phonemes ............................................................................ 28 Specification of the English Consonants in Terms of Articulator-based Feature Geometry............................... 29 The Phonetic Classification of the Slovak Consonant Sounds ................................................................................. 31 Feature Specification of the Slovak Consonant Phonemes ............................................................................ 32 The Specification of the Slovak Consonants in Terms of Articulator-based Feature Geometry............................... 34 Statistical Data about the Length of the Selected English Word-initial CC Sequences.................................... 40 Statistical Data about the Length of the Selected Slovak Word-initial CC Sequences ..................................... 41 The Sonority Hierarchy of the English Consonants ............ 70 The Sonority Hierarchy of the Slovak Consonants ............. 84 The Sonority-based Analysis of the English Word-initial CC Clusters .................................................... 88 The Sonority-based Analysis of the Slovak Word-initial CC Clusters .................................................... 89 Variation Specification of the Individual Sound Levels ... 105 The Hierarchy of the English Consonants according to the Degree of Stricture .................................................. 120 The Hierarchy of the Slovak Consonants according to the Degree of Stricture .................................................. 120 Syllabification of the English Sample Words ................... 148 Syllabification of the Slovak Sample Words .................... 149

LIST OF ABBREVIATIONS V C X FG Supra node Soft Pal s x Cv msec Ȉ ৻ N Cs v LP SSG SSA N N' N'' ı MSD OSP O L ES SPE SPT HPh Ph Phm MPhm HMPhm I U

vowel consonant vowel or consonant Feature Geometry Supralaryngeal node Soft Palate node standard deviation arithmetic mean, average variation coefficient millisecond sum of nasalised vowel nasal consonant complex segment, complex sound unstressed vowel Lexical Phonology Sonority Sequencing Generalization Syllable Structure Algorithm nucleus rhyme node syllable node syllable node Minimal Sonority Distance Obstruent Sequencing Principle obstruent liquid Exhaustive Syllabification Sound Pattern of English Synthetic Phonological Theory hypophone phone phoneme morphophoneme hypermorphophoneme the individual the universal

ACKNOWLEDGEMENTS

First, I would like to express my deep gratitude to Dr. H. C. Prof. PhDr. Ján Sabol, DrSc. from the Department of Slovak Studies, Slavonic Philologies, and Communication, Faculty of Arts, P. J. Šafárik University in Košice, Slovakia, for his professional guidance and constructive and useful comments on my scientific work. I also wish to thank to Prof. PhDr. Pavol Štekauer, DrSc. from the Department of British and American Studies, Faculty of Arts, P. J. Šafárik University in Košice, Slovakia, for a thought-provoking note that pushed me forward at the moment I got stuck in the writing of this monograph. And special thanks should be given to the proofreader Marlene deWilde for her excellent work. Of course, any errors are my own and should not stain the reputation of the publisher.

INTRODUCTION

In the first half of the 20th century, structuralism flourished in European linguistics. Linguistic structuralism1 as introduced by Ferdinand de Saussure (1922) manifested itself in the theories of several schools that differed from each other in their theoretical orientation. The most important schools of European structuralism are the Prague School of Linguistics, the Geneva School, the Copenhagen School, the Moscow School, and the London School (see, e.g., Albrecht 2011 for a detailed description of the individual schools). Considering the sound level of language, the Prague School of Linguistics with its functional approach to language analysis is the most inspiring one (cf., e.g., Sampson 1980; Sabol 1989; Goldsmith 2011). The contribution of the Prague Linguistic Circle to the field of phonology is immense and indisputable (see, e.g., ýerný 1996). The phonologists of the Prague School differentiated between phonetics and phonology and distinguished between sounds and phonemes. They believed the functions of the individual sounds played an important role, and emphasized the distinctive function of phonemes. Trubetzkoy’s theory of phonological oppositions is still valid without substantial changes (ibid., 155), and so are his concepts of neutralization and archiphoneme. Thus, for simplification and clarity, I have labelled the phonology of the Prague School of Linguistics (i.e., phonology developed within the framework of European structuralism) as “structuralist phonology” in this work.2 In the 1960s, Noam Chomsky presented his theory of generative grammar, which became the basis for the phonological theory known as generative phonology (Sampson 1986, 188). Generative phonology3 itself was the result of the work of Noam Chomsky and Morris Halle 1

Unless the attribute ‘European’ is used, the label ‘structuralism’ still refers to ‘European structuralism’ in this work. The survey of the fundamental tenets of European structuralism and the comparison of this linguistic stream to the socalled American structuralism can be found, for example, in Sampson (1980) or Albrecht (2011). 2 Detailing a depiction of structuralist phonology developed by the linguists in the main linguistic schools or, at least, co-operating with them, is not the aim of this work. That is why I think that this type of simplification is possible. 3 The core values and beliefs of generative phonology can be found, for example, in Sampson (1980) or in Goldsmith and Laks (2011).

xvi

Introduction

(Goldsmith and Laks 2011, 6). “Proponents of generative grammar […] believed that generative grammar was the first truly scientific account of language” (ibid., 7). Taking into account various extraneous conditions that play a role in the spreading of new ideas (for details, see Everaert and Reuland 2011), the generative approach to language received immediate attention in the USA and in the countries of Western Europe, while structuralism, including structuralist phonology, has remained popular in the territory of, geographically speaking, Central and Eastern Europe. The generative ideas were only marginal there (ibid.). One of the crucial notions in both structuralist and generative phonological theory is the notion of the syllable. (However, early generativism ignored the syllable as the unit of analysis, see Chapter 1.) The syllable as the basic sound unit of continuous speech encompasses all components of the speech signal: articulatory, acoustic, and perceptual (Sabol 1994, 217). It is a complex-sound unit connecting segmental and suprasegmental subsystems of a language (Sabol 1997, 27). Moreover, the syllable is the fundamental unit in phonological analysis (Blevins 1995, 206). For example, Pauliny (1979) sees the syllable as the smallest unit necessary for the identification of phonemes. And the syllable is used for the analysis of the phonological constraints of a given language. The syllable is thus also the central unit of phonotactic analysis. All in all, structuralist and generative linguistics agree the syllable is an important phonological constituent of a language. Much has been done in connection with the syllable in generative linguistics (see, e.g., Blevins 1997; Cairns and Raimy 2011; Goldsmith 2011) and in structuralist phonology (see, e.g., Sabol 1994). However, little has been done to compare or contrast structuralist and generative approaches to analysing syllable structure and function. There are two significant books on the subject. Brief notes about the understanding of the syllable in structuralism and in generative linguistics can be found in Cairns and Raimy’s (2011, 1–18) introduction to Handbook of the Syllable. The chapter, however, is descriptive rather than explanatory or comparative. A condensed classification of various approaches to the syllable, encompassing its position in structuralism and generativism, appears in Goldsmith’s (2011) chapter on the syllable in The Handbook of Phonological Theory. Otherwise, those whose attention is focused on generativism either ignore structuralist ideas or evaluate structuralism as an improper linguistic theory (Emonds, personal discussion; see also

The Generative and the Structuralist Approach to the Syllable

xvii

Turner 1987).4 I have been researching the syllable from the very beginning of my linguistic work (Gregová 1998; Gregová 2004). Studying linguistics in Central Europe gave me a structuralist linguistic education, and I became acquainted with the generative understanding of language through working with English linguistics. As I gradually specialized my research interest, examining comparative phonetics, phonology, and morphophonology, I saw many situations in which structuralism and generativism seemed to clash. Thus, the aim of this work is to compare and contrast the structuralist and the generative approaches to the syllable, which, as specified above, is the basic unit of phonological analysis. Since both linguistic streams manifest themselves in a number of schools and theories, a simplification is necessary, and only the most significant and typical representatives of each approach are evaluated here. Prior to the specification of the generative theories chosen for the analysis, the basic aspects of the syllable in generative phonology as a whole are presented in chapter 1. Generative phonology originated as an enlargement of Roman Jakobson’s work on phonological universals (Sampson 1980, 188), an underlying phonological system common for all languages. The essence of Chomsky’s theoryʊas explained in detail in chapter 1ʊis the belief about linguistic universals in syntax (ibid., 131). This idea was further developed in the theory of universal grammar that achieved incredible popularity among both linguists and non-linguists. Generativists and generative phonologists concerned with language universals usually characterize their theories as valid across languages (see, e.g., Chapter 2). Therefore, the secondary aim of this work is to evaluate the alleged universal nature of the generative theories presented here. Due to the reasons explained at the end of chapter 1, the generative approach to the syllable is exemplified by two theories, the CVX theory of syllable (Chapter 2) and Lexical Phonology (Chapter 3). The outcomes of the analyses of the syllable from the point of view of these two generative theories are summarized in chapter 4, where also a re-evaluation of the tenets of the generative approach to the syllable introduced in chapter 1 is made. Chapter 5 encompasses the most important issues of the syllable in structuralism represented by one theory, the synthetic phonological theory, which is introduced in chapter 6. Chapter 7 offers the comparison of the 4

The proponents of the structuralist approach to language seem to be less radical in advocating their linguistic beliefs and ideas. They simply disregard generativism and do not comment on it.

xviii

Introduction

generative and the structuralist approaches to the syllable typical of CVX syllable theory, Lexical Phonology, and the synthetic phonological theory. The outcome of the work can be found in the Conclusions. All three syllable theories introduced in this book are thoroughly examined using data from two languages – English and Slovak – which differ from the viewpoint of morphological typology. English is an analytic language belonging to the West Germanic branch of the IndoEuropean language family, and Slovak is a synthetic language from the West Slavic branch of the same family (for details, see, e.g., Brown and Ogilvie 2009).5 One can object that the comparative analysis of two languages is not sufficient when evaluating the universality of a certain theory or at least the tendency towards universality. In other words, the limited extent of the research material may give the impression that the results of the comparative and contrastive analyses presented in the Conclusions are nothing more than mere speculation. The studies of the syllable usually adopt one of two approaches: either a rather superficial survey of a large number of languages or an exhaustive analysis of a small number of languages (Duanmu 2009, 3). In order to be consistent with the authors whose theories are included in this work, I have chosen the latter approach. The CVX syllable theory is applied to five languages (Duanmu 2009); the initial ideas of the Syllable Structure Algorithm in Lexical Phonology were presented in the analysis of one language only (Mohanan 1982), and the basic aspects of the synthetic phonological theory were also exhibited on only one language (Sabol 1989). The ordering of the theories is based on the principle of a reverse chronology – the CVX theory was launched in 2009. Although the description of the syllable from the point of view of Lexical Phonology was introduced in 1982 (see above), the detailed algorithm of the syllable structure used in this work was created by Rubach (1993). The synthetic phonological theory originated in 1989. The reverse chronological order simultaneously follows the principle of the inner logical connection between these theories: the CVX theory is the peak of the generative understanding about language, Lexical Phonology is the gate between pregenerative structuralist phonology and classical generative phonology (Goldsmith and Laks 2011, 14), and synthetic phonological theory can be characterized as the summary of the structuralist approach to language. All analyses and evaluations included under the roof of the individual syllable theories are supported by the comprehensive data that can be 5

Despite an obvious genetic kinship, the languages have different vocabulary, syntax, and, of course, phonology.

The Generative and the Structuralist Approach to the Syllable

xix

found in the Appendices (Appendix 1–8). All English and Slovak consonant clusters6 and the results of their various examinations are included in this work to avoid objections that the author presents only those data that follow the given theory and ignores those that probably do not. For example, Scheer (2012) argues that the CVX theory of syllable “is not tested on the grounds of language data” (ibid., 719). Some consonant clusters are encompassed in the analysis and others are simply left unmentioned. A similar objection is made to Sagey’s evaluation of consonant clusters as complex segments (see Chapter 2). In brief, this work – as indicated by its title – presents comparative and contrastive analyses of the generative and structuralist approaches to the syllable on the basis of a detailed investigation of data from English and Slovak according to three syllable theories. The theories are described and their principles and rules are applied to the data. The results of the thorough analysis are assessed and compared. The conclusions may only be indicated since the work, of course, does not aspire to solve the clash between structuralism and generativism. Finding the truth is outside the scope of the study. The goal of this book is to point out the differences and find possible similarities between different methods of analysing the syllable and the issues associated with it.

6

Or rather, most consonant clusters from both languages are included in all detailed investigations. Taking into account the fact that language is a vivid system, only the most productive clusters from both languages were analysed (see also note 15 in Chapter 2).

CHAPTER ONE THE SYLLABLE IN GENERATIVE PHONOLOGY

In Chomsky and Halle’s (1968) Sound Pattern of English (henceforth, SPE), the assumption was that the individual speech segments, boundaries, and rules governing possible combinations of these segments in morphemes and words were satisfactory for the description of the sound systems of languages (Katamba 1989, 164). Thus, although “the syllable is one of the oldest constructs in the study of language, and most studies of phonology have found a place for the syllable within them” (Goldsmith 2011, 214), the syllable is not referred to as a formal phonological unit in early generative phonology. Prosodic structure is governed by the principles of universal grammar (see, e.g., Goldsmith and Laks 2011) and the existence of the syllable can only be deduced from the rules that seem to take this unit as the domain of their application (Sampson 1980, 193). Goldsmith, in Goldsmith and Laks (2011), gives two of the most important principles of generative linguistics reflected in the syllable theories appearing in the decades after SPE: (1) the main goal of a linguistic theory is the development of a formal theory, and (2) attention should be paid to the algorithmic explanation of the generation of surface forms on the basis of underlying forms (ibid., 9–10, see also, below). In brief, formalism and algorithmic explanations are the fundamental notions in generative linguistics, that is, also in generative phonology. This “abstract formalist logic” (ibid.) of classical generative phonology as presented in the SPE was questioned by the representatives of natural phonology (or natural generative phonology) in the 1970s (for details, see, e.g., Donegan and Stampe 2009). However, although criticising SPE as a kind of artificial phonology, natural phonology continued its surfaceoriented analysis, based on phonological derivations and ordered rules. One of the significant differences between the SPE and natural phonology is that the SPE dismisses the syllable while natural phonology understands the syllable as a unit of phonological analysis (Goldsmith and Laks 2011, 12, 16). It became clear that many phonological processes affect syllable structure rather than the structure of the morpheme or single segments. For example, the concept of the syllable has been used to account for such

2

Chapter One

prosodic features as stress, tone, and duration (Hooper 1972, 531, 533). Hooper (1976, 525) further demonstrates that “the syllable can be formally and universally defined by a rule, which inserts […] syllable boundaries between certain sequences of segments”. Consequently, the aim of numerous phonological theories within generative phonology has been to propose a universal definition of the syllable in terms “of conventions for the placement of syllable boundaries” (ibid.). The study of the syllable and its structure has become the essential part of phonological theories evolved within the framework of generative phonology (cf., e.g., Goldsmith 2011; Goldsmith and Laks 2011).1 The most influential were the following: autosegmental phonology2 and metrical phonology3 in the 1970s; Lexical Phonology,4 three-dimensional phonology5 and CV-phonology6 in the 1980s; and Optimality Theory7 in the 1990s (see also note 1).

1.1 The Internal Structure of the Syllable One of the features distinguishing among various syllable theories is whether they postulate flat syllables or an internal syllable hierarchy (Cairns and Raimy 2011, 15). The flat syllable structure (Fig. 1-1) in which there are only syllable segments with the identical direct relationship to the syllable node and no sub-constituents, can be found, for example, in Kahn (1976) or in Clements and Keyser (1983). Nevertheless, following the basic ideas and principles of the generative approach to language (see, e.g., Rocca 2013, and see also, above), the flat, linear representation of the syllable as the unit of phonological analysis is not sufficient, and a non-linear, hierarchical representation is preferred (Fig. 12, 1-3). This means that in generative phonology, the syllable is seen as a multitiered unit with an internal hierarchically organized structure. For example, Blevins (1995, 210) provides several pieces of sonority-based, feature-based, and position-based evidence in favour of the hierarchical internal structure of the syllable. The syllable models that have binary 1

The detailed account of the history of the syllable before and after the arrival of the generative approach to language as well as the classification and description of various approaches to the syllable can be found, for example, in Goldsmith 2011 or in Cairns and Raimy 2011. 2 Goldsmith (1976). 3 Liberman (1975). 4 Mohanan (1982). 5 Halle and Vergnaud (1980). 6 Clements and Keyser (1983). 7 Prince and Smolensky (1993).

The Syllable in Generative Phonology

3

branching either with body (Fig. 1-2) or with rhyme (Fig. 1-3) are the most frequently used (ibid.).

Figure 1-1 The Flat Syllable Structure

Figure 1-2 The Hierarchical Syllable Structure–binary Branching with Body

Figure 1-3 The Hierarchical Syllable Structure–binary Branching with Rhyme

4

Chapter One

In general, the syllable theory is based on two hypotheses: (1) universality, which assumes that there are syllables in all languages, and (2) exhaustiveness, which requires all segments in a word to belong to syllables (Cho and King 2003, 183). Another common feature of the syllable theories within generative phonology is that syllabification is not always exhaustive (ibid.; see also Cairns and Raimy 2011, 16) and many consonants are left unsyllabified, that is, they are extrasyllabic.

1.1.1 Extrasyllabic Consonants Syllable theories that accept the existence of unsyllabified consonants (e.g., Clements and Keyser 1983; Rubach 1993; Duanmu 2009) offer various solutions to account for these extra consonants (i.e., extrasyllabic consonants) at syllable edges. One of the easiest solutions is the distinction between the syllable core (or a core syllable) and the appendix, where the Sonority Sequencing Generalization (SSG, see section 3.1) holds only in the core syllable, and the segments violating the SSG are labelled as the appendix. For example, in English, the maximum number of segments at the onset of a core syllable is two and their sonority increases towards the peak. The sonority of segments at the coda decreases from left to right. Peak and coda form a phonological unit, which is called a rhyme. The maximum number of segments in rhymes is three. Any consonants that represent a violation of the pattern described above or to the sonority generalization are referred to as the appendix. The description of the word strings /strԘ‫و‬z/ in terms of its syllable structure is, then, as follows: [s]trԘ‫>و‬z], where [trԘ‫@و‬is the core syllable, the segment ‘s’ is the appendix to the onset, and the consonant ‘z’ is the appendix to the rhyme (Giegerich 1992, 146–50). Extrasyllabic consonants can also become syllabified through vowel epenthesis or may be incorporated into higher prosodic structures when they are extrasyllabic on the surface (Green 2003, 243). The other way of accounting for the extrasyllabic consonants is to remove them by Stray Erasure, which says if a segment cannot be incorporated into a syllable, it should be deleted (McCarthy and Prince 1995, 330). Blevins exemplifies the existence of the so-called stray consonant in English with the monosyllabic word damn. In English, stem C/ø alternations as in damn/damnation and hymn/hymnal can be accounted for by recognizing that *mn is an ill-formed coda

The Syllable in Generative Phonology

5

sequence in English, and hence, the pre-surface representation of /dæmn/ is /dæm.nƍ/ where Cƍ represents an unsyllabified C, which is deleted by stray erasure, resulting in [dæm]. (Blevins 1995, 218)

The latest approach to the problem of surface violations of the SSG by certain consonants is the idea of complex segments8 and the notion of semisyllables. It is supposed that “complex segments in several [...] languages obey both the SSP and ES9 once the notion of semisyllables are incorporated and proper morpheme analyses are given” (Cho and King 2003, 187). This means that the question of the syllable structure is solved together with the question of the morpheme structure because “morpheme boundaries play a crucial role in the distribution of semisyllables” (ibid., 199). Therefore, the following section concentrates on the morphological domain of syllabification.

1.2 Syllabification and Morphology There is no doubt that the syllable is a sound unit, and as a “phonological prime” (Jones 1976, 121), it is the field where many phonological processes take place. One cross-linguistic phenomenon occurs when the independent domain of syllabification is formed by each constituent in a compound and by prefixes; or, in other words, “the fact that wordinternal compound boundaries and prefix boundaries always form edges of a syllabification domain seems to be a universal tendency” (Rubach and Booij 1990, 45). Otherwise, syllabification does not have to respect the morphological structure of words (Laeufer 1995, 103). As for the suffixes, vowel-initial suffixes are syllabified with the preceding morpheme but consonant-initial suffixes are not, and they can (but do not have to) form an independent domain of syllabification. A detailed cross-linguistic comparison has proven that in languages with fixed stress (e.g., Czech and French), the word rather than the morpheme is the domain of syllabification. In contrast, in languages where morphology plays a role in stress assignment (e.g., English), the morpheme is the domain of syllabification (ibid., 118).

8 9

The details about the theory of a complex segment can be found in section 2.2. ES stands for exhaustive syllabification (note made by the author of this work).

6

Chapter One

1.3 The Underlying Representation and the Surface Representation In traditional generative phonology, a single underlying form is capable of accounting for the phonological variations that relate grammatical structures. For example, in the pair of words divine–divinity, there is a single, underlying representation – /divҸn/ – that, together with the rules relating this representation to its surface alternants, accounts for the native speaker’s awareness of the relationship between grammar and phonology (Crystal 2008, 475). The underlying representation or form of a language unit is thus understood as the form stored in the mental lexicon. It is the form used before the application of phonological rules and processes (Matthews 2007, 420). The actual pronunciation is then indicated as the surface form. For the current generative phonology, the underlying form is equal to the phonemic form of a word. For example, (1)

graphic form can’t ĺ underlying form //kænt//10ĺ surface form [NK ચ৉@

Considering this popular example, the underlying representation can be labelled as the phonemic representation of the given word, and the surface form is identical with its detailed phonetic form or representation. This idea of the difference between the underlying form and the surface form is also supported by Mannell, who mentions underlying forms are expressed in terms of phonemes and a surface form is the broad, phonemic transcription of how a word or morpheme is actually pronounced in a particular context, […] phonetic features are surface realizations of underlying phonological features. (Mannell 2008, 3)

A similar interpretation can be found in Odden’s introductory course to phonology. The author differentiates between the spelling of words, their phonetic representation, which is in square brackets, and the underlying forms placed in slanted brackets (Odden 2005, 20). However, several pages later, when refining the notion of underlying forms, the author provides a more precise explanation of what this concept indicates: “The underlying form of a word is whatever comes out of the morphology and is fed into the phonology, before any phonological rules have been 10 The underlying form is typically notated either as double slashes surrounding the phonemic form or in capital letters.

The Syllable in Generative Phonology

7

applied” (ibid., 72). The underlying form is thus seen as a component standing somewhere between morphology and phonology. Obviously, as indicated above, despite their importance in phonological theory, the content of these two terms – the underlying representation and the surface representation – has become slightly ambiguous. That is why it is necessary to explain what the underlying representation is in traditional generative phonology. Chomsky and Halle say in the Sound Pattern of English (1968) that the underlying lexical as well as phonological representation is abstract when compared with phonetic representation, although both are given in terms of phonetic features (Chomsky and Halle 1991, 8, 11). The symbols used in the underlying representation are understood as informal abbreviations for a certain set of phonological categories, that is, distinctive features (ibid., 12). The underlying representations are abstract and, in general, very close to conventional orthography (ibid., 46, 48). This means that sometimes the underlying representation has the form of the orthographic representation of the given word, for example, /erase/ (ibid., 27), but sometimes not, for example, /sinNg/ (ibid., 85). The cases where the underlying representations are identical with the spelling of the given words have led to misunderstanding the underlying forms as mere orthographic notations. Although, [...] English orthography in general, reflects the SPE “underlying forms” rather accurately [it] is not because (as Chomsky and Halle believe: SPE, p. 49) our spelling is a near-perfect encoding of the pronunciation of our words as they are stored in our (subconscious) minds, but because the underlying forms correspond phonetically to the distant ancestors of our modern words, and English spelling is highly conservative. (Sampson 1980, 203)

It seems that some generative theories mix up the graphic and the sound forms of a language. Recall now the word damn and its representation /dæmn/ used by Blevins (section 1.1.1 above). Is this really the form that the native speakers have in their mental lexicon? It is well known that the sound form (speech) is the primary means of communication. Writing is just a device people invented to preserve their thoughts and ideas, etc. (see, e.g., Lyons 1999, 14). It is well known that in languages, there is rarely a one-to-one correspondence between letters and phonemic (sound) units. The particular realization of the individual segments – vowels and consonants – in their linear sequences (words, phrases, and sentences) is influenced by various pronunciation and phonological rules of the given language (e.g., assimilation), and not all

8

Chapter One

combinations of graphemes are reflected in speech. Even native language users may have difficulties with the spelling of the words in their own language.11 The ‘n’ in the word damn is only a grapheme that has no realization in the pronunciation. Which form is part of the mental lexicon, /dæmn/ or /dæm/? Which one is the correct underlying form? The literate speaker will probably automatically connect the form /dæm/ with the orthographic form as an underlier. But what about the speaker that is illiterate? Although “it is not possible to provide direct psycholinguistic evidence for the underlying representations” (Giegerich 1999, 165), the nature of the relationship between the grapheme (orthography) and the sound form (pronunciation) of the word is one that promotes the latter. The spelling-pronunciation analysis has shown “that orthographic information in many cases facilitates speakers’ access to items stored in the mental lexicon” (ibid., 166). However, a number of research studies provide contrary results (Gibson et al. 1962). Analysing Chomsky and Halle’s (1968) Sound Pattern of English, it is clear that in generative phonology, there is no place for the third distinct level between the level of morphophonemes (underlying representations) and the phonetic level (surface representations), and thus “the phoneme level must go” (Sampson 1980, 200). Simply said, the level of phonemic representation is not included in the theory of generative phonology proposed by Chomsky and Halle (1968, as cited in Mohanan 1995, 26). However, apart from the abstract level of morphophonemes and the concrete level of physical phonetics, they too assumed the existence of “the level of universal binary distinctive features, or what Halle and Chomsky call ‘systematic’ (as opposed to ‘physical’) phonetics” (ibid.). The level of systematic phonetics12 encompassing the universally fixed and finite set of possible distinctive features, although not explicitly present in the SPE, is the level of phoneme. Thus, the discussion about the dichotomy of surface form – underlying form can be concluded by the assumption that in traditional generative phonology, surface forms represent the actual pronunciation and underlying forms refer to the abstract level of morphohoneme, with phonology hidden somewhere in between, as illustrated by Mohanan’s (1995) schema (Fig. 1-4).13 11

Writing is a matter of convention and has to be learnt. Some generative phonologists use the term “systematic phonemics” instead of this original term, “systematic phonetics”, and some of them use both to refer to different levels of representation in generative phonology (see, e.g., Crystal 2008). 13 It is important to note here that “phonemics” was not equivalent to “phonology” in early generative phonology (for details see, e.g., Mohanan 1995). 12

The Syllable in Generative Phonology

9

Figure 1-4 The Position of Phonology in Classical Phonemics and in the SPE (Mohanan 1995, 37)

Nevertheless, in the 1970s and 80s, the necessity to reintroduce the level of phonemic representation into phonology appeared. Lexical Phonology introduced the idea that “even though phonological rules/constraints themselves do not fall into types, the interaction between phonology and morphology leads to a level of phonological representation that is distinct from underlying and phonetic representations” (ibid., 26).

1.4 Universality and Diversity It is well known that we tend to look at other languages from the viewpoint of the language we speak; we tend to analyse other languages from the viewpoint of the categories suitable for the language of which we are native speakers. This tendency to analyse all languages in terms of the categories and principles valid in one’s own language is widespread in generative phonology. Chomsky’s universal grammar is based on the idea that part of our knowledge of language is innate. In detail, Chomsky argues that “the explanation for the fact that all languages of the world are cut to a common pattern (assuming that they are) is that the inherited structure of Man’s mind forces him to use languages of that particular type” (Sampson 1980, 147). And, thus, many linguists think that “according to Chomsky, a visiting Martian scientist would surely conclude that aside from their mutually unintelligible vocabularies, Earthlings speak a single language” (Pinker

10

Chapter One

1994, 232); that is, following generative tradition, there is a common belief that “all languages are English-like but with different sound systems and vocabularies” (Evans and Levinson 2009, 429). As a consequence, it has become popular among linguists to attribute the results of their research as language-universal, ignoring the fact that considerable “diversity can be found at almost every level of linguistic organization” (ibid., see also section 7.3 in this work). I agree with Sampson (1980, 148) when he says that Chomsky is right about the existence of a certain degree of universality in linguistic structure but “the existence of linguistic universal is, for Chomsky and his followers, not so much a finding, which has emerged from their research despite their expectations, but rather a guiding assumption, which determines the nature of the hypothesis they propose in order to account for data” (ibid.). Therefore, generative linguistics suggests “universalist” theories and terms that may have the “non-universalist” explanations if we wish. Sampson (1980) calls this tendency a “rush to universals” (149). It is widely known that although native language users can usually intuitively count the number of syllables in a word or a word-form, linguistic definitions of the substance, origin, structure, and function of this unit are not uniformly subscribed to (cf., e.g., Abercrombie 1967, 34; Romportl 1985, 107; Roach 2000, 70). The existence of the syllable as a sound unit can be considered a language-universal phenomenon (cf., e.g., DubČda 2005, 135). Most linguists agree that the theory of the syllable should be able to cover the wide extent to which syllable types vary crosslinguistically and, at the same time, account for those aspects of the syllable structure that remain constant across languages (Blevins 1995, 213). The relationship between the language-particular and languageuniversal features of syllables and the whole process of syllabification has been best expressed by Pulgram: If the syllable is an operative unit of all languages, it is also a universal of language. Its definition must be […] the same for all languages, regardless of the varying unit inventories in the different [languages] … there arises the interesting question whether it might not be possible to arrive at a phonotactic definition of the syllable, which […] does have universal validity for all languages. The question is, in other words, whether the phonotactic rules on syllabation might not be formulated in such a way that they are applicable to all languages, even though their implementations in the different languages must differ because of the underlying differences of phonotactics. I believe that such general phonotactic rules on syllabicity are not only possible but also necessary for the proper syllabation of any utterance in any language. (1970, 23)

The Syllable in Generative Phonology

11

These conventions governing the position of the syllable boundaries, or in other words, these syllabification algorithms, fall into two basic categories: (1) rule-based algorithms or approaches to the syllable structure, and (2) template-matching algorithms or approaches to syllabification (Blevins 1995). Template-matching algorithms for syllabification scan the segmental string in a fixed, language-particular direction (left to right, right to left), assigning successive segments to positions in a syllable template, always mapping to as many positions inside a given syllable template as possible. Rule-based algorithms posit an ordered set of structure-building rules which have similar status to that of other phonological rules: such rules may or may not apply directionally and do not require that syllable structure be maximalized in any sense from the start. (ibid., 217)

The most up-to-date, template-matching approach to syllable structure14 is represented by the CVX syllable theory by San Duanmu (2009). This theory comprises the ideas of several theories of phonological representation, especially prosodic phonology, autosegmental phonology, metrical phonology, and articulatory phonology (cf. ibid.). The CVX syllable theory can thus be seen as the present-day representative of the generative approach to syllable structure. This is also the reason why the theory has been chosen for the analysis and evaluation of the data in this work (Chapter 2). The most popular rule-based approach to syllabification in the framework of generative phonology is the Syllable Structure Algorithm (SSA) offered by Lexical Phonology that, although “proposed as the refinement of classical generative phonology” (Goldsmith and Laks 2011, 14), incorporates insights from pre-generative structuralist phonology and natural phonology (ibid.). Thanks to this “synthetic aspect” (ibid.), this generative approach to syllable structure is also presented in this work (Chapter 3).

14

Of course, “the most up-to-date” is relevant to the time of writing this work.

CHAPTER TWO THE CVX THEORY OF THE SYLLABLE

2.1 An Outline of the Theory The CVX theory of syllable structure (Duanmu 2009) assumes that the maximal syllable size in all languages is only CVX,1 that is, CVC or CVV (Figure 2-1), and accounts for any extra consonants at word edges either by morphology or by treating them as complex sounds.

Figure 2-1 The Structure of the Syllable in the CVX Theory

As the author specifies, this morphological approach to the syllable is based on the assumption that “there is a correlation between the 1 C stands for a consonant, V indicates a vowel, and X is a symbol denoting either a vowel or a consonant.

The CVX Theory of the Syllable

13

morphology of a language and the maximal syllable size” (ibid., 52). All word-edge consonants that exceed this three-slot syllable size can be explained by one of the three morphological concepts: the Affix Rule (section 2.1.1), the Potential Vowel Rule (section 2.1.2) and AntiAllomorphy (section 2.1.3). However, the most important role in this theory is played by the notion of complex sound (section 2.1.4).

2.1.1 The Affix Rule This rule states that “affix or affix-like sounds can be pronounced, whether they can fit into a syllable or not” (Duanmu 2009, 50). For example, the final consonant [s] in the English monosyllabic word pets [pets]2 is not part of the syllable whose structure is only CVC. This final [s] that represents the sound form of the real suffix -(e)s is solved by morphology (the Affix Rule). Similarly, the syllable structure of the word ax [æks] is only VC because the final [s] can be explained by morphology too. It is an affix-like sound or a “perceived suffix” covered by the Affix Rule (Duanmu 2010, 8): (1)

pets [pets] – syllabic structure [CVC]C ax [æks] – syllabic structure [VC]C

2.1.2 The Potential Vowel Rule The Potential Vowel Rule is also used for the morphological prediction of extra consonants at word edges: in languages having suffixes starting with a vowel, an extra consonant is allowed in a word-final position. This consonant can function as the onset of the suffix vowel. Analogically, in languages that have prefixes ending in a vowel, an extra consonant can be in a word-initial position to form a coda of the prefix with a vowel at its end (Duanmu 2009, 70, 150). For example, the final [p] in the word help [help] is an extra C when the word is in isolation (supported also by AntiAllomorphy; see below), but this [p] functions as the onset of the following V in the word helping – [hel]p, [hel][pܼƾ] (Duanmu 2010, 10): (2)

2

help [help] – syllabic structure [CVC]C helping [helpܼƾ] – syllabic structure [CVC][CVC]

In this chapter, the sound form of words and segments as well as the syllable structure of words is inserted within square brackets to be consistent with Duanmu’s theory (2009, 2010).

14

Chapter Two

2.1.3 Anti-Allomorphy Anti-Allomorphy suggests keeping “a morpheme in the same shape, regardless of the environment” (Duanmu 2009, 47). This rule supports the syllabification of help [help] as CVC + an extrasyllabic consonant – [hel]p (ibid., 47).

2.1.4 Complex Sounds In some languages, for example, Chinese, there are no prefixes. This means that the first consonant in the word-initial CC clusters cannot be accounted for by the Affix Rule or by the Potential Vowel Rule (Duanmu 2010, 13). To keep from extending the maximal syllable size of CVX, Duanmu proposes the representation of a sequence of sounds as one timing unit, that is, as a complex sound or a complex segment.

2.2 The General Theory of Complex Segments The notion of a complex segment has its stable position in non-linear phonology, which, contrary to linear phonology which prefers featurematrix representations, is based on a hierarchical representation of distinctive features and relations (for further details, see, e.g., McCarthy 1982). The simple definition of a complex segment says that it is a type of segment “with unordered or simultaneous multiple articulations” (Sagey 1986, 2). In detail, a complex segment is a segment with multiple articulations that has a phonetic duration of a single segment (Sagey 1986, 79; Newman 1997, 8) and phonemically occupies only one X-slot, that is, one timing unit in an autosegmental sense (Scheer 2012, 868), and thus it behaves like a simple sound (Newman 1997, 9). However, there is no general agreement on which sequences of sounds can be treated as complex segments and which cannot. Comparing several sources, one may come to the conclusion that what is a complex segment, for example, according to Sagey’s (1986) classification, can be treated as a consonant cluster in Duanmu’s (2009) approach and vice versa (for details on the discussion about defining a possible complex sound, see Scheer 2012). All authors agree that “a complex segment is a root node characterized by at least two different oral articulator features […]” (Clements and Hume 1995, 253). For example, the labio-coronal complex segment /p੾t/:

The CVX Theory of the Syllable

15

Figure 2-2 A Labio-Coronal Complex Segment /p੾t/ (Hall 2006, 331)

The specification of the other arguments for the monosegmental classification of traditionally named consonant clusters is considerably different (see, e.g., Sheer 2012 for further details). Because the main aim of this chapter is to verify the assumed universal nature of the CVX theory, which is based on the idea of a complex segment, Duanmu’s (2009) concept of a complex or a simplex character of the sound sequences is taken into consideration throughout the whole analysis. The surface representation of some consonant clusters as complex segments is based on Feature Geometry (for details, see, e.g., Newman 1997, Marlo 2004, or Scheer 2012). That is why in the following section I briefly introduce the main idea of this theory.

2.2.1 Feature Geometry Feature Geometry (FG) has its roots in autosegmental phonology – introduced by Goldsmith (1976) – which represents sounds, their features, and their phonological processes in a multitiered, hierarchical construct. It is based on the idea that features are grouped into higher-level functional units, the so-called constituents (Clement and Hume 1995, 247–48; see also Sagey 1986). Instead of the traditional method of placing features into matrices, in this approach, each segment is presented as a hierarchically organized node configuration whose terminal nodes represent feature values and whose intermediate nodes represent constituents (Clement and Hume 1995, 247–48), as illustrated by figure 2-3.

16

Chapter Two

Figure 2-3 A General Model of Feature Organization (Clement and Hume 1995, 248)

A root node (A) stands for the speech sound itself. The lower-level class nodes (B, C, D, E) designate functional feature groupings that include the laryngeal node and the place node (ibid.). Lower case letters (a, b, c...) stand for the individual features of a sound. This means that in FG, the individual properties of a sound are organized in a hierarchy. Each sound, that is, each root node, has a Laryngeal node (features connected with the state of the larynx) and a Supralaryngeal node (features associated with the state of the vocal tract above the larynx). A Supralaryngeal node (abbreviated Supra node) splits into the Soft Palate node (abbreviation, Soft Pal) – which carries information about the soft palate – and the Place node (related to the place of articulation). The Place node further branches into Labial, Coronal, and Dorsal3 nodes. While Labial covers bilabial and labiodental sounds, Coronal indicates interdental, dental, alveolar, palato-alveolar, and palatal sounds, and Dorsal applies to velar and uvular sounds (Newman 1997, 3– 7). The nodes mentioned so far are the so-called class nodes that may, but do not have to, be further extended. Features that appear under these nodes are known as terminal features because they do not expand further into other features. There is an important condition for the use of terminal features: “each terminal feature occurs under one and only one class node” (Newman 1997, 13; see also Sagey 1986, 207).

3 For the possibility of a more detailed specification of the place of articulation in Feature Geometry, see, e.g., Newman 1997, 6–7.

The CVX Theory of the Syllable

17

The terminal feature that appears under the Laryngeal node is [+/voice], which represents the presence or absence of voicing during articulation. This node also encompasses two other terminal features: [+/spread],4 referring to the spread glottis, and [+/-constricted], referring to the constricted glottis (Clements and Hume 1995, 258). The Soft Palate node is connected with the terminal feature [+/-nasal], which relates to the nasal or the oral character of the given speech sound (Newman 1997, 13–18). And finally, the remaining nodes are connected with the place of articulation: the Labial node contains the terminal feature [+/- round], which specifies production with or without the protrusion of the lips; the Coronal node branches into the terminal feature [+/-anterior], referring to the articulation in front of the alveolar ridge [+anterior] or behind it [-anterior], and [+/-distributed]5, which is associated with the timing of a constriction; the Dorsal node covers the features [+/-high], [+/low] and [+/-back] because all of them are carried out with the back of the tongue (Sagey 1986, 61). The features linked under the Place node in feature hierarchy are also called place features. They have their own node, that is, their own phonological representation (Labial, Coronal, Dorsal), and they do not operate on negative values; they are treated as privative (Clements and Hume 1995, 200). Other features fall into two categories. They are either articulatorbound features, that is, those located under the appropriate articulator node (e.g., [anterior], [distributed]) or articulator-free features that are not restricted to a specific articulator and indicate the degree of stricture. They are also known as stricture features. These are the features [+/continuant], [+/-sonorant], [+/-approximant], and [+/-consonantal]. The articulator-free features are usually placed higher in the hierarchy than articulator-bound features (Clements and Hume 1995, 250–63), as specified in figure 2-4. When describing the individual phonemes, only relevant class nodes and terminal nodes are represented, depending on the phonetic properties of sounds. The above-described theory of distinctive features is also known as an articulator-based feature theory “since the articulators play a fundamental role in the organization of segment structure” (ibid., 249). Articulators, as the movable parts in the vocal tract, participate in speech production, and

4

In Feature Geometry, the square brackets are used for the notation of terminal features. See also note 2. 5 The terminal feature [+/-distributed] is replaced with the feature [+/-lateral] in Marlo’s (2004, 80) approach.

18

Chapter Two

the gestures made by these articulators constitute features (Marlo 2004, 79).

Figure 2-4 A Formal Model of Articulator-based Feature Geometry (cf. Sagey 1986; Clements and Hume 1995; Marlo 2004)

2.2.2 A Complex Segment in Duanmu’s Theory As already mentioned (see above), the simple generally accepted definition of a complex segment is “a root node characterized by at least two different oral articulator features, representing a segment with two or more simultaneous oral tract constrictions” (Clements and Hume 1995, 253). Duanmu defines a speech sound, that is, a minimal time unit, by two important factors: “A speech sound is a set of features such that (a) the features are made in one unit of time and (b) no feature is made twice by the same articulator” (2009, 17). This delimitation of a sound is the basis for his specification of a complex segment.6 In Duanmu’s concept of a complex sound, the class nodes from Feature Geometry (Fig. 2-4) are named contour nodes or simply articulators and they are governed by the 6 Instead of the term “complex segment”, Duanmu (2009, 2010) uses the term “complex sound”.

The CVX Theory of the Syllable

19

No Contour Principle, according to which “an articulator cannot make the same feature twice within one sound” (2009, 174). In other words, “two sounds cannot form a complex sound if they have conflicting gestures” (Duanmu 2010, 15). This principle – as the author further adds – makes two predictions. The first one is that an articulator, or node, cannot make the opposite values of the same feature within one sound, that is, within one timing slot. It means that, for instance, a single sound cannot be characterised simultaneously by both [+nasal] and [-nasal] or by [+anterior] and [-anterior] because conflicting gestures cannot overlap; they must be made in sequence and require more than one timing slot (ibid., 16). For example, [b] is characterised by Labial [-nasal], while [m] is Labial [+nasal]; therefore [bm] cannot form a complex sound. On the other hand, [f] is [+fricative]7 and [r] is [-fricative], but the articulator for [f] is Labial and that for [r] is Coronal. There are no conflicting gestures because there are two articulators, and [f੾r] is a possible complex sound (ibid., 17). The second prediction is that all gestures within one sound are simultaneous; for example, [k੾p] is the same as [p੾k] (ibid., 17). This “simultaneity” makes it possible for a sequence of sounds to be realized within one timing unit. A complex sound, thus, is a merger (gestural overlap) of at least two sounds that are either produced by different articulators or by the same articulator, but without conflicting values of the same feature, for “if the articulatory gestures of two sounds can overlap, then they can form a complex sound” (ibid., 5). Overlapping gestures are made simultaneously; a complex sound then takes just one timing slot (Duanmu 2010). A special type of complex segment is represented by affricates, which are also traditionally treated as single segments (see, e.g., Sabol 1989). In Duanmu’s (2009) approach, affricate sounds /ࠔ/8 and /˔/ are made with the simultaneous features [+stop] [+fricative] with the single place articulator Coronal. This representation reflects the phonetic nature of affricate sounds (the combination of a closure and a narrowing) and gives them the status of being structurally complex (multiple articulations) but functionally simple segments (single element on the timing tier) (Sagey 1986, 93, 96). The same applies to the sounds /tr/ and /dr/, which are also 7

As already mentioned (see section 2.2.1), the so-called stricture features are considered to be articulator-free. However, Duanmu (2009, 23) supposes that “each articulator can have its own values for stricture features [stop] and [fricative]” and connects these features directly with the Place nodes. 8 Slanted brackets // are used for the notation of phonemes in this chapter (cf. section 1.1.1 and note 10 in Chapter 1).

20

Chapter Two

classified as affricates (phonetic affricates) by Duanmu (2009, 183). The sounds /ࠔ/, /˔/, /tr/ and /dr/ are affricates with the same articulator (cf. Table 2-1). The second group of affricates is created by those with different articulators, like /ps, /px/, /ks/ and /kߦ/. Duanmu (2009, 28) argues that “any stop can combine with any fricative to make an affricate […] [because] each articulator can have its own values of [stop] and [fricative]”. Thus, one articulator can make a stop gesture and another can make a fricative gesture (ibid., see also note 7). It is important to mention here that some authors, for example, Sagey (1986), whose delimitation of a complex segment is frequently quoted in a number of relevant works on non-linear phonology (see, e.g., Scheer 2012), talk about languages with complex segments (the Niger-Congo languages) and languages without complex segments (Sagey 1986), including English. But Duanmu (2009), whose CVX theory of the syllable is based predominantly on the possibility to represent consonant clusters as complex sounds, that is, segments filling in one timing slot only (for details, see section 2.4), considers all languages to be languages with complex sounds. However, the author’s analysis of English consonant clusters in terms of the notion of a complex segment is not sufficient. As mentioned by Scheer (2012) in his review of Duanmu’s book, “the reader is left guessing” (698) and there is no how-to that allows the reader to understand, for any given sound, which features exactly depend on which articulators. The internal structure of six consonants (p, m, s, ‫ݕ‬, t੾ߦ l) and four vowels (a, i, ܼ, u) is made explicit on page 24, and a few more identities are provided on the following pages. But in absence of the key to the distribution of features over articulators, the predictions of the system (i.e., which complex sounds should be possible and which ones are ruled out) cannot be computed. Blevins (2010: 289) makes the same observation. (701–2)

Because the simplification of some consonant clusters into complex segments is one of the basic pillars of the CVX theory, first it is important to find out – by the method of the phonemic and the phonetic analysisʊif the notion of a complex segment is a phenomenon connected with certain languages only (as suggested by Sagey 1986) or if it is a languageuniversal phenomenon (as supposed by Duanmu 2009). As specified in the Introduction to this work, two languages will be used for the analysis: the English language, whose consonant clusters seem to fulfil Duanmu’s phonemic criteria for a complex-segment analysis, and the Slovak

The CVX Theory of the Syllable

21

language, whose consonant clusters are heterogeneous and whose phonetic and phonemic specifications of speech segments are slightly different from those in English.

2.2.3 A Consonant Cluster or a Complex Segment? Duanmu’s (2009) assumption about the universal syllable structure, CVX, is based on the evaluation of the data from Standard Chinese, Shanghai Chinese, Jiarong – belonging to the Sino-Tibetan language family – and German and English, which are both West Germanic languages. It is well known that morphology and phonotactic possibilities and the constraints of genetically-related languages are usually at least partly akin. At this point the question arises if the CVX model of the syllable is applicable also to the other typologically different languages, for example, to the Slovak language, which is highly inflectional and allows relatively long sequences of consonants. Moreover, the articulatory basis of the Slovak language is completely different from that of Jiarong, for example. In addition, as mentioned in section 2.2.2, Duanmu (2009) does not specify which consonant sequences can be represented as complex segments. That is why, in order to find the key to this analysis, I decided to re-analyse all English word-initial CC clusters from the viewpoint of Duanmu’s conception of a complex sound. Then I applied the same procedure to the Slovak word-initial CC clusters. Consequently, a selected sample of the word-initial CC sequences from two languages9 that can phonemically be represented as complex segments are analysed phonetically, so as to prove or disprove the existence of complex sounds in English and in Slovak. The results of the phonemic and phonetic analyses of some onset CC sequences are commented on at the end of this section. However, prior to the complex-sound analysis of the English and the Slovak consonant clusters, it is necessary to describe the maximal phonological structure of the English and the Slovak syllable, that is, to delimit the onset and the coda clusters in both languages.

9

The complex picture of the maximal phonological structure of the English and the Slovak syllable can be found in section 2.3.

22

Chapter Two

2.3 Word-initial and Word-final Clusters in English and in Slovak 2.3.1 What is a Consonant Cluster? A consonant cluster is a group or sequence of consonants that appear together in the syllable without a vowel between them (see, e.g., Jones 1976). It can be studied in terms of graphemes, phones, and phonemes. The critical unit of a language for its production and perception is the syllable (Gibson et al. 1976), and the syllable is the fundamental unit of phonological analysis (cf., e.g., Jones 1976; Blevins 1995, see also Chapter 1). By implication, the following survey of the consonant clusters in English and in Slovak is based on the analysis of phonemes.10 The possible combinations of consonants are usually studied in three different positions: word-initial, word-medial, and word-final. In some languages, the word-medial position accepts combinations that are not plausible at the beginning or the end of a word. It seems that the middle part of words is the least “immune” (Sabol 1969a, 129) against the occurrence of consonant clusters (ibid.). Because the medial consonant clusters are usually “decomposable into legal initial or final clusters” (Hammond 1999, 69), they can be viewed as the combinations of the initials and the finals that provide a model “in terms of which medial clusters can be characterized” (Jones 1976, 121). This means that all initial and final clusters are tautosyllabic, but the word-medial are not. That is why the word-medial consonant sequences are treated separately in this analysis.

2.3.2 Consonant Clusters in English The comparison of several sources11 has shown that a word, that is, a syllable, in English can begin with a vowel or with one, two, or three 10 The transcription symbols valid in the given language are used for the graphic representation of phonemes in this work. The complete list of the English and the Slovak phonemic transcription symbols for consonants can be found in Appendix 1. 11 To get a complete picture of the initial (onset) and final (coda) clusters in English, the following sources have been compared: Giegerich (1992), who analyses consonant clusters in terms of generative phonology; Roach (2000), whose analysis of possible phoneme combinations is based on the traditional structural approach; Duanmu (2009), who supplements Giegerich’s and Roach’s descriptions of the phonotactic possibilities of English with aspects of the Optimality Theory and gives the reliable statistical data; and one internet source

The CVX Theory of the Syllable

23

consonants. No word (syllable) in English begins with more than three consonants (Roach 2000, 71). Thus the maximum number of segments in the word-initial consonant cluster in English is three. At the beginning of English words (syllables), there are 55 twoconsonant clusters. In many cases, the first element is /s/ and the second consonant is one of the approximants /l, r, w, j/ (Roach 2000, 73; Duanmu 2009, 160). Initial CC clusters in English starting with oral plosive: pr, pl, pj, pw12, pf, ps, pߦ, br, bl, bj, tr, tw, tj, dr, dj, dw, kr, kl, kw, kj, km, kn, kv, gr, gl, gw (26) starting with nasal plosive: nj, mj, mw (3) starting with fricative: fl, fr, fj, vj, vw, șr, șw, șj, st, sp, sk, sl, sw, sn, sm, sf, sj, sr, sv, zl, ߦr, ߦm, ߦn, ߦp, ߦw, hj (26)

The number of the initial three-consonant clusters in English is quite limited; there are nine of them, all starting with a consonant /s/. Initial CCC clusters in English starting with fricative /s/: spl, spr, spj, str, stj, skl, skr, skw, skj (9)

A word, that is, a syllable, in English can end with a vowel or with one, two, three, or four consonants (Roach 2000, 73). The maximum number of consonants in the final consonant cluster is four. Comparing the sources of consonant clusters (see note 11) shows that there are 55 final two-consonant clusters in English. They usually end with one of the consonants /s, z, t, d, ș/, which represent separate morphemes (Roach 2000, 73); /s, z/ are the sound forms of the ending -(e)s, while /t, d/ stand for the ending -(e)d.

(http://www.btinternet.com/~ted.power/clustersindex.html), which offers a practical list of some consonant clusters in the English language. 12 Many linguists claim that there is no ‘pw’ cluster in the word-initial position in English (see, e.g., Cho and King 2003, 185). However, Duanmu (2009, 160) includes this sequence of consonants in the list of the English consonant clusters. Because the aim of the analysis presented here is to be in accordance with Duanmu’s research, the given cluster is part of the analysis too.

Chapter Two

24

Final CC clusters in English ending with consonants [t]13 and [d]: pt, bd, kt, gd, mt, md, nt, nd, ƾt, ƾd, ft, vd, ðd, st, zd, ߦt, ‫ݤ‬d, lt, ld, ࠔt, ‫ݶ‬d (21) ending with consonants [s] and [z]: ps, bz, ts, dz, ks, gz, mz, ns, nz, ƾz, fs, vz, șs, ðz, ls, lz (16) ending with consonant [ș]: pș, tș, kș, nș, ƾș, fș, lș (7) ending with some other C: mp, mf, nࠔ, n‫ݶ‬, ƾk, ƾg, sp, sk, lp, lf, lk (11)

The final three-consonant clusters are quite numerous, too – there are 39 of them. They usually end with /s, z, t, d/. Final CCC clusters in English ending with consonants [t] and [d]: pst, tst, kst, mft, mst, nࠔt, n‫ݶ‬d, ƾst, lmd, lpt, lbd, lft, lvd, lnd, lࠔt, l‫ݶ‬d, lߦt, lkt (18) ending with consonants [s] and [z]: pșs, pts, tșs, kșs, kts, mps, mfs, mts, nts, ndz, ƾts, fșs, fts, sts, lmz, lps, lbz, lșs, lnz, ldz, lks (21)

The final four-consonant clusters in English are usually formed by the three-consonant cluster not containing final /s, z, t, d/ as separate morphemes and the sound form of the suffixes -(e)s and -(e)d. Final CCCC clusters in English ending with consonant [s]: ksșs, ksts, mpts, lfșs, ltst, lkts (6) ending with consonant [t]: ntst (1)

2.3.3 Consonant Clusters in Slovak The structure and the number of consonant clusters in Slovak are slightly different from the English consonant clusters. The enumeration of the Slovak consonant clusters is based on the work of Sabol (1969a, 1969b), Petrišþáková (2006), and the online version of The Short Dictionary of Slovak Language (KSSJ) as well as the Slovak National Corpus. A Slovak word (syllable) can begin with a vowel or with one, two, three, or four consonants. The maximal size of the word-initial consonant cluster in Slovak is CCCC. Five-consonant clusters are very rare and cannot be treated as monosyllabic. For example, the words zmrzlina “icecream” and štvrĢ “quarter” contain the combinations of five-consonant 13

Square brackets indicate a phoneme (see also note 2).

The CVX Theory of the Syllable

25

segments. In the word žblnknúĢ “to plop”, there is a sequence of six consonants but these sequences of consonants are legal clusters only on the graphematic level. The sounds r and l – which are consonants, phonetically – have a very high degree of sonority (see Table 3-2) and can form the centre of the syllable, that is, they function as vowels when surrounded by consonants. To put it simply, they are vowels, phonemically, such as in the words zmrzlina, štvrĢ, and žblnknúĢ. In the word nervstvo, “the nervous system”, the five-consonant cluster occurs in the word-medial position and is heterosyllabic14 too: (3)

zmr-zli-na: zm – onset, r – centre, zl – onset of the 2nd syllable štvrĢ: monosyllabic word, štv – onset, r – centre, Ģ – coda

In Slovak, there can be 116 two-consonant clusters, 23 three-consonant clusters, and two four-consonant clusters15 at the beginning of words (syllables). Initial CC clusters in Slovak starting with oral plosive: ps, pš, px, pn, pĖ, pl, bć, bz, bl, bĐ, br, tk, tx, tv, tm, tl, tĐ, tr, dv, dm, dn, dĖ, dl, dr, kt, kv, km, kn, kĖ, kr, kl, kĐ, gn, gĖ, gl, gĐ, gr (37) starting with nasal plosive: mn, mĖ, ml, mĐ, mr (5) starting with fricative: sp, st, sĢ, sk, sx, sv, sm, sn, sĖ, sl, sĐ, zb, zv, zn, zĖ, zl, zr, šp, št, šĢ, šk,šm, šn, šl, šĐ, šr, žv, žm, žĖ, žĐ, žr, hn, hĖ, hl, hĐ, hr, hm, hv, xc, xv, xm, xl, xĐ, xr, ft, fĢ, fþ, fs, fš, fl, fĐ, fr, vd, vz, vn, vĖ, vl, vr, lk, lž (60) starting with affricate: cv, cĢ, cm, cn, cĖ, cl, cĐ, þp, þv, þm, þn, þl, þĐ, þr (14)

Initial CCC clusters in Slovak starting with oral plosive: tkv (1) starting with nasal plosive: mdl (1)

14

Inasmuch as the different conception of consonant clusters on the graphematic and phonematic (and also phonetic) level requires phonological reconstruction of consonantal neutralizations, this issue remains open for further analysis. 15 The list does not report all possible consonant clusters in the Slovak language (see, e.g., Gregová 2008). To follow Duanmu’s (2009, 160–1) criteria specifying which consonant clusters should be treated as initial (see also section 2.2.1) and which as final (149), clusters that violate these criteria or are otherwise exceptional as well as less frequent sequences of consonants were excluded from the analysis.

Chapter Two

26

starting with fricative: vzd, vzl, str, skl, skĐ, skv, stv, smr, stl, zdr, zvl, zbr, zhl, škr, štv, škv, štr, špĐ, špr, hml, hmĐ, Đst (21)

Initial CCCC clusters in Slovak starting with oral plosive: pstr, pštr (2) 

The number of possible word-final consonant clusters is lower than the number of the word-initial clusters. There are 53 final two-consonant clusters and four word-final three-consonant clusters. The maximal size of a coda in Slovak is CCC. Final CC clusters in Slovak16 starting with oral plosive: ps, kt, ks (3) starting with nasal plosive: mp, mb, mf, nt, nd, nk, ng, nc, nþ, nš (10) starting with fricative: st, zd, sĢ, zć, sk, št, žd, šĢ, žć, ft, vk, xt, lp, lt, ld, lk, lc, ls, lz, lf, lm, rp, rt, rd, rk, rc, rþ, rs, rz, rš, rf, rv, rm, rn, rĖ, jt, jd, jk, jf, jn (40)17

Final CCC clusters in Slovak starting with nasal plosive: nkt (1) starting with fricative: jsĢ, jzć, jsk (3)

Having delimited the inventory of the possible word-initial and word-final consonant clusters in English and in Slovak, I can now go back to the phonemic and phonetic complex-sound analyses of the onset CC clusters, as specified in section 2.2.3.

16

Because none of these CC clusters or segments in them can be treated as affix or affix-like sounds, their division, contrary to English, is based on the manner of articulation of the individual consonants. This type of classification is more useful for a complex sound analysis that is necessary for the morphological explanation of word-final clusters in Slovak. 17 Most of these CC clusters are the result of the inflectional processes and contextual occurrences of words. Otherwise, the number of word-final CC clusters in the basic forms of Slovak words is slightly lower (cf. Gregová 2008).

The CVX Theory of the Syllable

27

2.4 A Phonemic Complex-sound Analysis of the Word-initial CC Clusters As explained above, a sequence of consonants can be treated as a complex segment when the consonants are either produced by different articulators (different class nodes or contour nodes) or are produced by the same articulator but without conflicting gestures. In both cases, the gestures, that is, the features (terminal features), can overlap and fit in one timing slot.

2.4.1 A Phonemic Complex-sound Analysis of the English Word-initial CC Clusters Distinctive features depend on the phonetic properties of a sound. The relevant phonetic characteristics of the English consonant phonemes are captured in table 2-1. Table 2-2 presents the feature specification of these phonemes. Because Duanmu (2009) does not specify which terminal features (gestures) under which node are considered relevant for the complex-sound analysis, the entire specification of the individual English consonants in terms of Articulator-based Feature Geometry can be found in table 2-3. Table 2-1 The Phonetic Classification of the English Consonant Sounds (cf., e.g., Giegerich 1992; Roach 2000) sound

place of articulation

manner of articulation

acoustic impression

p b m w

bilabial bilabial bilabial bilabial

plosive plosive nasal plosive approximant

f v ݇ ð t d n s z

labio-dental labio-dental dental dental alveolar alveolar alveolar alveolar alveolar

stop stop nasal stop frictionless continuant constrictive constrictive constrictive constrictive stop stop nasal stop constrictive constrictive

participation of vocal cords (voice) voiceless voiced voiced voiced

fricative fricative fricative fricative plosive plosive nasal plosive fricative fricative

voiceless voiced voiceless voiced voiceless voiced voiced voiceless voiced

Chapter Two

28 r

post-alveolar18

l ‫ݕ‬ ‫ݤ‬ ‫ݹ‬ ‫ݶ‬ j

alveolar post-alveolar19 palato-alveolar palato-alveolar palato-alveolar palatal

k g ƾ h

velar velar velar laryngeal/glottal

frictionless continuant lateral constrictive constrictive constrictive semi-occlusive frictionless continuant stop stop nasal stop constrictive

approximant

voiced

lateral fricative fricative affricate affricate approximant

voiced voiceless voiced voiceless voiced voiced

plosive plosive nasal plosive fricative

voiceless voiced voiced voiced

Table 2-2 Feature Specification of the English Consonant Phonemes20 (Halle 2006) phoneme stricture p b m w f v ݇ ð t d n s

18

[+stop] [-son] [-cont] [+stop] [-son] [-cont] [+stop] [+son] [-cont] [+son] [+cont] [+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [+stop] [-son] [-cont] [+stop] [-son] [-cont] [+stop] [+son] [-cont] [+fricative] [-son] [+cont]

distinctive features vocal velum cords [-voice] [-nasal] [+voice] [-nasal] [+voice] [+nasal] [+voice] [-nasal] [-voice] [-nasal]

Labial [-round] Labial [-round] Labial [-round] Labial [+round] Labial [-round]

[+voice]

[-nasal]

Labial [-round]

[-voice]

[-nasal]

Coronal [+anterior]

[+voice]

[-nasal]

Coronal [+anterior]

[-voice] [+voice] [+voice] [-voice]

[-nasal] [-nasal] [+nasal] [-nasal]

Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+anterior]

place

The place of articulation of the phoneme /r/ is not unified in British English. For a different classification, see, for example, McMahon 2002, 53. 19 Or palato-alveolar in the current IPA terminology (see also note 22 on the Slovak consonants). 20 The description of the distinctive features used in the matrix can be found in Appendix 2.

The CVX Theory of the Syllable [+fricative] [-son] [+cont] [+son] [+cont] [+son] [-cont]

z r l ‫ݕ‬

[+voice]

[-nasal]

Coronal [+anterior]

[+voice] [+voice]

[-nasal] [-nasal]

[-voice]

[-nasal]

Coronal [-anterior] Coronal [+ant] [+lateral] Coronal [-anterior]

[+voice]

[-nasal]

Coronal [-anterior]

‫ݹ‬ ‫ݶ‬ j

[+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [-son] [-cont] [-son] [-cont] [+son] [+cont]

[-voice] [+voice] [+voice]

[-nasal] [-nasal] [-nasal]

k

[+stop] [-son] [-cont]

[-voice]

[-nasal]

g

[+stop] [-son] [-cont]

[+voice]

[-nasal]

ƾ

[+stop] [+son] [-cont]

[+voice]

[+nasal]

h

[+fricative] [+son] [+cont]

[+voice]

[-nasal]

Coronal [-anterior] Coronal [-anterior] Coronal [-anterior] Dorsal [+high] Dorsal [+back] [+ high] Dorsal [+back] [+ high] Dorsal [+back] [+ high] Dorsal [+back]

‫ݤ‬



29





Table 2-3 Specification of the English Consonants in Terms of Articulator-based Feature Geometry class node Laryngeal

Soft Palate

phoneme

Labial

p b m w f

[-voice] [+voice] [+voice] [+voice] [-voice]

[-nasal] [-nasal] [+nasal] [-nasal] [-nasal]

[-round] [-round] [-round] [+round] [-round]

v

[+voice]

[-nasal]

[-round]

ș

[-voice]

[-nasal]

/

ð

[+voice]

[-nasal]

/

Place Coronal

/ / / / [+anterior] [+distributed] [+anterior] [+distributed] [+anterior] [+distributed] [+anterior] [+distributed]

Dorsal

/ / / / / / / /

Chapter Two

30 t

[-voice]

[-nasal]

/

d

[+voice]

[-nasal]

/

n

[+voice]

[+nasal]

/

s

[-voice]

[-nasal]

/

z

[+voice]

[-nasal]

/

l

[+voice]

[-nasal]

/

‫ݕ‬

[-voice]

[-nasal]

/

‫ݤ‬

[+voice]

[-nasal]

/

‫ݹ‬

[-voice]

[-nasal]

/

‫ݶ‬

[+voice]

[-nasal]

/

r

[+voice]

[-nasal]

/

j

[+voice]

[-nasal]

/

k

[-voice]

[-nasal]

/

[+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [-anterior] [+distributed] [-anterior] [+distributed] [-anterior] [+distributed] [-anterior] [+distributed] [-anterior] [+distributed] [-anterior] [+distributed] /

g

[+voice]

[-nasal]

/

/

ƾ

[+voice]

[+nasal]

/

/

h

[+voice]

[-nasal]

/

/

/ / / /

/ / / / / / [+high] [+high] [+back] [+high] [+back] [+high] [+back] [-high] [+back]

Thirty-five word-initial CC clusters (/pr, pl, pj, ps, pߦ, br, bl, bj, tw, tj, dj, dw, kr, kl, kw, kj, kv, gr, gl, rw, fl, fr, fj, rj, șw, șj, sp, sk, sw, sf, sj, sv, ߦp, ߦw, hj/) are possible complex sounds because they are produced by different articulators, and, thus, they may have any feature value – either the same or different (Duanmu 2009, 25). Three clusters (/pw, tr, dr/) are possible complex sounds, too. They involve the same articulator but without conflicting gestures. They have either different features or the same values of the same features. Seventeen clusters (/pf, km, kn, nj, mj, mw, vw, șr, st, sl, sn, sm, sr, zl, ߦr, ߦm, ߦn/) display different values of the same features done by the same

The CVX Theory of the Syllable

31

articulator and cannot be represented as complex segments (details about the complex-sound analysis of the English word-initial CC clusters are in Appendix 3). All in all, 38 word-initial CC clusters in English (69%) can be represented as complex sounds requiring one timing slot only. Clusters that cannot be accounted for as complex segments do not represent a violation of Duanmu’s (2009) theory because they have morphological solutions, as explained in section 2.6.

2.4.2 A Phonemic Complex-sound Analysis of the Slovak Word-initial CC Clusters To present the single-slot analysis of the Slovak onset CC clusters analogous to the single-slot analysis of the English word-initial CC clusters, the relevant phonetic characteristics of the Slovak consonant phonemes and the distinctive features of these phonemes are presented in table 2-4 and table 2-5, respectively. The description of the individual Slovak consonant phonemes from the viewpoint of Articulator-based Feature Geometry is in table 2-6.21 Table 2-4 The Phonetic Classification of the Slovak Consonant Sounds (KráĐ and Sabol 1989, 258–60) sound

place of articulation

manner of articulation

acoustic impression

p b m f v t d n s z c ‫ݤ‬

bilabial bilabial bilabial labio-dental labio-dental pre-alveolar pre-alveolar pre-alveolar pre-alveolar pre-alveolar pre-alveolar pre-alveolar

stop stop nasal stop constrictive constrictive stop stop nasal stop constrictive constrictive semi-occlusive semi-occlusive

plosive plosive nasal plosive fricative fricative plosive plosive nasal plosive fricative fricative affricate affricate

21

participation of vocal cords (voice) voiceless voiced voiced voiceless voiced voiceless voiced voiced voiceless voiced voiceless voiced

The Slovak transcription symbols used in this analysis can be found in Appendix 1.

Chapter Two

32 š ž þ ۤ r l Ģ ć Ė Đ j k g x h

postalveolar22 post-alveolar post-alveolar post-alveolar post-alveolar post-alveolar palatoalveolar palatoalveolar palatoalveolar palatoalveolar palatal velar velar velar laryngeal

constrictive

fricative

voiceless

constrictive semi-occlusive semi-occlusive constrictive constrictive stop

fricative affricate affricate fricative fricative plosive

voiced voiceless voiced voiced voiced voiceless

stop

plosive

voiced

nasal stop

nasal plosive

voiced

constrictive

fricative

voiced

constrictive stop stop constrictive constrictive

fricative plosive plosive fricative fricative

voiced voiceless voiced voiceless voiced

Table 2-5 Feature Specification of the Slovak Consonant Phonemes (Halle 2006) phoneme stricture

velum

place

[-nasal] [-nasal] [+nasal] [-nasal]

Labial [-round] Labial [-round] Labial [-round] Labial [-round]

[+voice]

[-nasal]

Labial [-round]

[-voice]

[-nasal]

Coronal [+anterior] Coronal [+anterior] Coronal [+anterior]

t

[+stop] [-son] [-cont] [+stop] [-son] [-cont] [+stop] [+son] [-cont] [+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [+stop] [-son] [-cont]

d

[+stop] [-son] [-cont]

[+voice]

[-nasal]

n

[+stop] [+son] [-cont]

[+voice]

[+nasal]

p b m f v

22

distinctive features vocal cords [-voice] [+voice] [+voice] [-voice]

In Slovak, the post-alveolar place of articulation is not identical with the palatoalveolar place of articulation as suggested by the current IPA terminology.

The CVX Theory of the Syllable

33

[-voice]

[-nasal]

[+voice]

[-nasal]

c

[+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [-son] [-cont]

[-voice]

[-nasal]

‫ݤ‬

[-son] [-cont]

[+voice]

[-nasal]

š

[-voice]

[-nasal]

[+voice]

[-nasal]

þ

[+fricative] [-son] [+cont] [+fricative] [-son] [+cont] [-son] [-cont]

[-voice]

[-nasal]

ۤ

[-son] [-cont]

[+voice]

[-nasal]

r

[+voice]

[-nasal]

[+voice]

[-nasal]

Ģ

[+fricative] [+son] [+cont] [+fricative] [+son] [+cont] [+stop] [-son] [-cont]

[-voice]

[-nasal]

ć

[+stop] [-son] [-cont]

[+voice]

[-nasal]

Ė

[+stop] [+son] [-cont]

[+voice]

[+nasal]

Đ

[+voice]

[-nasal]

[+voice]

[-nasal]

k

[+fricative] [+son] [+cont] [+fricative] [+son] [+cont] [+stop] [-son] [-cont]

[-voice]

[-nasal]

g

[+stop] [-son] [-cont]

[+voice]

[-nasal]

x

[+fricative] [+son] [+cont] [+fricative] [+son] [+cont]

[-voice]

[-nasal]

Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+anterior] Coronal [+ant] [-lateral] Coronal [+ant] [+lateral] Coronal [-anterior] Coronal [-anterior] Coronal [-anterior] Coronal [-ant] [+lateral] Coronal [-ant] [-lateral] Dorsal [+back] [+ high] Dorsal [+back] [+ high] Dorsal [+back]

[+voice]

[-nasal]

Dorsal [+back]

s z

ž

l

j

h

Chapter Two

34

Table 2-6 The Specification of the Slovak Consonants in Terms of Articulatorbased Feature Geometry class node Laryngeal

Soft Palate

phoneme

Labial

p b m f

[-voice] [+voice] [+voice] [-voice]

[-nasal] [-nasal] [+nasal] [-nasal]

[-round] [-round] [-round] [-round]

v

[+voice]

[-nasal]

[-round]

t

[-voice]

[-nasal]

/

d

[+voice]

[-nasal]

/

n

[+voice]

[+nasal]

/

s

[-voice]

[-nasal]

/

z

[+voice]

[-nasal]

/

c

[-voice]

[-nasal]

/

‫ݤ‬

[+voice]

[-nasal]

/

š

[-voice]

[-nasal]

/

ž

[+voice]

[-nasal]

/

þ

[-voice]

[-nasal]

/

ۤ

[+voice]

[-nasal]

/

r

[+voice]

[-nasal]

/

l

[+voice]

[-nasal]

/

Ģ

[-voice]

[-nasal]

/

ć

[+voice]

[-nasal]

/

Place Coronal

/ / / [+anterior] [+distributed] [+anterior] [+distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distibuted] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [+anterior] [-distributed] [-anterior] [+distributed] [-anterior] [+distributed]

Dorsal

/ / / / / / / / /

/ / / / / / / / / /

The CVX Theory of the Syllable

35

Ė

[+voice]

[+nasal]

/

Đ

[+voice]

[-nasal]

/

j

[+voice]

[-nasal]

/

k

[-voice]

[-nasal]

/

[-anterior] [+distributed] [-anterior] [+distributed] [-anterior] [+distributed] /

g

[+voice]

[-nasal]

/

/

x

[-voice]

[-nasal]

/

/

h

[+voice]

[-nasal]

/

/

/ / [+high] [+high] [+back] [+high] [+back] [+high] [+back] [-high] [+back]

Fifty-seven CC clusters /pn, pĖ, tm, tl, tĐ, dm, dn, dĖ, dl, km, kn, kĖ, gn, gĖ, ml, mĐ, mr, st, sĢ, sm, sn, sĖ, sl, sĐ, zn, zĖ, zl, zr, št, šĢ, šm, šn, šl, šĐ, šr, žm, žĖ, žĐ, žr, hn, hĖ, hm, xm, vn, vĖ, lž, cĢ, cm, cn, cĖ, cl, cĐ, þm, þn, þl, þĐ, þr/ have conflicting gestures by the same articulator and cannot be represented as complex segments, that is, segments requiring one timing slot only. Segments in 57 word-initial CC clusters /ps, pš, px, bć, pl, bz, bl, bĐ, br, tk, tx, tv, dv, kt, kv, kr, kl, kĐ, gl, gĐ, gr, mn, mĖ, sp, sk, sx, sv, zb, zv, šp, šk, žv, hl, hĐ, hr, hv, xc, xv, xl, xĐ, xr, ft, fĢ, fþ, fš, fš, fl, fĐ, fr, vd, vz, vl, vr, lk, cv, þp, þv/ are produced by different articulators and these sequences of consonants can thus be treated as complex segments. And two clusters /tr, dr/ represent a special type of complex segment–affricate because they are characterized by the simultaneous features [+stop] [+fricative] of the Coronal articulator (the whole detailed complex-sound analysis of the Slovak word-initial CC clusters is in Appendix 4). To all appearances, some consonant clusters in English (Appendix 3) and in Slovak (Appendix 4) can be represented as complex segments because they fulfil Duanmu’s (2009) phonemic criteria for the simplification of consonant clusters into single-segment elements.

2.5 A Phonetic Complex-sound Analysis of the Word-initial CC Clusters Following Sagey (1990), Newman (1997) defines a complex segment as a doubly articulated consonant that must be considered a single segment and not a cluster (Newman 1997, 8). The author then states that “there is evidence from studies on the timing of these segments in actual speech”

36

Chapter Two

(ibid.), and quotes Sagey, who studied the duration of Igbo /g੾b/ and of the consonants /g/ and /b/. She came to the conclusion that they “were all of roughly equal duration” (Newman 1997, 9): /g/ = 90 msec, /b / = 110 msec and /g੾b/ = 112 msec. The data thus support a single unit status of the complex stop /g੾b/. Sagey (1986), in her work on the representation of distinctive features in non-linear phonology, mentions “indeed […] complex segments have the phonetic durations of single segments’ (7880). Later in her book, she provides further data about the length of affricates and stops supporting this assumption. However, no details about the material point of departure are specified in Sagey’s or in Newman’s book. Duanmu (2010, 15) also claims that the notion of a complex sound “usually refers to two sounds that share one timing slot”; in other words, the length of a complex segment is approximately the same as the length of a simple segment. However, this assumption is, again, not supported by any arguments. To provide the relevant acoustic data affirming or refuting the claim that the length of a complex segment is approximately the same as the length of a single segment in a given language, research has been carried out in both English and Slovak languages. The length of the individual speech sounds (i.e., vowels and consonants) may vary to a large extent, depending on various factors and criteria. The duration of a speech segment is different when the given speech sound is produced in isolation or when it is pronounced in a word. When a given sound is produced in the word, the length of the given word influences the length of its individual elements. This means that the number of the syllables is important. The transgressive character of consonants and the inherence of vowels also play an important role in the acoustic analysis of the segment’s duration (for details, see Romportl 1976 and section 6.3.2 in this work). The word does not have to be pronounced in isolation. It can be part of a larger unit. Thus the speech sound, whose length we want to measure, can be part of the word connected to speech. Still other criteria influencing the duration of segments can appear here: speech style, stressed or unstressed character of the given syllable, the occurrence before a pause, etc. (for details, see Gregová 2012b). To avoid all the difficulties, or, better, to include all the factors that may influence the duration of speech sounds, the material starting point for the acoustic analysis was the recordings of news broadcasts from radio stations whose pronunciation styles can be considered standard in both languages. The recordings in standard accented British English are those

The CVX Theory of the Syllable

37

from the BBC because of its use of Received Pronunciation (see, e.g., Wells 1997). Those in Slovak are from the Rádio Slovensko because of its use of Standard Slovak. News texts of journalistic style are classified as objective styles of speech communication (Mistrík 1982). This objectivity results from the basic conditions for the creation of news texts – general comprehensibility and impartiality. Facts, along with the manner of their presentation, should be announced and not commented on. Radio utterances have all the peculiarities of spoken utterances: they disappear in time and space, and the speech is perceived at the moment of its creation. The uniqueness of radio communication is a certain communicative distance between the speaker and the hearer: the speaker cannot slow an utterance down and the addressee has no opportunity to tell the speaker that he or she cannot be understood (Mistrík 1997, 462–65). As a result, the basic condition for successful radio communication is perfect pronunciation – the faultless production of segmental sound elements and of those suprasegmentals with a phonologically relevant function (Sabol 2003, 34). Thus, journalistic style offers the most suitable material for the comparative analysis of language sound systems (see also Gregová 2008). Five-minute news broadcasts from both languages were recorded by Audacity software that enables the recording and transformation of an analogous sound into its digital form, which can then be stored in a computer. The recordings of radio broadcasts were segmented into sentences. The length of the individual consonants and consonant clusters was measured by Speech Analyser audio software – a well-known computer program used for the acoustic analysis of speech sounds. Oscillograph curves – composite curves with the parameters of time duration and momentary amplitude (see also note 14 in Chapter 6) – were displayed on the computer screen. The graphic representation of the recorded sentences was segmented into the individual words, and then the consonants (consonant clusters) in these words were cut off (see Figure 2-5). The beginning and the end of the segmented consonants were defined visually and perceptually. Finally, the data about the time duration of the segmented consonants in milliseconds were read from the screen (Fig. 25). The boundaries of the measured sounds were delimited to exactly 1/40 000 of a second because the sampling frequency 40 kHz was used when digitising the speech signal (cf. Sabol and Zimmermann 1994, 34).

38

Chapter Two

Figure 2-5 The Oscillogram of the Phrase “is almost the same”, the Word “same”, and the Segmented Consonant [s]

When comparing the length of a single consonant with the length of a consonant cluster, the average values are used. However, the length of a segment can vary to a certain extent depending on its position in a word or its sound environment (as explained above). The differences between the data are sometimes very great. For example, the timing of the English sibilants can be 112 msec, but also 62 msec. Thus, to find out whether the average duration for each consonant and consonant cluster, which is a phonemically possible complex sound occurring in the analysed acoustic material, can be considered reliable,23 all measured numerical information about the duration of consonants and consonant clusters (or complex segments in Duanmus’ view) in English and in Slovak was statistically processed so as to calculate the so-called measures of variability – the standard deviation (s) and the variation coefficient (Cv). These two 23

That is, if the data can be treated as statistically significant evidence because of the given items’ considerable variance.

The CVX Theory of the Syllable

39

variables give us a more precise picture of the distribution of the statistical item (length) around the mean value (i.e., around the arithmetic mean, x). The mean is the average value of the processed statistical item, in this case, length, and it is given as the total sum of the values divided by the number of values. The standard deviation shows the variation or dispersion of an item from the average. These two are used for the counting of the variation coefficient: Cv = s/x*100. The variation coefficient is given as a percentage. If the value of the variation coefficient for the given statistical set is higher than 50%, the statistical set is considerably dispersed, and its arithmetic mean should not be used (Sabol and Zimmermann, 1986, 27). The variance gives information about how far the set of numbers is spread out (ibid.).

2.5.1 A Phonetic Complex-sound Analysis of the English Word-initial CC Clusters Based on the data provided by the sound material, six consonant clusters that seem to fulfil the phonemic conditions for being evaluated as possible complex sounds were chosen for the phonetic analysis. Three consonant sequences, /pr, kw, sk/, involve different articulators, and three consonant clusters, /tr, dr, hj/, involve the same articulator with different features (see Appendix 3). The assumption was that the length of complex segments would be approximately the same as the length of simple segments, despite the multiple articulation of the complex segments, because they occupy one timing unit and not two timing units as consonant clusters (see also Sagey 1986). The data are presented in table 2-7. Comparing the data about the length of the word-initial CC sequences with the length of the individual simple segments, it is clear that none of the possible complex sounds supports the assumption about the similarity of the phonetic length of these multiple segments to the length of single consonants. This finding – together with the considerable discrepancy between Duanmu’s (2009) definition of a complex sound and the author’s own evaluation of consonant sequences as complex segments (Appendix 3) – questions whether English two-consonant clusters can be represented as complex segments (i.e., as segments with multiple articulation and simplesegment timing).

Chapter Two

40

Table 2-7 Statistical Data about the Length of the Selected English Word-initial CC Sequences word-initial CC sequence/complex segment

mean [x]/msec.

standard deviation [s]

variation coefficient [Cv]/%

pr25

111.1

42.2

37.8

kw

98.4

28.3

29

sk

172.2

18.2

10.6

tr

101.6

17.56

17.3

dr

96.8

14.89

15.4

hj

126.1

28.76

23.0

length of the individual segments/msec.24 p – 60.1 r – 56.1 ™ = 116.2 k – 61.2 w – 59.9 ™ =121.1 s – 81.0 k – 61.2 ™ = 142.2 t – 59.6 r – 56.1 ™ = 115.7 d – 61.7 r – 56.1 ™ = 117.8 h – 64.8 j – 70.5 ™ = 135.3

2.5.2 A Phonetic Complex-sound Analysis of the Slovak Word-initial CC Clusters Analogous to the English language, six sequences of consonants, which can be evaluated as complex sounds, have been chosen for phonetic analysis – /ps, tx, gl, šp, tr, þv/. The selection of the acoustic material as well as the statistical processing of the data gained is also identical with those used in the analysis of the English complex segments (see above). The data about the length of the selected Slovak complex sounds and their statistical data are in table 2-8. It is obvious that the length of the Slovak CC sequences, following Duanmu’s (2009) approach to the notion of a complex segment, can be 24

The data about the length of the individual English and Slovak consonant sounds were also statistically evaluated. The calculated measures of variability prove their statistical reliability (Gregová 2015). 25 A sequence of consonants that can be represented as a complex sound is written down either as CҮ C or as Cs, depending on the author.

The CVX Theory of the Syllable

41

represented as phonemically complex sounds, and are twice as long as individual consonants. This means that the consonants involved are articulated in a sequence and not simultaneously, as supposed by the complex-segment theory. Table 2-8 Statistical Data about the Length of the Selected Slovak Word-initial CC Sequences word-initial CC sequence/complex segment

mean [x]/msec.

standard deviation [s]

variation coefficient [Cv]/%

pl

100.4

7.98

7.95

bz

120.4

23.66

19.7

tv

121

13.2

10.9

dr

118.4

26.26

22.2

fš

129.6

18.52

14.29

cv

121.6

13.1

10.8

length of the individual segments/msec. p – 59.6 l – 51.3 ™ = 118.9 b – 58.2 z – 62.9 ™ = 121.1 t – 61.9 v – 52.7 ™ = 114.6 d – 59.9 r – 52 ™ = 111.9 f – 56.7 š – 69.3 ™ = 126 c – 66.6 v – 52.7 ™ = 119.3

2.5.3 Do Complex Sounds Exist? The idea of a complex segment is based on the notion of multiple articulation. Traditional phonetics makes a difference between double articulation and secondary articulation, where the term double articulation indicates “two simultaneous articulations of the same degree of stricture” (Ladefoged and Maddieson 1996, 328). Secondary articulation is used for the articulation of segments in which “two co-occurring articulations have different degrees of stricture, the one with the greater stricture is labeled primary and the lesser one is labeled secondary” (ibid.).

42

Chapter Two

Non-linear phonology prefers the term multiple articulation or multiple articulatory gestures26 to the previous traditional terms, which should be used in connection with segments produced by “multiple articulations that are effectively simultaneous from the phonetic point of view” (ibid.). Ladefoged and Maddieson (1996, 328–29) further argue it is the phonetic simultaneity that should be taken into account when talking about single segments with multiple articulation. Many authors have already proven that multiple articulated single segments are not part of the phonological inventory of all languages (see, e.g., Sagey 1986; KráĐ and Sabol 1989; Scheer 2012). For example, Sagey (1986) exemplifies the complex-segment theory on the Niger-Congo and other African languages. Ladefoged and Maddieson (1996), in their account of multiply articulated segments with the duration comparable to the length of single segments, also mention, for example, that bilabialvelar stops (/k੾p, g੾b/) are typical of West Africa and northern Central Africa languages, and they can also be found in several New Guinea languages. This means that labio-velar stops occur in Niger-Congo and other African languages. The existence of complex segments in some languages is thus indisputable. The methods of experimental and physiological phonetics have attested that in many Niger-Congo and Tsimshianic languages, there are labio-coronal, labio-dorsal, and corono-dorsal articulations (Hall 2006, 331). For example, [p] and [t] are pronounced nearly simultaneously in these languages, resulting in a labio-coronal sound /p੾t/. However, it seems that in English, the pronunciation is not simultaneous but in sequence (Hall 2003, 331). Physiological phonetics provides evidence that the articulatory basis of these languages is completely different from the articulatory basis of the languages spoken in Europe. Sometimes it is simply not possible to compare two distinct sounds systems at all (Sampson 1980, 60). The application of Duanmu’s (2009) phonological criteria to the wordinitial, traditionally called two-consonant clusters, seems to result in the existence of complex sounds in both analysed languages (Appendix 3, Appendix 4). But there is no phonetic evidence for this interpretation (section 2.5.2) and even phonological analysis itself is full of uncertainties and ambiguities. Similar conclusions have been made by Ladefoged and Maddieson (1996):

26 ‘Gesture’ means ‘feature’ in Articulator-based Feature Geometry (see also section 2.2.1).

The CVX Theory of the Syllable

43

We are not aware of any language which has plosive or nasal sounds involving simultaneous Coronal and Dorsal articulation, such as alveolarvelars. Some accounts of Kinyarwanda and of certain dialects of Shona have been interpreted as indicating the occurrence of segments such as t੾k, d੾g. Although phonological arguments can be marshalled in favor of treating these events as single segments (Sagey 1986, 1990) they are phonetically quite unambiguous. They are sequences of two stops. (345)

The phonemic and the phonetic analysis of the English and the Slovak word-initial two-consonant clusters as possible complex segments casts doubt on the universal validity of the CVX theory of the syllable. Anyhow, although the word-edge consonants cannot be simplified into complex sounds that require one timing slot only, there are morphological rules (sections 2.1.2–2.1.3) that explain the extra consonants. That is why the following parts of this chapter are dedicated to the single-slot analysis of all word-initial, word-final, and word-medial consonant clusters in English and in Slovak.

2.6 The CVX Syllable Theory and the English Language As indicated in section 2.3.2, traditionally, the maximal size of the English syllable is CCCVVCCCC (see, e.g., Roach 2000; McMahon 2002):

Figure 2-6 The Maximal Phonological Structure of the English Syllable

The CVX theory allows only one slot for the onset structure. Thus, when there is zero onset or only one consonant before a vowel, this one timing unit is preserved. For example, (4)

am [ m] – syllabic structure [VC], zero onset pot [p‫ܥ‬t] – syllabic structure [CVC], onset [C]

But two- and three-consonant clusters in the word-initial position violate the single-slot analysis as proposed by the CVX theory, and they

44

Chapter Two

are either represented as complex sounds or are accounted for by morphology. As for the structure of rhyme in English, the highest number of consonants in the word-final consonant cluster (i.e., in coda) is four (section 2.3.2). The centre of the syllable can be a short monophthong (representing one timing slot, V), a long monophthong (represented by two timing slots, VV), or a diphthong (represented by two timing slots, VV).27 This means that the maximal structure of the English word-final rhyme is [VVCCCC]. But Duanmu’s (2009) theory proposes only the rhyme structure, [VV]ʊlong vowel, or diphthong, and no coda or [VC]ʊshort vowel and one consonant in coda. As a consequence, it is necessary to reduce the rhyme structure into two slots only. Extra sounds at word edges should be somehow accounted for, too.

2.6.1 A Single-slot Analysis of the Word-initial Consonant Clusters in English In terms of Articulator-based Feature Geometry, 38 word-initial CC clusters can be represented as complex sounds (section 2.4.1). They either involve different articulators or the same ones but without conflicting gestures. Seventeen clusters cannot be simplified into a complex segment and need two timing slots but the first consonant in these clusters is not part of the syllable; it does not belong to the syllable onset. These unsyllabified consonants are simply accounted for by morphologyʊthe Potential Vowel Rule can be applied because English is a language with prefixes (section 2.1.2): (5)

throb [șr‫ܥ‬b] – syllable structure ș[r‫ܥ‬b] – C[CVC]

As for the word-initial CCC clusters, in Duanmu’s (2009) view the initial /s/ can simply be excluded due to its violation of the sonority requirements (see also 3.2.1.1) and all onset clusters either form a complex sound (they are produced with a different articulator, cf. Duanmu 2009, 43–44) or they are predictable by morphology as real or potential affixes and, therefore, they are solved by the Affix Rule (for details, cf. Duanmu 2009). For example, (6) 27

strip [strܼp] – syllable structure s[trܼp] – C[CsVC]

The analysis of syllables with syllabic consonants in their centre is identical with the analysis of syllables with short monophthongs (both monophthongs as well as syllabic consonants represent one timing slot).

The CVX Theory of the Syllable

45

By implication, the analysis of the data from the English language supports “the CVX theory in which there is only one onset slot” (ibid., 179).

2.6.2 The Analysis of the Word-final Rhymes in English In English, there are 24 consonantal phonemes (cf., e.g., Roach 2000; McMahon 2002), and any consonant can be a final except /h, w, j/ (Roach 2000, 59). If there is only one segment in coda and the syllable centre is formed by a short vowel, the structure of the rhyme proposed by Duanmu (2009) is preservedʊit is [VC]. For instance, (7)

but [bࣱt] – syllable structure [CVC], rhyme [VC]

If a long vowel or a diphthong create the peak (e.g., mine [maԘn], part [pľ:t]), the rhyme structure is [V:C] (i.e., [VVC]). This special case is solved by the simplification of [V:C] into [VC], where “V is tense and short, but still distinct from a lax vowel” (Duanmu 2009, 45).28 If this analysis is applied, the rhyme size is [VX]29: (8)

mine [maܼn] – syllable structure [CVC], rhyme [VC] part [p‫ܤ‬:t] – syllable structure [CVC], rhyme [VC]

Most of the word-final two-consonant clustersʊ44 of themʊend with one of the coronals /t, d, s, z, ș/ as the final one (section 2.3.2). These clusters have a morphological solutionʊthe final consonants can be accounted for by the Affix Rule (section 2.1.1), which covers both real and potential affixes. For example, blessed [blest] is a monosyllabic word whose rhyme exceeds [VX]ʊit is [VCC]. The final C [t] is a real suffix. In the monosyllabic word want [wLJnt], the final C is a potential suffix. These real and potential suffixes are extrasyllabic. They are not part of the syllable and the rhyme size does not exceed [VX]:

28

This simplification is possible in languages like English whose vowels are different in terms of length (quantity) and quality, too (e.g., Giegerich 1992). 29 Analogically, the assumption that long vowels (i.e., long monophthongs and diphthongs) “can be accounted for by the idea that [V:C] can be analysed as [VC]” (Duanmu 2009, 158) is applied also to the rhymes with two-, three-, and fourconsonant clusters, which are then not treated separately here.

46 (9)

Chapter Two blessed [blest] – syllable structure [bles]t, rhyme [es] – [VC] want [w‫ܥ‬nt] – syllable structure [w‫ܥ‬n]t, rhyme [‫ܥ‬n] – [VC]

(For details, see Duanmu 2009, 154–55.) In the case of the 11 two-consonant clusters (/mp, mf, nࠔ, n˔, ƾk, ƾg, sp, sk, lp, lf, lk/), the Affix Rule cannot be applied (e.g., in the word camp [kæmp]); the extra consonant at the word edge is supported by the Potential Vowel Rule (section 2.1.2) or Anti-Allomorphy (section 2.1.3), or both: (10)

camp [kæmp] – syllable structure [CVC]C ĺ camping [kæmpܼƾ] – syllable structure [CVC][CVC] (ibid.)

The final three-consonant clusters end with /s, z, t, d/, which, as already mentioned, can easily be accounted for by morphology because they represent separate morphemes. For example, the rhyme structure in the word depths [depșs] is [VCCC]. The final [s] can be solved by the Affix Rule. The remaining segments still exceed the structure [VX]. The consonant cluster [pș] is a good complex sound because the gestures of [p], which is Labial, and [ș], which is Dental, are independent and can overlap, filling in only one timing slot (see Appendix 3). Thus, the word depths [depșs] is syllabified as follows: [depș[s] with the rhyme structure [VC] (i.e., [epș] and unsyllabified [s]). The schematic structure of the word-final rhyme in arranged [̾reԘn˔d] is [VVCCC]. The final C [d] is solved by the Affix Rule, the penultimate C [˔] is extrasyllabic, too, accounted for by the Potential Vowel Rule. What remains is [eԘn – VVN]. In Duanmu’s (2009) approach, “[VVN] can be analysed as [ࣿࣿ]” (158). The maximal size of the wordfinal rhyme in arranged is then still VX; it is [ࣿࣿ]. The largest size of the word-final rhyme VX is preserved also in words that end with one of four-consonant clusters (section 2.2.3). Word-final consonants are explained by morphologyʊthe Affix and Potential Vowel Rules, Anti-Allomorphy, or the notion of a complex sound can be applied. For example, the word texts [teksts] has the rhyme [VCCCC]. The final coronal [s] is solved by the Affix Rule. The penultimate [t] can be accounted for by Anti-Allomorphy or the Potential Vowel Rule. [ks] is a good complex sound. The rhyme structure is then [VC] – [eks] (see also Duanmu 2009, 154–56).

The CVX Theory of the Syllable

47

2.6.3 The Syllable Boundaries in the English Word-medial Consonant Clusters As already explained in section 2.3.1, the word-medial clusters30 are usually not monosyllabic and they can be described from the viewpoint of the patterns provided by the initial and the final clusters. That is probably why the CVX theory pays very little attention to the syllabification of the word-medial clusters; although, generally, the placement of the syllable boundary in consonant clusters that are in intervocalic position causes the most difficulties. When evaluating different approaches to the syllabification of the sequence CVCVC (Duanmu 2009, 52–59), the author offers his own principle, which is considered generally accepted. The Weight-Stress Principle specifies that a. The stressed syllable should be heavy (rhyme being VX) b. Unstressed syllables should be light (rhyme being v31 or a syllabic C) (Duanmu 2009, 59)

However, the examples provided say nothing about the syllabification of word-medial clusters. Some brief notes concerning the analysis of these clusters can be found within the analysis of the onset clusters in non-final positions (see ibid., 165) and non-final rhymes (see ibid., 149–54). For example, the syllable boundary is placed between the cluster /sk/ in mascot /mæsk̽t/, where /s/ forms the coda of the preceding syllable and /k/ is the onset of the following syllable: (11)

mascot [mæs.k‫ۑ‬t]32 – syllabification [CVC][CVC]

Or the four-consonant cluster in hamstring [hæmstrԘƾ] is syllabified in a way where /m/ is in the coda of the first and /str/ in the onset of the second syllable: (12)

hamstring [hæmstrܼƾ] – syllabification [CVC] C[Cs33VC]

If a cluster seems to be problematic, a solution is offered by a perceived affix that has the same form as the real affix. For example, “tic” 30

The complete list of the word-medial consonant clusters is in Appendix 5. v is an unstressed vowel. 32 Dot [.] indicates the syllable boundary. 33 s C stands for a complex sound. 31

48

Chapter Two

in drama-tic is a real suffix, while in arctic and deictic, it is a perceived suffix: (13)

dramatic [dr‫ۑ‬mætܼk] – syllabification [CVC][CV][CVC] – [dr‫ۑ‬.mæ.tܼk] arctic [‫ܤ‬:ktܼk] – syllabification [CVC][CVC] – [‫ܤ‬:k.tܼk] deictic [daܼktܼk ] – syllabification [CVC][CVC]34 – [daܼk.tܼk] (Duanmu 2009, 152–53)

The author comes to the conclusion that the VX limit of non-final rhymes is quite strict in English. One timing slot for onset and rhyme structure [VX] is preserved in word-medial position in English, too (ibid., 154).

2.6.4 Summary The analysis of the data from the English language indicates that despite the long consonant sequences in onsets and codas (Figure 2-6), it is possible to reduce the structure of the syllable to the CVX template only (Figure 2-1). Consonants at word edges are either represented as complex segments (sections 2.1.1, 2.1.2) or they remain extrasyllabic because they have morphological solutions (section 2.6.3). If the rhyme structure still exceeds the limit [VX], it is possible to represent a long vowel as a short one (i.e., [VV] = [V]) because English vowels differ in quantity but especially in quality. Thus, “even if a tense vowel is shortened it remains distinct from a lax one” (Duanmu 2009, 152). Or the sequence of a vowel and a nasal consonant can be represented as one timing unit (i.e., [VN] = [ࣿ]). Duanmu argues “phonetically, it is well known that English vowels are nasalised when they follow a nasal coda […]” (ibid.).

34

See section 2.6.2 for the representation of a diphthong as [V].

The CVX Theory of the Syllable

49

2.7 The CVX Syllable Theory and the Slovak Language The maximal syllable size in Slovak is CCCCVVCCC (section 2.3.3):

Figure 2-7 The Maximal Phonological Structure of the Slovak Syllable

Similar to the English language, if there is no consonant at the beginning of the syllable, there is no onset. If the syllable begins with one consonant only, one slot for onset is preserved, as illustrated in (14): (14)

ak ‘if’ – syllable structure [VC], zero onset pas ‘passport’ – syllable structure [CVC], onset [C]

The highest number of consonants in the final position of Slovak words (i.e., in syllable coda) is three (section 2.3.3). Likewise in English, the syllabic nucleus can be a short monophthong (one timing slot, V), a long monophthong (two timing slots, VV), or a diphthong (two timing slots, VV).35 The maximal size of the word-final rhyme is [VVCCC] in Slovak (Gregová 2010). Thus, to preserve the rhyme structure [VX], certain simplifications are necessary.

2.7.1 A Single-slot Analysis of the Word-initial Consonant Clusters in Slovak In Duanmu’s (2010) theory, initial syllables are only those that are “[…] the first after a word boundary, whether there is a prefix or not. For example, in sprinkle, [sprԘƾ] is initial. In resprinkle, both [ri] and [sprԘƾ] are initial” (160–61). That is why all traditionally delimited initial consonant clusters in Slovak (Gregová 2010) were reanalysed in terms of this approach. Some consonant clusters were excluded from further 35

The centre of the Slovak syllable can also be formed by a syllabic consonant functioning as a vowel that can be either short (one timing slot) or long (two timing slots).

50

Chapter Two

analysis because they are not monomorphemic (e.g., the originally delimited two-consonant cluster ‘zr’ in the word zrub ‘chalet’ is decomposable into a prefix z- and a stem -rub). On the other hand, some two-consonant clusters were included as the result of the decomposition of three or four-consonant clusters (e.g., the original three-consonant cluster /vzb/ in the word vzbura ‘revolt’, morphematic structure vz-bur-a, contains a prefix vz- and, thus, only vz- can be treated as the initial cluster). But the four-consonant cluster /fspl/ʊas in the word vzplanúĢ ‘to flare up‘, the sound form [fsplanúĢ]ʊis not monomorphemic. The morpheme structure of the word is fs-plan-ú-Ģ; therefore, both fs- and -pl- are initial. The complex-sound analysis of the word-initial CC clusters has shown that 57 CC clusters in Slovak cannot form a complex sound (see section 2.4.2 and Appendix 4). This represents almost 49% of all possible CC word-initial clusters. Duanmu argues that every consonant should be accounted forʊalthough not syllabified (Duanmu 2010, 5) if it can serve as an affix (the Affix Rule, section 2.1.1). In Slovak, the Affix Rule accounts for the word-initial [s], [z], [v], and [f] that are separate prefixes. For instance, (15)

pamätaĢ ‘to remember’ – s-pamätaĢ ‘to recover from st’ niesĢ ‘to carry’ – z-niesĢ ‘to tolerate’ rub ‘reverse side’ – v-rub ‘notch’ þas ‘time’ – v-þas [f-þas] ‘in time’

Thus, these segments can be treated as affix-like sounds also in words where they do not represent separate morphemes. For example, [s] in snár ‘dream book’ is not a prefix but it can exist as a prefix-like sound accounted for by morphology: (16)

snár ‘dream book’ – syllable structure s[nár] – C[CVVC]

If the monomorphemic initial CC clusters starting with /s, z, v, f/ that cannot be represented as complex sounds (2.1.1) but can be solved by the Affix Rule are excluded, there are still 39 (34%) initial CC clusters that cannot be interpreted as complex sounds; neither can the first sound in these clusters be classified as an affix-like consonant. The CVX theory of the syllable offers another solution for the unsyllabified consonants at word edges, the Potential Vowel Rule (section 2.1.2): “[...] a word-initial C can serve as the coda of a potential V, which may come with a V-final prefix” (Duanmu 2010, 11). Slovak is a language with prefixes ending with a vowel; thus, this rule is applicable:

The CVX Theory of the Syllable (17)

51

tlak ‘pressure’ – syllable structure C[CVC], onset [C] ĺ útlak ‘oppression’– syllabification [VC][CVC] – [út.lak] tma ‘darkness’ – syllable structure C[CV], onset [C]36 ĺ polotma ‘semidarkness’ – syllabification [CV][CVC][CV] – [po.lot.ma]

Seemingly, all Slovak initial two-consonant clusters can fit the singleslot analysis. As for the monomorphemic word-initial three-consonant clusters, there are 23 types of these clusters in Slovak (section 2.3.3), and 12 of them can be solved by morphology: the first segment is one of the consonants /s, z, v, f/ and can be explained by morphology as a potential affix (applying the Affix or Affix-like Rule). Most of the CC clusters that remain after the deletion of the first segment (/tr, kl, kĐ, kv, tv, dr, vl, br, hl/) are possible complex sounds (Appendix 4). The consonant sequences, /zl, mr, tl/, are not possible complex sounds but they have a morphological solution. The Affix and Potential Vowel Rules can be applied. For instance, (18)

sklad ‘store’ – syllable structure C[CsVC], onset [Cs] smrek ‘spruce’ – syllable structure C[C][CVC], onset [C]

Seven word-initial three-consonant clusters (tkv, škr, štv, škv, štr, špĐ, špr) are easily reducible to one timing slot, too. The first consonant can be explained due to the Potential Vowel Rule, and all remaining twoconsonant clusters are possible complex sounds (see Appendix 4), as illustrated below: (19)

škrek ‘scream’ – syllable structure C[CsVC], onset [Cs]

Four word-initial clusters, /hml, hmĐ, mdl, Đst/, cause difficulty for a single-slot analysis. Although there is the initial CC cluster /hm/ in Slovak, it is in a root morpheme and not in a prefix. This cluster cannot be accounted for by the Affix Rule: neither /hm/ nor /h/ can be interpreted as an affix-like (or, better, a prefix-like) sound. Neither /hm/ nor /ml/ (or /mĐ/) is a good complex sound (Appendix 4). It seems that at least a twoslot structure is required for the analysis of these clusters: [h] can be accounted for by the Potential Vowel Rule and will, thus, remain 36

Duanmu, in his CVX theory, considers the Potential Vowel Rule generally applicable, that is, it is not restricted to cases where the initial consonant remains extrasyllabic only if the prefix with a vowel at its end can be attached to the given monosyllabic word.

52

Chapter Two

extrasyllabic, and sequence ml/Đ will occupy two slots, as illustrated in the following example: (20)

hmla ‘fog’ – syllable structure C[CCV], onset [CC]

The situation with the clusters /mdl/ and /Đst/ is similar. The Affix Rule is not applicable. A complex-sound analysis and the Potential Vowel Rule result in two-slot structuresʊneither /md/ nor /dl/ in /mdl/ is a possible complex sound, nor can /Đs/ or /st/ in /Đst/ be a complex segment (see Appendix 4): (21)

mdlý [mdli:]37 ‘feeble’– syllable structure C[CCVV], onset [CC] Đstivo ‘trickily’ – syllable structure C[CCV][CV], first onset [CC]

Four-consonant clusters /pstr/ and /pštr/ also require two slots. Neither /ps/ nor /pš/ can be explained as affix-like sounds but they are good complex sounds, and /tr/ is a possible complex sound, too (see Appendix 4): (22)

pstruh [pstrux] ‘trout’ – syllable structure [CsCsVC], onset [CsCs] pštros ‘ostrich’ – syllable structure [CsCsVC], onset [CsCs]

2.7.2 The Analysis of the Word-final Rhymes in Slovak38 In the Slovak language, there are 27 consonantal phonemes /p, b, m, f, v, t, d, n, s, z, c, ঱, r, l, š, ž, þ, শ Ģ, ć, Ė, Đ, j, k, g, x, h/ (see Appendix 1 for the transcription symbols) and all of them can occur in the final position of a word. If the syllable nucleus is a short vowel, the rhyme structure is [VC]: (23)

dom ‘house’ – syllable structure CVC, rhyme [VC]

If a long monophthong or a diphthong creates a centre, the final C is accounted for by morphology (the Potential Vowel Rule, AntiAllomorphy, or both) and the rhyme size [VX] is preserved, for example, the monosyllabic word pár ‘couple’, with the segmental structure CVVC:

37 When the sound form of a Slovak word is different from its graphic form, the pronunciation is in brackets. When there is no difference between the spelling and the pronunciation of Slovak words, only the grapheme is used. 38 The analysis covers both word-final rhymes without inflection, i.e., the basic form of a word, and with inflection.

The CVX Theory of the Syllable (24)

53

pár ‘couple’ – syllable structure [CVV]C, rhyme [VV] ĺ párový ‘geminate’ – syllable structure [CVV][CV][CVV]

In terms of Articulator-based Feature Geometry, 19 word-final CC clusters (see 2.3.3 for the complete list of word-final clusters) /ps, kt, ks, sk, žd, ft, vk, xt, lp, lk, lf, rp, rt, rd, rk, rf, rv, jk, jf/ can be represented as complex sounds (Appendix 4). If any of these 19 CC clusters forms a rhyme whose peak is a short vowel, the rhyme structure is [VC], where C is represented by a complex sound. For example, (25)

stisk [sĢisk] ‘squeeze’ – rhyme structure VC [isk] necht [Ėext] ‘nail’ – rhyme structure VC [ext]

Cluster /sk/ occurs in words where it forms a rhyme with a long vowel, for example, ložísk < G pl. of ložisko ‘lead’, mravenísk < G pl. of mravenisko ‘anthill’, and bars ‘betide’. As explained in section 2.6.2, Duanmu (2009) proposes the simplification of [VVC] into [VC]. But in Slovak, it is impossible to represent the difference between vowels belonging to the same vocalic timbre without referring to their length. The sound form of Slovak monophthongs and diphthongs is relatively stable; they are not reduced but are fully pronounced in every syllable (KráĐ and Sabol 1989). Moreover, from the aspect of the close-open quality of vowels, the Slovak vocalic sounds are neutral. The vowel pairs in Slovak do not have a different quality; they differ only in their quantityʊvowels are either short or long (ibid., 200–1). The length of vowels, or vocalic quantity, has a phonologically distinctive function in Standard Slovak; it differentiates the meaning of words and word forms (Sabol 1989, 141). For example, sud ‘barrel’ vs. súd ‘tribunal’, udaj ‘give’ vs. údaj ‘data’, etc. Thus, Slovak long monophthongs and diphthongs always have to be represented by two timing slots [VV]. This indicates that at least one more slot is necessary for word-final rhymes in Slovak: (26)

másk < G pl. of maska ‘masque’ – rhyme [VVCs]

Word-final two-consonant clusters /mp, mb, mf, nt, nd, nk, ng, nc, nþ, nš, st, sĢ, št, šĢ, zd, zć, žć, ld, ls, lz, lm, tc, rþ, rz, rs, rš, rm, rn, rĖ, jt, jd, jn/ are not good complex sounds (Appendix 4). The final C in these clusters can be accounted for by the Potential Vowel Rule (section 2.1.2) because in Slovak, there are suffixes starting with a vowel. An extrasyllabic consonant in the word-final position can serve as the onset for the suffix at its beginning. When these clusters are preceded by a short

54

Chapter Two

vowel, the rhyme size is in accordance with the CVX theory [VX]. For example, (27)

vinš ‘verse’ – syllable structure [CVC]C, rhyme [VC] dosĢ ‘enough’ – syllable structure [CVC]C, rhyme [VC]

Moreover, clusters /nš, sĢ, šĢ/ can also be easily accounted for by the Affix Rule (or, better, the Affix-like Rule) because consonants /š/ and /Ģ/ function as grammatical suffixes in Slovak. However, 21 two-consonant clusters from those that are not good complex sounds /mp, mb, nt, nd, nk, ng, st, zd, sĢ, zć, št, žd, šĢ, žć, lt, ld, lm, rt, rd, rm, jd/ occur in syllables where the nucleus is a long vowel or a diphthong. The traditional scheme of the rhyme is then [VVCC]. Clusters /sĢ, šĢ/ have a morphological solution. For example, the final C in the word plášĢ [pla:šĢ] ‘overcoat’ can be solved by the Affix Rule and the penultimate C can be accounted for by the Potential Vowel Rule; the rhyme is then only [VV]: (28)

plášĢ [pla:šĢ] ‘overcoat’ – syllable structure C[CVV]C]C

Two-consonant clusters starting with nasal consonants do not violate the rhyme size [VX] either, even if they are preceded by a long vowel or a diphthong. For example, the final C in the word žúmp [žu:mp] < G pl. of žumpa ‘cesspit’ can be explained by the Potential Vowel Rule, and the rhyme [VVC] with the nasal consonant can be solved by the analysis of [VN] as [ࣿ] (for details see section 2.6.2). The rhyme structure is only [ࣿࣿ]: (29)

žúmp ‘of cesspits’ – syllable structure [Cࣿࣿ]C

This solution is slightly artificial for the Slovak language. The representation of [VN] as [ࣿ] has no phonological justification because there are no nasal vocalic phonemes in Slovak. The recent phonetic experiment has proven some traces of nasalization in the acoustic spectrum of vowels in the neighbourhood of nasal consonants but it does not seem to be a merger of a vowel and the following nasal consonant (see also section 6.3.3). The clusters /st, zd, zć, št, žć, lt, ld, lm, rm, jd/ represent a violation of the CVX theory. Only the final C can be solved by morphology as extrasyllabic (the Potential Vowel Rule), and one more slot is necessary for the rhyme. For example,

The CVX Theory of the Syllable

(30)

55

hviezd [hviest] < G pl. of hviezda ‘star’ – syllable structure [CCVVC]C, rhyme [VVC]

As for the three-consonant word-final clusters in Slovak (section 2.3.3), the cluster /nkt/ occurs only in rhymes with a short monophthong. For example, in the word punkt ‘dot’ – CVCCC, [kt] is a good complex sound (Appendix 4) and the remaining [un] can be solved by the analysis of [VN] as [ࣿ]. The rhyme structure is [ࣿC], where [ࣿ] stands for [࠿] and C is the complex sound [kt]: (31)

punkt ‘dot’ – syllable structure [CࣿCs], rhyme [VC]

Although the cluster /jsk/ is rare and occurs only in rhymes with a short vowel as a peak, it does not fit the rhyme structure [VX]. For example, in the word vojsk < G pl. of vojsko ‘military forces’, the rhyme is [VCCC] and the final C can be accounted for by the Potential Vowel Rule, but [js] cannot be treated as a complex sound because of the conflicting gestures ([j] is Coronal [-anterior] and [s] is Coronal [+anterior], cf. Table 2-5). At least one more timing slot is required for this word-final cluster: (32)

vojsk ‘of military forces’ – syllable structure [CVCC]C, rhyme [VCC]

The cluster /jsĢ/ and its contextual version /jzć/ fit the rhyme size proposed by Duanmu’s (2009) theory. The final C [Ģ]/[ć] has a morphological solution (the Affix Rule). Neither [js] nor [jz] can be complex sounds because of the already mentioned conflicting gestures (feature specification of /z/ is identical with the feature specification of /s/; they differ only by the presence or the absence of the feature [voice] (see Table 2-5). But the penultimate consonants /s/ and /z/ can be accounted for by the Potential Vowel Rule. The rhyme size is [VC]. In occurrences with long vowels (e.g., in verbs dôjsĢ [duojsĢ] ‘to arrive, to come’, nájsĢ [na:jsĢ] ‘to find’), the final [Ģ] can similarly be unsyllabified (the Affix Rule) and the penultimate C is solved by the Potential Vowel Rule: (33)

dôjsĢ [duojsĢ] ‘to arrive’ – syllable structure CVVC[C[C], rhyme size [VVC] nájsĢ [na:jsĢ] ‘to find’ – syllable structure CVVC[C[C], rhyme size [VVC]

56

Chapter Two

2.7.3 The Syllable Boundaries in the Slovak Word-medial Consonant Clusters The combinatory possibilities of the Slovak consonant sounds are very rich in word-medial position, that is, in intervocalic position. The research aimed at the frequency of the occurrence of the consonant clusters (Sabol 1969a, 1969b) has shown that two-, three-, and four-consonant clusters can be found inside a word in Slovak (the complete list of the most frequent Slovak word-medial consonant clusters is in Appendix 5). Word-medial clusters are usually decomposable into word-initial and word-final clusters (sections 2.3.1, 2.6.3). In the CVX theory, the clusters in the word-medial position are the combinations of the clusters in nonfinal rhymes analysed as word-final rhymes (VV + no consonant or VC) and the clusters in non-initial onsets analysed as the word-initial clusters (no consonant or one consonant is allowed in the syllable-initial position). There are 346 CC clusters (Appendix 5) whose syllabification can be V.CC, C.C, or CC.V. For example, (34)

kvapka ‘drop’ – [kva.pka], [kvap.ka], [kvapk.a] dobre ‘well’ – [ do.bre], [dob.re], [dobr.e] ujma ‘harm’ – [u.jma], [uj.ma], [ujm.a]

The sources used report that 250 CCC clusters (Appendix 5), theoretically, have four possible ways of syllabification: V.CCC, C.CC, CC.C, and CCC.V: (35)

ostro ‘sharply’– [o.stro], [os.tro], [ost.ro], [ostr.o] zajtra ‘tomorrow‘ – [za.jtra], [zaj.tra]. [zajt.ra], [zajtr.a] lístkom ‘with the ticket’ – [li:.stkom], [li:s.tkom], [li:st.kom], [li:stk.om] partner ‘partner’ – [pa.rtner], [par.tner], [part.ner], [partn.er]

And there are 26 frequent CCCC clusters (Appendix 5) that can be syllabified as follows: V.CCCC, C.CCC, CC.CC, CCC.C, or CCCC.V. Although traditionally Slovak does not allow more than three consonants in the final position, the syllabification CCCC.V is included here because some of the clusters can be represented as complex segments that require one timing slot only or extrasyllabic consonants that have a morphological solution: (36)

majstrovstvo ‘artistry’ – [ma.jstro.vstvo], [maj.strov.stvo], [majs.trovs.tvo], [majst.rovst.vo], [majstr.ovstv.o] funkþný ‘functional‘ – [fu.nkþni:], [fun.kþni:], [funk.þni:], [funkþn.i:]

The CVX Theory of the Syllable

57

Duanmu (2009) proposes the so-called Weight-Stress Principle (section 2.6.3) for the syllabification of the consonant clusters in intervocalic position. In the Slovak language, the word stress is fixed, and it is on the first syllable of the word. This means that all word-initial syllables should have the rhyme structure VX and the remaining syllables of polysyllabic words should have to have the rhyme structure V or a syllabic C (ibid., 59). In line with this, CC clusters should be syllabified as C.C and the syllabifications C[C].C and C[CC].C should be chosen for the three- and four-consonant clusters, respectively. Now, it is possible to syllabify the words from (34)–(36): (37)

kvapka ‘drop’ – syllabification [k]vap.ka], 1st rhyme [VC], 2nd rhyme [V] dobre ‘well’ – syllabification, [dob.re], 1st rhyme [VC], 2nd rhyme [V] ujma ‘harm’ – syllabification [uj.ma], 1st rhyme [VC], 2nd rhyme [V]

(38)

ostro ‘sharply’ – syllabification [os[t].ro],39 1st rhyme [VC], 2nd rhyme [V] zajtra ‘tomorrow‘ – syllabification [zaj[t].ra], 1st rhyme [VC], 2nd rhyme [V] lístkom ‘with the ticket’ – syllabification [li:s[t].ko[m], 1st rhyme [VVC],40 2nd rhyme [V] partner ‘partner’ – syllabification [par[t].ne[r], 1st rhyme [VC], 2nd rhyme [V]

(39)

majstrovstvo ‘artistry’ – syllabification [maj[st].ro[vst].vo, 1st rhyme [VC], 2nd rhyme [V], 3rd rhyme [V] funkþný ‘functional‘ – syllabification [fun[kþ].nii, 1st rhyme [VC], 2nd rhyme [VV]41

2.7.4 Summary As a result of the analysis of the word-initial consonant clusters in Slovak (section 2.7.1), not all of these clusters can be explained by morphological concepts from CVX theory or they do not form a complex sound, and, thus, the one-slot onset template is not enough for the Slovak language (section 2.7.1). 39

The possibility of some consonant clusters to be phonemically represented as complex sounds (see Appendix 4) is not included in the syllabification because, as has been shown above, the so-called complex segments do not fulfil the phonetic requirements for their single-slot representation (section 2.5.2). 40 See also section 2.7.2. 41 See section 2.7.2.

58

Chapter Two

Similarly, the proposed two-slot template for the syllable rhymes does not seem to be suitable for Slovak. The maximal structure of the Slovak word-final rhymes is [VVCCC], but neither long monophthongs nor diphthongs can be represented by one timing slot only, as suggested by Duanmu (2009) above. There is no quality difference between related short and long vowels. Moreover, quantity has a phonological-distinctive function in Slovak. The difference of the vowels belonging to the same vocalic timbre has to be represented according to their length. The application of the Affix Rule is limited (section 2.7.2), and not all consonant clusters can be analysed as complex sounds. The possible structure of word-final rhymes in Slovak can be summarized as follows: (1) short monophthong + C (any consonant) – rhyme [VC] (2) long monophthong/diphthong + C (any consonant) – final C solved by morphology (the Potential Vowel Rule, Anti-Allomorphy) – rhyme [VV] (3) short monophthong + CC (complex sound) – rhyme [VCs] (4) long monophthong/ diphthong + CC (complex sound) – rhyme [VVCs] (5) short monophthong + CC (not complex sound) – final C solved by morphology (the Potential Vowel Rule, Anti-Allomorphy) – rhyme [VC] (6) long monophthong/ diphthong + CC (not complex sound): (a) final C possible suffix – final C solved by morphology (the Affix Rule), penultimate C accounted for by the Potential Vowel Rule or Anti-Allomorphy (or both) – rhyme [VV] (b) penultimate C is nasal – final C explained by the Potential Vowel Rule or Anti-Allomorphy (or both), [VVN] analysed as [ࣿࣿ] – rhyme [ࣿࣿ] (c) final C solved by morphology (the Potential Vowel Rule, AntiAllomorphy) – rhyme [VVC] (7) short monophthong + CCC (N + CC as possible complex sound) – final CC represents a complex sound, [VN] simplified as [ࣿ@ – rhyme [ࣿCs] (8) short monophthong + CCC (not complex sound) – final C solved by morphology (the Potential Vowel Rule, Anti-Allomorphy) – rhyme [VCC] (9) short monophthong + CCC (final C is [Ģ]) – final C accounted for by the Affix Rule, penultimate C accounted for by the Potential Vowel Rule – rhyme [VC]

The CVX Theory of the Syllable

59

(10) long monophthong/diphthong + CCC (final C is [Ģ]) – final C accounted for by the Affix Rule, penultimate C accounted for by the Potential Vowel Rule – rhyme [VVC] Clearly, although the highest number of consonants in the final position of Slovak words (syllables) is only three, not all of them can be solved by morphological concepts of Duanmu’s (2009) theory (4, 6c, 8 and 10).

CHAPTER THREE THE SYLLABLE STRUCTURE IN LEXICAL PHONOLOGY

3.1 The Three-dimensional Phonological Representation Lexical Phonology analyses the structure of the syllable on the basis of the so-called three-dimensional phonological representation that was originally proposed by Halle and Vergnaud (1980). The phonological representationʊas indicated by its nameʊshould have three tiers: 1. melodic 2. skeletal 3. syllable These tiers are independent from one another in such a way that the sequence of elements on one tier does not have to correspond linearly to the sequence of elements on the other tier (for other possibilities of the three-tier syllable representation, see, e.g., Clements and Keyser 1983). The number of syllables is identified on the syllable tier. The number of the individual segments (vowels and consonants) is identified on the skeletal tier. The melodic tier determines the characteristic features represented by articulatory movements. In other words, the melodic tier comprises the sequences of the phonetic features, which are connected to the articulatory nodes (see Fig. 2-4). The transcription symbols used on the melodic tier are the abbreviations for the sequences of these features (Halle and Vergnaud 1980, 83–105). For example, a long [a:] is represented as [a] on the melodic tier because during the production of [a:], the position of speech organs is identical with the position of articulators during the production of [a] and is unchanging within the whole articulation. The difference between a long and a short vowel is represented on the skeletal tier. A long vowel is interpreted as one melodic unit (melodic tier) connected to two X-slots (skeletal tier), and it is counted as two segments, as illustrated in figure 3-1.

The Syllable Structure in Lexical Phonology

61

Figure 3-1 Phonological Representations in Three-dimensional Phonology (Rubach 1993, 21)

This three-dimensional representation1 overcomes the difficulties that usually accompany the usage of the [syllabic] feature, which stands somehow apart from the other traditional phonetic features. It is not possible to delimit the syllabic feature in the same way as, for example, the [labial] or [nasal] features, that is, by the creation of an articulatory correlate (Rubach 1993, 22). The main notion for the syllable structure is the sonority hierarchy: on one side of the scale, low vowels are the most sonorous segments. On the other side, consonants (especially stops) are the least sonorous sounds. The most sonorous elements, vowels, are in the centre of the syllable. The other segments to the left of this centre (i.e., in onset) and to the right of the centre (i.e., in coda) are arranged according to the rule known as the Sonority Sequencing Generalization (SSG): “the sonority of segments must decrease towards the syllable margin, according to the following scale: nucleus – liquids – nasals – fricatives – stops” (Rubach 1993, 24). The syllable does not have to contain all classes of the segments mentioned in the SSG but the sequence of the included elements must correspond to this rule. The SSG as the universal rule is completed by the restrictions given by the rules of the given language. These restrictions either violate the SSG (e.g., in Polish, fricatives and stops can combine even vice versa) or make it stricter (in English, there cannot be the combination stop + nasal in onset) (ibid.).

1

In the earlier versions of his work, Rubach uses the term “segmental tier” instead of his later preferred label “melodic tier” (cf. Kenstowicz and Rubach 1987).

62

Chapter Three

3.2 The Syllable Structure Algorithm The syllable structure is not present in the underlying representation. It is determined by the set of rules summarised in the Syllable Structure Algorithm (SSA) that consists of two universal rules, the N-Placement and the CV-Rule, and involves language-specific rules used for the delimitation of onsets and codas: the Onset Rule, the Coda Rule, and the Complex Coda Rule (Rubach 1993, 211). The rule of N-Placement applies to every segment with the feature [consonantal].2 This specification enables the use of the N-Placement Rule in the languages with glides [j] and [u੷], represented as //i// and //u// on the melodic tier. Although phonetically these two glides are very similar to vowels /i/ and /u/, their representation on the melodic tier is identical with the representation of vowels ‘i’ and ‘u’ (see, e.g., Figure 3-12), and phonemically there is a substantial difference between them: glides are never the centre of the syllable. They belong either to onset or to coda. This means that they cannot carry ‘N’. Obviously, the N-Placement Rule overgenerates. Consequently, if there are glides in a language, these are governed by the so-called gliding rulesʊthe Progressive Gliding Rule and the Regressive Gliding Rule. In a word, the N-Placement Rule determines the syllable centre:

Figure 3-2 The N-Placement Rule (Rubach 1993, 221)

Two gliding rulesʊthe Progressive Gliding and the Regressive Glidingʊaccount for glides in a language.

2

The feature [-consonantal] applies to all vowels and glides /j/ and /w/ (e.g., Giegerich 1992).

The Syllable Structure in Lexical Phonology

63

Figure 3-3 The Gliding Rules (Rubach 1993, 229)

The CV-Rule divides sequences of segments into syllables according to the universal generalization that the combination VCV is always syllabified as V.CV and never as VC.V. This rule is also known as the Right-hand Rule (Roach 2000, see also section 6.3.4).

Figure 3-4 The CV-Rule (Rubach 1993, 25)

The Onset Rule and the Coda Rule insert the so-called “rhyme node” (N') between the N and the N''.3 The Onset Rule is actually the Complex Onset Rule because the simple syllable beginnings are derived by the CVRule.

3

N'' stands for a syllable node that is sometimes abbreviated as the sigma sign, ı.

64

Chapter Three

Figure 3-5 The Onset Rule (ibid.)

Figure 3-6 The Coda Rule (ibid.)

The Coda Rule corresponds with the CV-Rule by attaching segments neighbouring the nucleus, as illustrated in the following example:

The Syllable Structure in Lexical Phonology

65

Figure 3-7 The Syllable Structure Algorithm (Rubach 1993, 23)

If there are codas with more than one member, the Complex Coda Rule is used.

Figure 3-8 The Complex Coda Rule (Rubach 1993, 23)

The Onset Rule and the Coda Rule are specific for every language. Sometimes they can be omitted completely. In some languages, there are restrictions concerning the type and the number of the segments that form the marginal syllable parts.

66

Chapter Three

The reason why some rules are applicable only in some languages is that not all languages have a complex onset (i.e., different from that given by the CV-Rule) or a complex coda. But each language has to have the CV-Rule because this rule determines the syllabic node N'' (or ı), so it must be applied, although there is no consonant before a vowel (e.g., in the English monosyllabic word ant. See Fig. 3-9).

Figure 3-9 The Syllabification of the Word ant /ænt/

The formal representation of all mentioned rules (the N-Placement Rule, the CV-Rule, the Onset Rule, and the Coda Rule) is identical in all languages (see above). However, the sequence in which they are applied is different in every language (Rubach 1993, 21–28). As can be seen in figure 3-1, the syllabic structure is created on the skeletal tier and not directly on the melodic representation. The melodic tier makes no difference between, for example, short and long vowels, or the vowels ‘i’, ’u’, and semivowels ‘j’ and ‘w’ because, as explained above, their feature representation at this level is identical. All in all, “[…] the task of the SSA is to group skeletal Xs into syllables. Melodic representations play an indirect role in the sense that the SSA checks them only in order to determine whether a violation of the SSG does not take place. If it does, then the segment remains unsyllabified; that is, it remains extrasyllabic” (Rubach 1993, 219). Following from this is that only those consonants can be attached either to the syllable onset (the Onset Rule) or to the syllable coda (the Coda Rule), which do not disturb the SSG. If their attachment to the given syllable part would mean the violation of the SSG, the consonants are put aside. They cannot be part of the syllable structure. These extrasyllabic consonants are then dealt with differently depending on the language. They can be erased, tolerated, or given a nucleus node, as, for example, in English (ibid.) or in any other language in which sonorants create the centre of a syllable.4 The rule that links extrasyllabic consonants to the N is called Sonorant Syllabification: 4 For further details about the extrasyllabic consonants in generative phonology, see section 1.1.1.

The Syllable Structure in Lexical Phonology

67

Figure 3-10 The Sonorant Syllabification Rule (Rubach 1993, 219)

The solving of extrasyllabic consonants by making them syllabic is a language-specific phenomenon. The whole process is based on the prediction that extrasyllabic consonants themselves can form the centre of the syllable. Those extra consonants that cannot be the syllable centre remain unsyllabified. After Sonorant Syllabification, the SSA recurs and the CV-Rule acts as a resyllabification rule. “The recurrence of the SSA after every phonological (and word-formation) rule is a universal property of a syllable structure algorithm” (Rubach 1993, 27).

3.2.1 The Syllable Structure Algorithm in English As indicated by Rubach (1993) in the analysis of the word trip (Fig. 3-7), the Syllable Structure Algorithm for English proceeds as follows: the NPlacement (each peak of sonority receives its N node), the CV-Rule assigning the syllabification V.CV (instead of VC.V), the Onset Rule (forming the beginning of the syllable, provided that the SSG is followed), the Coda Rule (adding consonants to the end of the syllable, providing that the SSG is not violated), and the Complex Coda Rule, if necessary (Fig. 37, see also Fig. 3-11). Because the beginning of the syllable is created by the CV-Rule, the Onset Rule can sometimes be omitted, depending on the SSG, as will be specified in section 3.2.1.1. Moreover, there are glides in English and, thus, two gliding rules are applied when /j/ or /w/ appear in the sequence of sounds (Fig. 3-12). Sonorant Syllabification accounts for the extrasyllabic consonants (Fig. 3-13).

68

Chapter Three

Figure 3-11 The Syllabification of the Word extra /ekstr̾/

Figure 3-12 The Syllabification of the Word yogurt /jLJg̾t/

The Syllable Structure in Lexical Phonology

69

Figure 3-13 The Syllabification of the Word bottle /bLJtl/

3.2.1.1. The Analysis The order of the rules within the SSA depends on the sonority of segments because the rules should operate so as not to violate the SSG. A significant exception to the order of segments within a syllable on the basis of their sonority is represented by coronal consonants and by the consonant /s/, whose occurrence at the beginning of onsets seems to violate the SSG in many languages (for details, see McMahon 2002, 108),5 as I also exemplify later in my analysis (3.2.3.2.1–2). The sonority of a sound is “its relative loudness compared to other sounds, everything else […] being equal” (Giegerich 1992, 132). In English, as well as in the other languages, there are several less or more detailed versions of the sonority scale. They all agree vowels are the most sonorous sounds, that is, they are always at the top of the hierarchy, while obstruents, that is, fricatives and stops as the less sonorous sounds, are at its end. For example, McMahon’s (2002, 107) hierarchy consists of: low 5

Duanmu assumes not only the special status of the consonant /[s]/ but also of the consonant /[ߦ]/ as already indicated in 2.6.1 (for details, see also, Duanmu 2009, 169–71). However, none of the authors specifies the reason for this “exceptional behaviour”.

Chapter Three

70

vowels–high vowels–glides–liquids–nasals–voiced fricatives–voiceless fricatives–voiced plosives–voiceless plosives, whereas the slightly different Zec’s (2006, 178) scale has a different ordering of fricatives and stops: low vowels–mid vowels–high vowels–rhotics–laterals–nasals– voiced fricatives–voiced stops–voiceless fricatives–voiceless stops. On the other hand, Kenstowicz (1994, 52) introduced a scale of only four degrees of consonant sonority: glides (degree of sonority 4)–liquids (degree of sonority 3)–nasals (degree of sonority 2)–obstruents (degree of sonority 1). To be in accord with the SSG as specified by Rubach (1993) (section 3.1), the sonority hierarchy of the English consonants presented in table 3-1 is taken into account in the following analysis. Table 3-1 The Sonority Hierarchy of the English Consonants sonority degree

type of consonant

8

semivowels [j, w]6

7

liquids [l, r]

6

nasals [m, n, ƾ]

5

voiced fricatives [v, ð, z, ‫ݤ‬, h]

4

voiceless fricatives [f, ș, s, ߦ]

3

voiced affricates [‫]ݶ‬

2

voiceless affricates [ࠔ@

1

voiced plosives [b, d, g]

0

voiceless plosives [p, t, k]

The minimal sonority distance (MSD) for the word-initial (i.e., syllable-initial) onset clusters in English is 2 (Baertsch 2012, 5). This means that the sonority of the second consonant in the CC cluster should be “at least two degrees higher than that of the first” (Duanmu 2009, 168). For example, in the CC cluster [pr]: the sonority degree of [p] is 0, the sonority degree of [r] is 7, and the MSD is 7; that is, [pr] is a good onset 6

In the following parts of this chapter, when a sound is treated as a particular speech sound, square bracketsʊphonetic notationʊare used. When referring to a sound as the element of an abstract system, slanted bracketsʊphonemic notationʊare used.

The Syllable Structure in Lexical Phonology

71

cluster. But, for example, the CC cluster [۟h] does not satisfy the MSD: the sonority degree of [۟] is 4 and the sonority degree of [h] is 5 so the MSD of this cluster is only 1. The number of the word-medial consonant clusters in Englishʊsimilar to other languagesʊis quite high (see Appendix 5). In the case of twoconsonant clusters, theoretically there are two possible syllabifications: when the Onset Rule is not included in the SSA, the syllabification is C1.C2, and when the SSA includes the Onset Rule, the syllabification is V.C1C2. The latter is possible when the SSA is not violated in the sequence V.C1C2; that is, when the degree of the sonority of the C1 is at least two degrees lower than the sonority degree of the C2. The total number of the most frequent two-consonant word-medial clusters in English is 128 (Appendix 5). The sonority-based analysis of these clusters (Appendix 6) has shown that the MSD of the 64 medial CC clusters is below 2 and these clusters are not suitable for creating onsets. That is why they should be syllabified in accordance with the template C1.C2. This type of the syllabification of the CC clusters in a word-medial position, as mentioned above, is given when the Onset Rule is not included in the SSA. For example,

Following the sonority-based analysis, the other 64 word-medial twoconsonant clusters can be syllabified as V.C1C2. The MSD of these clusters is 2 or more (Appendix 6) and the SSG specified for the English onset is not violated by this syllabification. As it is clear, the Onset Rule is included in the SSA when dividing these clusters into syllables. For example,7

7

The aim of this part of the work is not to evaluate the correct or incorrect syllabification of the word-medial clusters. The aim of the analysis in this section is only to show the syllable division according to the given approach.

72

Chapter Three

As for the word-medial three-consonant clusters, if the SSA includes the Onset Rule (i.e., the ordering of the rules is as follows: the NPlacement, the CV-Rule, the Onset Rule, the Coda Rule, the Complex Coda Rule), the word-medial three-consonant clusters are syllabified as C1.C2C3. This type of syllabification is acceptable when the two-consonant clusters in the second syllable follow the MSD for the English initial clusters, which is 2. If the Onset Rule is omitted, the syllabification is C1C2.C3. English has 45 of the most common medial CCC clusters. The syllabification C1.C2C3 (the SSA with the Onset Rule) is suitable for almost all of these clusters. Most of the two-consonant clustersʊ26 of themʊthat arise as the onset of the second syllable have an MSD of at least 2 (see Appendix 6):

Six CC clusters that become the onset of the second syllable when the syllabification is C1.C2C3 start with the fricative /s/, and, thus, they do not obey the sonority rule. However, the irregular behaviour of the consonant /s/ is characteristic for many languages (see section 3.2.1.1), and, thus, these clusters are acceptable as the initial clusters as well:

The Syllable Structure in Lexical Phonology

73

Due to the phonotactic rules of the English language, eight wordmedial three-consonant clusters, those starting with /ƾg/ and /ƾk/, should be syllabified as C1C2.C3 (see Appendix 6); that is, without the Onset Rule in the SSA. For instance,

And, finally, five medial CCC clusters (/njh, mpt, mpk, mps, nࠔb/) also require the division C1C2.C3 because two-consonant clusters given by the syllabification of these consonant sequences as C1.C2C3 do not follow the SSG (Appendix 6):

Only one of the word-medial three-consonant clusters, the cluster /njh/, contains the glide element /j/. This means that the gliding rules are not applicable to CCC word-medial clusters in English. As for the four-consonant word-medial clusters in English, only six of them are considered to be the most frequent (see Appendix 5 and Appendix 6). Four of these clusters can be syllabified as C1C2.C3C4. This type of syllabification fulfils SSG requirements. The MSD for the final clusters in English is less strict than the MSD for the initial clusters (see above). In the coda position, it is enough to follow the SSG. In the sequence C1C2, the sonority degree of the C1 has to be higher than the sonority degree of C2. The syllabification C1C2.C3C4 is given by the SSA that includes the Onset Rule. For example,

The syllabification C1C2.C3C4 is not suitable for two CCCC clusters (Appendix 6). This means that the SSA used for the division of these clusters into syllables should be without the Onset Rule. When the Onset

74

Chapter Three

Rule is not included in the process of syllabification, the third consonant in the sequence C1C2C3C4 remains unsyllabified; however, not in a sense that it is extrasyllabic and can become the centre of a syllable after Sonorant Syllabification is applied (see section 3.2). This is illustrated in the following example8:

3.2.2 The Syllable Structure Algorithm in Slovak In Slovak, SSG is replaced with the Obstruent Sequencing Principle (OSP): “With obstruents there is no requirement of sonority distance” (Rubach 1993, 213). Following OSP, fricatives and stops in consonant clusters can combine in any order and in random mutual combinations, and obstruents of the same class may also combine. This principle has two restrictions: first, the sequences of identical intervocalic consonants are always heterosyllabic. The word brutto is then syllabified as brut.to, not bru.tto. Second, in Slovak, neither identical nor almost identical consonants can form onset and coda. No word in Slovak begins with combinations sš, zž, vf, or gk (ibid., 213–14), as Rubach (1993) concludes9 in his account of the Slovak language in terms of Lexical Phonology. The Syllable Structure Algorithm in Slovak proceeds as the NPlacement, the CV-Rule, the Coda Rule, the Onset Rule and the Complex Coda Rule. The SSA for Slovak is exemplified on the syllabification of the word kompromis ‘compromise’. The application of the rules in the order specified above leads to the syllabification kom.pro.mis (Fig. 3-14). When the Complex Coda Rule is applied before the Onset Rule, the syllable structure of the word would be komp.ro.mis. In the author’s view, this is evidence that the given order of the rules in the SSA specified for Slovak is correct (Rubach 1993, 217–18).

8 Recall that *C in Lexical Phonology indicates that the given consonant is extrasyllabic (see section 3.2). 9 This assumption about the phonotactic possibilities of the Slovak language can be slightly modified. The cluster ‘zž‘ is permissible in the word-initial position. For example, zžiĢ sa ‘to get used to something’ and zženštilý ‘effeminate’.

The Syllable Structure in Lexical Phonology

75

Figure 3-14 The Syllabification of the Word kompromis ‘compromise’ (ibid., 217)

Although the syllabification kom.pro.mis seems to be correct, the word itself does not belong to the centre of the Slovak vocabulary. It is a foreign word and, thus, its syllabification should not be taken as confirmation of the right order of the rules within the SSA specified for Slovak. That is why the SSA is applied to the words that fall in the centre of the Slovak lexicon. Three words with a CCC cluster in the word-medial position were chosen for this sample analysis: mastný ‘greasy’, dáždnik ‘umbrella’ and ostroha ‘spur’. The four-consonant clusters in an intervocalic position are represented by the words klamstvo ‘falsehood’ and horstvo ‘mountains’. The SSA (the N-Placement, the CV-Rule, the Coda Rule, the Onset Rule and the Complex Coda Rule) leads to the syllabifications10:

10

Relevant parts are highlighted.

76

Chapter Three

Despite a different sound environment and the number of syllables, the word-medial three-consonant clusters are always syllabified as C1.C2C3 (see also section 3.2.2.2 below). Rubach’s arguments for the syllabifications in (9) are based on the results of the questionnaire, which are “in agreement with the standard descriptive grammars […]” (Rubach 1993, 216). Nevertheless, standard grammars account for the syllabification of the graphic form of words at the end of the line when writing. The basic rule is that at least one consonant should belong to the previous syllable, which means it should remain at the end of the line (see Pravidlá slovenského pravopisu). These are the orthographic rules, and the syllable is a sound unit. The results of the questionnaire simply prove that native language speakers have a good command of these rules. The native speaker’s judgement on syllabification cannot be considered a relevant criterion. First, it is the information from the graphic level of a language that usually does not correspond with the sound level (see also section 1.3) and second, it has been shown by various research and studies that native language speakers are able to delimit the number of syllables in words (section 1.4) but they have difficulties determining the precise structure of these syllables. Simply said, speakers’ intuition is usually not reliable, as captured in the descriptivists’ slogan: “Accept everything a native speaker says in his language and nothing he says about it” (Sampson 1980, 64).

The Syllable Structure in Lexical Phonology

77

The same criterionʊfrom the results of the questionnaireʊis applied when explaining the role of morpheme boundaries in syllabification. On the basis of native-speakers’ judgement, Rubach (1993, 220) rejects Stanislav’s (1977) recommendation that the morpheme boundary should be respected when dividing words into syllables. However, the morphological structure as a guideline for syllabification is accepted in the case of words with prefixes. When the syllable boundary is put at the boundary between the prefix and the stem, the CV-Rule is violated. The degree of this violation depends on the degree of the transparency of the morphological structure of the given word. Following from this, if the morpheme boundary between the prefix and the stem is now fully lexicalized and it is not clear, the CV-Rule applies across this prefix juncture. For instance, (10)

rozumieĢ ‘to understand’ – morpheme structure roz-um-ie: Ģ (Sokolová et al. 1999) – syllable structure ro.zu.mie (cf. Rubach 1993, 227)

The syllabification of the words with prefixes is governed by the Prosodification Constraint: “Derivation of prosodic structure is blocked by the constituency bracket” (ibid., 228). The CV-Rule is blocked at prefix-stem structures, such as [bez[oký]] ‘eyeless’, but also in compounds [[šesĢ][uholník]] ‘hexagon’ (ibid.). Figure 3-15 shows the syllabification of the word bezoký without the application of the Prosodification Constraint, and figure 3-16 displays the syllabification of the same word where the CV-Rule is blocked by the constituency bracket:

Figure 3-15 The Syllabification of the Word with Prefix following the Basic SSA

78

Chapter Three

Figure 3-16 The Syllabification of the Word with Prefix following the Basic SSA and the Prosodification Constraint

Analogically to the division of the sequences CCC described above, the syllabification of CCCC clusters in the intervocalic position derived by the SSA as specified by Rubach (1993) is always C1C2.C3C4, unless one of the consonants is /j/, as explained below. This means that the first two consonants are part of the coda in the first syllable, and the other two consonants form the onset in the following syllable. The word klamstvo ‘falsehood’ can be used as an example (see also section 3.2.2.2 below):

Figure 3-17 The Syllabification of the Word klamstvo ‘falsehood’

3.2.2.1. Some Special Issues of Slovak Phonology In this section I will return to the problem of the so-called glides in the Slovak language, as indicated above. Within the SSA, a special

The Syllable Structure in Lexical Phonology

79

mechanism is developed for the languages with glides (section 3.2.1). In Rubach’s (1993) view, these are [j] and [u੷] in Slovak. 3.2.2.1.1 Is Slovak ‘ j’ a glide? Accounting for the sound system of Slovak in Lexical Phonology, the generalization is made that “[j] occurs when adjacent to a vowel” (Rubach 1993, 228). Moreover, Rubach sees only one difference between [i] and [j], their function in the syllable; [i] creates a nucleus while [j] is always either in onset or in coda (ibid., 229). However, it has been proven by extensive phonetic research that the articulatory-acoustic properties of the Slovak [j] are different from the articulatory-acoustic properties of the Slovak [i]. [j] is a consonant not only phonemically, as assumed by Rubach (1993), but also phonetically; it is a fricative sound during whose production the back of the tongue creates a narrowing with the hard palate (Dvonþová et al. 1969; KráĐ and Sabol 1987, 250–51). In brief, contrary to English, where ‘j’ is classified as a semivowel due to its phonetic similarity to vowel [i], the Slovak ‘j’ is a consonant phonemically (it occurs in syllable margins) as well as phonetically (a type of obstacle occurring during its production). The phonetic misinterpretation of the Slovak ‘j’ leads to its incorrect representation on the melodic tier in the three-dimensional phonological model. Before the N-Placement Rule is applied, ‘j’ is noted down as //i//. The melodic tier includes the sum of phonetic features of sounds that are abbreviated by transcription symbols (section 3.1). In the Slovak language, symbol [i] stands for a front high vowel and symbol [j] stands for a sonorous, palatal, fricative consonant (Appendix 1). The representation of these two sounds by the same set of features that are abbreviated as [i] (Rubach 1993, 22) results from an incorrect identification of Slovak ‘j’ as a glide. Considering this (details about the phonetic and the phonemic specification of Slovak ‘j’ can be found in Table 2-4 and Table 2-5, respectively), it is obvious that there are no glides in Slovak (see, e.g., KráĐ and Sabol 1987, 275 for further details). And, consequently, the postulation of the gliding rules (section 3.2) for SSA in Slovak is not necessary. This conclusion will be illustrated by the syllabification of the word jajþaĢ ‘to whine’ that is used as an example of the effectiveness of the application of two gliding rules by Rubach (1993, 230).11 Figure 3-18 11 The word jajþaĢ ‘to whine’ is an expressive word that does not belong to the centre of the Slovak lexicon.

80

Chapter Three

displays the syllabification of this word when ‘j’ is treated as a fricative consonant. The result of the application of SSA does not violate the SSG as the basic notion in the process of the syllabification of words in Lexical Phonology.

Figure 3-18 The Syllabification of the Word jajþaĢ ‘to whine’

3.2.2.1.2 The Problem of the Slovak ‘v’ The classification of ‘v’ as a labiodental approximant at the underlying level (Rubach 1993, 244) is another misinterpretation of the Slovak sound resulting from the mixing of the sound systems of (al least) two languages. A ‘v’ in syllable coda is classified as a glide even though there are no glides in the present-day Standard Slovak (Sabol 1989, 275). From the viewpoint of the phonetic representation, Rubach (1993) delimits four different types of ‘v’ in Slovak: (11)

[v] – labiodental approximant,12 e.g., voda ‘water’ [u੷] – labiovelar glide, e.g., stav ‘state’ [v] – labiodental voiced fricative, e.g., vzor ‘pattern’ [f] – labiodental voiceless fricative, e.g., vtip ‘wit’ (ibid., 244)

The author is right about the fact that there are several different phonetic realizations of the phoneme /v/. But the phonetic classification of the first two segments is slightly different from the one provided by Rubach (1993): (12)

[v] – labiodental fricative sonorant, e.g., voda ‘water’ [u੷] – bilabial voiced fricative, e.g., stav ‘state’ [w] – labiodental voiced fricative, e.g., vzor ‘pattern’

12 There are no approximants in Slovak. This term was probably adopted here on the basis of the classification of the English speech segments.

The Syllable Structure in Lexical Phonology

81

[f] – labiodental voiceless fricative, e.g., vtip ‘wit’ (Sabol 1989, 127–28)

Comparing the data in (11) and (12), the phonetic misclassification is clear. The misinterpretation continues. In the analysed syllable theory, “[…] the glide [u੷] differs from the vowel [u] in exactly the same way as [j] differs from [i]. That is, [u੷] and [u] are represented in the same way at the melodic tier but in different ways at the syllabic tier: [u੷] is either an onset or a coda while [u] is a nucleus” (Rubach 1993, 245). Foremost, the identical representation on the melodic tier is incorrect because [u] is a vowel (high back rounded) while [u੷] is a consonant (bilabial voiced fricative). These two sounds have different feature specifications and that is why they should be represented by different symbols (on the difference between [i] and [j], see above.) Second, it is true that [u] occurs in the nucleus but [u੷] can never be found in the onset. It occurs only in the postvocalic position as a combinatory variant of a phoneme /v/13’ that is, only in coda. Moreover, almost all examples used to illustrate the assumed special status of [u] and [u੷] as well as [i] and [j], as described above, are words of foreign origin. They are borrowed from other languages: duel ‘duel’, individuálny ‘individual’, evakuovaĢ ‘to evacuate’, mauzóleum ‘mausoleum’, druid ‘druid’, kotail ‘cocktail’, egoista ‘egoist’, etc. (see Rubach 1993, 212–44). Comparing their Slovak form and the English translation, it is clear that the segmental structure of these words does not reflect the real character of the Slovak phonotactics, and it can also be said that it hardly reflects the rules of the phonotactics of Slavonic languages in general. All these words can be found in the dictionary of foreign words (cf. http://slovniky.korpus.sk). To conclude, in Slovak, the syllabification of words is governed by the Syllable Structure Algorithm that follows the universal Sonority Sequencing Generalization and the heterosyllabic nature of geminate clusters. The universal SSG is modified by the language-specific Obstruent Sequencing Principle, the j-constraint, and the Prosodification Constraint (Rubach 1993, 292). The Prosodification Constraint solves syllabification at the prefix-stem juncture and the syllable boundary in compounds. The function of the j-constraint is to account for glides,

13 The phonemic identification of [u࡬ ] to /v/ is indisputable in Slovak phonology (KráĐ and Sabol 1987, 300).

82

Chapter Three

although, as explained above, there are no glides in Slovak.14 If both ‘j’ and ‘v’ are treated as fricative consonants phonetically and phonemically, as they should be, no further rules and principles in the SSA for Slovak are necessary. Because both ‘j’ and ‘v’ in Slovak display phonetic properties and phonemic behaviour in accordance with the other obstruents, the syllabification of all CC, CCC, and CCCC clusters in the word-medial position (the prefix-stem juncture and the compounds are excluded from this account) is always the same—C1.C2, C1.C2C3 and C1C2.C3C4. Paradoxically, this schematismʊalthough not very convenient for the clusters starting with stops and fricatives (see Appendix 5 and Appendix 6)ʊoffers a satisfactory solution for all two-, three-, and four-consonant intervocalic clusters that start with sonorants /j/ and /v/ (see below). 3.2.2.2. The Analysis All Slovak monomorphemic medial clusters (the whole list is in Appendix 5) are now analysed in terms of the SSA delimited for this language. The SSA includes only the N-Placement, the CV-Rule, the Onset Rule, the Coda Rule, and the Complex Coda Rule. It should be recalled here that the gliding rules are not necessary because there are no glides in Slovak. The Prosodification Constraint can be omitted too due to the character of sample words from which the Slovak consonant clusters15 have been excerpted. Following the order of the rules within the SSA for Slovak, all 346 word-medial two-consonant clusters are syllabified as C1.C2. For example,

14

The special principle should also be applied to the syllabification of the clusters with liquids ‘r’ and ‘l’. The so-called Liquid Syllabification is not mentioned here because it is not relevant from the viewpoint of this work. For details, see Rubach 1993, 232–34. 15 The selected Slovak medial clusters are not the result of the combination prefix + word-formative base or of the combination word-formative base + wordformative base.

The Syllable Structure in Lexical Phonology

83

The syllabification of the three-consonant clusters is C1.C2C3. For instance,

The arising of two-consonant clusters, which become the onset of the following syllable, require the sonority-based analysis whose aim is to verify the correct order of the rules within the SSA specified for Slovak because the SSA should follow the SSG. The sonority hierarchy of the Slovak consonant phonemes is presented in table 3-2. Slovak, similarly to Bulgarian, allows the combinations OO, ON, NN, NL, and OL16 in onset clusters, and it can be characterized as a language whose MSD is 0 (Zec 2006, 189). This means that in both onsets and codas, there can be a sequence of consonants with the same degree of sonority.

16

O stands for obstruent, N means nasal, and L is liquid.

Chapter Three

84

Table 3-2 The Sonority Hierarchy of the Slovak Consonants sonority degree

type of consonant

8

sonorants [r, l]

7

sonorants [m, n]

6

sonorants [j, Đ, Ė]

5

voiced fricatives [v, z, ž, h]

4

voiceless fricatives [f, s, š, x]

3

voiced affricates [‫ݤ‬, ۤ]

2

voiceless affricates [c, þ]

1

voiced plosives [b, d, ć, g]

0

voiceless plosives [p, t, Ģ, k]

After the syllabification of the medial CCC clusters as C1.C2C3, all emerging CC clusters starting with ‘s’ can be considered well formed from the point of view of the SSG due to the cross-language special status of the consonant /s/ in onset clusters (for details, see 3.2.1.1). Then there are only 24 C2C3 clusters out of 250 (Appendix 6, part 2, underlined clusters) that do not satisfy the SSG. Three-consonant clusters with a C2C3 sequence can be syllabified by the SSA without the Onset Rule17 as C1C2.C3. For example, in zoženšĢuje ‘she (or) he makes effeminate’, the syllabification -n.šĢ- violates the SSG but the syllabification -nš.Ģ- satisfies it. The syllabification -j.šk- in the word zovĖajšku ‘to the appearance’ also violates the SSG but the syllabification -jš.k- (when the Onset Rule is not part of the SSA) is in accordance with it. The syllabification of the four-consonant word-medial clusters by the SSA that involves the Onset Rule results in the syllable structure C1C2.C3C4. This type of syllable division seems to be correct for almost all 17

It is important to remember here that Lexical Phonology delimits the SSA on the basis of the SSG. However, it is not specified when the Onset Rule can be omitted and when it is an inevitable part of the algorithm. It is only a prediction that if the syllabification resulting from the SSA with the Onset Rule is not correct from the aspect of the SSG, the Onset Rule can be skipped in the process of syllabification (see also section 3.2.1.1).

The Syllable Structure in Lexical Phonology

85

CCCC medial consonant clusters because the resulting coda and onset clusters follow the MSD specified for Slovak (see Appendix 6, part 3). For example,

Eight final two-consonant clusters that arise during the syllabification C1C2.C3C4 violate the SSG (Appendix 6). The only possibility to syllabify these clusters is by the SSA without the Onset Rule. The syllabification is then C1C2.*C3.C4. This means that the third consonant in the sequence C1C2C3C4 remains unsyllabified.18 For instance,

However, the data in (15) and (16) show that neither the syllabification C1C2.C3C4 nor the syllabification C1C2.*C3.C4 solve the syllable-final C1C2, which does not follow the MSD. The only word-medial five-consonant cluster occurring in the word nervstvo ‘nervous system’ requires the SSA in its whole form (i.e., with the Onset Rule). Even this “full” version of the SSA accounts for only four consonants and thus one consonant remains unsyllabified. In Lexical Phonology, segments in the extreme situation like this “are forcefully adjoined to the syllable node (the Initial Adjunction Rule) or to the phonological word node (the Postcyclic Adjunction Rule)” (Rubach 1993, 292).

18

Recall here that in Lexical Phonology, extrasyllabic consonants that can be the centre of the syllable are accounted for by Sonorant Syllabification (section 3.2), otherwise this consonant remains unsyllabified. It is simply left outside the structure of the syllable.

86

Chapter Three

Figure 3-19 The Postcyclic Adjunction Rule (Rubach 1993, 250)

The results of this “unique syllabification”, as Rubach (1993, 250) mentions, are illustrated in figure 3-20.

Figure 3-20 The Syllabification of the Word nervstvo ‘nervous system’ (Rubach 1993, 250)

The extrasyllabic consonant ‘v’ is linked to the phonological word node that is also called the mot node (Rubach 1993, 249–51).

3.2.3 The Sonority-based Analysis As illustrated above, Lexical Phonology with SSA tries to find the syllable boundary in a word-medial position on the basis of SSG. However, the analysis of the English and the Slovak word-medial consonant clusters has shown that the organization of consonantal segments in the medial position of words is not always in accordance with the principles of SSG. There are many exceptions and irregularities. A kind of criticism of the

The Syllable Structure in Lexical Phonology

87

analysis of (especially) onset clusters19 in terms of the notion of sonority can also be found in Duanmu’s (2009) CVX theory of the syllable (Chapter 2). After reviewing Kenstowicz’s (1994) and Gouskova’s (2004) sonority-based analyses of onset clusters in English, Duanmu (2009) comes to the conclusion that all sonority-based analyses require two additional assumptions: (a) restrictions on sound sequences (phonotactics) besides sonority, such as the prohibition against two alveolar sounds, and (b) the special status of [s] and [ߦ]. Even so, there are missing clusters that are predicted to be good and occurring clusters that are predicted to be bad. (173–74)

Consequently, the author rejects the idea that the segments are combined in the syllable according to their sonority and adopts the idea of the complex-sound analysis (sections 2.4 and 2.5). The sonority-based analysis of all word-medial consonant clusters in English and in Slovak has been presented above. Now, to complete the picture of the relation between the sonority of segments and the way they are organized in the syllable, I briefly introduce the sonority-based analysis of the word-initial clusters in both languages in the following sections. 3.2.3.1. The Sonority-based Analysis of the Word-initial CC and CCC Consonant Clusters in English The list of the English word-initial consonant clusters can be found in section 2.3.2. The sonority hierarchy of the English consonants is specified in table 3-1; the basic principles of the SSG and the specification of the MSD for English onsets are explained in section 3.2.1.1. There are only six word-initial two-consonant clusters in English, [st, sp, sk, sv, sf, ߦp], that do not satisfy the MSD (Table 3-3). However, if the exceptional behaviour of /[s]/ and /[ߦ]/ (2.6.1, 3.2.1.1, cf. also Kenstowicz 1994, 171) is accepted, all English word-initial CC clusters meet the principle of the MSD. Similarly, if the initial [s] is excluded from the analysis of the wordinitial three-consonant clusters ([spl, spr, spj, str, stj, skl, skr, skw, skj]), they all satisfy the MSD.

19

The beginning of the syllable (i.e., its onset) is more sensitive to the principles of SSG than the end of the syllable (i.e., its coda), as can also be seen in sections 3.2.1.1 and 3.2.2.2.

Chapter Three

88

Table 3-3 The Sonority-based Analysis of the English Word-initial CC Clusters the value of the MSD

consonant clusters

8 7 6 5 4

pj, pw, tw, tj, kw, kj pr, pl, bj, tr, dj, dw, kr, kl, gw br, bl, dr, km, kn, gr, gl kv pf, ps, pߦ, tj, șw, șj, w, sj, ߦw

3

fl, fr, vj,vw, șr, sl, sr, ߦr, hj

2

nj, mj, mw, sn, sm, zl, ߦm, ߦn sv sf st, sp, sk, ߦp

1 0 -4

3.2.3.2 The Sonority-based Analysis of the Word-initial Consonant Clusters in Slovak The inventory of consonant phonemes in Slovak as well as their phonetic specification and the subsequent distinctive features specification is slightly different from that in English (cf. Tables 2-1, 2-2 and Tables 2-4, 2-5). The sonority hierarchy of the Slovak consonant sounds is presented in table 3-2. The MSD for Slovak onsets is 0 (section 3.2.2.2). The whole list of the Slovak word-initial consonant clusters can be found in section 2.3.3. 3.2.3.2.1 The Sonority-based Analysis of the Initial CC Clusters in Slovak The final results of the sonority-based analysis of the Slovak word-initial two-consonant clusters are presented in table 3-4. Ninety-five word-initial CC clusters report the excellent value in the SSG; the MSD is not below 0. Twenty-one word-initial two-consonant clusters, [mĖ, mĐ, cx, fþ, cĢ, þp, lž, cv, sp, st, sĢ, sk, zb, šp, št, šĢ, šk, ft, fĢ, vd, lk], do not follow the MSD specified for Slovak. Eight of these clusters can be excluded from further analysis because they start with [s] or [ߖ] (see also section 3.2.3.1). The clusters [mĖ, mĐ] and the clusters [lk, lž] have a relatively low frequency of occurrence (http://korpus.juls.savba.sk). All in all, there are 116 word-initial two-consonant clusters in Slovak. The sonority-based analysis of these clusters has revealed that 21 of them do not have a proper sonority rise. But, if the irregular (starting with [s]

The Syllable Structure in Lexical Phonology

89

and [ߖ]) and rare clusters are excluded, there are only eight CC clusters left that do not follow the MSD 0 required for this language and cannot be accounted for by the sonority-based analysis. Table 3-4 The Sonority-based Analysis of the Slovak Word-initial CC Clusters the value of the MSD 8 7 6 5 4 3 2 1 0 -1 -2 -3 -4 -8

consonant clusters pl, tl, tr, kr, kl pn, bl, br, tm, dl, dr, km, kn, gl, gr pĖ, tĐ, dm, dn, kĖ, kĐ, gr, cl, þl, þr bĐ, tr, dĖ, kr, gĖ, gĐ, cm, cn, cĐ, þm, þn ps, pš, px, bz, dv, sl, šl, šr, xl, xr, fl, fr, cĖ, þĐ tx, sm, sn, zl, zr, šm, šn, žr, hl, hr, xm, vl, vr, þv sĖ, sĐ, zn, šĐ, žm, hn, hm, xĐ, fĐ, vn ml, mr, sv, zĖ, žĖ, žĐ, hĖ, hĐ, xv, vĖ bć, tk, kt, mn, sx, zv, žv, hv, fs, fš, vz mĖ, mĐ xc, fþ, cĢ, þp lž, cv sp, st, sĢ, sk, zb, šp, št, šĢ, šk, ft, fĢ, vd lk

3.2.3.2.2 The Sonority-based Analysis of the initial CCC Clusters in Slovak There are two requirements for three-consonant clusters: (1) the MSD between two adjacent segments should be at least 0, and (2) the closer to the nucleus the segment is, the higher its sonority. Simply said, sonority should increase toward the syllabic nucleus or at least flat sonority (MSD 0) is necessary. But, of course, the sonority of onset elements has to be lower than the sonority of the nucleus. This part of the work deals with onsets only. The minimal requirement of the SSG for Slovak is risingʊor at least flatʊsonority between neighbouring elements (Zec 2006, 190). There is one CCC cluster starting with an oral plosive [tkv] that fulfils the minimal requirement of the SSG, that is, flat sonority between the first two consonants and increasing sonority between the second and the third consonant. One CCC cluster starts with nasal plosive [mdl]. This cluster violates the MSD principle because nasal [m] has a higher sonority degree than voiced plosive [d]. Most of the Slovak word-initial CCC clusters start with a fricative. Twelve onset CCC clusters start with voiceless fricative /[s]/ or /[š]/ and violate the MSD, but the remaining CC clusters (after removing these initial fricatives) show a good sonority distance. This

90

Chapter Three

finding proves the already indicated special status of /s/ and /ߦ/ (i.e., /š/ in Slovak, see also 2.6.1, 3.2.1.1 and Appendix 1) in the sound systems of languages (cf., e.g., McMahon 2002; Duanmu 2009).20 The clusters [vzl, zvl, zhl, hml, hmĐ] satisfy the sonority-based analysis. They have a flat sonority at the beginning that rises toward the nucleus. The cluster [vzd] has flat sonority at the beginning, but the sonority degree of both /v/ and /z/ is higher than that of /d/. The clusters [zdr, zbr] do not meet the sonority requirements. The segments that are the closest to the syllable peak have the highest sonority degree instead of the lowest one. This completely violates the SSG. And the cluster [Đst] has falling sonority instead of flat and rising. However, the frequency of occurrence of this cluster is very low. To sum up, there are 23 onset CCC clusters in Slovak; 12 of them start with consonants /s/ or /‫ݕ‬/ and can be excluded from further evaluation. One cluster can be excluded due to its low frequency of occurrence. Six clusters have good sonority rise. Thus, there are only four CCC clusters that violate the MSD principle. 3.2.3.2.3 The Sonority-based Analysis of the initial CCCC Clusters in Slovak There are only two word-initial four-consonant clusters in Slovak, [pstr, pštr]. They both show rising sonority when analysed as a combination of two CC clusters (see Table 3-4); otherwise, these sequences of consonants violate the SSG principle. But the cluster [pstr] occurs only in a word pstruh ‘trout’ and its derivatives. The cluster [pštr] can also be found in one word onlyʊin the word pštros ‘ostrich’ and its derivatives.

3.2.4 The Sonority and the Structure of the Syllable The sonority-based analysis of the English and the Slovak word-medial (sections 3.2.1.1, 3.2.2.2) and word-initial (section 3.2.3) consonant

20

As already mentioned in note 5, the authors speak about the cross-linguistic exceptional behaviour of the consonants /s/ and /‫ݕ‬/ in terms of the sonority requirements, but the reason for this exemption is not clear; yet. I think that to find the sufficient solution for the problem of sonority sequencing and the sounds /s/ and /‫ݕ‬/, a kind of in-depth, historical, cross-linguistic analysis is necessary. The present-day situation seems to be the consequence of a process (or processes) that happened during the development of languages. Now this topic remains open for further research.

The Syllable Structure in Lexical Phonology

91

clusters21 indicates that the organization of speech segments in syllable onsets and codas is not always fully in accordance with the idea of the SSG. Thus taking the SSG as the only principle when delimiting the syllable structure does not account for all segments in the syllabified word, as can be seen in the example (16) above. Basing the SSA on the sonority requirements only does not solve all consonants in a word, which is now also illustrated by the analysis of the monosyllabic words. This analysis will help delimit the maximal phonological structure of the syllable accepted in Lexical Phonology.

3.2.5 The SSA and Monosyllabic Words 3.2.5.1 The Analysis of the English Monosyllabic Words Traditionally, the maximal phonological structure of the English syllable is CCCVVCCCC, that is, three consonants in onset and four consonants in coda (Fig. 2-6). But the SSA for English only accounts for two consonants in onset. The first consonant in the sequence C1C2C3 remains unsyllabified, as illustrated in figure 3-21. The maximal number of coda consonants solved by the SSA is two as well. This means that the third consonant in the final three-consonant cluster C1C2C3 (Fig. 3-22) and the last two consonants in the word-final four-consonant cluster are not part of the syllable (Fig. 3-23). The maximal structure of the English syllable in Lexical Phonology is CCVCC.22

Figure 3-21 The Syllabification of the English Word-initial Three-consonant Cluster 21

The word (i.e., syllable) final clusters are not included in the sonority-based analysis because the sonority requirements at the end of words (syllables) are less strict than in the initial position of words (syllables). See also section 3.2.1.1. 22 V can be a short monophthong, a long monophthong, or a diphthong.

92

Chapter Three

Figure 3-22 The Syllabification of the English Word-final Three-consonant Cluster

Figure 3-23 The Syllabification of the English Word-final Four-consonant Cluster

3.2.5.2 The Analysis of the Slovak Monosyllabic Words Recall here that in Slovak, the maximal structure of the syllable is CCCCVVCCC (Fig. 2-7). Nevertheless, the SSA delimited for Slovak accounts only for two clusters in onset (Fig. 3-24, Fig. 3-25) and two in coda (Fig. 3-26). All consonants above the structure CCVVCC remain unsyllabified.

The Syllable Structure in Lexical Phonology

93

Figure 3-24 The Syllabification of the Slovak Word-initial Three-consonant Cluster

Figure 3-25 The Syllabification of the Slovak Word-initial Four-consonant Cluster

Figure 3-26 The Syllabification of the Slovak Word-final Three-consonant Cluster

94

Chapter Three

3.3 Summary The sample analysis of the English and the Slovak consonant clusters in the initial, medial, and final position of words (syllables) shows that the SSA allows only two consonants in onset and two consonants in coda in both languages. The maximal phonological structure of the syllable in Lexical Phonology is CCVVCC (Fig. 3-27).

Figure 3-27 The Structure of the Syllable in Lexical Phonology

CHAPTER FOUR THE SYLLABLE IN GENERATIVE PHONOLOGY AGAIN

4.1 On the Universality of the CVX Syllable Theory According to the CVX syllable theory, the maximal syllable template in all languages is only CVX (i.e., CVV or CVC), and any consonants at word edges that extend this model remain extrasyllabic. As already explained (section 1.1.1), the consonants that are part of the word but not part of the syllable can be accounted for in several different ways. The CVX syllable theory assumes that these consonants can be treated as complex segments (section 2.2) or they are explained by the morphological rules (sections 2.1.1–2.1.3). However, the sample phonemic and phonetic analysis of English and the Slovak consonant sequences, in terms of the possibility of the surface representation of some consonant clusters as complex segments (sections 2.4–2.5), questions the cross-language occurrence of complex sounds. Considering the outcomes of this analysis, it seems that in its effort to reduce all syllables to the CVX model, a template-matching CVX syllable theory does not respect the considerable articulatory (and acoustic) diversity among languages in the sense that what is articulatorily possible and acoustically acceptable in one language (multiple articulation of a single timing unit or nasalization of vowels in the neighbourhood of nasal consonants) does not have to fit the articulatory-acoustic features of another language (section 2.5.3). As for the morphological solution of the extrasyllabicity of the consonants at word edges, the basic concepts of the CVX theory (the Affix Rule, the Potential Vowel Rule, and Anti-Allomorphy) do not seem to account for all segments in Slovak words and word forms (section 2.7). And, as proven by the detailed analysis of the word-initial and wordmedial consonant clusters, at least one more timing slot is necessary for both the onsets and rhymes of Slovak syllables (section 2.7.4). Otherwise, many consonants remain unsyllabified either because they do not fit the CVX syllable model or because they do not have a morphological

96

Chapter Four

solution. Moreover, the simplification of the traditionally large consonant sequences into one timing slot is possible in monosyllabic words, but it does not solve the syllabification of the consonant clusters in word-medial positions (section 2.7.3). All these facts challenge the assumed universal nature of the CVX syllable model. Furthermore, the basic assumption of the theory is the possibility that the surface simplification of sound sequences includes vowels. Duanmu (2009, 152) argues that the sequence “vowel + nasal consonant” can be simplified as a nasalised vowel (section 2.6.4) because English vowels are phonetically nasalised when following a nasal coda. That is true; however, taking into account the standard accents of English, there are no nasalised vowels. Speakers and listeners perceive a certain degree of a nasal colouring of vowels in the neighbourhood of nasals (see also sections 6.3.2, 6.3.3) but they do not evaluate those vowels as nasal (see, e.g., Roach 2000). The other example of the overprediction made by this theory is the representation of the long vowel as a short one, as proposed by Duanmu (section 2.6.4). But the reduction of [VV] into [V] means the loss of the distinctive function of quantity in Slovak (section 2.7.2); that is, the violation of its phonological use. It can be assumed that the same would be true for all the other languages in which the quantity (the length of syllabic nuclei) has the ability to distinguish meaning. The syllable is seen as a phonological unit in the CVX theory. The sequences of sounds are represented as the complex sounds on the surface level. Considering this, it is not very clear how the author understands the terms “underlying” and “surface”, “phonemic” and “phonetic” (see also Scheer 2012). When affirming the universal nature of his CVX syllable template, Duanmu (2009), a follower and supporter of the Chomskyan belief of universality (section 1.4), questions Port and Leary (2005), who maintain that there are no universal phonological categories in languages. On the contrary, each language forms its own categories, which differ from those in other languages (Duanmu 2009, 240). Duanmu’s argument against this claim is “if each language makes up its own syllable structure, we expect syllables to differ from language to language, probably dramatically” (ibid.). And on the basis of Chomsky and Halle’s division of linguistic universals,1 the author adds the maximal syllable size CVX for all languages to the list of formal or substantive universals (ibid., 241). 1

Formal universals “determine the structure of grammars and the form and organization of rules” (Chomsky and Halle 1968, 4). Substantive universals “determine the set of elements that may figure in particular grammars” (ibid., see also Duanmu 2009, 241).

The Syllable in Generative Phonology Again

97

So, is the CVX theory of the syllable acceptable cross-linguistically, as claimed by its author? The answer is yes; however, not on the surface, as supposed by the author, but on the underlying level as understood by traditional generative grammar as it is part of our mental lexicon and innate universal grammar (section 1.4). Generative linguistics sees phonology as “the study of the mental representations of sound structures” (Blevins 1995, 160), and these mental or phonological representations “provide input to the phonetic interpretative component” (178). Considering this, as well as the indisputable mutual relationship between phonetics and phonology (see Chapter 6), it can be concluded (although the author of the CVX theory would definitely not agree with this conclusion) that the CVX syllable model is acceptable as an abstract model that reflects the origin and the very substance of the syllable in languages (this is the universal aspect), and the number of segments in onsets and rhymes differs from language to language, depending on the phonological possibilities of the given language (this is a language-particular aspect of the CVX syllable theory). Finally, the role of morphology in the process of syllabification (section 1.2) is not very transparent in the CVX theory. Because “the CVX theory derives the variations from morphology” (Duanmu 2009, 51), the author labels the theory “the morphological approach”. However, the difference between the syllable and the morpheme is not clear throughout the whole book, as exemplified by the author’s account of the possible and actually-used syllables in languages: “The set of syllables a language needs is rather small. For example, English and Chinese each have about 10 000 morphemes of which only half are frequently used” (Duanmu 2009, 252; italics my emphasis).

4.2 On the Universality of the SSA in Lexical Phonology Taking into account what has been said about the delimitation of the syllable in Lexical Phonology (Chapter 3), it is obvious that the SSA (the set of rules determining the structure of the syllable) is less strict than the CVX syllable model when assuming the universality of the syllable and its structure. Although Lexical Phonology “was proposed as a refinement of classical generative phonology” (Goldsmith and Laks 2011, 14), it also incorporates ideas from pre-generative structuralist phonology and natural phonology (ibid.). Thus, in Lexical Phonology, the syllable is determined to follow the universal principles connected with the syllable but it also accepts a certain degree of diversity among languages. Consequently, the SSA is set up to follow the ideas that are supposed to occur across

98

Chapter Four

languagesʊthere are syllables in each language, and the basic syllable model is CV. The ordering of segments in the syllable should follow the sonority of segments in a way that the sonority must decrease toward the margins of the syllable (sections 3.1–3.2). SSA also follows the rules and principles that occur in the individual languages onlyʊthe degree of the sonority of the individual speech sounds and the minimal sonority distance between elements in the syllable onsets and codas (cf. sections 3.2.1, 3.2.2). Because there are many combinations of speech segments in languages that violate the principle of SSG, the syllabification resulting from this rule-based approach (cf. Blevins 1995, 171) is not exhaustive, and the segments that do not follow SSG remain extrasyllabic. The unsyllabified consonants (except those which themselves can be the centre of the syllable, that is, syllabic consonants) are not given any special solution; they are simply tolerated (section 3.2, see also section 1.1.1). However, despite the acceptance of the universality as well as the diversity in the structure of the syllable, the application of the rules included in the SSA specified for the given language (section 3.2.5) allows a maximum of two segments in onset and two in coda in both languages analysed (section 3.3). The great advantage of Lexical Phonology is the use of a three-tier representation of the syllable that enables linguists to “see” not only the structure of the syllable (which is determined on the surface2) and its constituents but also the structure of the individual segments (section 3.1, see also Mohanan 1982, 158). The aim of Lexical Phonologyʊto find out how phonology and morphology interact under various phonological rulesʊis reflected also in the process of the specification of the rules for the delimitation of the syllable. Contrary to the CVX syllable theory, where the relation “syllable–morpheme” is not always clear (section 4.1), the morphological structure of words is respected in cases of words with a prefix and compounds (the Prosodification Constraint, Fig. 3-16).

2

The surface level, or the surface representation, known in LP as the phonetic representation, is used to refer to the output of all the phonological rules. The term underlying representation refers to the phonological representation of morphemes in the set of lexical entries, whereas the phonological representation of words in the set of lexical entries is called the lexical representation in Lexical Phonology (Mohanan 1982, 100).

CHAPTER FIVE THE SYLLABLE IN STRUCTURALISM

Structuralism1 in linguistics is, of course, connected with the name of its founder, Ferdinand de Saussure (section 5.1) and especially with the Prague School of Linguistics (for details, see, e.g., ýerný 1996, 148–64); that is, inter alia, well known due to the “evolvement of a compact theory of modern phonology” (145). Two key phonologists must be mentioned in connection with phonology of the Prague School and its understanding of the syllable—N. S. Trubetzkoy and R. O. Jakobson (section 5.2).

5.1 Ferdinand de Saussure and the Syllable In Saussure’s (1922, 10, 103) phonology, the syllable is a phonological unit given by the subdivision of a spoken chain because “syllables are easier to identify than their sounds” (50). And the structure of the syllable is discussed in terms of so-called explosion–implosion combinations. In 2003, Laks commented, It is the temporal integration in the succession of implosions and explosions that defines, constricts and structures the syllable. In Saussurian phonotactics, it is the localization of the syllabic point within the succession of explosive and implosive chunks that orders the division of syllables and regulates their internal organization. (209)

Both explosion and implosion differ from the viewpoint about their length. Explosion is very rapid and never makes a vocalic impression. On the other hand, implosion is measurable and can make a vocalic impression. A sound that makes a vocalic impression is a vocalic peak. Vocalic peaks have also been called sonants, and all other sounds in the same syllable con-sonants [consonants]. Vowels and con-sonants designate different

1

See note 1 in Introduction.

100

Chapter Five species […]; sonants and con-sonants, on the other hand, designate functions within syllables. (Saussure 1959, 57–58)

Saussure divided speech segments into three groups of sounds: consonants (known as obstruents in modern linguistics), vowels (these are non-high vowels in modern linguistics), and sonorants, that is, sounds realized in one way or the other depending on the context (Goldsmith 2011, 167). The realization of sonorants is governed by a set of four rules (for details, see Goldsmith 2011) that must be applied from right to left, that is, from the end of the word to its beginning. Saussure (1959, 49–64) also sees the syllable as a unit necessary for the identification of phonemes.

5.2 The Syllable in the Prague School of Linguistics Both Saussure’s description of the syllable as the succession of the explosion and the implosion and his understanding of the syllable as a basic unit of phonological organization were ignored by the phonologists from the Prague School. Trubetzkoyʊone of the founders of phonology in the Prague Schoolʊhad to reinvent the syllable when concentrating on the suprasegmental features with potentially distinctive functions, that is, on stress and tone (Cairns and Raimy 2011, 10). He identified the syllable or, more precisely, the syllabic nucleus “as the locus of distinctive prosodic properties” (ibid.). Thus, for Trubetzkoy, the syllable was the unit important for the identification of prosodic features and the basis for the delimitation of the prosodic unit (Vachek 2003, 153). As for the relationship between the syllable and the prosodic unit, Trubetzkoy (1939) says, Even from a purely phonetic point of view, the “syllable” is basically something quite different from a combination of vowels and consonants. The phonological prosodic unit is, of course, not simply identical with the ‘syllable’ (in the phonetic sense). However, it is either a specific segment of the syllable or the entire sequence of syllables. (95)

Trubetzkoy (1939) was also the first to see the syllable as a unit with an internal structure, a unit that is crucial for describing phonemes and phoneme sequences (for details, see 46–65). When discussing syllabic nuclei, Trubetzkoy (1939) admits the existence of the syllable as the bearer of distinctive prosodic properties of language universally; however, the syllabic nucleus “must be defined differently for various languages” (170). That is, the structure of the syllable depends on “the laws of the particular language” (ibid.).

The Syllable in Structuralism

101

For Jakobson, the syllable is the elementary pattern underlying any grouping of phonemes (Jakobson and Halle 1956, 20). It is a unit necessary for the delimitation of distinctive features: “The pivotal principle of syllable structure is the contrast of successive features within the syllable” (ibid.). The authors also introduce an interesting typology of syllable inventories, and maintain that CV is the universal syllable type present in all languages. “There are languages lacking syllables with initial vowels and/or syllables with final consonants, or both, but there are no languages devoid of syllables with initial consonants or of syllables with final vowels” (Jakobson 1962, 526). The schemes CVC, V, or VC can also be found in many languages. As for the structure of this basic CV syllable model, Jakobson and Halle (1956, 21) say, “Both part C and part V may contain more than one phoneme. The phonemes constituting parts V and C of the syllable are termed crest phonemes and slope phonemes respectively”. All in all, Prague School phonologists emphasize the function of the syllable as the unit necessary for the realization and perception of prosodic features and for the identification of phonemes. Prague School phonology, following the langue–parole dichotomy, also makes a difference between the syllable as a langue (phonological) and parole (phonetic) entity (Cairns and Raimy 2011, 10). At the level of speech (parole), the syllable was described predominantly in terms of the articulatory and acoustic differences between the individual sounds in the syllable centre and its margins (see, e.g., Hála 1956). At the abstract level, the level of language (langue), the emphasis was put on understanding the syllable as a unit necessary for the identification of phonemes (see, e.g., Pauliny 1979). A kind of synthesis of this functional-structural approach to the syllable can be found in the Synthetic Phonological Theory (SPT) by J. Sabol (1989). Although it goes back 27 years, it can still be seen as a unique representation of European structuralism because of its views on the structure and function of the syllable. An in-depth description of the syllable within the SPT can be found in the following chapter.

CHAPTER SIX THE SYNTHETIC PHONOLOGICAL THEORY

6.1 General Outline The Synthetic Phonological Theory (SPT) was developed on the basis of analysing the sound material of Standard Slovak. But due to its general linguistic subsoil its main ideas and principles can be applied to other languages, too, as shown in this chapter. The SPT comprises the theories of the Prague School and the Moscow Phonological School with the aim, as its author mentions, of avoiding “mistakes that accompany some phonological conceptions: ‘forced’ adherence to ‘their’ level of abstraction when defining sound elements, inability to look at sounds in complexity and, thus, to have their own ‘partial’ truth” (Sabol 1989, 6). The philosophical point of departure of the SPT is the existence of the different levels of the relationship between the individual and the universal, that is, between the concrete and the abstract. The linguistic shaping of these gnoseological categories is reflected in the oppositions langue–parole, language–speech, paradigm–syntagma, and sociolect– idiolect as well as phone–phoneme, sentence–utterance, connotation– denotation, etc. (ibid., 14–15). Phonetic elements represent the first level of abstraction, and the phonemic elements are from the second level of abstraction. The units of the language system can move in both directions of these levels (categories). The movement from the individual to the universal can be exemplified by the phonologization of sound elements or by the extension of the word meaning. The reverse movement from the universal to the individual, that is, from the abstract to the concrete, can be exemplified by dephonologization of sound units or by the narrowing of word meanings. Of course, at each level of abstraction, there is “a dialectic link between the universal and the individual, only their mutual rearrangement changes” (ibid., 17), as illustrated in figure 6-1.

The Synthetic Phonological Theory

103

Figure 6-1 The Relation between the Individual (I) and the Universal (U) (see also, Sabol 1989, 20)

In the direction from the individual to the universal, the author delimits five levels of the relation between the variant and the invariant, or the concrete and the abstract: (1) hypophone (HPh) – the concrete physical speech segment, the concrete unreproducible sound that can be pronounced only once; (2) phone (Ph) – a sound, a speech signal segment; (3) phoneme (Phm) – an abstract sound unit materialized as the bundle of distinctive features, a sound element able to distinguish the meaning of words and word forms; (4) morphophoneme (MPhm) – the sound “construct” delimited in the morpheme as the smallest bearer of the elementary or complex meaning; (5) hypermorphophoneme (HMPhm) – the sound “pre-basis” of the morpheme from which the other units were generated as they originated historically (ibid., 18–19). In this conception of the individual sound levels, the higher-level unit represents the universal against the lower-level unit, which is the

104

Chapter Six

individual. For example, the phone is the individual against the phoneme, but it is the universal against the hypophone. The phone “takes” only those features from the hypophone that become relevant, that is, the ones entering the phoneme as the universal, and also those features that are irrelevant from the viewpoint of the phoneme. This means that the higher the abstraction, the higher the universal, that is, the typology in the sound systems of languages (see also Table 6-1). As for the relations among the three main levels of this abstraction, the level of the phone, the phoneme, and the morphophoneme, according to the extent of these units, there is the inclusion: Ph < Phm < MPhm. When the content is the criterion taken into account, there is the reverse inclusion: MPhm < Phm < Ph. The basic relations between Ph, Phm, and MPhm can be determined as follows: the phone is identical with the phoneme or differs from it by the phonologically irrelevant feature or features; the phoneme is identical with the morphophoneme or differs from it by the phonologically relevant feature or features (for details, see Sabol 1989, 22–27). The above delimited relationships result from uniting the relationship between the individual and the universal, the concrete and the abstract (i.e., the variant and the invariant) and can be described in terms of certain variations of form that are divided into four groups: (a) the variations of the 1st degree – the individual articulatory-acoustic characteristics (b) the variations of the 2nd degree – the specific orthophonic phenomena (c) the variations of the 3rd degree – neutralizations1 (d) the variations of the 4th degree – alternations2 1

Neutralization, “the loss of distinction between two phonemes in a particular linguistic environment” (Štekauer 1995, 286), is a change of a phoneme triggered by another phoneme. For example, in Slovak, a West Slavonic language, the change of alveolar /n/ into alveo-palatal /Ė/ when inflecting the noun vrana ‘a crow’: Nominative vran-a ‘a crow’> Dative vran-e /vraĖ-e/ ‘to a crow’ is caused by /e/ as a phoneme. This is a phonological process. 2 Alternation, a variation in a phonological realization of a phoneme or a morpheme (e.g., Štekauer 1995), is a change induced by another morpheme (Sabol 1989). For example, in Slovak, the change of alveolar /n/ into alveo-palatal /Ė/ when creating the diminutive form of the noun vrana ‘a crow’: vran-a ‘a crow’> DIM vráĖ-a ‘a little crow’ is caused by /a/ as a diminutive morpheme. The same diminutive morpheme in this example is also responsible for the change of a short vowel ‘a’ in vran-a ‘a crow’ into a long vowel ‘á’ in vráĖ-a ‘a little crow’. This is a morphological process.

The Synthetic Phonological Theory

105

The specification of the individual sound levels (units) in terms of these variations can be found in table 6-1. Table 6-1 Variation Specification of the Individual Sound Levels level of sound

hypophone phone phoneme morphophoneme hypermorphophoneme

variations of the 1st degree + -

variations of the 2nd degree + + -

variations of the 3rd degree + + + -

variations of the 4th degree + + + + -

6.2 The Syllable and the Synthetic Phonological Theory Within the framework of the SPT, the syllable is a complex sound unit that connects the phone, the phoneme, and the morphophoneme with the higher-level phonic elements (units of rhythm and their groups) as well as the semantic elements (morpheme, word). The complex nature of this sound unit is proven by the fact that its function interferes with all three parts of the communication act: articulatory (the syllable reflects the physiology of speech sounds in the order stricture + aperture, i.e., C + V), acoustic (acoustically, the syllable represents the transition from silence to sound), and perceptual (the better the onset phase of the syllable is phonetically delimited, the better we perceive the syllable) (Sabol 1994, 217–18). The syllable is defined as “the basic syntagmatic-paradigmatic and phonic-rhythmical unit with one peak of sonority in which the contrast and distinctive features of phonic elements are manifested and which is the bearer of all suprasegmental phenomena (prosodemes) of a given language” (ibid., 217). The complex (or synthetic) approach is also applied when looking for the syllable boundary in the word-medial, that is, inter-sonantic, position. The following criteria can be used for the syllabification of polysyllabic words with large consonant clusters: 1st criterion – the interface of the semantic and the phonic part of a linguistic sign: In many languages, the syllable (as a sound unit) is often identical with the morpheme (as the unit of the meaning). Thus, when there is a consonant cluster in the word-medial position, the syllable boundary can be placed at the morpheme boundary. This criterion is based on the well-known symmetry between the formal (phonic) and the meaningful component of a language, or of a linguistic sign.

106

Chapter Six

2nd criterion – the transgressive nature of consonants: The acoustic features of certain consonants have their reflection on the formant field of the following vowels. This “overlap” of acoustic properties represents a kind of acoustic “symbiosis” of the neighbouring speech sounds. Consonants in such a position (i.e., before vowels) are then part of the onset of a given syllable. 3rd criterion – the coarticulation: The coarticulation can be seen as an articulatory “version” of the acoustic symbiosis of the neighbouring sounds. 4th criterion – the degree of stricture: This criterion follows the claim of B. Hála (1956) that the higher the degree of stricture in a sound’s articulation, the better it is for creating the beginning of the syllable, that is, its onset. 5th criterion – the phonotactics of the given language + Kury‫ן‬owicz’s rule: If the beginning of the word is at the same time the beginning of the first syllable of the word, the end of the word is at the same time the end of the last syllable, and in the given language no word begins with more than, for example, three consonants, then the sequence CCVC1C2C3C4V will be syllabified as CCVC1.C2C3C4V. 6th criterion – the frequency criterion: This is the information about the frequency of the occurrence of the word-initial and the word-final consonant clusters in a given language. 7th criterion – the power of the syllable, ‘welds’3: This criterion results from the delimitation of the syllable as a minimal contrastive unit of a language. The optimal contrast is given by the natural ordering of sounds in the syllable, that is, consonant + vowel (Pauliny 1979, 169–70). This is the strongest ‘weld’ because it has the highest number of contrasts (Sabol 1994, 219–21). In the following part of the work, first, all these criteria for the delimitation of the syllable boundary within a word are applied to Slovak and to English. Then the criteria are ranked on the axis of “the individual ļ the universal”. The results of the syllabification of Slovak and English word-medial consonant clusters are summarized and commented upon at the end of this chapter.

3

The word ‘weld’ was used by Pauliny (1979) in his account of the syllable. It indicates the connection between the adjacent elements or the juncture.

The Synthetic Phonological Theory

107

6.3 The Criteria for the Syllabification 6.3.1 The Interface of the Semantic and the Phonic Part of a Linguistic Sign The determination of the syllable boundary at the place of the morpheme boundary comes from the relationship between the form and the meaning in the structure of a linguistic sign. This relation is very flexible and dynamic: the symmetry is applied along with the asymmetry. The relationship of the symmetry and the asymmetry between the units of the form and the content in a language can be illustrated by the relationship between the syllable as the sound unit and the morpheme as the smallest part of a word that is the bearer of basic or complex meaning (Ondruš and Sabol 1987, 164). In cases where there is no significant difference between the syllable boundary and the morpheme boundary, the syllable boundary is placed at the morpheme juncture. This type of syllabification proves the symmetry in the relationship between two facets of the linguistic sign (Sabol 2011, 35). At the same time, it also confirms another universal truth—in the relationship between sound and meaning (the form and the content), meaning is the decisive one (see also 6.4) In Slovak, the semantic part of the language represented by the morpheme can determine one third of the syllables identified in the individual lexical units (Gregová 1998). On the other hand, considerable tension between the syllabic and morphemic boundaries can be found at two positions, as specified by Sabol (2011, 35): (1) at the borderline between the root morpheme and the derivational morpheme; the basic phonematic structure of the root morpheme in Slovak is CVC and the basic structure of derivational suffixes is VC (see also, Sokolová et al. 1989); and (2) at the borderline between the word-formative base ending with a consonant, which creates the coda of the last or the penultimate syllable of the word-form, and a grammatical morpheme with the basic phonematic structure V, VCV, VC (Sabol 2011, 35). For example, the syllabification of the word odžiĢ ‘live out’ is od.žiĢ and the morpheme structure of this word is od-ži-Ģ.4 The structure of the first syllable is identical to the structure of the derivational prefix. When the syllable boundary is at a different place than the morpheme boundary, that is, the relation between these two units is asymmetric, the syllable (a sound unit) is either inside the morpheme structure (the unit of meaning) ʊthat is, the syllable is part of the morpheme, for example, 4

od- is a derivational prefix, ži- is a root morpheme and -Ģ is a form morpheme that expresses verb forms (see, e.g., Ondruš and Sabol 1987, 166).

108

Chapter Six

(1)

oáza ‘oasis’ – syllable structure o.á.za, morpheme structure oáz-a5

(2)

obar ‘cooked pig`s head’ – syllable structure o.bar, morpheme structure obar-ø

ʊor the syllable consists of two or more morphemes, for example, (3)

objednávka ‘order’ – syllable structure ob.je.dnáv.ka,7 morpheme structure ob-jedn-áv-k-a8

(4)

odchlípnúĢ ‘to sip’ – syllable structure od.chlip.núĢ, morpheme structure od-chlíp-n-ú-Ģ9

At this stage of the analysis, the structure of a derivational affix usually helps identify the first and last syllable of a word, as in the examples (3) and (4) above (see also sections 6.2.1.1 and 6.2.1.2.). The division of words into syllables according to the morpheme structure is connected with the phonological phenomena that occur at the borderlines of morphemes; these are neutralizations and alternations10 (for details, see Pauliny 1979, 183–87, see also notes 1 and 2). This criterion can be characterized as a phonological or morphophonological one.11 5

oáz- is a root morpheme and -a is a grammatical morpheme. The structure obar-ø means root morpheme + zero grammatical morpheme. 7 The syllabification ob.jed.náv.ka is also possible. However, the syllabification ob.je.dnáv.ka is preferred. For details, see below. 8 The morpheme structure ob-jedn-áv-k-a means derivational prefix–root morpheme– derivational morpheme–derivational morpheme–grammatical morpheme. 9 The morpheme structure od-chlíp-n-ú-Ģ indicates derivational prefix–root morpheme–derivational morpheme–thematic morph–form morpheme. 10 Derivational and inflectional processes may put together such combinations of sounds (phonemes) that are not typical of positions inside syllables and morphemes, and cause articulatory difficulties. That is why the morpheme boundary is a place of various consonant changes, that is, assimilations. The aim of these changes is to exclude these combinations of sounds (phonemes) on the morpheme boundary that do not occur inside the morpheme or the syllable (for details, see KráĐ 2005, 49). 11 Although the syllable is a sound unit of language (speech), its structure cannot be described without meaning, which has its reflection in the CVX theory (Chapter 2), as well as in Lexical Phonology (Chapter 3). However, in the CVX theory, the morpheme dominates over the syllable in a way that destroys the basic structure (some consonants remain outside the syllable structure) and function of the syllable as the unit of the formal part of a language, and, thus, it also violates the symbiosis between the form and the meaning of a linguistic sign in language. 6

The Synthetic Phonological Theory

109

6.3.1.1 The Sample Analysis of the Slovak Language The syllabification of words on the basis of morpheme structure is not new in Slovak linguistics (see, e.g., Pauliny 1979 for further details). The distinct morpheme boundary is between (a) word bases in compound words (e.g., desaĢ-roþný ‘decennial’, vlakvedúci ‘conductor’) (b) a derivational prefix12 and a word base (e.g., ob-zvlášĢ ‘particularly’, roz-dvojiĢ ‘bifurcate’) (c) a word base and a derivational (rarely inflectional) suffix (e.g., horstvo ‘mountains’, žen-ský ‘female’) The placement of the syllable boundary at the morpheme juncture confirms the correspondence between the morpheme structure and the syllable structure of the given language, that is, the symmetry between the form and the meaning. The words mentioned above can be syllabified as follows: (5)

desaĢ-roþný ‘decennial’ – morpheme structure desaĢ-roþ-n-ý – syllable structure de.saĢ.roþ.ný

(6)

vlak-vedúci ‘conductor’ – morpheme structure vlak-ved-úc-i – syllable structure vlak.ve.dú.ci

(7)

ob-zvlášĢ ‘particularly’ – morpheme structure ob-zvlášĢ – syllable structure ob.zvlášĢ

(8)

roz-dvojiĢ ‘bifurcate’– morpheme structure roz-dv-oj-i-Ģ – syllable structure roz.dvo.jiĢ

(9)

hor-stvo ‘mountains’ – morpheme structure hor-stv-o – syllable structure hor.stvo

(10)

žen-ský ‘female’ – morpheme structure žen-sk-ý – syllable structure žen.ský

Taking into account the most productive derivational affixes in Slovak, the idea about the identity of the syllable boundary and the morpheme boundary accounts for the syllabification of the words containing the derivational suffixes -stvo and -ský. Four-consonant clusters with the combination ‘stv’ should be syllabified as C.stv. This solves the 12

There are no inflectional prefixes in Slovak.

110

Chapter Six

syllabification of 15 word-medial four-consonant clusters and the rare five-consonant cluster (see Appendix 5). Analogically, all three-consonant clusters with the combination ‘st’ can be syllabified as C.st. 6.3.1.2 The Sample Analysis of the English Language The relation between the morpheme structure and the syllable structure can be used as a tool for syllabification in English, too. Similar to Slovak, only the sample analysis is presented here. Examples below show words with prefixes, their morpheme structure, and a suggested syllable structure. The application of the different criteria may lead to the different syllabification, as illustrated: (11)

instead /Ѣnsted/ – morpheme structure Ѣn-sted – syllable structure Ѣn.sted

(12)

inactive /Ѣn ktѢv/ – morpheme structure Ѣn- kt-Ѣv – syllable structure Ѣn. k.tѢv

(13)

import /Ѣmpэ:t/ – morpheme structure Ѣm-pэ:t – syllable structure Ѣm.pэ:t

(14)

postpone /pΩ҂stpΩ҂n/ – morpheme structure pΩ҂st-pΩ҂n – syllable structure pΩ҂st.pΩ҂n

(15)

transform /tr nsfɲ:m/ – morpheme structure tr ns-fэ:m – syllable structure tr ns.fэ:m

As can be seen in examples (11)–(15), in comparison with Slovak, the correspondence between the morpheme boundary and the syllable boundary is even more obvious in English. The reason is the morphological difference between English and Slovak. In English, an analytic language, only root morphemes and derivational morphemes help solve syllabification. In Slovak, a synthetic language, the morpheme structure is much richer. Apart from root morphemes and derivational morphemes, there are also inflectional morphemes, modificational morphemes, form morphemes, and congruation morphemes (for details, see Ondruš and Sabol 1987, 163–69). Now, a few examples show the morpheme and syllable structure of words with derivational suffixes in English:

The Synthetic Phonological Theory (16)

mouthful /ma҂șf҂l/ – morpheme structure ma҂ș-f҂l – syllable structure ma҂ș.f҂l

(17)

wisdom /wѢzdΩm/ – morpheme structure wѢz-dΩm – syllable structure wѢz.dΩm

(18)

kindness /kaѢndnΩs/ – morpheme structure kaѢnd-nΩs – syllable structure kaѢnd.nΩs

(19)

endless /endlΩs/ – morpheme structure end-lΩs – syllable structure end.lΩs

111

To illustrate the close relation between the structure of the syllable and the morpheme in English, consider the syllabification and the morpheme analysis of the word in example (20): (20)

unselfconscious /ࣱnselfkLJnߦ̾s/ – morpheme structure ࣱn-self-kLJnߦ̾s, historically ‘conscious’ was formed on ‘con + sci’, that is, the morpheme structure ࣱn-self-kLJn-ߦ̾s can be acceptable (Hoad 1986) – syllable structure ࣱn.self.kLJn.ߦ̾s

6.3.2 The Transgressive Nature of Consonants The opposition of transgression–inherence that was originally introduced by Romportl (1976) is very important in the acoustic analysis, based on the segmentation of the flow of speech. Consonants are labelled as “transgressive” because certain features of their acoustic spectrum overlap into the sound spectrum of the neighbouring sound, which is then necessary for the correct identification of the given consonant. On the other hand, vowels are called “inherent” because all important acoustic features are realized within their own sound spectrum (Romportl 1962, 284). When segmenting continuous speech, sometimes it is not possible to find the precise border between the neighbouring sounds. A case in point is the combination nasal consonant + vowel. Because the important acoustic features of the nasal consonant infiltrate the acoustic spectrum of the following vowel, it is quite difficult to find the accurate borderline between these two sounds.13 Consequently, a consonant that needs the 13

Probably this finding led Duanmu (2009) to the conclusion that the combination ‘vowel + nasal consonant’ is actually the nasal vowel, that is, two sounds are perceived as one (Chapter 2; section 4.1). However, the mutual–either articulatory or acoustic (or both)–phonetic interconnection between the neighbouring speech

112

Chapter Six

neighbouring vowel for its acoustic realization forms a syllable with the given vowel. Similar situations can be observed during the segmentation of geminate consonants or when parsing the combinations sonorant + vowel (ji), sonorant + sonorant (rn), stop + sonorant (pl, pr), stop + stop (st), etc. These “tight bonds” between neighbouring sounds have their reflection in the structure of the syllable. If speech sounds are acoustically interconnected, they form the onset or the coda of the same syllable (see also Vachek 1989, 37). When segmenting the flow of speech into smaller units, that is, speech segments (vowels and consonants), the graphic representations of a sound in the form of an oscillogram14 or a spectrogram15 are usually used. The major categories of sounds are characterized by a certain shape of an oscillographic curve and a spectrogram. The differences in the sound wave or in the structure of the spectrogram given by the differences in the acoustic structure (depending of the differences in the production of sounds, as specified below) serve as a tool for the segmentation of continuous speech into smaller units. Keeping in mind the basic information about the spectral characteristics of each category of sound, it is not difficult to identify individual speech segments in the flow of speech. Nevertheless, when the acoustic structures of the neighbouring sounds are interconnected, it is also reflected in the form of an oscillogram or sonogram. In such cases, it is very difficult or even impossible to delimit the boundaries between these sounds (Figure 6-2). The examples proving the validity of this assumption as well as the material for the sample analysis of the Slovak and the English language were taken from the sound material that was used in the complex-sound analysis (section 2.5).

sounds does not have to mean that these sounds cannot be perceived as distinct and discrete (see also section 6.3.3). 14 An oscillogram is a visual representation of a sound wave that gives information about the time a sound takes (on the x axis) and its amplitude (on the y axis). See, for example, figure 6-2. 15 A spectrogram is a graphic representation of a sound whose x axis represents the time duration of the sound. The y axis represents the frequency and the intensity of the sound at any particular time as indicated by the colour of the spectrogram at that point (Johnson 2008, 42–44). See, for example, figure 6-3.

The Synthetic Phonological Theory

113

Figure 6-2 The Oscillogram and the Sonogram of the Word krajiny ‘countries’

6.3.2.1 The Sample Analysis of the Slovak Language The acoustic analysis of the sound material of the Slovak language has shown that the combination sonorant + vowel forms one acoustic unit; the acoustic features overlap and the borderline between these two speech segments is blurred. The shape of the oscillographic wave and the spectrogram of the word krajiny ‘countries’ (Figure 6-2) exemplify this situation. Nasals are relatively short in an intervocalic position compared to their occurrence in other sound environments, which also proves their transgressive nature. The acoustic analysis also reveals some other combinations that are difficult to divide: [ćĖ], [pr], and [zć] as well as [mn] or [ml], as illustrated in figure 6-3.

114

Chapter Six

Figure 6-3 The Oscillogram and the Sonogram of the Cluster [mn]

If the assumption considered in this section is correct (and the sound material indicates this is true at least for the acoustic “co-existence” of sonorants), word-medial two-consonant clusters consisting of sonorants (Appendix 5) should together be the onset of the following syllable because the acoustic spectra of the neighbouring nasals overlap. For example, (21)

humno ‘barn-floor’ – syllable structure hu.mno prítomný ‘present’ – syllable structure prí.to.mný

This criterion can be used in the syllabification of three-, four-, and five-consonant clusters with the sequence of sonorants as well: (22)

rozmlátiĢ ‘to trash’ – syllable structure roz.mlá.tiĢ nadvláda ‘domination’ – syllable structure nad.vlá.da

(23)

bernský ‘of Bern’ – syllable structure bern.ský

(24)

nervstvo ‘nervous system’ – syllable structure nerv.stvo

6.3.2.2 The Sample Analysis of the English Language The analysis of the sound material of English shows that, similar to Slovak, sonorants in the intervocalic position are shorter than sonorants in

The Synthetic Phonological Theory

115

other sound environments due to the opposition of transgression– inherence. However, generally speaking, the borderlines between the individual segments are more distinct in English than in Slovak (see, e.g., Gregová 2012b). Following this, the transgressive character of the Slovak consonants seems to be stronger than the transgression of the English consonantal segments. Nevertheless, the acoustic “fusion” of neighbouring sonorants, predominantly nasals, is observable in English, too (see note 13). The observation is reflected in the syllabification of the words with the sequence sonorant + sonorant: (25)

randomly [r ndΩmlѢ] – syllable structure r n.dΩ.mlѢ16

(26)

firmly [fє:mlѢ] – syllable structure fє:.mlѢ

6.3.3 The Criterion of Coarticulation Coarticulation is a well-known phenomenon of many languages. It is a type of articulation during which there is a reciprocal influence of the articulatory (motoric) movements of sounds (Dvonþová 1980, 76).17 This articulatory influence of neighbouring sounds, or in other words, the transition of a certain articulatory feature of one sound to another, results in the modification of certain phases of sound production. For example, in all languages, one can observe “some degree of coarticulatory nasalisation of vowels adjacent to nasal consonants” (http://www.personal.rdg.ac.uk/~llsroach/phon2/asscoareli-into.htm). All “coarticulatory processes are more likely to be looking for effects, which are found (not necessarily in exactly the same form) in all languages because they are due to mechanical and biological limits on what the articulators can do in a given amount of time” (ibid.). Coarticulation is connected with the syllable in the sense that the articulatory orders are given not for the production of the individual 16 Recall again that the chapters on the different approaches to the syllable structure analysis (Chapter 2, 3, and 6) do not solve the problem of syllabification in the words with consonant clusters. The aim of these chapters is to show different possibilities of the division of words into syllables when applying different criteria. 17 Recent phonetic research tends to use the term coarticulation as a synonym for the term assimilation. Nevertheless, in this work, these are two notions referring to different processes. Coarticulation can simply be defined as the mutual articulatory influence of neighbouring sounds, and assimilation is a process by which a certain feature of a segment is modified by its neighbours (for details, see, e.g., http://www.personal.rdg.ac.uk/~llsroach/phon2/asscoareli-into.htm).

116

Chapter Six

speech sounds but for the production of the higher-level units, that is, syllables (KráĐ and Sabol 1989, 146). Considering the physiology of speech organs, the combination consonant + vowel is the most natural ordering of sounds within the syllable because the final phase of the production of a consonant is at the same time the so-called transient phase to vowel, at the beginning of which there are changes of vocalic formants that help identify the consonant (Romportl 1985, 100). Taking into account this phonetic necessity of speech production, the syllabification of the sequence VCV should always be V.CV and not VC.V, as accepted by some syllable theories (see, e.g., Duanmu 2009).18 According to Recasens (1999, 94), “[…] data reveal a trend for the vowel to show more tongue body repositioning than the consonant in [...], thus indicating the existence of stronger C-to-V than V-to-C effects”. The phenomenon of coarticulation has its roots in the way the human brain controls the production of speech. When we speak, many muscles are active at the same time and sometimes the brain wants them to make sudden changes that they are not capable of. For example, in the word none [nࣱn], the vowel is normally an oral sound pronounced with the soft palate raised to prevent the escape of air through the nose, while the two ‘n’ sounds must have the soft palate lowered because /[n]/ is a nasal sound during whose production the air escapes through the nasal cavity. But the soft palate cannot be lowered and then raised and lowered again as quickly as necessary in the given string of sounds so the vowel is likely to be pronounced with the soft palate still lowered, giving a nasalized quality to the vowel. The nasalization is a coarticulation effect caused by the nasal consonant environment (Roach 2000). This is the so-called velopharyngeal or nasal coarticulation (Chafcouloff and Marchal 1999). The thorough cross-language research of articulatory control in speech production reports the existence of the other types of coarticulationʊlingual, laryngeal, labial, and velar (Hardcastle and Hewlett 1999). But only the 18

It has been proven by various experimental and theoretical research (e.g., Ladefoged and Maddieson 1996; Sagey 1986; Roach 2000) that both phonetic phenomena briefly described here (acoustic symbiosis of sounds in section 6.2.2 and coarticulation in section 6.2.3) occur in all languages; however, the degree to which they apply is language specific (see, e.g., Dvonþová 1980 for details on the so-called genetic coarticulation). The notion of coarticulation and its acoustic consequences as specified here is the basis for the theory of the existence of complex segments in languages. However, this phonetic or articulatory-acoustic property of speech sounds is asserted to a different degree in different languagesʊas has been shown in the complex-sound analysis of English and Slovak consonant clusters (sections 2.4–2.5).

The Synthetic Phonological Theory

117

nasal coarticulation is directly connected with the structure of the syllable; that is why only this type of coarticulation is exemplified in both languages. 6.3.3.1 The Sample Analysis of the Slovak Language The extensive cross-linguistic research in the given field (see Chafcouloff and Marchal 1999 for further details) provides physiological, acoustic, and perceptual evidence that nasal sounds are strongly co-articulated with the adjacent segments, especially vowels (73). Two-consonant clusters with the structure VNC or CNV should be syllabified as VN.C and C.NV, respectively. This type of syllabification accounts for 92 word-medial twoconsonant clusters in Slovak (Appendix 5). For example, (27)

obmena ‘change’ – syllable structure ob.mena verejný ‘public’ – syllable structure ve.rej.ný

(28)

þinka ‘dumb-bell’ – syllable structure þin.ka minca ‘coin’ – syllable structure min.ca

Eight word-medial CC clusters have the structure NN and can be syllabified as N.N or the cluster as a whole forms the onset of the following syllable19, as exemplified below: (29)

þinnosĢ ‘activity’ – syllable structure þin.nosĢ zimný ‘winteradj’ – syllable structure zim.ný or zi.mný

One hundred and twenty-two CCC word-medial clusters contain a nasal sound that can be in a different position in the cluster. The structures NCC, CCN, or NCN dominate (Appendix 5). According to the principles of nasal coarticulation, the nasal consonant should form articulatory unity with the adjacent vowel. This leads to the syllabification N.CC, CC.N, and N.CN/NC.N. For example, (30)

omdlieĢ ‘to faint’ – syllable structure om.dlieĢ

(31)

vlastnosĢ ‘quality’– syllable structure vlast.nosĢ

(32)

dominantný ‘dominant’ – syllable structure do.minan.tný or do.minant.ný

19 This type of syllabification is not possible in the case of non-nasal geminate consonants.

118

Chapter Six

Four-consonant clusters with the nasal consonant in the initial position prevail among the four-consonant word-medial clusters in Slovak (Appendix 5). These clusters can be syllabified as N.CCC: (33)

panstvo ‘nobility’ – syllable structure pan.stvo

6.3.3.2 The Sample Analysis of the English Language The nasal consonants are also frequent in the English word-medial clusters. In the word-medial consonant clusters adopted for the purposes of this work (Appendix 5), there are 47 clusters with the structure NC or CN and two NN clusters. The syllable boundary can be placed between the consonants in all these clusters: (34)

landed /l ndԘd/ – syllable structure l n.dԘd

(35)

badminton /b dmԘnt̾n/ – syllable structure b d.mԘn.t̾n

(36)

environment /ԘnvaԘ̾r̾nm̾nt/ – syllable structure Ԙn.vaԘ̾r̾n.m̾nt

Forty three-consonant clusters with the nasal consonant have the structure NCC, and they can be syllabified as N.CC: (37)

apparently /̾p rΩntlԘ/ – syllable structure ̾p rΩn.tlԘ

In four-consonant clusters, the nasal consonant is the first one and the clusters can be syllabified as N.CCC: (38)

Sanskrit /s nskrԘt/– syllable structure s n.skrԘt

However, the cluster /‫و‬kst/ should be syllabified as NC.CC due to the context-dependent velar coarticulation (Hardcastle and Hewlett 1999): (39)

gangster /g ‫و‬kstΩ/ – syllable structure g ‫و‬k.stΩ

6.3.4 The Degree of Stricture The degree of stricture as a factor that determines the structure of the syllable was originally formulated by B. Hála (1956) in his syllable theory. The starting point for Hála’s delimitation of the syllable is the assumption

The Synthetic Phonological Theory

119

that the syllable is not a contemporary “product” but the consequence of a tradition inherited during the millennia. At the beginning of the development of human speech, the most important thing was to produce a sound and release the cavities above the glottis so the sound could get out of the body (Hála 1956, 276). Phonationʊthe creation of a soundʊpresents itself by the same properties as any other sound, by intensity, tone, time, and timbre. These modulations are realized within the syllable as the unit of speech: intensity increases from the beginning of the syllable toward its centre then it decreases; tone usually slightly rises toward the syllable centre and then it falls, and timbre is best perceived in the syllable centre, too. The length of the syllable means the length of its centre (ibid.). Hála (1956, 279) defines a syllable as “the transition of speech organs from stricture to aperture while the pure laryngeal sound forms”. When there is no stricture, there is no syllable. If a word, that is, a syllable, begins with a vowel, there is the physiological stricture in the form of the glottal stop, the most complete physiological stricture.20 The differences between speech sounds are given by the modification of the phonation stream. Vowels are given by the modification of aperture, and consonants are the result of the modification of stricture (ibid., 281). This means that the stronger the stricture, the better the beginning of the syllable is perceived. The syllable boundary is at the place of the strongest stricture so that the stricture creates the beginning of the following syllable. A consonant with considerable aperture (e.g., a sonorant) is part of the previous syllable if the consonant that follows this apertural one is considerably strictured (ibid., 288–89). For example, the word nájde ‘she (or he) will find’ is syllabified as náj.de because /j/, as a sonorant, has a lower degree of stricture than the voiced plosive /d/. This articulatory property of sounds (the degree of stricture) is reflected in their acoustic characteristics as the degree of sonority. Sounds produced by stricture (obstacle) have low degrees of sonority. The stronger the stricture during the articulation of the sound, the lower the sonority of the given sound. Thus, sounds with the lowest degree of sonority (i.e., with the strongest stricture) are the most suitable for creating the beginning of a syllable. The ranking of the English and the Slovak consonants according to the degree of stricture during their production is in table 6-2 and table 6-3, 20

For example, the syllable structure of the Slovak word ak ‘if’ is /VC/ at the level of phoneme, but at the level of phone it is [CVC] because the basic phonation position of the glottis is represented by the glottal stop [‫ސ[ – ]ސ‬ak]. Similarly, the syllable structure of the English word if is as follows: [‫ܼސ‬f] – [CVC], /ܼf/ – /VC/.

Chapter Six

120

respectively (see also Table 3-1 and Table 3-2 for the classification of consonants in terms of their sonority). Table 6-2 The Hierarchy of the English Consonants according to the Degree of Stricture degree of stricture

type of consonant

8

voiceless plosives p, t, k

7

voiced plosives b, d, g

6

voiceless affricates ࠔ

5

voiced affricates ‫ݶ‬

4

voiceless fricatives f, ș, s, ߦ

3

voiced fricatives v, ð, z, ‫ݤ‬, h

2

nasals m, n, ƾ

1

liquids l, r

0

semivowels j, w

Table 6-3 The Hierarchy of the Slovak Consonants according to the Degree of Stricture degree of stricture

type of consonant

8

voiceless plosives p, t, Ģ, k

7

voiced plosives b, d, ć, g

6

voiceless affricates c, þ

5

voiced affricates ‫ݤ‬, ۤ

4

voiceless fricatives f, s, š, x

3

voiced fricatives v, z, ž, h

2

sonorants j, Đ, Ė

1

sonorants m, n

0

sonorants r, l

The Synthetic Phonological Theory

121

The sample analysis of the English and the Slovak consonant clusters in accordance with this phonetic criterion about the degree of stricture is presented in the following subsections. 6.3.4.1 The Sample Analysis of the Slovak Language One hundred and ninety-two word-medial two-consonant clusters21 can be treated as heterosyllabic; that is, they can be deconstructed into two syllables (C1.C2) either because the stricture of the second consonant in the sequence is higher than the stricture of the first consonant or both consonants have the same degree of stricture and the syllabification C1.C2 is acceptable. For example, (40)

ústav ‘institute’ – syllable structure ús.tav cenný ‘valuable’– syllable structure cen.ný

The remaining 155 CC clusters should form the onset of the following syllable22 because the first consonant in these clusters is more strictured (i.e., less sonorous), and, thus, more suitable for creating the beginning of the syllable. For example, (41)

jednota ‘unity’ – syllable structure je.dno.ta kufre ‘cases’ – syllable structure ku.fre

Following this criterion about the syllable boundary at the place of the narrowest stricture, 40 CCC word-medial clusters in the sequence VCCCV 21 Two-consonant word-medial clusters that can be syllabified as C1.C2: /pt, pĢ, pk, bd, bć, tp, tĢ, tk, db, dć, dg, Ģp, Ģt, ĢĢ, Ģk, ćb, dd, dć, kp, kt, kĢ, kk, gd, gć, gb, cp, ck, cc, cþ, þp, þt, ck, þĢ, þk, þþ, ‫ݤ‬b, ‫ݤ‬d, ‫ݤ‬ć, ‫ݤ‬ž, ۤb, ۤg, sp, st, sĢ, sk, sc, sþ, ss, šc, sš, sf, zb, zd, zć, zg, zz, zž, zh, zv, šp, št, šĢ, šk, þc, šš, žb, žd, žć, žv, hd, hž, hv, xt, xĢ, xk, xc, xþ, xš, fp, ft, fĢ, fk, fš, vb, vd, vć, vk, vc, vþ, vۤ, vz, vh, mp, mh, bx, mf, mv, mm, mn, mĖ, mĐ, nt, nĢ, nd, nć, nk, ng, nc, nþ, n‫ݤ‬, nۤ, ns, nš, nz, nž, nh, nx, nf, nv, nm, nn, nĖ, nĐ, nj, Ėb, Ėh, lp, lb, lt, ld, lć, lk, lg, lc, lþ, lš, lz, lh, lx, lf, lv, lm, ln, ll, lr, Đb, Đt, Đd, Đć, Đk, Đþ, Đš, Đž, Đh, Đv, Đn, ĐĖ, rp, rb, rz, rž, rh, rx, rf, rv, rm, rn, rĖ, rl, rĐ, rr, rj, jp, jb, jt, jd, jĢ, jć, jk, jg, jc, jþ, js, jz, jš, jh, jx, jf, jv, jĖ, jĐ, jj/. 22 The word-medial CC clusters that should be together in the onset of the following syllable: /pc, pþ, ps, pš, px, pm, pn, pĖ, pl, pĐ, pr, bz, bž, bh, bv, bm, bn, bĖ, bl, bĐ, br, ts, tš, tx, tf, tv, tm, tn, tĖ, tl, tĐ, tr, dh, dv, dm, dn, dĖ, dl, dĐ, dr, dj, ćh, ćm, ćn, ćl, ćr, kc, kþ, ks, kš, kv, km, kn, kĖ, kr, gm, gz, gž, gh, gv, gn, gĖ, gl, gĐ, gr, cs, cx, cf, cv, cm, cn, cĖ, cl, cĐ, þv, þm, þn, þĖ, þl, þĐ, þr, ‫ݤ‬ž, ‫ݤ‬h, ‫ݤ‬v, ‫ݤ‬m, ‫ݤ‬n, ‫ݤ‬Ė, ‫ݤ‬Đ, ‫ݤ‬r, ‫ݤ‬j, ۤh, ۤr, sv, sm, sn, sĖ, sl, sĐ, sr, zm, zn, zĖ, zl, zĐ, zj, šv, šm, šn, šĖ, šl, šĐ, šr, žm, žn, žĖ, žĐ, žr, hv, hm, hn, hĖ, hl, hĐ, hr, xv, xm, xn, xĖ, xl, xĐ, xr, fĐ, fn, fĖ, fl, fr, vn, vĖ, vl, vĐ, vr, bx, ml, mr, nr, jm, jn, jl, jr/.

Chapter Six

122

(/psk, pkl, psl, psĐ, pkr, pšĢ, pxl, pxn, pxĖ, tkl, tkĐ, tpr, tpl, tkn, tkĖ, tkr, tkv, tpĐ, tfþ, txn, tsĢ, tsv, ttl, tšk, tfš, txl, txĐ, txv, tsl, tsp, tšm, tšt, tšĢ, Ģkr, ktr, ktn, ktĖ, kþn, zmĐ, hmĐ/) can be syllabified as V.C1C2C3 because the first consonant in these clusters is the most strictured one. In addition, all word-medial CCC clusters in Slovak, besides 25 of them, /vzd, vžd, nšĢ, vzb, nšp, nšt, tfþ, jšk, jšĢ, mšt, mšĢ, pšĢ, tšk, vzć, dvć, jvć, nþĢ, tšĢ, Ģšt, fšĢ, ršt, jft, jfĢ/, can be syllabified as C1.C2C3 (if the exceptional behaviour of /s/ is taken into account, as specified in Appendix 6 and already commented upon in sections 3.2.1.1 and 3.2.2.2). The 25 irregular CCC clusters should be syllabified as C1C2.C3 to follow the rule analysed here. For example, (42)

dotknúĢ ‘to touch’ – syllable structure do.tknúĢ or dot.knúĢ týždne /tíždĖe/ ‘weeks’ – syllable structure tíž.dĖe

(43)

inšpektor ‘inspector’ – syllable structure inš.pek.tor/inš.pe.ktor

If four-consonant word-medial clusters are to be syllabified at the place of the narrowest stricture, 17 of them, /nstv, rstv, Đstv, mstv, šstv, nštr, lstv, cstv, jstr, vstv, jstv, þstv, jstn, jzdr, fskĐ, jskr, rstr/, should be syllabified as C1C2.C3C4. For example, (44)

þlenstvo ‘membership’ – syllable structure þlens.tvo

Seven four-consonant clusters have two consonants representing the minimum of stricture (/tstr, pstr, pstv, dzbr, tškr, kstr, kspl/), and, thus, there are two syllabifications possible in the sequence VCCCCCʊeither V.C1C2C3C4 or C1C2.C3C4. For example, (45)

biskupstvo ‘episcopacy’ – syllable structure bis.ku.pstvo or bis.kups.tvo

The cluster /rnsk/ should be syllabified as C1C2C3.C4: (46)

kalifornský ‘Californian’ – syllable structure ka.li.forns.ký

and the cluster /nkþn/ needs the syllable boundary after the first consonant C1.C2C3C4: (47)

funkþný ‘functional’ – syllable structure fun.kþný (cf. Table 6-3)

The Synthetic Phonological Theory

123

In the five-consonant word-medial cluster /rvstv/, the consonant /t/ represents the narrowest stricture. This means that the syllable boundary should be placed before this consonant: (48)

nervstvo ‘nervous system’ – syllable structure nervs.tvo

The syllable boundary at the place of the narrowest stricture can be proven, even visually, by means of the so-called sonority graph (Fig. 6-4, 6-5). The mirror inscription of the individual segments as they appear in the graph depends on the degree of their sonority and reliably indicates the place with the lowest sonority, that is, the narrowest stricture. Figure 6-4 visualizes the syllabification of the word pestrý ‘colourful’, which encompasses a three-consonant cluster [str]. Figure 6-5 confirms that the syllabification of the word nervstvo ‘nervous system’ in (48) is correct.

Figure 6-4 The Sonority Graph of the Word pestrý ‘colourful’23

23

The sonority hierarchy of the Slovak consonants can be found in Table 3-2. The 10th degree belongs to vowels, and the 9th degree is typical of the so-called semivowels [i࡬ ] and [u࡬࡬ ] (the first part of Slovak diphthongs).

124

Chapter Six

Figure 6-5 The Sonority Graph of the Word nervstvo ‘nervous system’

6.3.4.2 The Sample Analysis of the English Language If the aim of syllabification is to put the syllable boundary at the place of the narrowest stricture, 52 two-consonant word-medial clusters in English, /nd, nt, mp, mb, ƾg, st, ns, ƾk, kt, nv, lk, lt, sp, nz, nf, lb, ls, lv, lm, lf, mz, mf, pt, ft, mn, nm, ms, lg, lp, n۟, zb, ln, l‫ݶ‬, nࠔ, nh, nߦ, sf, mk, mt, mߦ, md, z˔, gb, pk, tk, lࠔ, lr, lߦ, ߦt, dp, f۟, s۟/, should be syllabified as C1.C2. The second consonant has a narrower stricture than the first one (Table 6-2). Or both consonants in the sequence display the same value of stricture, and then the syllabification C1.C2 is acceptable. For example, (49)

intentionally /ѢntenѻΩnΩlѢ / – syllable structure Ѣn.ten.ѻΩ.nΩ.lѢ hospital /hыspѢtΩl/ – syllable structure hыs.pѢ.tΩl

Seventy-six CC clusters, /ks, nj, sk, lj, kj, mj, gm, ts, tr, kn, zm, tj, zl, gn, gj, pj, bj, br, fr, kr, dl, sj, fl, kߦ, ml, gl, pr, ps, tl, vr, dn, fj, sm, kw, km, kl, kd, zj, zn, gz, gr, pn, tn, tf, vj, lw, bl, b˔, bs, dm, dj, kb, mr, gw, pm, pl, tw, tm, vl, ߦm, ߦn, ߦr, nl, ۟m, ۟l, ۟j, bn, dr, dv, ðm, fg, fn, sw, sl, sn, sb/, require the syllabification V.C1C2 because the most strictured segment is the first consonant in the sequence. For example, (50)

absolute / bs̾lu:t /– syllable structure  bs̾lu:t sharpness /ߦľ:pn̾s/ – syllable structure ߦľ:.pn̾s (see also Fig. 6-6)

The Synthetic Phonological Theory

125

Figure 6-6 The Sonority Graph of the Word sharpness [ߦľ:pn̾s]24

In the group of the word-medial three-consonant clusters, there are several possibilities for syllabification, depending on the position of the most strictured consonant in the CCC sequence. Thirty of these consonant clusters, /ndr, ntr, str, mbr, ‫و‬kw, mpr, mpl, nfr, nkr, ltr, ‫و‬gw, ‫و‬gr, ‫و‬kl, mdr, mfr, mkw, mbl, mfl, ‫و‬gh, ‫و‬gs, mps, ldr, lfr, lgr, lpr, nsl, ntl, n۟r, ntm, spr/, are syllabified as C1.C2C3, as illustrated in (51) and attested by the sonority graph of the word esprit (Fig. 6-7): (51) 



ultra- /ࣱltr̾/ – syllable structure ࣱl.tr̾ esprit /espri:/ – syllable structure es.pri:

Clusters /njh, nst, lst, ‫و‬gt, nࠔb, nsk/ need the division C1C2.C3 because the last segment is the least sonorous (i.e., the most strictured). For example, (52)

24

constitute /kLJnstԘtju:t/ – syllable structure kLJns.tԘtju:t

The sonority hierarchy of the English consonants is in table 3-1. Vowels have either a degree of 10 (open vowels) or 9 (closed vowels).

126

Chapter Six

Two clusters (/ksw, dzw/) should as a whole form the onset of the following syllable, as illustrated below: (53)

waxwork /w kswͩ:k/ – syllable structure w kswͩ:k

Clusters /‫و‬kt, mpt, mpk/ can be syllabified either as C1.C2C3 or C1C2.C3; cluster /ktr/ can either form the onset of the second syllable or can be divided into C1.C2C3. Finally, consonant sequences /kst, ksp, pst/ can be syllabified as C1C2.C3 or the whole CCC sequence may be in the onset of the second syllable. For example, (54)

prompting /prLJmptԘ‫و‬/ – syllable structure prLJm.ptԘ‫و‬orprLJmp.tԘ‫و‬ expert /ekspͩt/ – syllable structure e.kspͩt25 or eks.pͩt

Figure 6-7 The Sonority Graph of the Word esprit [espri:]

Taking into account the criterion about the degree of stricture, the syllabification possibilities of the most common word-medial fourconsonant clusters in English are rich, too. The syllabification C1C2.C3C4 is suitable for the clusters /nstr, mstr, nskr/, as can be seen in figure 6-8 (see also Table 6-2). The clusters /kstr, kskl/ can be divided either as V.C1C2C3C4 or as C1C2.C3C4, and the divisions C1C2.C3C4 or C1C2C3.C4 are good for the cluster /‫و‬kst/. For example, 25

This syllabification of the word expert is only theoretically possible.

The Synthetic Phonological Theory

(55)

Sanskrit /s nskrԘt/ – syllable structure s ns.krԘt

(56)

exclaim /ԘkskleԘm/ – syllable structure ԘkskleԘm26 or Ԙks.kleԘm

127

Figure 6-8 The Sonority Graph of the Word Sanskrit [sænskrܼt]

6.3.5 The Phonotactics of the given Language and Kury‫ן‬owicz's Rule The syllable is the basic unit of phonotactic analysis; that is, the syllable is the basic unit that determines the possible combinations of phonemes in a given language.27 English allows maximum clusters of three consonants in onset and four consonants in coda (Fig. 2-6). In the onset consisting of three consonants, only /s/ can be the first, and consonants /l, r, w, j/ are good as the third ones in the sequence CCC. Consonants /v, ˙, z, ঱never occur in onset clusters. Consonants /t, d, ۟/ cannot form an onset cluster with /l/. As for the English coda, an English syllable can end with any

26 27

This type of the syllabification is only theoretically possible. For the view that phonotactics is not syllable based, see Blevins 2003.

128

Chapter Six

consonant except /h, r, w, j/. Consonants /s, z, t, d, ۟/ are good for ending any type of coda cluster. /lg/ is not a possible coda cluster and the sequence nasal stop + oral stop is acceptable only if these consonants have the same place of articulation (for details, see, e.g., Giegerich 1992; Roach 2000; McMahon 2002). In Slovak, a maximum of four consonants can combine in onset and three in the syllable coda (Fig. 2-7). The four-consonant onset clusters are relatively rare, and they usually start with combinations /fs/ and /ps/. The most frequent three-consonant clusters start with the consonant /s/ (Sabol 1969a, 1969b). There are no two-consonant onset clusters consisting of the sequence sonorant + other type of consonant (i.e., obstruent). Two sonorants can combine in onset only if the first one is /m/. In coda, there cannot be the combination of another type of consonant (obstruent) + sonorant (Sabol 1994, 222). The most frequent three-consonant coda cluster is /jsĢ/. The most natural order of sounds in both onsets and codas is fricative consonant + occlusive consonant (for details, see Sabol 1975, 204–8). Kuryáowicz's (1948) idea that the beginning of the word is at the same time the beginning of the first syllable and the end of the word is at the same time the end of the last syllable makes two predictions: (1) word boundary determines the syllable boundary28 and (2) the structure of a monosyllabic word helps identify the structure of the syllables wordinternally. For example, the English word extra /ekstrԥ/ can be syllabified as /ek.strԥ/ because the maximal structure of the English syllable onset is CCC with /s/ as the pre-initial consonant. The Slovak word sestra ‘sister’ can be syllabified as se.stra because Slovak allows three-consonant onset clusters starting with /s/. 6.3.5.1 The Sample Analysis of the Slovak Language In short, Kuryáowicz's (ibid.) rule together with the rules of phonotactics determines the number and the structure of the possible combinations of phonemes at the beginning of word-initial syllables (i.e., in syllable onsets) and at the end of word-final syllables (i.e., in syllable codas). Word-medial clusters are then syllabified in terms of the model provided by the word-initial and word-final clusters (Jones 1976, 121, see also section 2.3.1). If the given combination of consonants occurs wordinitially, it will also occur word-medially, and if the given consonant clusters appear word-finally, they can also appear word-medially. 28 For the possibility of the phonetic delimitation of syllables across word boundaries, see, for example, Sabol 1997.

The Synthetic Phonological Theory

129

Thirty-five word-medial CC clusters, /ps, pš, px, pn, pĖ, pl, bć, bz, bl, bĐ, br, tk, tx, tv, tm, tl, tĐ, tr, dv, dm, dn, dĖ, dl, dr, kt, kv, km, kn, kĖ, kr, gn, gĖ, gl, gĐ, gr/, occur word-initially.29 They do not have to be deconstructed and form the onset of the following syllable, as illustrated below: (57)

zápcha /zápxa/ ‘constipation’ – syllable structure zá.pxa vlákno ‘fibre’– syllable structure vlá.kno

The rest of the word-medial two-consonant clusters (Appendix 5) should be syllabified as C1.C2. For example, (58)

benzín ‘petrol’– syllable structure ben.zín vojvoda ‘duke’ – syllable structure voj.vo.da

Although the clusters /ps, kt, ks, mp, mf, nt, nd, nk, ng, nc, nþ, nš, st, zd, sĢ, zć, sk, št, žd, šĢ, žć, ft, vk, xt, lp, lt, ld, lk, lc, lf, lm, rp, rz, rf, rv, rm, rn, rĖ, jt, jd, jk, jf, jn/ occur word-finally, they cannot form the coda of the previous syllable because such a syllabification would violate the socalled Right-hand Rule (section 3.2). For example, (59)

hĐadisko ‘viewpoint’ – syllable structure hĐa.di.sko, not hĐa.disk.o

Word-medial three-consonant clusters that occur word-initially (section 2.3.3 and Appendix 5) cannot be in the coda of the previous syllable for the same reason: in the sequence VCCCV, at least one consonant should be the member of the right-hand syllable. These consonant clusters have to be divided, too. The syllabification would be C1C2.C3 if the arising CC cluster occurs word-finally. Ninety word-medial CCC clusters (see Appendix 7) fulfil this condition. For example, (60)

európsky ‘European’ – syllable structure e.u.róps.ky elektrina ‘electricity’ – syllable structure e.lekt.rina

Word-medial three-consonant clusters should be syllabified as C1.C2C3 if the emerging CC cluster can occur word-initially. There are 200 clusters that can be divided in this way (Appendix 7). For instance, (61) 29

hocktorý ‘any‘– syllable structure hoc.kto.rý výzbroj ‘weaponry’ – syllable structure výz.broj

The list of the possible word-initial two-consonant clusters in Slovak is in section 2.3.3.

130

Chapter Six

The syllabification in which the whole CCC cluster creates the onset of the right-hand syllable is allowed when the given three-consonant cluster occurs word-initially, too. There are 19 CCC clusters following this condition (Appendix 7). For example, (62)

registrovaĢ ‘to register’ – syllable structure re.gi.stro.vaĢ puzdro ‘casket’ – syllable structure pu.zdro

To preserve the rule so that the sequence VCV is syllabified as V.CV (see section 5.2), the evaluation of the possible syllabifications of the fourconsonant word-medial clusters should start from right to left. This means that CCCC clusters can be syllabified as C1C2C3.C4 if the arising threeconsonant cluster also occurs word-finally. The cluster /jskr/ can be analysed in this way, as can be seen below: (63)

najskromnejší ‘the most modest’ – syllable structure najsk.rom.nej.ší

The syllabification C1C2.C3C4 is possible when the emerging C1C2 clusters can be found word-finally (those at the end of the first syllable) and the clusters C3C4 word-initially (in the onset of the following syllable). This type of syllabification is suitable for the clusters /rstv, nštr, lstv, pstr, pstv, rnsk, kstr, nkþn, kspl, rstr/. For instance, (64)

kalifornský ‘Californian’ – syllable structure ka.li.forn.ský funkþný ‘functional’ – syllable structure funk.þný

Analogically, the division C1.C2C3C4 is acceptable if the CCC clusters that come into existence also occur word-initially. Only two clusters, /rnsk/ and /nkþn/, do not fulfil this condition, but they can be divided into C1C2.C3C4. The remaining 24 word-medial four-consonant clusters can be syllabified in this way (see Appendix 5 for the whole list of the wordmedial clusters and section 2.3.3 for the list of the word-initial and the word-final clusters). For example, (65)

klamstvo ‘lie’ – syllable structure klam.stvo konštrukcia ‘construction’ – syllable structure kon.štruk.ci.a

Because the maximal number of consonants in Slovak onsets is four, the syllabification V.C1C2C3C4 is possible when the given four-consonant cluster can be found word-initially:

The Synthetic Phonological Theory (66)

131

abstrakt /apstrakt/ ‘abstract’ – syllable structure a.pstrakt

The only five-consonant medial cluster, /rvstv/, can be syllabified as C1C2.C3C4C5 because there is the word-final CC cluster /rv/ and the wordinitial CCC cluster /stv/ (cf. section 2.3.3): (67)

nervstvo ‘nervous system’ – syllable structure nerv.stvo

6.3.5.2 The Sample Analysis of the English Language On the basis of the model of the word-initial and the word-final clusters in English (section 2.3.2), the word-medial clusters (Appendix 5) can be syllabified as follows: Two-consonant clusters that also occur word-initially, /pr, pl, pj, ps, br, bl, bj, tr, tw, tj, dr, dj, kr, kl, kw, kj, km, kn, gr, gl, gw, mj, fl, fr, fj, vj, ۟j, st, sp, sk, sl, sw, sn, sm, sf, sj, zl, ߦr, ߦm, ߦn/, can form the onset of the following syllable, as can be seen in the words below: (68)

western /westΩn/ – syllable structure we.stΩn foolishness /fu:lѢѻnΩs/ – syllable structure fu:.lѢ.ѻnΩs produce /prΩdju:s/ – syllable structure prΩ.dju:s

The rest of the CC clusters (Appendix 5) should be syllabified as C1.C2. The syllabification in which the whole two-consonant word-medial cluster will be in the coda of the preceding syllable is not possible for the reasons already explained in section 6.2.5.1 (the Right-hand Rule): (69)

founder /fa҂ndΩ/ – syllable structure fa҂n.dΩ advantage /Ωdva:ntѢҜ/ – syllable structure Ωd.va:n.tѢҜ

Three-consonant word-medial clusters are syllabified as C1C2.C3 when the arising CC cluster can occur word-finally. Thirty-six medial CCC clusters can be divided in this way (Appendix 7). For example, (70)

introduce /intrΩdju:s/ – syllable structure int.rΩ.dju:s

Thirty-four medial CCC clusters (Appendix 7) can be divided as C1.C2C3 because the arising CC cluster also occurs word-initially: (71)

imprint /ѢmprѢnt/ – syllable structure Ѣm.prѢnt

132

Chapter Six

And two word-medial three-consonant clusters, /spr/ and /str/, allow the syllabification V.C1C2C3 because these clusters occur word-initially. For instance, (72)

ancestral / nsestrΩl/ – syllable structure  n.se.strΩl disprove /dѢspru:v/ – syllable structure dѢ.spru:v

The sources used (see note 11 in Chapter 2) report six of the most common word-medial four-consonant clusters in English (see also Appendix 5)—/nstr, mstr, kstr, ‫و‬kst, nskr, kskl/. The division C1C2C3.C4 is suitable for the clusters /mstr, kstr/; the combinations /mst/ and /kst/ are possible word-finally (section 2.3.2): (73)

extreme /Ѣkstri:m/– syllable structure Ѣkst.ri:m

The syllabification C1C2.C3C4 is allowed when the emerging C1C2 clusters can be found word-finally and C3C4 clusters word-initially. This seems to be applicable to the clusters /nstr, kstr, nskr, kskl, ‫و‬kst/. For example, (74)

exclusive /Ѣksklu:sѢv/ – syllable structure Ѣks.klu:.sѢv inkstand /Ѣƾkst nd/ – syllable structure Ѣƾk.st nd

And the syllable structure C1.C2C3C4 is acceptable when the CCC cluster occurs word-initially. All medial CCCC clusters, except one, /‫و‬kst/, can be divided in this way, as illustrated below: (75)

hamstring /h mstrѢƾ/ – syllable structure h m.strѢƾ construct /kΩnstr҄kt/ – syllable structure kΩn.str҄kt (Appendix 7)

6.3.6 The Frequency Criterion The information about the phonotactic possibilities and constraints of a given language, as illustrated in sections 6.2.5.1 and 6.2.5.2, can sometimes offer several solutions, in which case the one with the higher frequency of occurrence should be preferred. For example, the Slovak word sestra ‘sister’ can be syllabified either as ses.tra or as se.stra. Both solutions are correct because the word (syllable) in Slovak can start with the two-consonant cluster /tr/ but also with the three-consonant cluster /str/ (section 2.3.3). Here, the frequency criterion “advises” syllabifying the word sestra as se.stra because the frequency of the occurrence of the

The Synthetic Phonological Theory

133

cluster /str/ word-initially is higher than the occurrence of the cluster /tr/ (korpus.juls.savba.sk). The other situation in which this criterion can be used for syllabification is when there is a two-consonant cluster in intervocalic, that is, in word-medial, position. The placement of the syllable boundary between these two consonants is theoretically possible. The syllable structure of, for example, a two-syllabic word would be …VC1.C2V…. But if the sequence C1C2 has a high frequency of occurrence in the wordinitial position, the syllable boundary can be before this sequence. For example, the clusters consisting of two obstruents /-sk-, -kt-, -sp-, -st-/ have relatively high frequencies of occurrence in the word-initial position (section 2.3.3); that is why the syllable boundary is placed before the given clusters that form the onset of the second syllable, testovaĢ ‘to test’ (syllable structure te.sto.vaĢ). On the contrary, the CC clusters /-pþ-, -bd-, pk-/ are not very frequent at the beginning of words and thus the syllable boundary is between them: obþan /opþan/ ‘citizen’ (syllable structure /op.þan/). 6.3.6.1 The Sample Analysis of the Slovak Language Considering the syllabification of the Slovak word-medial consonant clusters from the viewpoint of the Slovak phonotactics (section 6.2.5.1, Appendix 7), it is clear that many medial consonant clusters allow (at least) two syllabifications. For example, the words in (60) can also be syllabified as: (76)

európsky ‘European’ – syllable structure e.u.róp.sky elektrina ‘electricity’ – syllable structure e.lek.tri.na

A comparison of the data about the frequency of the occurrence of the consonant clusters resulting from the division in (60) and in (76) (see, e.g., the Slovak National Corpus [online]) shows that the syllabification in (76) should be preferred. All in all, there are 68 word-medial three-consonant clusters that can be syllabified either as C1C2.C3 or as C1.C2C3 (Appendix 7). For example, (77)

Ondrej ‘Anrew’ – syllabification ond.rej or on.drej pozdrav ‘greeting’ – syllable structure pozd.rav or poz.drav

Words in (62) are examples of word-medial three-consonant clusters that also occur word-initially, and, thus, they can form the onset of the

134

Chapter Six

following syllables. However, there are also other possibilities for their syllabification: (78)

registrovaĢ ‘to register’ – syllable structure re.gis.tro.vaĢ puzdro ‘casket’ – syllable structure puz.dro

Following the frequency criterion, the syllabification of the words in (77) should be on.drej and poz.drav, and the words registrovaĢ ‘to register’ and puzdro ‘casket’ can be divided as illustrated in (78) and not (62). The division of three- and four-consonant clusters also has several possibilities. The selection of the right one in accordance with the phonotactic possibilities of the Slovak phonemes depends on the frequency of the occurrence of the clusters arising during syllabification. 6.3.6.2 The Sample Analysis of the English Language The analysis of the English word-medial two-, three-, and four-consonant clusters (section 2.3.2, Appendix 7) showed that the phonotactic possibilities of this language sometimes lead to several options for placing the syllable boundary. For example, the syllabification of the word introduce in (70) is only theoretically possible. The word-final sequence /nt/ is acceptable but it is less frequent than the word-initial cluster /tr/. The syllabification then should be: (79)

introduce /ѢntrΩdju:s/ – syllable structure Ѣn.trΩ.dju:s

When comparing possible syllabifications of the word-medial threeconsonant clusters, it is clear that 26 clusters allow both syllabifications, C1C2.C3 and C1.C2C3 (Appendix 7). For example, (80)

angry / ƾgrѢ/– syllable structure  ƾ.grѢ or  ƾg.rѢ expect /Ѣkspekt/ – syllable structure Ѣk.spekt or Ѣks.pekt

The information about the frequency of the occurrence of the emerging clusters (see, e.g., British National Corpus [online]) supports  ƾ.grѢ and Ѣk.spekt as the correct divisions.30

30

“Correct” from the viewpoint of the given criterion.

The Synthetic Phonological Theory

135

6.3.7 The Power of Syllable Welds The strongest syllable connection is at the place of the highest number of contrasts, and the syllable boundary is usually at the place of the lowest number of contrasts (Pauliny 1979, 169–70). The contrasts between two neighbouring segments are counted on the basis of the similarities and differences between their distinctive features. This means that the number of contrasts depends on the articulatory and acoustic properties of the segments. Distinctive features, although they are theoretical constructs, result from the articulatory and acoustic properties of sounds (Giegerich 1992, 89). Consequently, the highest number of contrasts is in the sequence consonant + vowel; that is why C+V represents the strongest syllable connection.31 On the other hand, the more similar the sounds are phonetically, the lower the number of contrasts that can be found between them. For example, in the word isto ‘for sure’, there is a two-consonant cluster, ‘st’, in an intervocalic position. The contrast between these two segments is very low. The only difference is that the consonant /s/ is constrictive and the consonant /t/ is a stop. The given word should then be syllabified as is.to (for details, see Pauliny 1979, 171–82). 6.3.7.1 The Sample Analysis of the Slovak Language As explained above, the degree of the contrast between the adjacent elements in a consonant cluster depends on the distinctive features of the individual phonemes in the given cluster. The survey of the distinctive features of Slovak consonant phonemes can be found in table 2-5. The contrast is expressed by the number. This number indicates the number of different distinctive features of the compared phonemes. For example, /t/ is [+stop], [-sonorant], [-continuant], [-voice], [-nasal] and Coronal [+anterior]; /n/ is [+stop], [+sonorant], [-continuant], [+voice], [+nasal] and Coronal [+anterior]. Three distinctive featuresʊ[sonorant], [voice], [nasal]ʊhave different +/- specifications. That is why the number of the differences, that is, the degree of the contrast between the phoneme /t/ and the phoneme /n/, is three. If one phoneme has the place feature, for example, Coronal [+anterior], and the other phoneme is specified as Dorsal [+back], [+high], the difference is counted as one, representing the 31 The reverse order, vowel + consonant, displays the same type and number of contrasts; however, in Pauliny’s view, the syllable is determined from the end of the word. For example, in the word rada ‘advice’ the last syllable -da is determined first and then the first one (for details, see Pauliny 1979, 179–80; see also section 6.2.5.1).

136

Chapter Six

different place of articulation. If one phoneme has stricture feature specifications, for example, [+stop], [-sonorant], [-continuant], and the other phoneme’s stricture features are [+sonorant], [-continuant], the difference is counted as two. This means that the missing feature is also included in the evaluation of the similarities and differences. The entire analysis of the Slovak word-medial consonant clusters from the point of view of the contrast between neighbouring phonemes in the cluster is in Appendix 8. The syllabification of the Slovak word-medial clusters on the basis of the similarities and differences between phonemes in a given cluster brings the following results: 13 word-medial two-consonant clusters have identical distinctive features, that is, there is no contrast between them and they should be syllabified as C1.C2 (Appendix 8). For example, (81)

preddavok ‘cash advance’ – syllable structure pred.davok vyšší ‘higher’ – syllable structure vyš.ší

One hundred and seven medial CC clusters have one or two distinctive features. This means that there is some contrast between these segments but it is still very low and the syllable boundary can be placed between themʊC1.C2. For instance, the clusters /cþ/ and /mn/ have one degree of contrast: (82)

viacþlenný ‘consisting of more members’ – syllable structure viac.þlen.ný zimný ‘winter’– syllable structure zim.ný

And 226 word-medial CC clusters have more than two distinctive features, and they can be together in the onset of the following syllable (Appendix 8), as illustrated below by the clusters /tm/ and /xþ/: (83)

rytmus ‘rhythm’ – syllable structure ryt.mus dychþaĢ /dixþaĢ/ ‘wheeze’ – syllable structure dix.þaĢ

The degree of the contrasts between the neighbouring segments in the word-medial three-consonant clusters calculated on the basis of the distinctive features (Table 2-5) of the individual consonants in the clusters and the suggested syllabification of these clusters is in Appendix 8. If a syllable boundary is at the place of the lowest number of contrasts (Pauliny 1979, 169–70), then, for example, the cluster /jst/ should be syllabified as js.t because the number of the differences (i.e., the number of the different distinctive features) between /j/ and /s/ is four. The contrast is thus higher than between /s/ and /t/, where it is two. This means that the

The Synthetic Phonological Theory

137

syllable connection between the phonemes /j/ and /s/ is stronger than the one between the phonemes /s/ and /t/: (84)

majster ‘master’ – syllable structure majs.ter

However, the calculation of the degree of the contrast between consonants in a consonant cluster shows that in 38 cases (Appendix 8), it is not possible to find the place of the lowest contrast because the number of different distinctive features between consonants in a three-consonant cluster is identical. For example, there are four different features between /m/ and /k/ and four differences between /k/ and /n/ in the cluster /mkn/. There are two possible syllabifications: m.kn or mk.n. But Pauliny (1979, 177, see also note 117) says that the syllable is determined from the end of the word; that is why the division mk.n should be preferred. The first syllable weld is between /k/ and /n/: (85)

zamknúĢ ‘to lock’ – syllable structure zamk.núĢ

Fifteen word-medial four-consonant clusters, /nstv, rstv, Đstv, mstv, nštr, lstv, jstr, jstv, pstr, jstr, pstv, jskr, kstr, nkþn, rstr/, should be divided into C1C2.C3C4 because the lowest degree of the contrast is between the second and the third consonant in the sequence (Appendix 8). For example, (86)

tajomstvo ‘secret’ – syllable structure ta.joms.tvo chlapstvo /xlapstvo/ ‘manhood’ – syllable structure xlaps.tvo

Four clusters, /šstv, dzbr, tskr, fskĐ/, have the lowest number of differences between the first and the second consonant (Appendix 8) and their division should be C1.C2C3C4. For instance: (87)

odzbrojiĢ ‘disarm’ – syllable structure od.zbro.jiĢ velikášstvo ‘megalomania’ – syllable structure ve.li.káš.stvo

The cluster /rnsk/ requires the syllabification C1C2C3.C4 because the lowest number of different features is between /s/ and /k/: (88)

kalifornský ‘Californian’ – syllable structure ka.li.forns.ký

Six clusters, /cstv, vstv, tstr, þstr, jzdr, kspl/, have two possible places for the syllable boundary (Appendix 8)—C1.C2C3C4 or C1C2.C3C4.

138

Chapter Six

Recalling that syllables are identified from the end of the word, the correct division is C1C2.C3C4. For example, (89)

najzdravší ‘the healthiest’ – the syllable structure najz.drav.ší

The word-medial five-consonant cluster /rvstv/ has three places with the identical number of the different distinctive features between the neighbouring consonants. This means that there are three possible syllabifications— C1.C2C3C4C5, C1C2.C3C4C5, or C1C2C3.C4C5 (Appendix 8). The delimitation of the syllables from the end of the word results in the division C1C2C3.C4C5: (90)

nervstvo ‘nervous system’ – syllable structure nervs.tvo

6.3.7.2 The Sample Analysis of the English Language The distinctive features of the English consonant phonemes are in table 22. The whole analysis of the English word-medial consonant clusters in terms of the degree of the contrast specified above is in Appendix 8. Twenty word-medial two-consonant clusters allow the syllabification C1.C2 because the number of the different features between them is not higher than two. For instance, (91)

number /n҄mbΩ/ – syllable structure n҄m.bΩ optimum /ыptѢmΩm/ – syllable structure ыp.tѢ.mΩm

One hundred and eight medial CC clusters can be in the onset of the following syllable because the degree of the contrast between them is higher than three. The more differences between the adjacent elements, the stronger the syllable connection between them, as illustrated below: (92)

suspect /sΩspekt/ – syllable structure sΩ.spekt address /Ωdres/ – syllable structure Ω.dres

The syllabification of 41 word-medial three-consonant clusters is clear. The syllable boundary is at the place of the lowest number of the differences between the neighbouring consonants, that is, at the place of the lowest degree of the syllable contrast (see Appendix 8). For example, in the cluster /ƾgr/, there are two differences between /ƾ/ and /g/ and four differences between /g/ and /r/; that is why the syllabification should be ƾ.gr:

The Synthetic Phonological Theory

(93)

139

angry / ƾgrѢ/ – syllable structure  ƾ.grѢ

The number of the different distinctive features between the adjacent elements in the clusters /mfl, mps, nsl, ksp/ is identical and that is why their syllabification can be either C1.C2C3 or C1C2.C3. However, having adopted Pauliny’s (1979) view that syllables in a word should be identified from the end of the given word (see above), the syllabification C1C2.C3 is preferred. For instance, in the cluster /ksp/, there are three differences between /k/ and /s/ and three between /s/ and /p/, but, counting from the end, the first place of the lowest contrast (lower than the one between /k/ or /p/ and the adjacent vowel) is between /s/ and /p/. For example, (94)

expert /ekspє:t/ – syllable structure eks.pє:t

As for the medial four-consonant clusters, considering the position of the lowest degree of the syllable contrast, they can all be syllabified as C1C2.C3C4. One of these clusters, the cluster /kskl/, seems to have two places with the identical degree of contrast but, as already explained, the syllabification of words starts at the end. That is why this cluster can be divided into C1C2.C3C4 as well: (95)

exclude /Ѣksklu:d/ – syllable structure Ѣks.klu:d

6.4 The Complex (Synthetic) Approach to the Syllable Structure Considering the examples in section 6.2 and Appendices 6–8, it is obvious that one and the same consonant cluster, often in one and the same word, can be syllabified differently, depending on individual criteria. For example, the Slovak word-medial two-consonant cluster /mn/ can either be in the onset of the following syllable (the 2nd criterion), or syllabified as m.n according to the 3rd or 5th criterion. The three-consonant cluster /str/ is syllabified as s.tr when the 3rd or 4th criterion is applied or the cluster as a whole forms the onset of the following syllable when following the 5th criterion. The four-consonant cluster /rnsk/ is divided into rn.sk according to the 2nd criterion or the syllabification is rns.k when applying the 4th or 7th criterion. The English consonant clusters have several possible syllabifications, too. For example, the two-consonant cluster /sp/ is divided as s.p according to the 7th criterion, or the whole cluster is in the onset of the adjacent syllable when the 4th criterion is followed. The English word-

140

Chapter Six

medial three-consonant cluster /nst/ is syllabified as n.st when applying the 1st criterion or the syllabification is ns.t according to the 4th criterion. The cluster /nskr/ is syllabified either as n.skr or as ns.kr when the 2nd and the 3rd criteria are used, respectively. The answer to the simple question, Which syllabification is the correct one? is quite elaborate. The synthetic phonological theory sees the syllable as a complex sound unit that interconnects the segmental and the suprasegmental subsystems of a language, and this reflects the natural human ability to produce sounds and combine them into higher-level units. In addition to this universal aspect of the syllable and its structure, there is also the typological dimension given by the configuration of sound elements in the structure of the syllable as part of the form of a linguistic sign (Sabol 1997, 27). The syllable as a sound unit is incorporated into the structure of a word (and a word form) as a unit of meaning, and “this bilateral unit provides the framework for the delimitation of the syllable” (ibid.). As already mentioned (6.3.1), in the relation between sound and meaning, the meaning is decisive. This claim is proven by the possibility to delimit three combinations of sound and meaning identity and diversity within language units: (1) suppletion – in the individual forms of one and the same semantic element, there are different form-creating bases, (2) synonymy – different forms with (almost) identical meaning, and (3) homonymy – identical forms with different meanings (Sabol 1989, 52– 53). And as the author further adds, “[…] only through meaning it is possible to think about the combinations in the connection between sound and semantic elements, only meaning makes it possible to delimit these combinations. The sound thus works in the name of the meaning, the form in the service of the content, not vice versa” (Sabol 1989, 59). More evidence of the priority of meaning over sound is the possibility of “forcing” the syllable function upon certain combinations of phonemes (Kury‫ן‬owicz 1948; Sabol 1994). For example, the Slovak sequence of consonant phonemes /pst/ is perceived as the syllable, or more precisely, as the syllaboid or quasi-syllable when standing alone as the interjection pst ‘hush’. This group of consonants is not a syllabic part of Slovak words or word forms, as, for example, in the word pstruh ‘trout’, but it is perceived as the syllable in the word pstkaĢ ‘to say sh’, which was derived from the interjection pst (Sabol 1994). The English sh /ߦ/ with the meaning ‘hush, silence’ can be seen as the quasi-syllable, too. Therefore, if there is no important difference between the morpheme structure and the syllable structure, the syllable boundary should be at the place of the morpheme boundary, as can be seen in examples (1)–(20)

The Synthetic Phonological Theory

141

above. The delimitation of the syllable as a sound unit in such a way proves the dominance of the meaning over the sound as well as the symmetry between the sound (form) and the meaning. The 5th and 6th criteria, that is, the phonotactics of the given language and Kuryáowicz’s (1948) rule supplemented by the frequency criterion, are directly connected with the meaning as well through the word as a unit of meaning. Thus, if the 1st criterion cannot be applied, the syllabification of word-medial clusters should follow the phonotactic possibilities at the beginning and at the end of words (syllables) in the given language. Because phonotactics offers several syllabification options, those with the higher frequency of occurrence should be used, as illustrated by examples (57)–(80). The phonological criterion that comes from the degree of contrast between the adjacent elements in the syllable (the 7th criterion) can be used when its application does not violate the phonotactic possibilities and constraints of the given languageʊsuch as the higher-level criteria (see below)ʊbecause the delimitation of the syllable on the basis of the distinctive features of phonic elements emphasises the syllable as the unit crucial for the analysis of speech sounds in various phonological oppositions and contrasts [see examples (81)–(95)].32 The 2nd and the 3rd criteria, which are concerned about the acoustic and the articulatory symbiosis between the elements in the syllable, can be characterised as phonetic ones overlapping into phonology because consonant features that need the acoustic spectrum of the neighbouring sound depend on the acoustic or articulatory (i.e., phonetic) characteristics of these sounds. The syllable boundary can be delimited on the basis of these two criteria [see examples (21)–(39)] when the syllabification does not violate phonology (i.e., the previously mentioned criteria). Finally, purely phonetic criterion that delimits the syllable as a unit reflecting the physiology of sound production (the 4th criterion) can be considered the most universal in the sense that the physiology of speech organs is identical in all languages [see examples (40)–(56)]. All the criteria can be ranked into the phonetic-phonology hierarchy (Fig. 6-9) that illustrates the language-specific or language-universal character of the given criterion at the same time, depending on the degree of the individual and the universal; that is, the concrete and the abstract, as specified by the synthetic phonological theory.

32

Of course, the degree of the contrast between the neighbouring elements in the syllable depends on the phonological theory used for the specification of distinctive features (Sabol 1994, 221).

142

Chapter Six

Figure 6-9 The Hierarchy of the Criteria Used for the Syllabification in the SPT

6.4.1 The Sample Complex Analysis The complex synthetic approach to the syllable structure specified above is now illustrated by the analysis of one Slovak and one English word with a consonant cluster in the word-medial position. 6.4.1.1 The Analysis of the Slovak Word The maximum number of consonants in the word-medial position in Slovak is five. There is only one five-consonant medial cluster /rvstv/; however, this cluster occurs in the word nervstvo ‘nervous system’, which belongs to the central vocabulary. The morpheme structure of this word is nerv-stv-o, where nerv- is the root morpheme, -stv- is the derivational morpheme, and -o is the grammatical morpheme. Looking at the morpheme structure of this word, it is obvious that the syllabification nerv.stvo is possible. Phonotactics of the Slovak language together with Kuryáowicz’s (1948) rule also syllabify this word as nerv.stvo. Moreover, the word-initial (syllable-initial) three-consonant cluster /stv/ is frequent in Slovak. According to the degree of contrasts between the neighbouring

The Synthetic Phonological Theory

143

consonants in the cluster as well as the degree of the stricture, the syllabification of the word nervstvo ‘nervous system’ should be nervs.tvo. However, the criteria on the borderline between phonetics and phonology, the acoustic symbiosis and coarticulation, support the division nerv.stvo. All in all, taking into account the information about the dominance of meaning over sound as well as the facts about the mutual relation between phonetics and phonology, the Slovak word nervstvo ‘nervous system’, containing a five-consonant medial cluster, is syllabified as nerv.stvo. It seems that this division violates the phonetic nature of the syllable as a unit reflecting the basic principles of the physiology of speech but the degree of the stricture of /s/ is only slightly lower than the degree of the stricture of the consonant /t/ (Table 6-3). One may conclude that not only phonemic (phonological) but also phonetic properties of the syllable are preserved by the syllabification nerv.stvo, and, despite the exceptionally high number of elements in the cluster under investigation, all consonants are syllabified and none evaluated as extrasyllabic. 6.4.1.2 The Analysis of the English Word The highest number of consonants in the word-medial position of the English words is four, as can be seen, for instance, in the word extra /ekstrԥ/. Theoretically, there are five possible ways of syllabification (Roach 2000, 77): (1) (2) (3) (4) (5)

e.kstrԥ ek.strԥ eks.trԥ ekst.rԥ ekstr.ԥ

The fifth possibility can be excluded due to the Right-hand Rule (section 6.3.5.1) or the so-called CV-Rule in Lexical Phonology (section 3.2). The first possibility violates the phonotactics of the English language, also known as the Maximal Onset Principle, saying that the maximal number of consonants in English onsets is three (e.g., McMahon 2002, 110). The syllabification in (4) is against the phonetic understanding of the syllable as a unit formed in accordance with the principles of phonation (section 6.3.4). Thus, when there is the option, the beginning of the syllable should be created by a strictured sound as much as possible. /r/ as a sonorant is not very suitable for a syllable onset (Table 6-2). There are two possibilities left: (2) and (3), ek.strΩ and eks.trΩ. From the viewpoint of the degree of contrasts, there are three differences between /k/ and /s/

144

Chapter Six

and two different features between /s/ and /t/ (Appendix 8, part 5). Thus, the syllable boundary should be between /s/ and /t/; that is, eks.trΩ. This syllabification is also supported by the morpheme structure because in English there is the prefix ex- /eks/. Or, to be consistent with the sample analysis of the Slovak word, the criterion based on the symbiosis between the morpheme and the syllable leads to the syllabification eks.trΩ. Kuryáowicz’s (1948) rule and the frequency of phonotactically allowed consonant sequences divide the given word as eks.trΩ. As to the degree of contrast, the syllabification should be eks.trΩ. The acoustic symbiosis and the coarticulation between the adjacent segments lead to the division eks.trΩ, and the degree of stricture offers two possible syllabifications: ek.strΩ or eks.trΩ. To sum up, considering all the information about the possible syllable boundary in the word extra /ekstrΩ/, it seems that the syllabification eks.trΩ is the most appropriate.

6.4.2 The Syllable and the Different Levels of Abstraction As can be seen in the previous analyses, synthetic phonological theory offers the manifold phonetic, phonemic, and morphophonemic approach to the syllabification of words. Moreover, this theory enables the analysis of the sound structure of the syllable itself through the levels of the phone, the phoneme, and the morphophoneme, as defined in section 6.1.33 According to the SPT, a detailed analysis of these levels identifies the vertical structure of the sound unit. This means the view on the sound segment from the differentiated level of the relation between the individual and the universal (Figure 6-1) and the horizontal dimension of the relation Ph–Phm–MPhm is within the given morpheme. These are the results of neutralization (see note 1) and alternation (see note 2) processes in the morpheme. The levels of Ph, Phm, and MPhm are fully reflected in the phonic structure of morphemes and create their corresponding phonic layers: the level of the Ph-morpheme shows the sound structure of the morpheme at the level of the phone, the level of the Phm-morpheme displays the sound structure of the morpheme at the level of the phoneme, and the level of the MPhm-morpheme shows the sound structure of the morpheme at the level of the morphophoneme (for details, see Sabol 1989, 132–33). The same procedure can be applied to the analysis of the phonic structure of the syllable. 33

The possible connection between these three levels of abstraction and the underlying and surface forms as specified in generative phonology (section 1.3) can be found in section 7.2.

The Synthetic Phonological Theory

145

The complex analysis of the structure of the syllable through the level of Ph, Phm, and MPhm results in a Ph-syllable (the syllable structure at the level of the phone), a Phm-syllable (the syllable structure at the level of the phoneme), and a MPhm-syllable (the syllable structure at the level of the morphophoneme). The level of the phone captures the phonetic structure of lexical units influenced by the orthophony of the given language, neutralizations (see note 1), and alternations (see note 2), and is noted down in phonetic symbols notated in square brackets. The level of the phoneme reflects the phonemic structure of lexical units influenced by the neutralization and alternation rules and is transcribed in phonemic symbols notated in slanted brackets. The level of the morphophoneme means the morphophonemic structure of the given lexical unit, which is affected only by alternations and is written down in capital letters (see also Table 6-1). The morphophoneme is delimited in the morpheme as the smallest bearer of elementary or complex meaning on the basis of the position of the maximum phonological differentiation (ibid., 105). Because the neutralization and alternation processes as well as the reconstruction of the morphophoneme are beyond the scope of the interest of this work, the three-level analysis of the morpheme structure and the syllable structure in both languages is illustrated here only by one simple Slovak example to show the possibilities of this type of analysis. The Slovak word navždy ‘forever’ has the morpheme structure navždy (one morpheme) and is syllabified as na.vždy. The complex sound structure of the morpheme is: Ph-morpheme Phm-morpheme MPhm-morpheme

[nawždi]34 /navždi/ NAVŽDI

The phonic structure of the syllables delimited in the given monomorphemic word at different levels of abstraction is as follows: Ph-syllable Phm-syllable MPhm-syllable

[na.wždi] /na.vždi/ NA.VŽDI35

34 [w] is a bilabial fricative voiced oral sound; it is a combinatory variant of the phoneme /v/ occurring before the so-called pair consonants (Sabol 1989, 77). 35 This is only a hypothetical sound structure of the syllables at the level of the morphophoneme because the neutralizations of distinctive features have to be taken into account (for details, see Sabol 1989, 105–33).

CHAPTER SEVEN THE GENERATIVE AND THE STRUCTURALIST APPROACH TO THE SYLLABLE STRUCTURE

Synthetic phonological theory is a complex approach to the sound level of language(s) that encompasses all the fundamental principles and ideas of European structuralism (see note 1 in Introduction). Syllabification is based on several principles (for details, see section 6.4) that have their roots in the interconnection between phonetics and phonology. Assuming the dominance of meaning over sound in language (section 6.3.1), the syllable boundary is at the place of the morpheme boundary if there is no discrepancy between the structure of the syllable as a sound unit and the structure of the morpheme as a unit of meaning. This applies to the prefix-stem juncture and the compound juncture. A similar idea of the influence of the morpheme structure of words on their syllabification can be found in the SSA delimited by Lexical Phonology (see section 3.2.2 for the so-called Prosodification Constraint) as well as in CVX syllable theory. In CVX theory, morphologically complex words, that is, words with a morpheme boundary, are usually excluded from analysis (see Duanmu 2009, 212) because the theory concentrates on the syllable structure of monomorphemic words only. This theory also accepts the idea of morpheme-structure versus syllable-structure relations in a different manner. It sees the correspondence between the morphology of the given language and the maximal size of its syllables (section 2.1) in such a way that the morphological structure of affixes accounts for extra consonants at word edges (sections 2.1.1–2.1.3). Phonotactics (possible phoneme sequences) play a crucial role in the three approaches examined in this work, too. Though the criteria used for the evaluation of the well-formed and ill-formed combinations of consonants are different in different approaches, CVX syllable theory accepts only those clusters of segments in onsets and codas that can be treated as complex sounds (Duanmu 2010, 13; see also section 2.2). In Lexical Phonology, onset and coda sequences are governed by the principle of sonority (sections 3.2, 3.2.5). In the SPT, more phenomena are

The Generative and the Structuralist approach to the Syllable Structure 147

taken into account when talking about the syllable as the basic unit of a phonotactic analysis. It is assumed that the composition of consonant clusters reflects articulatory and the acoustic principles (a minimum articulatory effort on the side of the speaker and the easiest perception on the side of the hearer) as well as the systemic, functional classification of sounds (Sabol 1975, 204). For the SPT, the syllable is a phonological constituent necessary for the delimitation of phonemes and their features (section 6.3.5). This universal syllable function is accompanied by the language-specific nuances. In detail, the basic syllable contrast is CV; this is also the basic syllable model. The further structure of the syllable in the individual languages then differs depending on the distinctive phonemic features, the features comprising the common base, the morphological structure of the given language, etc. (Pauliny 1979, 104). The CVX syllable theory rejects the CV-only approach (Duanmu 2010, 6) but the role of distinctive features is unchallenged because only those sequences of segments are part of the CVX template whose features are made simultaneously. There is no plus-minus specification of the same distinctive feature and the whole sequence can be treated as a complex segment (section 2.2.2). The theory of the syllable in LP as introduced by Mohanan (1982) follows Halle and Vergnaud’s (1980) three-dimensional representation of the syllable (section 3.1), whose one tier, the so-called melodic tier, shows distinctive feature specifications of the individual segments (cf. Mohanan 1982, 158). In the SPT, the phonetic aspects of the syllable as a sound unit are represented by the principles of the articulatory and the acoustic symbiosis between the elements in the syllable onsets and codas, and by admitting that CV, as the fundamental syllable model, reflects the mechanism of the production of human speech (sections 6.3.2–6.3.4). All these phonetic facets seem to be ignored by the CVX theory and the SSA in LP. In modern generative phonology, not all phonological concepts have phonetic correlates. As for the syllable, it is seen as a phonological construct, defined over sequences of discrete phonological segments rather than over phonetic primes as such. At this level of abstraction (which includes most of phonology), few constructs have direct phonetic definitions. (Clements 2006, 1)

Nevertheless, although the syllable is referred to as the phonological syllable, “the discussion is not entirely abstract, […] the phonological syllable is related in a very natural way to the phonetic syllable” (Hooper 1972, 534). The idea of the existence of complex segments, which is the

Chapter Seven

148

core of CVX theory, can be classified as a kind of coarticulation and the subsequent acoustic interconnection between neighbouring segments (sections 2.4–2.5). Lexical Phonology sees CV as the basic syllable model and it emphasizes sonority, which is a phenomenon whose phonetic content is undisputable (section 3.2).1

7.1 The Syllabification of the Sample Words The differences between the theories of generative phonology and structuralism as displayed through the analyses produced in this work (cf. Chapter 1 and Chapter 5) are not so distant from each other and are now illustrated by the syllabification of a sample of English and Slovak words with one consonant occurring word-medially and with two-, three-, and four-consonant clusters in the word-medial position. In Slovak, fiveconsonant cluster is also included in the analysis. These words, which have already been used as examples when analysing the individual theories and their principles, are re-syllabified according to the CVX syllable theory, the SSA as specified in Lexical Phonology, and according to the criteria from SPT. The results of these three types of syllabification are presented in table 7-1 (English words) and table 7-2 (Slovak words). The syllable boundaries and the unsyllabified consonants (if accepted by the theory) are indicated in accordance with the given approach (see Chapter 2 for the CVX syllable theory, Chapter 3 for SSA, and Chapter 6 for syllabification in SPT). Table 7-1 Syllabification of the English Sample Words syllabification following

words

CVX theory

SSA in LP

structuralist theory SPT

sܼt.i ræb.aܼ ԥ.tæk

sܼ.ti ræ.baܼ ԥ.tæk

sܼ.ti ræ.baܼ ԥ.tæk

wܼs.pԥ k‫ܥ‬k.teܼ[l] æd.v‫ܮ‬:[b]

wܼs.pԥ k‫ܥ‬k.teܼl æ.dv‫ܮ‬:b

wܼs.pԥ k‫ܥ‬k.teܼl æd.v‫ܮ‬:b

generative theories with C word-medially city /‫ޖ‬cܼti/ rabbi /‫ޖ‬ræbaܼ/ attack /ԥ‫ޖ‬tæk/ with CC word-medially whisper /‫ޖ‬wܼspԥ/ cocktail /‫ޖ‬k‫ܥ‬kteܼl/ adverb /‫ޖ‬ædv‫ܮ‬:b/ 1

For the discussion about the phonetic basis of sonority, see Clements 2006.

The Generative and the Structuralist approach to the Syllable Structure 149 crafty /‫ޖ‬kr‫ܤ‬:fti/ fitment /‫ޖ‬fܼtmԥnt/ lopsided /l‫ܥ‬p‫ޖ‬saܼdܼd/ episcopacy /ܼ‫ޖ‬pܼskԥpԥsi/ with CCC word-medially empty /‫ޖ‬empti/ anthem /‫ޖ‬æntșԥm/ hundred /‫ޖ‬h‫ݞ‬ndrԥd / upstairs /‫ݞ‬p‫ޖ‬steԥz/ twinkling /‫ޖ‬twܼƾklܼƾ/ pumpkin /‫ޖ‬p‫ݞ‬mpkܼn/ softly /‫ޖ‬s‫ܥ‬ftli/ monthly /‫ޖ‬m‫ݞ‬nșli/ introduction /ܼntrԥ‫ޖ‬dju:s/ friendship/‫ޖ‬frend‫ܼݕ‬p/ with CCCC wordmedially instruct /ܼn‫ޖ‬str‫ݞ‬kt/ extra /‫ޖ‬ekstrԥ/

[k]r‫ܤ‬:.f.ti fܼt.mԥ[n[t] l‫[ܥ‬p].saܼ.dܼ[d] ܼ.pܼs.kԥ.pԥ.si

*kr‫ܤ‬:f.ti fܼ.tmԥnt l‫ܥ‬p.saܼ.dܼd ܼ.pܼs.kԥ.pԥ.si

kr‫ܤ‬:f.ti fܼt.mԥnt l‫ܥ‬p.saܼ.dܼd ܼ.pܼs.kԥ.pԥ.si

em.[p]ti æn.[t]șԥ[m] h‫ݞ‬n.[d]rԥ[d] ‫[ݞ‬p].[s]teԥz [t]wܼƾ.[k]lܼ[ƾ]/ [t]wܼƾ.kl.ܼƾ p‫ݞ‬m.[p]kܼ[n] s‫ܥ‬f.[t]li m‫ݞ‬n.[ș]li ܼ.[n][t]rԥ.[d]ju:[s] [f]ren.[d]‫[ܼݕ‬p]

emp.ti æn.tșԥm h‫ݞ‬n.drԥd ‫ݞ‬p.steԥz twܼƾk.lܼƾ

emp.ti ænt. șԥm h‫ݞ‬n.drԥd ‫ݞ‬p.steԥz twܼƾk.lܼƾ

p‫ݞ‬mp.kܼn s‫ܥ‬f.tli m‫ݞ‬n.șli ܼn.trԥ.dju:s fren.d‫ܼݕ‬p

p‫ݞ‬mp.kܼn s‫ܥ‬f.tli m‫ݞ‬nș.li ܼn.trԥ.dju:s frend.‫ܼݕ‬p

ܼ[n].[s[t]r‫ݞ‬k[t] ek.[s[t]rԥ

ܼns.tr‫ݞ‬kt eks.*t.rԥ

ܼns.tr‫ݞ‬kt eks.trԥ

Table 7-2 Syllabification of the Slovak Sample Words syllabification following

words

CVX theory

SSA in LP

structuralist theory SPT

ruk.a/ru.ka pia.tok/piat.ok

ru.ka pia.tok

ru.ka pia.tok

[k]vap.ka dob.re uj.ma lop.ta [z]rej.me lek.va[a[r]

kvap.ka dob.re uj.ma lop.ta zrej.me lek.vár

kvap.ka do.bre uj.ma lop.ta zrej.me le.kvár

zaj[t].ra lii[s[t].ko[m]

zaj.tra lís.tkom

zaj.tra lís.tkom

generative theories with C word-medially ‫ޖ‬ruka ‘hand’ ‫ޖ‬piatok ‘Friday’ with CC wordmedially ‫ޖ‬kvapka ‘drop’ ‫ޖ‬dobre ‘well’ ‫ޖ‬ujma ‘harm’ ‫ޖ‬lopta ‘ball’ ‫ޖ‬zrejme ‘apparently’ ‫ޖ‬lekvár ‘jam’ with CCC wordmedially ‫ޖ‬zajtra ‘tomorrow’ ‫ޖ‬lístkom ‘with the ticket’

Chapter Seven

150 ‫ޖ‬partner ‘partner’ ‫ޖ‬mastný ‘greasy’ ‫ޖ‬vlastný ‘own’ ‫ޖ‬dáždnik ‘umbrella’ ‫ޖ‬ostroha ‘spur’ ‫ޖ‬zajtra ‘tomorrow’ ‫ޖ‬obrovský ‘huge’ with CCCC wordmedially ‫ޖ‬majstrovstvo ‘artistry’ ‫ޖ‬funkþný ‘functional‘ ‫ޖ‬klamstvo ‘falsehood’ ‫ޖ‬šialenstvo ‘insanity’ ‫ޖ‬monštrum ‘monster‘ ‫ޖ‬biskupstvo ‘episcopacy’ with CCCCC wordmedially ‫ޖ‬nervstvo ‘nervous system’

par[t].ne[r] mas.[t]ni[i] [v]las.[t]nii daa.[ž[d]ni[k] os.[t]ro.ha zaj.[t]ra ob.ro.[v[s]ki[i]

par.tner mas.tný vlas.tný dáž.dnik os.tro.ha zaj.tra ob.rov.ský

par.tner mast.ný vla.stný dážd.nik o.stro.ha zaj.tra o.brov.ský

maj[st].ro[vst].vo fun[kþ].ni[i] [k]lam.[s[t]vo šia.le.[n[s][t]vo mon.[š[t]ru[m] bis.ku.[p[s[t]vo

majs.trovs.tvo funk.þný klams.tvo šia.len.stvo monš.trum bis.kups.*t.vo

maj.strov.stvo funk.þný klam.stvo šia.len.stvo monš.trum bis.kup.stvo

ner.[v[s[t]vo

ner.*v.stvo

nerv.stvo

The tables show that the syllabification of the selected English (Table 7-1) and Slovak words (Table 7-2) is almost identical in all three theories analysed in this work. The only difference is that the CVX theory not only accepts the syllabification VC.V but also the V.CV typical of SSA in Lexical Phonology and SPT. Furthermore, the CVX syllable theory leaves more consonants unsyllabified. Surprisingly, the disagreement over the placement of the syllable boundary can be observed not between the CVX theory and SSA on one side and SPT on the otherʊas expected initiallyʊbut between CVX on one side and SSA and SPT on the other, especially due to the limited number of segments in the syllable within the CVX syllable model. Thus, is the generative approach to the syllable really so different from the structuralist approach to this sound unit, or are these two approaches basically the same?

7.2 Different and the Same The comparison of the generative and the structuralist syllabifications has revealed that there are only minimal differences between these two approaches to the syllable and its structure. The generative syllabification

The Generative and the Structuralist approach to the Syllable Structure 151

also2 admits the parsing VC.V and is not exhaustive; on the other hand, the structuralist syllabification is exhaustive and accepts CV as the basic syllable model. To complete the picture of the differences and similarities between the conception of the syllable in generative and structuralist phonology, the syllable at the different levels of abstraction is now juxtaposed. Drawing parallels between the data in section 6.4.2 and the information about the underlying and the surface form in section 1.3, it is apparent that the level of the morphophoneme is equivalent to the underlying form, and the level of the phone refers to the surface form. The correspondence between these two levels is assigned by the set of phonological rules, which yield an intermediate level of representation (Roca 1994, 44). This intermediate representation level, that is, the level of the phoneme, is either formally absent (because it is incorporated in the surface form) or it is labelled as a pre-surface form in traditional generative phonology (see, e.g., Blevins 1995, 218), as illustrated by figure 7-1. As already indicated in section 1.3, certain principles and ideas of traditional generative phonology introduced by Chomsky and Halle (1968) in the Sound Pattern of English have undergone so many interpretations and, oftentimes, even misinterpretations (see, e.g., Sampson 1980) that the content of the notions of the underlying and the surface form (or representation) is not completely unequivocal in different theories within the framework of generative phonology. Taking into account the schema (b) from figure 7-1, one can conclude that when making connections between the different levels of abstraction in representing the sound structure of language, only different terminology separates structuralist and generative phonology. The fundamental difference between the structuralist (represented by the SPT) and the generative approach (represented by the CVX syllable theory and Lexical Phonology) to the syllable is the fact that while the SPT enables linguists to analyse the syllable through the different levels of abstractionʊthe level of the phone, the phoneme, and the morphophoneme (section 6.4.2)ʊboth the CVX and the SSA in Lexical Phonology see the syllable as a sound unit on the surface only. In generative phonology, “syllable structure is not present in underlying representation” (Blevins 1995, 216) because “the underlying forms in many languages are so abstract that they do not contain a unit definable as a syllable” (Hooper 1975, 534). 2

This adverb is very important here because the SSA in Lexical Phonology sees CV as the fundamental syllable template, too.

152

Chapter Seven

Figure 7-1 The Generative and Structuralist Levels of Abstraction

7.3 Final Remarks The existence of certain features or phenomena that are common to all, or at least to most, world languages3 is a lucid fact. For example, most of the researchers in the field of phonological typology concur with the idea that all languages have vowels and that there are no languages without stops. Borrowing Hyman’s terminology (2007), these are the so-called absolute universals, sometimes also classified as descriptive universals; that is, those “where the effect of different theoretical frameworks is minimized” 3

Despite the ideas offered by typologists, the communication system as a whole seems to be essentially variable at all its levels as well as in its sound, meaning and syntactic organization in the world’s 6,000 to 8,000 languages (Evans and Levins 2009, 429).

The Generative and the Structuralist approach to the Syllable Structure 153

(345). In addition, universals dealing with syllable structure usually fall into the category of analytic universals, that is, those that are specifically theory-dependent (ibid.). Taking into account the existence of “absolute” and “analytic” universals or, in Hyman’s terminology, “descriptive” universals and “theory-driven” or “architectural” universals (ibid., 345, 362), what does it mean when an author indicates a theory is universal? To answer this question, first it is necessary to determine what is meant by the term phonological universal, and, second the author should “determine the intention behind” the alleged universal(s) (ibid., 347). Thus, if the term “phonological universal” indicates a feature that is to occur in every phonological system (see, e.g., Crystal 2005), then, for example, Duanmu’s (2009, 241) claim about the universality of the CVX syllable theory means that this syllable model occurs cross-linguistically, without exception. However, as shown by the detailed analysis in chapter 2, this is not true. But if the “intention behind” was specified, then the CVX syllable model would be plausible as a kind of architectural universal. Nevertheless, considering all three theories presented in this work and all descriptions, analyses, and comparisons, the existence of one wellknown and absolute universal has been attested: all languages have syllables.4 The particular structure of the syllable in the particular language can, but does not have to, be in accordance with the particular structure of the particular syllable in some other particular language because, apart from the features, which are similar, identical, or, we may dare to say, universal, there are also features and processes that differ from language to language. One can simply say that linguistic universality goes hand in hand with linguistic diversity. At some levels of linguistic organization, universality prevails, but at the others, it is shifted to the background and the diversity among languages can be observed. Whether we accept the existence of CV syllables or the existence of VC syllables in languages,5 the syllable as a sound unit exists; however, the number of segments in its parts, especially in onsets and codas (i.e., the structure of consonant clusters occurring syllable-initially, -medially and -finally) varies across languages. 4 Of course, the attribute ‘absolute’ in connection with the term ‘universal’ means they occur in most of the world languages, not in all of them. The reason for this is quite simple: an in-depth investigation of two languages is nonetheless not enough for the assertion that it is a ‘universal universal’. Any kind of universal claim can be challenged (for details, see Hyman 2007). 5 The relevant linguistic literature provides the arguments for the occurrence of both CV and VC models in languages (see, e.g., Blevins 1995; Hyman 2007).

154

Chapter Seven Consonant clusters in languages were formed on the basis of the lowest effort of the speech organs during their production and at the same time these clusters were created on the background of the linguistic changes triggered by the language as a system. New language, that is, also new combinatory peculiarities of the given language system up to its presentday dynamic state, have been created through these changes and with them.6 (Sabol 1975, 204)

6

This quotation has been translated by the author of this work.

CONCLUSIONS

The syllable has been the centre of my linguistic research for a long time (see, e.g., Gregová 1998; Gregová 2004). I acquired a structuralist linguistic background1 due to my education in Central Europe. However, the syllable theories in English linguistics are deeply rooted in generativism. While structuralists usually pay no attention to the generative theory of language, many generativists assess structuralism as an inadequate theory (see, e.g., Holland 1992; Koster 1996). The first account of language able “to develop something that could be called a theory” is the generativist approach (Goldsmith and Laks 2011, 7). Although early generative phonology discarded the syllable (which was referred to only indirectly), it has been the fundamental element of phonological analysis, with its hierarchical internal structure (cf., e.g., Hooper 1972), since the 1970s. Generally speaking, generative phonology assumes that there is one syllable model (or one set of rules for syllabification) valid in all languages and that the syllabification is not exhaustive (see Chapter 1 for details). However, structuralism accepts the universality of this sound unit only in the sense that there are syllables in all languages. It accepts the mutual interconnection between the form and the meaning in a linguistic sign, that is, in language, claiming syllabification is always exhaustive (Chapter 5). Some time ago I came across the book The Lexical Phonology of Slovak by J. Rubach (1993) in which I found the syllabification of Slovak words based on the so-called Syllable Structure Algorithm. The SSA is considered to be language-universal but it also accepts a certain degree of uniqueness in individual languages (Chapter 3). A couple of years later, I was asked to review the book Syllable Structure: The Limits of Variation by S. Duanmu (2009). The author, a generativist, connects the most influential generative syllable theories into one CVX syllable theory, and claims it is the maximal structure of the syllable in all languages. He

1

Despite the generativist claim that structuralism is an “outdated paradigm” (Albrecht 2011, 821), the basic ideas of Saussurrean structuralism have remained very popular in Slovak and in the whole of Slavonic linguistics (see, e.g., Horecký 1975).

156

Conclusions

argues, inter alia, if the CVX structure is suitable for English, it must be suitable for any other language (for details, see Chapter 2). The description and understanding of the syllable in these two approaches–Lexical Phonology with its Syllable Structure Algorithm and the CVX syllable theory–seemed to contradict the basic tenets of the structuralist approach to the syllable as presented in The Synthetic Phonological Theory by J. Sabol (1989) (see Chapter 6 for details), which was one of the pillars of my linguistic training. Having in mind all the positive and negative—even insulting, (see, e.g., Turner 1987)— evaluations of the Saussurean and Chomskyan concepts, I have decided to compare these two linguistic streams on the basis of their view of the syllable, which is—and here agreement can be found between structuralism and generativism—the crucial unit of a phonological analysis. Because generative phonology is only a cover term for a number of phonological theories, I chose the above-mentioned syllable theories within a generative approach to language for the comparison: CVX syllable theory (Chapter 2) is the most current approach to the syllable, while syllabification according to Lexical Phonology is the most influential (Chapter 3). Structuralism demonstrates itself in a number of theories, too. The synthetic phonological theoryʊas indicated by its name—represents the synthesis of most of the structuralist views of the syllable, and that is why this approach was chosen as representative of structuralism for the analysis. The aim of the workʊon the basis of an in-depth analysis of the data from two different languages, English and Slovakʊwas to compare and contrast these two approaches to the syllable, to point out the differences between them, and to find the similarities, if there are any. At the same time, the book has aimed to evaluate the assumed universal character of both generative theories of the syllable selected for the analysis, CVX theory and the Syllable Structure Algorithm of Lexical Phonology. Following reverse chronological order, chapter 1 specified the key features of the syllable in generative phonology. Then, attention was paid to the most up-to-date representative of the generative approach, CVX syllable theory (Chapter 2) and to the most popular generative theory of the syllable, the Syllable Structure Algorithm in Lexical Phonology (Chapter 3). The outcome of the thorough analysis of English and Slovak data from the viewpoints of these two syllable theories were evaluated and summarized in chapter 4. The CVX syllable theory allows the maximal syllable structure CVC or CVV for all languages, assuming that those consonants at the syllable edges, which do not fit this template, can be left extrasyllabic. These

The Generative and the Structuralist Approach to the Syllable

157

unsyllabified consonants can either be accounted for by one of three morphological concepts of this theory (the Affix Rule, the Potential Vowel Rule and Anti-Allomorphy), or they can be treated as complex segments (sections 2.1–2.2.3). However, as specified in detail in section 4.1, there are many controversial points in this theory. The core of the CVX approach, the notion of a complex sound, cannot be seen as a phenomenon occurring cross-linguistically, as claimed by Duanmu (2009). Treating certain sequences of sounds as segments with multiple articulation that have a phonetic duration of single segments and phonemically occupy only one X-slotʊthat is, one timing unit in an autosegmental sense and thus behave like simple soundsʊis not a universal phenomenon. Although some sequences of consonants seem to fulfil phonemic criteria for their representation as complex segments, there is no phonetic evidence for their single-slot representation. Obviously, in English and in Slovak, it is not simultaneous pronunciation but pronunciation in a sequence (for details, see sections 2.4–2.5). The morphological concepts of this approach do not display a universal nature,2 either. The maximal size of the English syllable, CCCVVCCCC, can be reduced to the required model, CVX (section 2.6); however, it does not seem to be possible to reduce the maximal size of the Slovak syllable, CCCCVVCCC, into this template (section 2.7). Duanmu tested his theory on the data from five languages: English, German, Standard Chinese, Shanghai Chinese, and Jiarong (section 2.2.3). English and German were chosen for their large consonant clusters, according to Duanmu (2009, 71) but, for example, Slovak allows more consonants in onsets than English or German, which both have the maximal number of three in initial clusters (cf. Duanmu 2009, 207). Considering the relation between the underlying representation and the surface representation as understood by generative phonology (section 1.3) and the thorough investigation of the English and the Slovak consonant clusters according to CVX syllable theory, the CVX approach to the syllable is acceptable cross-linguistically. Not on the surface, however, as claimed by its author, but on the underlying level as indicated by the results of the analysis (section 4.1).

2

Recall here that the limited sample of the analysed data cannot prove or disprove the universal validity of a certain notion or an idea. The outcome of such an analysis—limited in the number of languages but otherwise offering a comprehensive quantitative data (see also, Duanmu 2009, 51)—can question the postulate universality, which then opens the possibilities for more extensive research. See also, Introduction.

158

Conclusions

As for Lexical Phonology (Chapter 3),3 the syllable is delimited within the three-tier phonological representation (section 3.1), which reveals not only the structure of the syllable itself but also the detailed phonetic structure of the individual segments in the syllable. The SSA strictly follows the sonority of segments and their ordering in the syllable on the basis of the Sonority Sequencing Generalization (section 3.2). Once the rules for the delimitation of the syllable in the given language are ordered, that is, the SSA for the given language is specified (sections 3.2.1 and 3.2.2), it is applied to the syllabification of all words regardless of the other features that may govern the phonotactics of a given language (cf. sections 3.2.1.1 and 3.2.2.1). Simply said, although the SSA accepts the universality and the diversity among languages (at least in the field of the syllable structure),4 the delimitation of the individual rules within the SSA presumes the cross-language occurrence of certain linguistic phenomena; for example, the existence of glides in languages (cf. sections 3.2 and 3.2.2). This uniformity in the evaluation of sound elements leads to the schema of the syllable structure delimitation. In brief, the identical syllable structure, CCVVCC, for both languages analysed (section 3.3) results from this rule-based approach to the syllable, providing the consonants in onsets and codas follow the principle of the SSG. All consonants that somehow violate the SSG are simply left unsyllabified and are tolerated without any special treatment (see section 4.2 for details). The details of Sabol’s (1989) synthetic phonological theory and its understanding of the syllable, as well as the exhaustive analysis of English and the Slovak data, were found in chapter 6. Within the framework of the SPT that departs from the existence of different levels of the relationship between the individual and the universal (i.e., between the concrete and the abstract, as reflected in the oppositions “langue–parole” and “language–speech”), the syllable is seen as a complex unit, too. It is the fundamental syntagmatic-paradigmatic, phonic-rhythmical unit of a language with one peak of sonority. This sound unit functions as the place for manifesting the contrast and the distinctive features of phonic elements. It is also the bearer of all prosodic features of a given language 3

As already mentioned above, the SSA was specified for Slovak by J. Rubach (1993). The author’s understanding of some linguistic phenomena is not fully consistent with their function in the Slovak language. This not absolutely correct interpretation of certain phonic features is reflected in the results of the syllabification of Slovak words, as seen in chapter 3. 4 It is important to remember here that the goal of this book was not to concentrate on Lexical Phonology as a whole. I have paid full and exclusive attention to the approach of LP to delimit the syllable on the basis of the SSA.

The Generative and the Structuralist Approach to the Syllable

159

(Sabol 1994). The complex (or synthetic) approach is also applied when looking for syllable boundaries. The polysyllabic words with large consonant clusters can be syllabified in accordance with one of seven phonetic, phonemic, or morphophonemic criteria, depending on which aspect of the syllable is highlighted. All the criteria for syllable structure delimitation can be positioned into the hierarchy (Fig. 6-9), which displays the degree of phonetic and phonemic criteria. At the same time, this ranking shows the language-specific and language-universal nature of these criteria. The higher the degree of the phonemic criterion on the axis, the higher the degree of the individual, or concrete, aspect of the criterion; that is, the stronger the language-specific nature of the criterion. The higher the degree of the phonetic character of the criterion, the stronger the abstract, or universal, aspect; that is, the given criterion is more languageuniversal (section 6.4). Despite all the differences specified above, the final comparison of the fundamental principles of all three syllable theories (Chapter 7) has shown that both generative and structuralist syllabifications follow the same principles when parsing words into syllables: the analogy between the syllabic and morphemic boundaries and the phonotactic rules. Although ill-formed and well-formed consonant clusters are determined by different criteria in the individual theories (cf. Chapter 2, 3, 6), the final sample syllabification of the English and the Slovak words with large consonant clusters (section 7.1) proves that—contrary to initial expectations—there are only a few differences between the generative and the structuralist syllabification (Table 7-1 and Table 7-2). Obviously, generativism, as represented by CVX syllable theory and the SSA of Lexical Phonology, puts emphasis on universality. Sometimes the languages are forced to fit the theory and that is why many consonants remain unsyllabified since they either do not fit the given syllable model or they violate the rule. The other observation resulting from syllabification in accordance with the generative approach is that the syllable as a fundamental unit of phonological analysis is understood as the element independent of the higher-level unit, the word,5 of which the syllable is part. The structuralist syllabification, exemplified by the synthetic phonological theory, accepts the universal aspects in creating syllables (those connected with the way humans produce speech, that is,

5

None of the theories presented here accepts the syllabification across word boundaries, as is possible in the purely phonetic delimitation of the syllable (see, e.g., Pauliny, Ružiþka, and Štolc 1968).

160

Conclusions

those connected with the phonetic substance of the syllable6), but also the cross-linguistic variability (the possible combinations of sounds and the number of segments in onsets and codas are typical of a particular language only). And although the SPT sees the syllable as a sound unit, this theory emphasizes the priority of meaning over sound in language as a whole when building up the syllable. And, finally, it seems that the structuralist levels of abstractionʊthe level of the phone, the phoneme, and the morphophoneme (section 6.4.2) ʊmore or less correspond to the generative levels of abstraction, the surface form, the pre-surface form, and the underlying form (Fig. 7-1). To conclude, this book has aimed to compare and contrast the generative and the structuralist approaches to the syllable, the important unit of a phonological analysis. Taking into account the facts that structuralism was subjected to severe criticism by many generativists (see, e.g., Koster 1966; Turner 1987) and Slavonic linguistics, which is mainly structuralist, pays no attention to generative ideas, I initially expected a great discrepancy in the syllabification based on these two approaches. But the detailed comparison of the fundamental ideas, principles, and methods of the analysis as well as the analysis of data from English and Slovak has revealed that the dissimilarities are only minor: There is a different terminology and a difference of opinion regarding whether the syllabification should be exhaustive or not. Linguistically speaking, this work is a small attempt to bridge the gap between structuralist and generative linguistics thatʊdue to various nonlinguistic reasonsʊseems to be so diverse. Philosophically speaking, the outcome of the book again verifies one well-known truth: Reality (and everything in it) remains the same, no matter how we look at it.

6

The universality of phonetic description is mentioned, for example, by Haspelmath (2007) in his discussion on the existence of the pre-established categories in language. Assuming that the pre-established categories do not exist, the author says that the cross-linguistic comparison in phonology “must be phonetically based” (Haspelmath 2007, 6), because substance and not the category is universal. See also Chapter 7 for the position of phonetics in generative and structuralist delimitation of the syllable.

APPENDIX 1 THE LIST OF THE TRANSCRIPTION SYMBOLS FOR CONSONANT PHONEMES

1. The list of the English transcription symbols for consonant phonemes: /p/ /t/ /k/ /b/ /d/ /g/ /f/ /v/ /۟/ /˙/ /s/ /z/

pot tea key bee dog gun fun very thick that say zip

/ߦ/ show /঱/ measure /ࠔ/ chin /˔/June /m/ may /n/ nose /ƾ/ king /r/ raw /l/ late /w/ wet /j/ yet /h/ ham

2. The list of the Slovak transcription symbols for consonant phonemes1: /p/ /t/ /k/ /b/ /d/ /g/ 1

pero‘pen’ ten ‘this’ kniha ‘book’ byĢ ‘to be’ daĢ ‘to give’ gaštan ‘chestnut’

/c/ cena ‘price’ ঱FXG G]t‘foreign’ /þ/ þas ‘time’ /শ/ džem ‘jam’ /m/ maĢ ‘to have’ /n/ nos ‘nose’

The transcription symbols of post-alveolar and alveopalatal consonants in Slovak are different from those used by the IPA. The Slovak /[š, ž, þ, Ģ, ć, Ė, Đ]/ stand for the IPA /[‫ݕ‬, ‫ݤ‬, ‫ݹ‬, tj, dj, nj, lj]/, respectively.

162

/Ģ/ ĢahaĢ ‘to pull’ /˄/ ćakovaĢ‘to thank’ /f/ fúkaĢ ‘to blow’ /v/ vietor ‘wind’ /s/ sneh ‘snow’ /z/ zima‘winter’ /š/ šaty ‘dress’ /ž/ žaba‘frog’

Appendix 1

/Ė/ /r/ /l/ /Đ/ /j/ /x/ /h/

tuleĖ ‘seal’ rok ‘year’ lano ‘rope’ Đan ‘flax’ jar ‘spring’ chmeĐ ‘hop’ hodina ‘hour’

APPENDIX 2 THE DISTINCTIVE FEATURES OF CONSONANT PHONEMES USED IN THE ANALYSIS

Major class features: [consonantal ] – [+consonantal] are segments produced with an obstruction to the air flow, i.e., all consonants; [-consonantal] are /j, w/. [sonorant] – [+sonorant] are sonorant consonants, i.e., approximants and nasals; [-sonorant] are obstruents. Laryngeal features: [voice] – [+ voice] indicates active participation of vocal cords; [-voice] means no vocal cords vibration [+/í spread glottis]. Manner features: [continuant] – [+continuant] sounds are produced without any obstruction to the air flow (approximants and fricatives); [-continuant] are oral and nasal stops. [nasal] – during the production of [+nasal] sound the velum is lowered and the air escapes through the nasal cavity; a sound is [-nasal] when the soft palate closes the entrance into the nasal cavity. [stop] – [+stop] sounds are produced with a complete closure. [fricative] – [+fricative] consonants are produced by narrowing. [lateral] – [+lateral] sound is produced by lowering the mid section of the tongue, only /l/ in English; [-lateral] – all the rest. Place features: [Labial] – [+labial] are articulated with lips, i.e. bilabial and labio-dental consonants. [round] – [+round] are produced with rounded lips (/w/ in English). [Coronal] – [+coronal] sounds are produced by raising the tongue blade (dentals, alveolars, palato-alveolars, palatals); [-coronal] segments are labials, velars, laryngeals.

164

Appendix 2

[anterior] – [+anterior] sounds are articulated in front of the palatoalveolar region (bilabial, labio-dental, dental, alveolar); palato-alveolar, palatal, velar and laryngeal sounds are [-anterior]. [distributed] – [+distributed] are dental, palato-alveolar and palatal sounds. The tongue is extended for some distance in the mouth during their articulation. [Dorsal] – articulated with the back of the tongue, i.e. palatal and velar consonants. Depending on the position of the dorsum, they can be: [high] – [+high] when the tongue is raised, [low] – [+low] when the tongue is lowered, [back] – [+back] when the tongue body is retracted from the neutral position. (Halle and Clements 1983; Giegerich 1992, 93–129)

APPENDIX 3 A COMPLEX-SOUND ANALYSIS OF THE ENGLISH WORD-INITIAL CC CLUSTERS

/pr/1 /p/ – Labial [-round] /r/ – Coronal [-anterior] /pr/ is a possible complex sound (/pr/) characterized by different articulators. /pl/ /p/ – Labial [-round] /l/ – Coronal [+anterior], [+lateral] /pl/ is a possible complex sound (/pl/) characterized by different articulators. /pj/ /p/ – Labial [-round] /j/ – Coronal [-anterior] /pj/ is a possible complex sound (/pj/) characterized by different articulators. /pw/ /p/ – Labial [+stop], [-fricative]2, [-round] /w/ – Labial [-stop], [-fricative], [+round] /pw/ is a possible complex sound (/pw/) characterized by the same articulator, however without conflicting gestures; relevant feature for /p/ is 1

Based on the theory of underspecification, only relevant, i.e., contrastive, features of articulators are specified. For example, /p/ is unspecified for Coronal and /r/ is unspecified for Labial (Duanmu 2010, 18). 2 The delimitation of the stricture features [stop] and [fricative] is based on the theory of underspecification which assumes that “any sound unspecified for [fricative] is [-fricative] and any sound unspecified for [stop] is [-stop]” (Duanmu 2009, 24; see also Hall 2006, 313).

166

Appendix 3

Labial [+stop] and for /w/ Labial [+round] (Duanmu 2009, 175). /pw/ is the so-called round (labialized) labial (Hall 1986, 322) occurring, for example, in the Nupe language (Chomsky and Hall 1991, 311).3 /pf/ /p/ – Labial [+stop], [-fricative], [-round] /f/ – Labial [-stop], [+fricative], [-round] /pf/ cannot be a complex sound. The individual consonants are produced by the same articulator with conflicting stricture gestures (features). /ps/ /p/ – Labial [-round] /s/ – Coronal [+anterior] /ps/ is a possible complex sound (/ps/) characterized by two different articulators. /pߦ/ /p/ – Labial [-round] /ߦ/ – Coronal [-anterior] /pߦ/ is a possible complex sound (/pߦ/) characterized by two different articulators. /br/ /b/ – Labial [-round] /r/ – Coronal [+anterior] /br/ is a possible complex sound (/bU/) characterized by two different articulators. /bl/ /b/ – Labial [-round] /l/ – Coronal [+anterior] /bl/ is a possible complex sound (/bO/) characterized by two different articulators.

3

If a complex segment exists in some other languages (e.g., /pw/ in Nupe), Duanmu accepts its existence in English, too.

The Generative and the Structuralist Approach to the Syllable

167

/bj/ /b/ – Labial [-round] /j/ – Coronal [-anterior] /bj/ is a possible complex sound (/bM/) characterized by two different articulators. /tr/ /t/ – Coronal [+anterior], [+stop] /r/ – Coronal [-anterior], [-fricative] /tr/ is an affricate with the same articulator (see also section 2.2.2), i.e. a complex segment. Following several authors (e.g., Jones, Abercrombie, Gimson), Duanmu classifies /tr/ as an affricate (Duanmu 2009, 175). /tw/ /t/ – Coronal /w/ – Labial /tw/ is a good complex segment (/tw/), there are two different articulators. /tj/ /t/ – Coronal [+anterior] /j/ – Coronal [-anterior], but also Dorsal [+high] /tj/ can be classified as a complex segment (/tj/) since /j/ is Coronal, but also Dorsal, and thus there is a different articulator (Duanmu 2010, 175). /dr/ /d/ – Coronal [+anterior], [+stop] /r/ – Coronal [-anterior], [-fricative] /dr/ is an affricate with the same articulator (see also section 2.2.2), i.e., a complex segment (see also the evaluation of the cluster /tr/ above). /dj/ /t/ – Coronal [+anterior] /j/ – Coronal [-anterior], but also Dorsal [+high] /dj/ can be–similarly to /tj/–classified as a complex segment (/dj/). /dw/ /d/ – Coronal /w/ – Labial /tw/ is a good complex segment (/tw/) characterized by different articulators.

168

Appendix 3

/kr/ /k/ – Dorsal /r/ – Coronal /kr/ can be represented as a complex segment (/kr/) and it involves different articulators. /kl/ /k/ – Dorsal /l/ – Coronal /kl/ is a possible complex sound (/kl/) with different articulators. /kw/ /k/ – Dorsal /w/ – Labial /kw/ is a good complex sound (/kw/) with different articulators. /kj/ /k/ – Dorsal [+high], [+back] /j/ – Coronal [-anterior], Dorsal [+high] /kj/ can be a complex sound (/kj/) either produced by different articulators, Coronal and Dorsal, or by the same one, but without a conflicting Dorsal gesture [+high]. /km/ /k/ – Dorsal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /km/ cannot be a complex segment due to the conflicting gestures by the Soft Palate.4 /kn/ /k/ – Dorsal, Soft Palate [-nasal] /m/ – Coronal, Soft Palate [+nasal] /km/ is not a possible complex sound since there are the conflicting gestures under the Soft Palate articulator /kv/ /k/ – Dorsal /v/ – Labial /kv/ is a good complex sound (/kv/) involving different articulators. 4

For the possibility of a complex-segment status of the so-called pre-nasalized stops, see, for example, Sheer 2012.

The Generative and the Structuralist Approach to the Syllable

169

/gr/ /g/ – Dorsal / r/ – Coronal /gr/ is a complex sound (/gr/), there are two different articulators. /gl/ /g/ – Dorsal / l/ – Coronal /gl/ is a possible complex sound (/gl/) since it involves two different articulators. /gw/ /g/ – Dorsal / w/ – Labial /gw/ can be a complex segment (/gw/) produced by different articulators. /nj/ /n/ – Coronal [+anterior], Soft Palate [+nasal] /j /– Coronal [-anterior], Dorsal [+high], Soft Palate [-nasal] Following Duanmu’s criteria for the classification of a consonant cluster as a complex sound, /nj/ cannot be a complex segment due to the conflicting gestures by the same articulators Coronal and Soft Palate. However, Duanmu lists /nj/ among the clusters that can be represented as complex sounds (Duanmu 2009, 178), but without any further explanation. Since this simplification contradicts the No Contour Principle (section 2.2.2), the corner-stone of the CVX theory of the syllable, this cluster will not be treated as a possible complex sound. /mj/ /m/ – Labial, Soft Palate [+nasal] / j/ – Coronal, Soft Palate [-nasal] /mj/ cannot be a complex segment due to the conflicting gestures by the Soft Palate. /mw/ /m/ – Labial [-round], Soft Palate [+nasal] /w/ – Labial [+round], Soft Palate [-nasal] /mw/ is not a good complex sound since there are conflicting gestures by the Labial and Soft Palate articulators. Nevertheless, Duanmu–without any supporting arguments–classifies this cluster as a complex segment (Duanmu 2009, 175), thus contradicting his own theory. To preserve the

170

Appendix 3

continuity with the previous analysis, this cluster will be classified as an impossible complex segment. /fl/ /f/ – Labial /l/ – Coronal /fl/ is a good complex segment (/fl/) involving different articulators. /fr/ /f/ – Labial /r/ – Coronal /fr/ can be represented as a complex sound (/fr/) since there are different articulators. /fj/ /f/ – Labial / j/ – Coronal /fj/ can be a complex sound (/fj/) since different articulators are involved. /vj/ /v/ – Labial /j/ – Coronal /vj/ is a possible complex segment (/vj/) since both its parts have different articulators. /vw/ /v/ – Labial [-round] /w/ – Labial [+round] Although Duanmu (2009) includes this cluster among the possible complex segments (Duanmu 2009, 175), it cannot be presented as a complex sound due to the conflicting gestures by the same Labial articulator. /șr/ /ș/ – Coronal [+anterior] /r / – Coronal [-anterior] /șr/ is not a good complex sound since there are the conflicting gestures by the Coronal articulator.

The Generative and the Structuralist Approach to the Syllable

171

/șw/ /ș/ – Coronal /w/ – Labial /șw/ can be a complex segment (/șw/) involving different articulators. /șj/ /ș/ – Coronal [+anterior] /j/ – Coronal [-anterior], Dorsal [-high] /șj/ can be a complex sound since /j/ is not only Coronal but also Dorsal, and thus there are two different articulators. /st/ /s/ – Coronal [+anterior], [-son], [+continuant] /t/ – Coronal [+anterior], [-son], [-continuant] /st/ cannot be a complex sound because of the conflicting gesture of stricture [+/-continuant]. /sp/ /s/ – Coronal [+fricative], [-son], [+continuant] /p/ – Labial [+stop], [-sonorant], [-continuant] /sp/ is a good complex segment (/sp/) involving different articulators. Contrary to the complex-segment evaluation of the clusters /mw/ or /vw/, Duanmu (2009) does not classify the cluster /sp/ as a possible complex segment. This, again, contradicts the No Contour Principle and the author’s own delimitation of a complex sound. /sk/ /s/ – Coronal /k/ – Dorsal /sk/ can be represented as a complex segment (/sk/) produced by different articulators. However, Duanmu does not include this cluster among the possible complex sounds. /sl/ /s/ – Coronal [+anterior],[-lateral] /l/ – Coronal [+anterior], [+lateral] /sl/ cannot be a complex sound because of the conflicting gesture [+/lateral] made by the same articulator.

172

Appendix 3

/sw/ /s/ – Coronal /w/ – Labial /sw/ is a possible complex sound (/sw/) with different articulators. /sn/ /s/ – Coronal [+anterior], Soft Palate [-nasal] /n/ – Coronal [+anterior], Soft Palate [+nasal] /sn/ is not a good complex segment; there are conflicting gestures by the Soft Palate. /sm/ /s/ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /sm/ cannot be a complex segment because of the conflicting gestures by the Soft Palate. /sf/ /s/ – Coronal /f/ – Labial /sf/ seems to be a possible complex segment (/sf/) involving different articulators. This cluster is another example of the discrepancy between Duanmu’s theory of a complex sound and his analysis of the English consonant clusters as complex segments (Duanmu 2009). /sj/ /s/ – Coronal [+anterior] /j/ – Coronal [-anterior], Dorsal [+high] /sj/ is a possible complex segment (/sj/) with different Coronal and Dorsal articulators. /sr/ /s/ – Coronal [+anterior] /r/ – Coronal [-anterior] /sr/ cannot be a complex segment; there are conflicting gestures by the same articulator.

The Generative and the Structuralist Approach to the Syllable

173

/sv/ /s/ – Coronal /v/ – Labial /sv/ is a possible complex segment (/sv/) involving a different articulator. /zl/ /z/ – Coronal [+anterior], [-son], [+continuant] /l/ – Coronal [+anterior], [+son], [-continuant] /zl/ cannot be a complex sound because of the conflicting gestures of stricture by the same articulator. ߦr ߦ – Coronal [-anterior], [+fricative] /r/ – Coronal [-anterior], [-fricative] (see note 18 in Chapter 2) ߦU/ cannot be a complex segment because of the conflicting stricture gestures. ߦm/ ߦ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] ߦP/ is not a good complex sound; there are conflicting gestures by the Soft Palate.  ߦn/ ߦ – Coronal [-anterior], Soft Palate [-nasal] /n/ – Coronal [+anterior], Soft Palate [+nasal] ߦn/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. ߦp/ ߦ– Coronal /p/ – Labial ߦS/ is a possible complex sound ߦS) with a different articulator.

174

Appendix 3

ߦw/ ߦ– Coronal / w/ – Labial ߦZ/ can be a complex segment ߦZ) involving different articulators. /hj/ /h/ – Dorsal /j/ – Coronal /hj/ is a complex sound (/hj/) with a different articulator.

APPENDIX 4 A COMPLEX-SOUND ANALYSIS OF THE SLOVAK WORD-INITIAL CC CLUSTERS

/ps/ /p/ – Labial /s/ – Coronal /ps/ is a possible complex segment (/ps/) involving different articulators. /pš/ /p/ – Labial /š/ – Coronal /pš/ can be a complex sound (/pš/); there are two different articulators. /px/ /p/ – Labial /x/ – Dorsal /px/ is a good complex segment (/px/) because of the different articulators used. /pn/ /p/ – Labial, Soft Palate [-nasal] /n/ – Coronal, Soft Palate [+nasal] /pn/ cannot be a complex sound due to the conflicting gestures by the Soft Palate articulator. /pĖ/ /p/ – Labial, Soft Palate [-nasal] /Ė/ – Coronal, Soft Palate [+nasal] /pĖ/ is bad in a complex-sound analysis because of the conflicting values of the gesture [nasal].

176

Appendix 4

/pl/ /p/ – Labial /l/ – Coronal /pl/ can be represented as a complex sound (/pl/) with two different articulators. /bć/ /b/ – Labial /ć/ – Coronal /bć/ is a complex segment (/bć/) with different articulators. /bz/ /b/ – Labial /z/ – Coronal /bz/ is a possible complex sound (/bz/); there are different articulators. /bl/ /b/ – Labial /l/ – Coronal /bl/ can be represented as a complex sound (/bl/) produced by different articulators. /bĐ/ /b/ – Labial /Đ/ – Coronal /bĐ/ is a possible complex sound (/bĐ/) involving different articulators. /br/ /b/ – Labial /r/ – Coronal /br/ can be a complex segment (/br/); there are two different articulators. /tk/ /t/ – Coronal /k/ – Dorsal /tk/ is a complex sound (/tk/) with different articulators. /tx/ /t/ – Coronal /x/ – Dorsal /tx/ is a possible complex sound (/tx/) with different articulators.

The Generative and the Structuralist Approach to the Syllable

177

/tv/ /t/ – Coronal /v/ – Labial /tv/ is a good complex sound (/tv/) with different articulators. /tm/ /t/ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /tm/ is not suitable for a complex-sound analysis because of the conflicting features by the Soft Palate. /tl/ /t/ – Coronal [+anterior], [-lateral]1, [+stop] /l/ – Coronal [+anterior], [+lateral], [+fricative] /tl/ cannot be a complex sound because of the conflicting gestures by the same articulator. /tĐ/ /t/ – Coronal [+anterior], [-lateral], [+stop] /Đ/ – Coronal [-anterior], [+lateral], [+fricative] /tĐ/ is not a possible complex sound, there are conflicting gestures by the same articulator. /tr/ /t/ – Coronal [+anterior], [+stop] /r/ – Coronal [+anterior], [+fricative] /tr/, as in English, can be formally treated as an affricate because of the simultaneous features [+stop], [+fricative] by the Coronal articulator 2 (see also section 2.2.2). /dv/ /d/ – Coronal /v/ – Labial /dv/ is a good complex sound (/dv/) with different articulators.

1

The delimitation of this feature in phoneme /t/ is based on the theory of underspecification. See notes 1 and 2 in Appendix 3. 2 These are the so-called ‘contour segments’ in Sagey’s theory (1986).

178

Appendix 4

/dm/ /d/ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /dm/ is not suitable for a complex-sound analysis, there are conflicting features by the Soft Palate articulator. /dn/ /d/ – Coronal [+anterior], Soft Palate [-nasal] /n/ – Coronal [+anterior], Soft Palate [+nasal] /dn/ cannot be a complex sound because of the conflicting features by the Soft Palate. /dĖ/ /d/ – Coronal [+anterior], Soft Palate [-nasal] /Ė/ – Coronal [-anterior], Soft Palate [+nasal] /dĖ/ cannot be a complex sound because of the conflicting gestures by the Coronal and Soft Palate articulators. /dl/ /d/ – Coronal [+anterior], [-lateral] /l/ – Coronal [+anterior], [+lateral] /dl/ is not a complex sound because of the conflicting features [+/-lateral] /dĐ/ /d/ – Coronal [-anterior], [-lateral] /Đ/ – Coronal [+anterior], [+lateral] /dĐ/ is not a complex sound because of the conflicting features by the Coronal articulator. /dr/ /d/ – Coronal [+anterior], [+stop] /r/ – Coronal [+anterior], [+fricative] /dr/ is a special type of a complex segment–an affricate–characterized by the simultaneous features [+stop], [+fricative] with the Coronal articulator (see also the analysis of the cluster /tr/ above). /kt/ /k/ – Dorsal /t/ – Coronal /kt/ can be a complex segment (/kt/) involving different articulators.

The Generative and the Structuralist Approach to the Syllable

179

/kv/ /k/ – Dorsal /v/ – Labial /kv/ is a possible complex segment (/kv/) with two different articulators. /km/ /k/ – Dorsal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /km/ cannot be a complex segment because of the conflicting values of the feature [nasal]. /kn/ /k/ – Dorsal, Soft Palate [-nasal] /n/ – Coronal, Soft Palate [+nasal] /kn/ cannot be a complex segment; there are conflicting values of the feature [nasal]. /kĖ/ /k/ – Dorsal, Soft Palate [-nasal] /Ė/ – Coronal, Soft Palate [+nasal] /kĖ/ cannot be a complex segment because of the conflicting features by the Soft Palate articulator. /kr/ /k/ – Dorsal /r/ – Coronal /kr/ is a possible complex sound (/kr/) with different articulators. /kl/ /k/ – Dorsal /l/ – Coronal /kl/ can be a complex sound (/kl/) produced with different articulators. /kĐ/ /k/ – Dorsal /Đ/ – Coronal /kĐ / can be a complex sound (/kĐ/) involving different articulators.

180

Appendix 4

/gn/ /g/ – Dorsal, Soft Palate [-nasal] /n/ – Coronal, Soft Palate [+nasal] /gn/ cannot be a complex segment; there are the conflicting values of the feature [nasal]. /gĖ/ /g/ – Dorsal, Soft Palate [-nasal] /Ė/ – Coronal, Soft Palate [+nasal] /gĖ/ cannot be a complex segment because of the conflicting features by the Soft Palate articulator. /gl/ /g/ – Dorsal /l/ – Coronal /gl/ can be a complex sound (/gl/) produced with different articulators. /gĐ/ /g/ – Dorsal /Đ/ – Coronal /gĐ / can be a complex sound (/gĐ/) involving different articulators. /gr/ /g/ – Dorsal /r/ – Coronal /gr/ is a possible complex sound (/gr/) with different articulators. /mn/ /m/ – Labial, Soft Palate [+nasal] /n/ – Coronal, Soft Palate [+nasal] /mn/ can be a complex segment (/mn/) with different articulators. /mĖ/ /m/ – Labial, Soft Palate [+nasal] /Ė/ – Coronal, Soft Palate [+nasal] /mĖ/ is a possible complex segment (/mĖ/) with different articulators.

The Generative and the Structuralist Approach to the Syllable

181

/ml/ /m/ – Labial, Soft Palate [+nasal] /l/ – Coronal, Soft Palate [-nasal] /ml/ cannot be a complex segment because of the conflicting gestures by the Soft Palate. /mĐ/ /m/ – Labial, Soft Palate [+nasal] /Đ/ – Coronal, Soft Palate [-nasal] /mĐ/ is not a complex segment; there are the conflicting values of the feature [nasal]. /mr/ /m/ – Labial, Soft Palate [+nasal] /r/ – Coronal, Soft Palate [-nasal] /mr/ cannot be a complex segment because of the conflicting gestures by the Soft Palate articulator. /sp/ /s/ – Coronal /p/ – Labial /sp/ is a possible complex sound (/sp/) with different articulators. /st/ /s/ – Coronal [+anterior] /t/ – Coronal [-anterior] /st/ cannot be a complex sound because there are conflicting gestures by the same articulator. /sĢ/ /s/ – Coronal [+anterior] /Ģ/ – Coronal [-anterior] /sĢ/ cannot be a complex sound because of the conflicting gestures by the same articulator. /sk/ /s/ – Coronal /k/ – Dorsal /sk/ is a possible complex sound (/sk/) involving two different articulators.

182

Appendix 4

/sx/ /s/ – Coronal /x/ – Dorsal /sx/ is a possible complex sound (/sx/) with different articulators. /sv/ /s/ – Coronal /v/ – Labial /sv/ can be represented as a complex sound (/sv/) with different articulators. /sm/ /s/ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /sm/ is not a good complex sound; there are the conflicting gestures under the Soft Palate articulator /sn/ /s/ – Coronal, Soft Palate [-nasal] /n/ – Labial, Soft Palate [+nasal] /sn/ is not a good complex sound because of the conflicting gesture by the Soft Palate articulator. /sĖ/ /s/ – Coronal [+anterior], Soft Palate [-nasal] /Ė/ – Coronal [-anterior], Soft Palate [+nasal] /sĖ/ cannot be a good complex sound; there are the conflicting features by two articulators. /sl/ /s/ – Coronal [+anterior], [-lateral] /l/ – Coronal [+anterior], [+lateral] /sl/ is not a good complex sound because of the conflicting values of the [lateral] feature. /sĐ/ /s/ – Coronal [+anterior], [-lateral] /Đ/ – Coronal [-anterior], [+lateral] /sĐ/ cannot be a complex sound; there are the conflicting values of the [anterior] and [lateral] features.

The Generative and the Structuralist Approach to the Syllable

183

/zb/ /z/ – Coronal /b/ – Labial /zb/ is a possible complex sound (/zb/) with different articulators. /zv/ /z/ – Coronal /v/ – Labial /zv/ can be a complex sound /zv/; there are different articulators. /zn/ /z/ – Coronal, Soft Palate [-nasal] /n/ – Labial, Soft Palate [+nasal] /zn/ is not a good complex sound because of the conflicting gesture by the Soft Palate articulator. /zĖ/ /z/ – Coronal [+anterior], Soft Palate [-nasal] /Ė/ – Coronal [-anterior], Soft Palate [+nasal] /zĖ/ cannot be a good complex sound; there are the conflicting features by two articulators. /zl/ /z/ – Coronal [+anterior], [-lateral] /l/ – Coronal [+anterior], [+lateral] /zl/ is not a good complex sound, because of the conflicting values of the [lateral] feature. /zr/ /z/ – Coronal [+anterior], [+fricative], [-sonorant], [+continuant] /r/ – Coronal [+anterior], [+fricative], [+sonorant], [+continuant] /zr/ is not a good complex sound because of the conflicting values of the [sonorant] feature. /šp/ /š/ – Coronal /p/ – Labial /šp/ is a possible complex sound (/šp/) with different articulators.

184

Appendix 4

/št/ /š/ – Coronal [-anterior], [+fricative], [-sonorant], [+continuant] /t/ – Coronal [+anterior], [+stop], [-sonorant], [-continuant] /št/ cannot be a complex sound; there are the conflicting gestures by the same articulator. /šĢ/ /š/ – Coronal [-anterior], [+fricative], [-sonorant], [+continuant] /Ģ/– Coronal [-anterior], [+stop], [-sonorant], [-continuant] /št/ cannot be a complex sound; there are the conflicting stricture gestures by the same articulator. /šk/ /š/ – Coronal /k/ – Dorsal /šk/ is a possible complex sound (/šk/) involving two different articulators. /šm/ /š/ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /šm/ is not a good complex sound; there are the conflicting gestures under the Soft Palate articulator. /šn/ /š/ – Coronal, Soft Palate [-nasal] /n/ – Labial, Soft Palate [+nasal] /šn/ is not a good complex sound because of the conflicting gestures by the Soft Palate articulator. /šl/ /š/ – Coronal [-anterior], [-lateral] /l/ – Coronal [+anterior], [+lateral] /šl/ cannot be a good complex sound; there are the conflicting features by the same articulator. /šĐ/ /š/ – Coronal [-anterior], [-lateral] /Đ/ – Coronal [-anterior], [+lateral] /šĐ/ cannot be a complex sound; there are the conflicting values of the [lateral] feature.

The Generative and the Structuralist Approach to the Syllable

185

/šr/ /š/ – Coronal [-anterior], [+fricative], [-sonorant], [+continuant] /r/ – Coronal [+anterior], [+fricative], [+sonorant], [+continuant] /šr/ is not a good complex sound because of the conflicting values of the same features. /žv/ /ž/ – Coronal /v/ – Labial /žv/ is a good complex sound (/žv/) because there are different articulators. /žm/ /ž/ – Coronal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /žm/ is not a good complex sound; there are the conflicting gestures under the Soft Palate articulator. /žĖ/ /ž/ – Coronal [+anterior], Soft Palate [-nasal] /Ė/ – Coronal [-anterior], Soft Palate [+nasal] /žĖ/ cannot be a good complex sound; there are conflicting features by two articulators. /žĐ/ /ž/ – Coronal [+anterior], [-lateral] /Đ/ – Coronal [-anterior], [+lateral] /žĐ/ cannot be a good complex sound; there are two conflicting features by one articulator. /žr/ /ž/ – Coronal [-anterior], [+fricative], [-sonorant], [+continuant] /r/ – Coronal [+anterior], [+fricative], [+sonorant], [+continuant] /žr/ is not a good complex sound because of the conflicting values of the same features. /hn/ /h/ – Dorsal, Soft Palate [-nasal] /n/ – Coronal, Soft Palate [+nasal] /hn/ cannot be a complex sound because of the conflicting gestures by the Soft Palate.

186

Appendix 4

/hĖ/ /h/ – Dorsal, Soft Palate [-nasal] /Ė/ – Coronal, Soft Palate [+nasal] /hĖ/ cannot be a complex sound; there are the conflicting gestures by the Soft Palate. /hl/ /h/ – Dorsal /l/ – Coronal /hl/ is a possible complex sound (/hl/) with different articulators. /hĐ/ /h/ – Dorsal /Đ/ – Coronal /hĐ/ is a possible complex sound (/hĐ/) with two different articulators. /hr/ /h/ – Dorsal /r/ – Coronal /hr/ can be a complex sound (/hr/); there are different articulators. /hm/ /h/ – Dorsal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /hm/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. /hv/ /h/ – Dorsal /v/ – Labial /hv/ is a good complex sound (/hv/) because there are different articulators. /xc/ /x/ – Dorsal /c/ – Coronal /xc/ can be a complex segment (/xc/) with different articulators. /xv/ /x/ – Dorsal /v/ – Labial /xv/ can be a complex segment (/xv/) involving different articulators.

The Generative and the Structuralist Approach to the Syllable

187

/xm/ /x/ – Dorsal, Soft Palate [-nasal] /m/ – Labial, Soft Palate [+nasal] /xm/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. /xl/ /x/ – Dorsal /l/ – Coronal /xl/ is a possible complex sound (/xl/) with different articulators. /xĐ/ /x/ – Dorsal /Đ/ – Coronal /xĐ/ can be a complex sound (/xĐ/) with different articulators. /xr/ /h/ – Dorsal /r/ – Coronal /xr/ can be a complex sound (/xr/); there are different articulators. /ft/ /f/– Labial /t/ – Coronal /ft/ is possible complex sound (/ft/); there are different articulators. /fĢ/ /f/– Labial /Ģ/ – Coronal /fĢ/ can be a complex sound (/fĢ/) with different articulators. /fþ/ /f/– Labial /þ/ – Coronal /fþ/ can be a complex sound (/fþ/); there are different articulators. /fs/ /f/ – Labial /s/ – Coronal /fs/ is a possible complex sound (/fs/); there are different articulators.

188

Appendix 4

/fš/ /f/ – Labial /š/ – Coronal /fš/ can be a complex sound (/fš/) with different articulators. /fl/ /f/ – Labial /l/ – Coronal /fl/ is a possible complex sound (/fl/); there are different articulators.

/fĐ/ /f/ – Labial /Đ/ – Coronal /fĐ/ can be a complex sound (/fĐ/) with different articulators. /fr/ /f/ – Labial /r/ – Coronal /fr/ can be a complex segment (/fr/); there are two different articulators. /vd/ /v/ – Labial /d/ – Coronal /vd/ is a possible complex segment (/vd/) involving different articulators. /vz/ /v/ – Labial /z/ – Coronal /vz/ is a possible complex segment (/vz/) with different articulators. /vn/ /v/ – Labial, Soft Palate [-nasal] /n/ – Coronal, Soft Palate [+nasal] /vn/ cannot be a complex sound because of the conflicting gestures by the Soft Palate.

The Generative and the Structuralist Approach to the Syllable

189

/vĖ/ /v/ – Labial, Soft Palate [-nasal] /Ė/ – Coronal, Soft Palate [+nasal] /vĖ/ cannot be a complex sound; there are the conflicting gestures by the Soft Palate. /vl/ /v/ – Labial /l/ – Coronal /vl/ can be a complex sound (/vl/) with different articulators. /vr/ /v/ – Labial /r/ – Coronal /vr/ is a possible complex sound (/vr/) with different articulators. /lk/ /l/ – Coronal /k/ – Dorsal /lk/ can be a complex sound (/lk/) involving different articulators. /lž/ /l/ – Coronal [+anterior], [+lateral] /ž/ – Coronal [-anterior], [-lateral] /lž/ cannot be a complex sound; there are the conflicting features by the same articulator. /cv/ /c/ – Coronal /v/ – Labial /cv/ is a possible complex sound (/cv/) with different articulators. /cĢ/ /c/ – Coronal [+anterior] /Ģ/ – Coronal [-anterior] /cĢ/ cannot be a complex sound because of the conflicting gestures by the same articulator.

190

Appendix 4

/cm/ /c/ – Coronal, Soft Palate [-nasal] /m/ – Coronal, Soft Palate [+nasal] /cm/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. /cn/ /c/ – Coronal [+anterior], Soft Palate [-nasal] /n/ – Coronal [+anterior], Soft palate [+nasal] /cn/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. /cĖ/ /c/ – Coronal [+anterior], Soft Palate [-nasal] /Ė/ – Coronal [-anterior], Soft Palate [+nasal] /cĖ/ cannot be a complex sound, there are the conflicting gestures by two articulators. /cl/ /c/ – Coronal [+anterior], [-lateral] /l/ – Coronal [+anterior], [+lateral] /cl/ cannot be a complex sound because of the conflicting values of the [lateral] feature. /cĐ/ /c/ – Coronal [+anterior], [-lateral] /Đ/ – Coronal [-anterior], [+lateral] /cĐ/ cannot be a complex sound because of the conflicting values of the [anterior] and [lateral] features. /þp/ /þ/ – Coronal /p/ – Labial /þp/ can be a complex sound (/þp/); there are different articulators. /þv/ /þ/ – Coronal /v/ – Labial /þv/ can be a complex sound (/þv/) with different articulators.

The Generative and the Structuralist Approach to the Syllable

191

/þm/ /þ/ – Coronal, Soft Palate [-nasal] /m/ – Coronal, Soft Palate [+nasal] /þm/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. /þn/ /þ/ – Coronal [+anterior], Soft Palate [-nasal] /n/ – Coronal [+anterior], Soft Palate [+nasal] /þn/ cannot be a complex sound because of the conflicting gestures by the Soft Palate. /þl/ /þ/ – Coronal [+anterior], [-lateral] /l/ – Coronal [+anterior], [+lateral] /þl/ cannot be a complex sound because of the conflicting values of the [lateral] feature. /þĐ/ /þ/ – Coronal [+anterior], [-lateral] /Đ/ – Coronal [-anterior], [+lateral] /þĐ/ cannot be a complex sound because of the conflicting values of the [anterior] and [lateral] features. /þr/ /þ/ – Coronal [+anterior], [-sonorant], [-continuant] /r/ – Coronal [+anterior], [+sonorant], [+continuant] /þr/ cannot be a complex sound; there are the conflicting stricture gestures by the same articulator. A complex-sound analysis of the Slovak word-initial CC sequences that come into existence after the application of the morphological rules of the CVX theory but do not occur as independent CC clusters: /pĐ/ /p/ – Labial /Đ/ – Coronal [-anterior], [+lateral] /pĐ/ can be a complex sound (/pĐ/); there are different articulators.

192

Appendix 4

/pr/ /p/ – Labial /r/ – Coronal /pr/ is a possible complex sound produced by different articulators. A complex-sound analysis of the Slovak word-final CC clusters, which do not occur word-initially3: /ks/ /k/ – Dorsal /s/ – Coronal /ks/ is a possible complex sound (/ks/) involving a different articulator. /št/ /š/ – Coronal [+anterior] /t/ – Coronal [-anterior] /št/ cannot be a complex sound because there are conflicting gestures by the same articulator. /šĢ/ /š/ – Coronal [+anterior] /Ģ/ – Coronal [-anterior] /šĢ/ cannot be a complex sound because of the conflicting gestures by the same articulator. /xt/ /x/ – Dorsal /t/ – Coronal /xt/ is a good complex sound with different articulators. /rc/ /r/ – Coronal [+anterior], [+sonorant], [+continuant] /c/ – Coronal [+anterior], [-sonorant], [-continuant] /rc/ cannot be a complex sound; there are conflicting stricture gestures by the same articulator. 3

Duanmu sees a complex sound as a gestural merge or overlap of at least two sounds (section 2.2.2). This means that if the word-initial cluster /tr/ can be represented as a complex segment, its mirror cluster, cluster /rt/, can be a complex segment, too, since overlapping gestures are made simulaneously. That is why the word-final CC clusters, which are the ‘mirror‘ representations of the word-inital CC clusters, are not analysed separately here.

The Generative and the Structuralist Approach to the Syllable

193

/sr/ /s/ – Coronal [-anterior], [-sonorant], [+continuant] /r/ – Coronal [+anterior], [+sonorant], [+continuant] /sr/ is not a possible complex sound due to the conflicting gestures by the same articulator. /jt/ /j/ – Coronal [-anterior] /t/ – Coronal [+anterior] /jt/ cannot be a complex sound because of the conflicting values of the same feature. /jk/ /j/ – Coronal /k/ – Dorsal /kj/ can be a complex sound (/kj/); there are two different articulators. /jf/ /j/ – Coronal /f/ – Labial /jf/ is a possible complex segment (/jf/) with different articulators. /jn/ /j/ – Coronal [-anterior], Soft Palate [-nasal] /n/ – Coronal [+anterior], Soft Palate [+nasal] /jn/ is not a good complex sound, there are conflicting features by two articulators. /zd/ /z/ – Coronal [+anterior], [+fricative], [-sonorant], [+continuant] /d/ – Coronal [+ anterior], [+stop], [-sonorant], [-continunat] /zd/ cannot be a complex sound because of the conflicting stricture gestures. /zć/ /z/ – Coronal [+anterior], [+fricative], [-sonorant], [+continuant] /ć/ – Coronal [-anterior], [+stop], [-sonorant], [-continuant] /zć/ is not a possible complex segment; there are the conflicting values of several features.

194

Appendix 4

/žd/ /ž/ – Coronal [+anterior], [+fricative] /d/ – Coronal [+ anterior], [+stop] /žd/ can be treated as a special type of a complex segment–affricate– characterized by the simultaneous features [+stop], [+fricative] with the Coronal articulator. /žć/ /ž/ – Coronal [+anterior], [+fricative], [-sonorant], [+continuant] /ć/ – Coronal [-anterior], [+stop] [-sonorant], [-continuant] /žć/ is not a good complex sound; there are the conflicting gestures. /jd/ /j/ – Coronal [-anterior] /d/ – Coronal [+anterior] /jd/ cannot be a complex sound because of the conflicting values of the same feature.

APPENDIX 5 THE LIST OF THE WORD-MEDIAL CONSONANT CLUSTERS

1. The most common word-medial consonant clusters in English1 Word-medial CC Clusters (128) /nd, nt, mp, mb, ƾg, st, ks, ns, ƾk, nj, sk, kt, lj, kj, nv, lk, lt, mj, gm, ts, sp, tr, kn, zm, tj, lm, nz, nf, zl, gn, lb, ls, gj, pj, lv, lf, bj, br, fr, kr, mz, mf, pt, dl, ft, sj, fl, kߦ, ml, mn, nm, ms, gl, pr, ps, tl, vr, lg, lp, nș, dn, fj, sm, kw, km, kl, kd, zj, zb, zn, gz, gr, pn, tn, tf, vj, lw, ln, l‫ݶ‬, nࠔ, nh, nߦ, bl, b‫ݶ‬, bs, dm, dj, sf, kb, mk, mr, mt, mߦ, md, z‫ݶ‬, gw, gb, pk, pm, pl, tw, tk, tm, vl, lࠔ, lr, lߦ, ߦm, ߦn, ߦr, ߦt, nl, șm, șl, șj, bn, dp, dr, dv, ðm, fg, fn, fș, sw, sl, sn, sș, sb/ Word-medial CCC Clusters (45) /njh, ndr, ntr, str, mbr, ƾkw, kst, mpr, mpl, ƾkt, nfr, ƾkr, nst, mpt, lst, ltr, ƾgw2, ƾgr, ƾkl, ksw, ksp, ktr, mdr, mfr, mkw, mbl, mfl, ƾgh, ƾgt, ƾgs, mpk, mps, pst, nࠔb, ldr, lfr, lgr, lpr, nsl, ntl, nsk, nșr, ntm, dzw, spr/

1

The list of the medial consonant clusters is not completely exhaustive. Only those clusters that are considered the most common, i.e. the most frequent, in monomorphemic words are mentioned here. See also note 15 in Chapter 2. 2 The sequence of graphemes ‘ng’ is pronounced as /[ƾ]/ at the end of the morpheme. When it occurs inside the morpheme the pronunciation is /[ƾg]/ (Roach 2000, 59). The presented list of the English medial consonant clusters is based on Gorman’s analysis of monomorphemic words (Gorman 2013, 71). That is why both /ƾ/ and /g/ are included in the cluster which is then treated as the threeconsonant cluster. The delimitation of medial clusters in monomorphemic words is in accordance with the delimitation of initial and final clusters in both analysed languages (see also section 2.7.1).

196

Appendix 5

Word-medial CCCC Clusters (6) /nstr, mstr, kstr, ƾkst, nskr, kskl (Glowacka 2001; Gorman 2013) 2. The most common word-medial consonant clusters in Slovak Word-medial CC clusters (346) /pt, pĢ, pk, pc, pþ, ps, pš, px, pm, pn, pĖ, pl, pĐ, pr, bd, bć, bz, bž, bh, bv, bm, bn, bĖ, bl, bĐ, br, tp, tĢ, tk, ts, tš, tx, tf, tv, tm, tn, tĖ, tl, tĐ, tr, db, dć, dg, dh, dv, dm, dn, dĖ, dl, dĐ, dr, dj, Ģp, Ģt, ĢĢ, Ģk, Ģc, Ģþ, Ģs, Ģš, Ģm, ćb, dd, dć, ćh, ćm, ćn, ćl, ćr, kp, kt, kĢ, kk, kc, kþ, ks, kš, kv, km, kn, kĖ, kr, gd, gm, gć, gz, gž, gb, gh, gv, gn, gĖ, gl, gĐ, gr, cp, cc, cþ, cs, cx, cf, cv, cm, cn, cĖ, cl, cĐ, þp, þt, ck, þĢ, þk, þþ, þv, þm, þn, þĖ, þl, þĐ, þr, ‫ݤ‬b, ‫ݤ‬d, ‫ݤ‬ć, ‫ݤ‬ž, ‫ݤ‬h, ‫ݤ‬v, ‫ݤ‬m, ‫ݤ‬n, ‫ݤ‬Ė, ‫ݤ‬Đ, ‫ݤ‬r, ‫ݤ‬j, ۤb, ۤg, ۤh, ۤr, sp, st, sĢ, sk, sc, sþ, ss, šc, sš, sf, sv, sm, sn, sĖ, sl, sĐ, sr, zb, zd, zć, zg, zz, zž, zh, zv, zm, zn, zĖ, zl, zĐ, zj, šp, št, šĢ, šk, þc, šš, šv, šm, šn, šĖ, šl, šĐ, šr, žb, žd, žć, žv, žm, žn, žĖ, žĐ, žr, hd, hž, hv, hm, hn, hĖ, hl, hĐ, hr, xt, xĢ, xk, xc, xþ, xš, xv, xm, xn, xĖ, xl, xĐ, xr, fp, ft, fĢ, fk, fš, fĐ, fn, fĖ, fl, fr, vb, vd, vć, vk, vc, vþ, vۤ, vz, vh, vn, vĖ, vl, vĐ, vr, mp, mh, bx, mf, mv, mm, mn, mĖ, ml, mĐ, mr, nt, nĢ, nd, nć, nk, ng, nc, nþ, n‫ݤ‬, nۤ, ns, nš, nz, nž, nh, nx, nf, nv, nm, nn, nĖ, nĐ, nr, nj, Ėb, Ėh, Ėm, Ėn, lp, lb, lt, ld, lć, lk, lg, lc, lþ, lš, lz, lh, lx, lf, lv, lm, ln, ll, lr, Đb, Đt, Đd, Đć, Đk, Đþ, Đš, Đž, Đh, Đv, Đn, ĐĖ, Đr, Đj, rp, rb, rz, rž, rh, rx, rf, rv, rm, rn, rĖ, rl, rĐ, rr, rj, jp, jb, jt, jd, jĢ, jć, jk, jg, jc, jþ, js, jz, jš, jh, jx, jf, jv, jm, jn, jĖ, jl, jĐ, jr, jj/ Word-medial CCC clusters (250) /str, nsk, stn, stĖ, rsk, spr, vsk, zdr, ctv, jsk, stl, stĐ, zdn, zdĖ, stv, skr, skl, ndr, lsk, šsk, Đsk, jtr, stk, vzd, nkr, dvr, fst, fšĢ, zdv, bzr, skv, tkl, tkĐ, msk, ntr, psk, pst, tpr, mst, msĢ, jst, cst, škr, štn, štĖ, ntn, ntĖ, zhl, skĐ, vžd, ktr, ždĖ, tpl, zbr, skn, skĖ, tkn, tkĖ, jkr, Ģkr, tkr, zvr, mbr, pkl, spl, tkv, zćr, nsl, zhĐ, ntk, tpĐ, rkv, sxn, sxĖ, mdĐ, ršt, sxl, jft, jfĢ, ngl, psl, psĐ, rtv, dhr, nšĢ, vzb, jpr, jtm, nšp, nšt, zvl, þsk, jbl, jhl, jsl, mkl, nkc, zhr, dhl, dhĐ, mkn, mpl, mpĐ, ndĐ, ngĐ, sĢp, šsĢ, tfþ, zml, zmn, gzv, jsm, jšk, rsp, sxv, spt, zhĖ, dzĖ, jml, lsĢ, nbĐ, nþn, ngr, nsp, pkr, rtr, sxĐ, stm, štv, vzĐ, þšm, fpl, fsp, jšĢ, ldr, ndn, nkn, nkĖ, mšt, mšĢ, ndl, nkt, rkl, rtn, štr, txn, txĖ, vzn, vzĖ, vzr, zmĐ, ckt, dbl, dbĐ, dvl, dvĐ, fkr, hmĐ, jsv, nfl, ntn, ntĖ, pšĢ, rkn, rkĖ, škĐ, škv, tsĢ, bdr, jhn, jsx, jvh, jvl, ktn, ktĖ, lkl, ltr, mkĐ, pxl, rhr, rkr, sxr, skm, spĐ, sšĢ, tsv, ttl, tšk, vzć, bhr, ckr, dbr, dvć, dzn, ‫ݤ‬gć, fpĐ, fsĢ, xsĐ, xtl, xtĐ, jbr, jdr, jfš, jhr, jpl, jsp, jvć, jvn, jvĖ, jvz, jzh, kþn, ldĖ, lkn, lkĖ, lkr, mbl, mbĐ, mdl, mdĐ, mhĐ, mpk, mpr, mpt, nþĢ, ngv, nkl, pxn, pxĖ, rdn,

The Generative and the Structuralist Approach to the Syllable

197

rdĖ, rgr, rkm, rpr, rst, rĢĐ, smr, špĐ, tfš, txl, txĐ, txv, tsl, tsp, tšm, tšt, tšĢ, zbz, zgl, zgĐ/ Word-medial CCCC clusters (26) /nstv, rstv, Đstv, mstv, šstv, nštr, lstv, cstv, jstr, vstv, tstr, jstv, þstv, pstr, jstn, pstv, jzdr, dzbr, tškr, fskĐ, rnsk, jskr, kstr, nkþn, kspl, rstr/ Word-medial CCCCC cluster (1) /rvstv/ (Sabol 1969a; 1969b; 1975)

APPENDIX 6 A SONORITY-BASED ANALYSIS OF THE WORDMEDIAL CONSONANT CLUSTERS

1. The sonority-based analysis of the word-medial consonant clusters in English 1.1. the MSD of the word-medial CC clusters (128): MSD -7 liquid (r, l) + voiceless stop (p, t, k): lk, lt, lp (3) MSD -6 liquid (r, l) + voiced stop (b, d, g): lb, lg (2) nasal (m, n, ‫ )و‬+ voiceless stop (p, t, k): nt, mp, ƾk1, mt, mk, np (6) MSD -5 nasal (m, n, ‫ )و‬+ voiced stop (b, d, g): nd, mb, ƾg2, nb (4) liquids (r, l) + voiceless affricate (ࠔ): Oࠔ MSD -4 nasal (m, n, ‫ )و‬+ voiceless affricate (ࠔ): Qࠔ voiceless fricative (f, ș, s,ߦ) + voiceless stop (p, t, k): st, sk, sp, ft, ߦt(5) voiced fricative (v, ð, z, ‫ݤ‬, h) + voiced stop (b, d, g): zb (1) MSD -3 voiceless fricative (f, ș, s, ߦ) + voiced stop (b, d, g): fg, sb (2) liquids (r, l) + voiceless fricative (f, ș, s, ߦ): ls, lf, lߦ (3) 1

Even if the consonant cluster ƾk (and ƾg) fulfilled the MSD for English, it would not be possible to place the cluster in the syllable onset due to the phonotactic constrains of the English language. 2 See note 2 in Appendix 5.

The Generative and the Structuralist Approach to the Syllable

199

MSD -2 nasal (m, n, ƾ) + voiceless fricative (f, ș, s, ‫) ݕ‬: nf, mf, ms, nș, mߦ, nߦ, ns (7) voiced fricative (v, ð, z, ‫ݤ‬, h) + voiced affricate (˔): z˔(1) liquids (r, l) + voiced fricative (v, ð, z, ‫ݤ‬, h): lv (1) MSD -1 nasal (m, n, ƾ) + voiced fricative (v, ð, z, ‫ݤ‬, h): mh, mv, nv (3) voiced stop (b, d, g) + voiceless stop (p, t, k): dp (1) liquids (r, l) + nasals (m, n, ƾ): lm, ln (2) MSD 0 nasal (m, n, ƾ) + nasal (m, n, ƾ): mn, nm (2) voiceless stop (p, t, k) + voiceless stop (p, t, k): kt, pt, pk, tk (4) voiced stop (b, d, g) + voiced stop (b, d, g): gb (1) voiceless fricative (f, ș, s, ߦ) + voiceless fricative (f, ș, s, ߦ): sf, fș, sș (3) liquids (r, l) + liquids (r, l): lr (1) MSD 1 nasal (m, n, ƾ) + liquid (r, l): ml, mr, nl (3) voiceless stop (p, t, k) + voiced stop (b, d, g): kd, kb (2) voiced fricative (v, ð, z, ‫ݤ‬, h) + nasals (m, n, ƾ): zm, zn, ðm (3) liquids (r, l) + semivowel (j, w): lj, lw (2) MSD 2 nasal (m, n, ƾ) + semivowel (j, w): nj, mj (2) voiced stop (b, d, g) + voiced affricate (‫)ݶ‬: b‫( ݶ‬1) voiceless fricative (f, ș, s, ߦ) + nasals (m, n, ƾ): sm, ߦm, ߦn, șm, fn, sn (6) voiced fricative (v, ð, z, ‫ݤ‬, h) + liquids (r, l): zl, vr, vl (3) MSD 3 voiced stop (b, d, g) + voiceless fricative (f, ș, s, ߦ): bs (1) voiceless fricative (f, ș, s, ߦ) + liquids (r, l): fr, fl, ‫ݕ‬r, șl, sl (5) voiced fricative (v, ð, z, ‫ݤ‬, h) + semivowel (j, w): zj, vj (2) MSD 4 voiceless stop (p, t, k) + voiceless fricative (f, ș, s, ߦ): kߦ, ps, tf, ks, ts (5) voiced stop (b, d, g) + voiced fricative (v, ð, z, ‫ݤ‬, h): gz, dv (2) voiceless fricative (f, ș, s, ߦ) + semivowel (j, w): sj, fj, șj, sw (4)

200

Appendix 6

MSD 5 voiced stop (b, d, g) + nasals (m, n, ƾ): gm, gn, dn, dm, bn (5) MSD 6 voiceless stop (p, t, k) + nasals (m, n, ƾ): km, pn, tn, pm, tm (5) voiced stop (b, d, g) + liquids (r, l): dl, gl, gr, bl, dr, br (6) MSD 7 voiceless stop (p, t, k) + liquids (r, l): tl, pr, tr, kl, pl (5) voiced stop (b, d, g) + semivowel (j, w): gj, bj, dj, gw (4) MSD 8 voiceless stop (p, t, k) + semivowel (j, w): kj, pj, kw, tw, tj (5) 1.3 Word-medial CCC Clusters (45) 偪 clusters requiring the syllabification C1C2.C3 because of the phonotactic rules:

ƾkw, ƾkt, ƾgw, ƾgr, ƾkl, ƾgh, ƾgt, ƾgs 偪 the MSD of the C2C3 clusters when the syllabification is C1.C2C3

n.jh/ -3, n.dr/ 6, n.tr/ 7, s.tr/ 7, m.br/ 6, k.st/ -43, m.pr/ 7, m.pl/ 7, n.fr/ 3, n.kr/ 7, n.st/ -4, m.pt/ 0, l.st/ -4, l.tr/ 7, k.sw/ 4, k.sp/ -4, k.tr/ 7, m.dr/ 6, m.fr/ 3, m.kw/ 8, m.bl/ 6, m.fl/ 3, m. pk/ 0, m.ps/ -4, p.st/ -4, n.‫ݹ‬b/ 1, l.dr/ 6, l.fr/ 3, l.gr/ 6, l.pr/ 7, n.sl/ 3, n.tl/ 7, n.sk/ -4, n.șr/ 3, n.tm/ 6, d.zw/ 3, s.pr/ 7 1.3 Word-medial CCCC Clusters (6) 偪 the sonority-based analysis of the initial C3C4 clusters when the syllabification is C1C2.C3C4 o the MSD for each C3C4 cluster is in brackets following the cluster,

3

Text box indicates clusters whose two-consonant part does not fulfil the MSD requirement specified for English.

The Generative and the Structuralist Approach to the Syllable

201

o (Y) in brackets after the final cluster means that the cluster follows the SSG, (N) in brackets after the final cluster indicates that the SSG is violated ns(Y).tr(7) ms(Y).tr(7) ks(N).tr(7) ƾk(Y).st (-4)4 ns(Y).kr(7) ks(N).kl(7) 2. The sonority-based analysis of the word-medial consonant clusters in Slovak 2.1 Word-medial CCC Clusters (250) 偪 the MSD of the C2C3 clusters when the syllabification is C1.C2C3

s.tr/ 8, n.sk/ -45, s.tn/ 7, s.tĖ/ 6, r.sk/ -4, s.pr/ 8, v.sk/ -4, z.dr/ 7, c.tv/ 5, j.sk/ -4, s.tl/ 8, s.tĐ/ 6, z.dn/ 6, z.dĖ/ 5, s.tv/ 5, s.kr/ 8, s.kl/ 8, n.dr/ 7, l.sk/ 4, š.sk/ -4, Đ.sk/ -4, j.tr/ 8, s.tk/ 0, v.zd/ -46, n.kr/ 8, d.vr/ 3, f.st/ -4, f.šĢ/ -4, z.dv/ 4, b.zr/ 3, s.kv/ 5, t.kl/ 8, t.kĐ/ 6, m.sk/ -4, n.tr/ 8, p.sk/ -4, p.st/ -4, t.pr/ 8, m.st/ -4, m.sĢ/ -4, j.st/ -4, c.st/ -4, š.kr/ 8, š.tn/ 7, š.tĖ/ 6, n.tn/ 7, n.tĖ/ 6, z.hl/ 3, s.kĐ/ 6, v.žd/ -4, k.tr/ 8, ž.dĖ/ 5, t.pl/ 8, z.br/ 7, s.kn/ 7, s.kĖ/ 6, t.kn/ 7, t.kĖ/ 6, j.kr/ 8, Ģ.kr/ 8, t.kr/ 8, z.vr/ 3, m.br/ 7, p.kl/ 8, s.pl/ 8, t.kv/ 5, z.ćr/ 7, n.sl/ 4, z.hĐ/ 1, n.tk/ 0, t.pĐ/ 6, r.kv/ 5, s.xn/ 3, s.xĖ/ 2, m.dĐ/ 5, r.št/ -3, s.xl/ 4, j.ft/ -4, j.fĢ/ -4, n.gl/ 7, p.sl/ 4, p.sĐ/ 2, r.tv/ 5, d.hr/ 3, n.šĢ/ -4, v.zb/ -3, j.pr/ 8, j.tm/ 7, n.šp/ -4, n.št/ -4, z.vl/ 3, þ.sk/ -4, j.bl/ 7, j.hl/ 3, j.sl/ 4, m.kl/ 8, n.kc/ 2, z.hr/ 3, d.hl/ 3, d.hĐ/ 1, m.kn/ 7, m.pl/ 8, m.pĐ/ 6, n.dĐ/ 5, n.gĐ/ 5, s.Ģp/ 0, š.sĢ/ -4, t.fþ/ -2, z.ml/ 1, z.mn/ 0, g.zv/ 0, j.sm/ 3, j.šk/ -4, r.sp/ -4, s.xv/ 1, s.pt/ 0, z.hĖ/ 1, d.zĖ/ 1, j.ml/ 1, l.sĢ/ -4, n.bĐ/ 5, n.þn/ 5 , n.gr/ 7, n.sp/ -4, p.kr/ 8, r.tr/ 8, s.xĐ/ 2, s.tm/ 7, š.tv/ 5, v.zĐ/ 1, þ.šm/ 3, f.pl/ 8, f.sp/ -4, j.šĢ/ -4, l.dr/ 7, n.dn/ 6, n.kn/ 7, n.kĖ/ 6, m.št/ -4, m.šĢ/ -4, n.dl/ 7, n.kt/ 0, r.kl/ 8, r.tn/ 7, š.tr/ 8, t.xn/ 3, t.xĖ/ 1, v.zn/ 2, v.zĖ/ 1, v.zr/ 3, z.ml/ 1, z.mĐ/ -1, c.kt/ 0, d.bl/ 7, d.bĐ/ 5, d.vl/ 3, d.vĐ/ 1, f.kr/ 8, h.mĐ/ -1, j.sv/ 1, n.fl/ 8, n.tn/ 7, n.tĖ/ 6, p.šĢ/ -4, r.kn/ 7, r.kĖ/ 6, 4

See note 2 in Appendix 5. Text box indicates the clusters with the consonant /[s]/, whose exceptional behavior is well-known in many languages (see, e.g., section 3.2.2.1). 6 The clusters that do not follow the principle of the SSG are underlined. 5

202

Appendix 6

š.kĐ/ 6, š.kv/ 6, t.sĢ/ -4, b.dr/ 7, j.hn/ 2, j.sx/ 0, j.vh/ 0, j.vl/ 3, k.tn/ 7, k.tĖ/ 6, l.kl/ 8, l.tr/ 8, m.kl/ 8, m.kĐ/ 6, p.xl/ 2, r.hr/ 3, r.kr/ 8, s.xr/ 4, s.km/ 7, s.pĐ/ 6, s.šĢ/ -4, t.sv/ 1, t.tl/ 8, t.šk/ -4, v.zć/ -4, , b.hr/ 3, c.kr/ 8, d.br/ 7, d.vć/ -4, d.zn/ 2, ঱gć/ 0, f.pĐ/ 6, f.sĢ/ -4, x.sĐ/ 2, x.tl/ 8, x.tĐ/ 6, j.br/ 7, j.dr/ 7, j.fš/ 0, j.hr/ 3, j.pl/ 8, j.sp/ -4, j.vć/ -4, j.vn/ 2, j.vĖ/ 1, j.vz/ 0, j.zh/ 0, k.þn/ 5, l.dĖ/ 5, l.kn/ 7, l.kĖ/ 6, l.kr/ 8, m.bl/ 7, m.bĐ/ 5, m.dl/ 7, m.dĐ/ 5, m.hĐ/ 1, m.pk/ 0, m.pr/ 8, m.pt/ 0, n.þĢ/ -2, n.gv/ 4, n.kl/ 8, p.xn/ 3, p.xĖ/ 2, r.dn/ 6, r.dĖ/ 5, r.gr/ 7, r.km/ 7, r.pr/ 8, r.st/ -4, r.ĢĐ/ 6, s.mr/ 1, t.fš/ 0, t.xl/ 4, t.xĐ/ 2, t.xv/ 1, t.sl/ 4, t.sp/ -4, t.šm/ 3, t.št/ -4, t.šĢ/ -4, z.bz/ 4, z.gl/ 7, z.gĐ/ 5

Word-medial CCCC Clusters (26) 偪 the sonority-based analysis of the initial and the final CC clusters when the syllabification is C1C2.C3C4 o the MSD for each initial C3C4 cluster is in brackets following the cluster, o (Y) in brackets after the final cluster means that the cluster follows the SSG, (N) in brackets after the final cluster indicates that the SSG is violated. The clusters with (N) are in text box.

ns(Y).tv(5), rs(Y).tv(5), Đs(Y).tv(5), ms(Y).tv(5), šs(Y).tv(5), nš(Y).tr(8), ls(Y).tv(5), cs(N).tv(5), js(Y).tr(8), vs(Y).tv(5), ts(N).tr(8), js(Y).tv(5), þs(N).tv(5), ps(N).tr(8), js(Y).tn(7), ps(N).tv(5), jz(Y).dr(7), dz(N).br(7), tš(N).kr(8), fs(Y).kĐ(6), rn(Y).sk(-4), js(Y).kr(8), ks(Y).tr(8), nk(Y).þn(5), ks(N).pl(8), rs(Y).tr(8)

Word-medial CCCCC Cluster (1) rv.stv

APPENDIX 7 THE SYLLABIFICATION OF THE WORD-MEDIAL CONSONANT CLUSTERS ACCORDING TO THE PHONOTACTIC RULES

1. Slovak 1.1 Word-medial CC clusters that occur also word-initially (35) /ps, pš, px, pn, pĖ, pl, bć, bz, bl, bĐ, br, tk, tx, tv, tm, tl, tĐ, tr, dv, dm, dn, dĖ, dl, dr, kt, kv, km, kn, kĖ, kr, gn, gĖ, gl, gĐ, gr/ 1.2.1 Word-medial CCC clusters that can be syllabified as C1C2.C3 (90) /psk, pst, psl, psĐ, ktr, ktn, ktĖ, mpl, mpĐ, mpk, mpr, mpt, mbr, mbl, mbĐ, ndr, ndĐ, ndn, ndl, nkr, nkc, nkn, nkĖ, nkt, nkl, ngl, ngĐ, ngr, ngv, nþn, nþĢ, nšĢ, nšp, nšt, str, stn, stĖ, stl, stĐ, stv, stk, stm, zdr, zdn, zdĖ, zdv, sĢp, zćr, skr, skl, skv, skĐ, skn, skĖ, skm, štn, štĖ, štv, ždĖ, xtl, xtĐ, ltr, ldr, ldĖ, lkt, lkn, lkĖ, lkr, lsk, lsĢ, rpr, rtv, rtr, rtn, rdn, rdĖ, rkv, rkl, rkn, rkĖ, rkr, rkm, rsk, rsp, rst, ršt, jtr, jtm, jdr, jkr/ 1.2.2 Word-medial CCC clusters that can be syllabified as C1.C2C3 (200) /tpl, spl, mpl, fpl, jpl, jbl, dbl, mbl, nbĐ, dbĐ, mbĐ, zbr, mbr, dbr, jbr, stk, ntk, ctv, stv, rtv, štv, jtm, stm, stl, ttl, xtl, stĐ, xtĐ, str, jtr, ntr, ktr, rtr, štr, ltr, zdv, zdm, ndn, rdn, zdĖ, ždĖ, ldĖ, rdĖ, ndl, mdl, zdr, ndr, ldr, bdr, jdr, nkt, ckt, skv, tkv, rkv, škv, skm, rkm, skn, tkn, mkn, nkn, rkn, lkn, skĖ, tkĖ, nkĖ, rkĖ, lkĖ, skr, nkr, škr, jkr, Ģkr, tkr, pkr, fkr, rkr, ckr, lkr, skl, tkl, pkl, mkl, rkl, lkl, nkl, tkĐ, skĐ, škĐ, mkĐ, ngl, zgl, ngĐ, zgĐ, ngr, rgr, zmn, zml, jml, zmĐ, hmĐ, smr, rsp, nsp, fsp, jsp, tsp, fst, pst, mst, jst, cst, rst, msĢ, šsĢ, lsĢ, tsĢ, fsĢ, nsk, rsk, vsk, jsk, lsk, šsk, Đsk, msk, psk, þsk, jsx, jsv, tsv, jsm, nsl, psl, jsl, tsl, psĐ, xsĐ, vzb, gzv, vzn, dzn, vzĖ, dzĖ, bzr, vzr, nšp, ršt, nšt, mšt, tšt, fšĢ, nšĢ, jšĢ, mšĢ, pšĢ, sšĢ, tšĢ, jšk, tšk, þšm, tšm, jhn, zhĖ, zhl, jhl, dhl, zhĐ, dhĐ, mhĐ, dhr, zhr, rhr, bhr, jhr, sxv, txv, sxl, pxl, txl, sxĐ, txĐ, sxr, jft, jfĢ, tfþ, jfš, tfš, nfl, jvz, jvn, jvĖ, zvl, dvl, jvl, dvr, zvr, nþn, kþn/

204

Appendix 7

1.2.3 Word-medial CCC clusters that can be syllabified either as C1C2.C3 or as C1.C2C3 (68) /psk, pst, psĐ, ktr, mpl, mbr, mbl, mbĐ, ndr, ndn, ndl, nkr, nkn, nkĖ, nkt, nkl, ngl, ngĐ, ngr, nþn, nšĢ, nšp, nšt, str, stl, stĐ, stv, stk, stm, zdr, zdĖ, zdv, skr, skl, skĐ, skn, skĖ, skm, štv, ždĖ, xtl, xtĐ, ltr, ldr, ldĖ, lkn, lkĖ, lkr, lsk, lsĢ, rtv, rtr, rdn, rdĖ, rkv, rkl, rkn, rkĖ, rkr, rkm, rsk, rsp, rst, ršt, jtr, jtm, jdr, jkr/ 1.2.4 Word-medial CCC clusters that can be syllabified as V.C1C2C3 (19) /tkv, mdl, str, skl, skĐ, skv, smr, stv, stl, zdr, zvl, zbr, zhl, škr, štv, škr, štr, špĐ, hmĐ/

2. English 2.1 Word-medial CC clusters that occur also word-initially (40) /pr, pl, pj, ps, br, bl, bj, tr, tw, tj, dr, dj, kr, kl, kw, kj, km, kn, gr, gl, gw, mj, fl, fr, fj, vj, ۟j, st, sp, sk, sl, sw, sn, sm, sf, sj, zl, ߦr, ߦm, ߦn/ 2.2.1 Word-medial CCC clusters that can be syllabified as C1C2.C3 (36) /ktr, mdr, ntr, ntl, ntm, ndr, str, ltr, ldr, pst, dzw, kst, ksw, ksp, nst, nsl, nsk, lst, nșr, mpr, mpl, mpt, mpk, mps, n‫ݹ‬b, ƾkw, ƾkt, ƾgw, ƾkr, ƾgr, ƾkl, ƾgh, ƾgt, ƾgs, spr, lpr, lfr/ 2.2.2 Word-medial CCC clusters that can be syllabified as C1.C2C3 (34) /mpr, lpr, spr, mpl, mps, mbr, mbl, ntr, str, ltr, ktr, ndr, mdr, ldr, ƾkl, ƾkr, ƾkw, mkw, ƾgr, lgr, ƾgw, mfl, nfr, mfr, lfr, nșr, kst, nst, lst, pst, ksp, nsk, nsl, ksw/ 2.2.3 Word-medial CCC clusters that can be syllabified either as C1C2.C3 or as C1.C2C3 (26) /mpr, lpr, spr, mpl, mps, ntr, str, ltr, ktr, ndr, mdr, ldr, ƾkl, ƾkr, ƾkw, ƾgr, ƾgw, nșr, kst, nst, lst, pst, ksp, nsk, nsl, ksw/ 2.2.4 Word-medial CCC clusters that can be syllabified as V.C1C2C3 /spr, str/

APPENDIX 8 THE SYLLABIFICATION BASED ON THE POWER OF SYLLABIC WELDS

1. The Slovak word-medial CC clusters (347): 1. 1 The Slovak CC clusters that have the identical distinctive features (cf. Table 2-5) and should be syllabified as C1.C2 (13) ĢĢ, dd, kk, cc, þþ, ss, zz, šš, mm, nn, ll, rr, jj 1.2 The Slovak CC clusters that have one or two different distinctive features (cf. Table 2-5) and should be syllabified as C.C (107) pc(2)1, pþ(2), pt(1), pĢ(1), pk(1), bd(1), bć(1), bm(2), bv(2), tp(1), tĢ(1), tk(1), ts(2), tš(2), db(1), dć(1), dg(1), dn(2), Ģp(1), Ģt(1), Ģk(2), Ģþ(2), ćb(1), dć(1), kc(2), kþ(2), kp(1), kt(1), kĢ(2), gd(1), gć(1), gb(1), cp(2), cþ(1), cs(2), ck(2), þk(2), þp(2), þt(1), þĢ(2), ‫ݤ‬b(2), ‫ݤ‬d(1), ‫ݤ‬ć(2), ‫ݤ‬ž(2), ۤb(2), ۤg(2), st(2), sc(2), sþ(2), šc(2), sš(1), sv(2), sf(1), zh(2), zv(1), zd(2), zl(2), zž(1), šv(2), št(2), šr(2), žd(2), žv(1), žr(2), þc(1), hž(2), hv(2), hl(2), hĐ(1),hr(1), xr(2), xš(2), fp(2), fš(1), vb(2), vz(1), vh(2), vr(2), mn(1), mĖ(1), nd(2), nm(1), nĖ(1), Ėm(1), Ėn(1), lz(2), lh(2), lr(1), Đr(2), Đh(2), Đj(1), rv(2), rz(2), rž(2), rl(1), rĐ(2), rh(1), rx(2), rj(1), jl(2), jĐ(1), jr(1), jz(2),jh(1), jx(2), jv(2) 1.3 The Slovak CC clusters that have more than two different distinctive features (cf. Table 2-5) and can together be in the onset of the following syllable (226) ps(3)2, pš(3), px(4), pm(3), pn(4), pĖ(4), pl(6), pĐ(6), pr(5), bn(3), bĖ(3), bz(3), bž(3), bh(4), bl(5), bĐ(4), br(5), tx(4), tf(3), tv(4), tm(4), tn(3), tĖ(4), tl(5), tĐ(6), tr(4), dh(4), dv(3), dm(3), dĖ(3), dl(4), dĐ(5), dr(4), dj(4), Ģs(3), 1 2

The number of different distinctive features is in brackets after the cluster. See note 1.

206

Appendix 8

Ģš(3), Ģm(4), ćh(4), ćm(3),ćn(3), ćl(5), ćr(5), ks(3), kš(3), kv(4), km(4), kn(4), kĖ(4), kr(6), gm(3), gz(3), gž(3), gh(3), gv(3), gn(3), gĖ(3), gl(4), gĐ(4), gr(4), cx(4), cf(3), cv(4), cm(5), cn(3), cĖ(5), cl(5), cĐ(6), þv(4), þm(5), þn(4), þĖ(5), þl(5), þĐ(6), þr(4), ‫ݤ‬h(4), ‫ݤ‬v(3), ‫ݤ‬m(4), ‫ݤ‬n(3), ‫ݤ‬Ė(4), ‫ݤ‬Đ(5), ‫ݤ‬r(3), ‫ݤ‬j(4), ۤh(4), ۤr(3), sp(3), sk(3), sm(6), sn(5), sĖ(6), sl(3), sĐ(4), sr(3), sĢ(3), zb(3), zć(3), zg(3), zm(5), zn(4), zĖ(5), zĐ(3), zj(3), šp(3), šk(3), šm(6), šn(5), šĖ(6), šl(3), šĐ(4), šĢ(3), žb(3), žć(3), žm(5), žn(4), žĖ(5), žĐ(3), hd(4), hm(4), hn(4), hĖ(4), xt(4), xĢ(4), xc(4), xþ(4), xv(3), xm(5), xn(5), xĖ(5), xk(3), xl(3), xĐ(3), ft(3), fĢ(3), fĐ(3), fn(6), fĖ(6), fl(4), fr(3), fk(3), vd(3), vć(3), vk(4),vn(5), vĖ(5), vc(4), vþ(4), vۤ(3), vl(3), vĐ(3), mv(4), mp(3), mh(4), bx(4), mf(5), ml(5), mĐ(4), mr(4), nt(3), nĢ(5), nć(3), nv(5), nk(4), ng(3), nc(3), nþ(4), n‫(ݤ‬3), nۤ(3), ns(5), nš(4), nz(4), nž(4), nh(4), nx(5), nf(6), nĐ(5), nr(4), nj(4), Ėb(3), Ėh(4), lp(6), lb(5), lt(5), ld(4), lć(4), lk(5), lg(5), lc(5), lþ(5), lš(3), lx(3), lf(4), lv(3), lm(4), ln(4), Đb(4), Đt(6), Đd(5), Đć(4), Đk(5), Đþ(6), Đš(4), Đž(3), Đv(3), Đn(5), ĐĖ(4), rp(5), rb(5), rf(3), rm(4), rn(4), rĖ(4), jp(5), jb(4), jt(5), jd(4), jĢ(4), jć(3), jĖ(3), jk(6), jg(4), jc(5), jþ(5), js(4), jš(3), jf(3), jm(4), jn(4) 2. The Slovak word-medial CCC clusters (250) The number of differences between the neighbouring segments is in brackets after the clusters. For example, ‘str (2-4) – s.tr’ means that there are two different distinctive features between ‘s’ and ‘t’ and five differences between ‘t’ and ‘r’ (see Table 2-5). The syllable boundary should be between ‘s’ and ‘t’ str (2-4) – s.tr nsk (5-3) – ns.k stn (2-3) – s.tn stĖ (2- 5) – s.tĖ rsk (3-3) - r.sk or rs.k3 spr (3-5) – s.pr vsk (2-3) – v.sk zdr (2-4) – z.dr ctv (1-4) – c.tv jsk (4-3) – js.k stl (2-5) – s.tl stĐ (2-6) – s.tĐ 3

zdv (2-3) – z.dv bzr (3-2) – bz.r skv (3-4) – s.kv tkl (1-5) – t.kl tkĐ (1-5) – t.kĐ msk (6-3) – ms.k ntr (3-5) – n.tr psk (3-3) – ps.k or p.sk pst (3-2) – ps.t tpr (1-5) – t.pr mst (6-2) – ms.t msĢ (6-3) – ms.Ģ

In this section, text box indicates the clusters with the identical degree of contrast between the neighbouring segments. In these clusters, there are two possible places for the syllable boundary or three in cases of four-consonant clusters.

The Generative and the Structuralist Approach to the Syllable

zdn (2-2) – z.dn or zd.n zdĖ (2-3) – zd.Ė stv (2-4) – s.tv skr (3-6) – s.kr skl (3-5) – s.kl ndr (2-5) – n.dr lsk (4-3) – ls.k šsk (1-3) – š.sk Đsk (4-3) – Đs.k jtr (6-4) – jt.r stk (2-1) – st.k vzd (1-2) – v-zd nkr (4-6) – n.kr dvr (3-2) – dv.r fst (1-2) – f.st fšĢ (1-3) – f.šĢ tkn (1-4) – t.kn tkĖ (1-4) – t.kĖ jkr (6-6) – j.kr or jk.r Ģkr (2-6) – Ģ.kr tkr (1-6) – t.kr zvr (1-3) – z.vr mbr (2-4) – m.br pkl (1-6) – p.kl spl (3-6) – s.pl tkv (1-4) – t.kv zćr (3-5) – z.ćr nsl (5-3) – ns.l zhĐ (2-2) – z.hĐ or zh.Đ ntk (3-1) – nt.k tpĐ (1-6) – t.pĐ rkv (6-4) – rk.v sxn (2-5) – s.xn sxĖ (2-5) – s.xĖ mdĐ (3-5) – m.dĐ ršt (3-2) – rš.t sxl (2-3) – s.xl jft (3-3) – j.ft or jf.t jfĢ (3-3) – j.ft or j.fĢ ngl (3-5) – n.gl psl (3-3) – p.sl or ps.l

jst (4-2) – js.t cst (2-2) – c.st or cs.t škr (3-6) – š.kr štn (2-3) – š.tn štĖ (2-3) – š.tĖ ntn (3-3) – n.tn or nt.n ntĖ (3-5) – n.tĖ zhl (2-2) – z.hl or zh.l skĐ (3-5) – s.kĐ vžd (1-2) – v.žd ktr (1-5) – k.tr ždĖ (2-3) – ž.dĖ tpl (1-6) – t.pl zbr (3-5) – z.br skn (3-4) – s.kn skĖ (3-4) – s.kĖ nšĢ (4-3) – nš.Ģ vzb (1-3) – v.zb jpr (5-5) – j.pr or jp.r jtm (6-4) – jt.m nšp (5-3) – nš.p nšt (5-2) – nš.t zvl (1-3) – z.vl þsk (2-3) – þs.k jbl (5-5) – j.bl or jb.l jhl (1-2) – j.hl jsl (3-3) – j.sl or js.l mkl (4-6) – m.kl nkc (4-2) – nk.c zhr (2-1) – zh.r dhl (4-2) – dh.l dhĐ (4-1) – dh.Đ mkn (4-4) – m.kn or mk.n mpl (3-6) – m.pl mpĐ (3-6) – m.pĐ ndĐ (2-5) – n.dĐ ngĐ (3-4) – n.gĐ sĢp (3-1) – sĢ.p šsĢ (1-3) – š.sĢ tfþ (3-3) – t.fþ or tf.þ zml (5-4)- zm-l

207

208

psĐ (3-4) – p.sĐ rtv (4-4) – r.tv or rt.v dhr (4-2) – dh.r jšk (3-3) – j.šk or jš.k rsp (2-3) – r.sp sxv (2-3) – s.xv spt (3-1) – sp.t zhĖ (2-4) – z.hĖ dzĖ (2- 5) – d.zĖ jml (4-5) – j.ml lsĢ (3-3) – l.sĢ or ls.Ģ nbĐ (3-4) – n.bĐ nþn (4-4) – n.þn or nþ.n ngr (3-4) – n.gr nsp (5-3) – ns.p pkr (1-6) – p.kr rtr (4-4) – r.tr or rt.r sxĐ (2-3) – s.xĐ stm (2-4) – s.tm štv (2-4) – š.tv vzĐ (1-3) – v.zĐ þšm (2-6) – þ.šm fpl (2-6) – f.lp fsp (1-3) – f.sp jšĢ (3-3) – j.Ģ or jš.Ģ ldr (4-4) – l.dr or ld.r ndn (2-2) – n.dn or nd.n nkn (4-4) – n.kn or nk.n nkĖ (4-4) – n.kĖ or nk.Ė mšt (6-2) – mš.t škv (3-4) – š.kv bdr (1-5) – b.dr jhn (1-4) – j.hn jsx (4-2) – js.x jvh (2-2) – j.vh or jv.h jvl (2-3) – j.vl ktn (1-3) – k.tn ktĖ (1-4) – k.tĖ lkl (5-5) – l.kl or lk.l ltr (5-4) – lt.r mkĐ (4-5) – m.kĐ

Appendix 8

zmn (5-1) – zm.n gzv (3-1) – gz.v jsm (3-6) – j.sm mšĢ (6-3) – mš.Ģ ndl (2-4) – n.dl nkt (4-1) – nk.t rkl (6-5) – rk.l rtn (4-3) – rt.n štr (2-4) – š.tr txn (4-5) – t.xn txĖ (4-5) – t.xĖ vzn (1-4) – v.zn vzĖ (1-5) – v.zĖ vzr (1-2) – v.zr zmĐ (5-4) –zm.Đ ckt (2-1) – ck.t dbl (1-5) – d.bl dbĐ (1- 4) – d.bĐ dvl (3-3) – d.vl or dv.l dvĐ (3-3) – dv.Đ or d.vĐ fkr (3-6) – f.kr hmĐ (4-4) – h.mĐ or hm.Đ jsv (4-2) – js.v nfl (6-4) – nf.l ntn (3-3) – n.tn or nt.n ntĖ (3-5) – n.tĖ pšĢ (3-3) – p.šĢ or pš.Ģ rkn (6-4) – rk.n rkĖ (6-4) – rk.Ė škĐ (3-5) – š.kĐ tsĢ (2-3) – t.sĢ ‫ݤ‬gć (2-1) – ‫ݤ‬g.ć fpĐ (2-6) – f.pĐ fsĢ (1-3) – f.sĢ xsĐ (2-4) – x.sĐ xtl (4-5) – x.tl xtĐ (4-6) – x.tĐ jbr (5-5) – j.br or jb.r jdr (4-4) – j.dr or jd.r jfš (3-1) – jf.š jhr (1-1) – j.hr or j.hr

The Generative and the Structuralist Approach to the Syllable

pxl (4-3) – px.l rhr (1-1) – r.hr or rh.r rkr (6-6) – r.kr or rk.r sxr (2-2) – s.xr or sx.r skm (3-4) – s.km spĐ (3-6) – s.pĐ sšĢ (1-3) – s.šĢ tsv (2-2) – t.sv or ts.v ttl (0-5) – t.tl tšk (2-3) – t.šk vzć (1-3) – v.zć bhr (4-1) – bh.r ckr (2-6) – c.kr dbr (1-5) – d.br dvć (3-3) – d.vć or dv.ć dzn (2-5) – d.zn mhĐ (4-2) – mh.Đ mpr (3-5) – m.pr nþĢ (4-2) – nþ.Ģ ngv (3-3) – n.gv or ng.v nkl (4-6) – n.kl pxn (4-5) – p.xn pxĖ (4-5) – p.xĖ rdn (4-2) – rd.n rdĖ (4-3) – rd.Ė rgr (4-4) – r.gr or rg.r rkm (6-4) – rk.m rpr (5-5) – r.pr or rp.r rst (3-2) – rs.t rĢĐ (5-5) – r.ĢĐ or rĢ.Đ smr (6-4) – sm.r

209

jpl (5-6) – j.pl jsp (3-3) – j.sp or js.p jvć (2-3) – j.vć jvn (2-5) – j.vn jvĖ (2-5) – j.vĖ jvz (2-1) – jv.z jzh (2-2) – j.zh or jh.z kþn (2-4) – k.þn ldĖ (4-3) – ld.Ė lkn (1-4) – l.kn lkĖ (1-4) – l.kĖ lkr (1-6) – l.kr mbl (2-5) – m.bl mbĐ (2-4) – m.bĐ mdl (3-4) – m.dl mdĐ (3-5) – m.dĐ mpk (3-1) – mp.k mpt (3-1) – mp.t špĐ (3-6) – š.pĐ tfš (3-1) – tf.š txl (4-3) – tx.l txĐ (4-3) – tx.Đ txv (4-3) – tx.v tsl (2-3) – t.sl tsp (2-3) – t.sp tšm (2-6) – t.šm tšt (2-2) t.št or tš.t tšĢ (2-3) – t.šĢ zbz (3-3) – z.bz or zb.z zgl (3-5) – z.gl zgĐ (3-4) – z.gĐ

3. Slovak word-medial CCCC clusters (26) The number of differences between the neighbouring segments is in brackets after the clusters. For example, ‘nstv (5-2-4) – ns.tv’ means that there are five different distinctive features between ‘n’ and ‘s’, two different features between ‘s’ and ‘t’, and four differences between ‘t’ and ‘v ’ (see Table 2-5). The syllable boundary should be between ‘s’ and ‘t’:

210

nstv (5-2-4) – ns.tv rstv (3-2-4) – rs.tv Đstv (4-2-4) – Đs.tv mstv (6-2-4) – ms.tv šstv (1-2-4) – š.stv nštr (4-2-4) – nš.tr lstv (4-2-4) – ls.tv cstv (2-2-4) – c.stv or cs.tv4 jstr (4-2-4) – js.tr vstv (2-2-4) – v.stv or vs.tv tstr (2-2-4) – t.str or ts.tr jstv (4-2-4) – js.tv þstv (2-2-4) – þ.stv or þš.tv

Appendix 8

pstr (3-2-4) – ps.tr jstn (4-2-3) – js.tn pstv (3-2-4) – ps.tv jzdr (2-2-4) – j.zdr or jz.dr dzbr (2-3-5) – d.zbr tškr (2-3-6) – t.škr fskĐ (1-3-5) – f.skĐ rnsk (4-5-3) – rns.k jskr (4-3-6) – js.kr kstr (3-2-4) – ks.tr nkþn (4-2-4) – nk.þn kspl (3-3-6) – k.spl or ks.pl rstr (3-2-4) – rs.tr

4. The Slovak word-medial CCCCC cluster (1) rvstv (2-2-2-4) – r.vstv or rv.stv or rvs.tv 5. The English word-medial CC clusters (128) 5.1 The English CC clusters that have one or two different distinctive features (cf. Table 2-2) and should be syllabified as C1.C2 (20) nd(2)5, mb(2), ƾg(2), st(2), kt(1), ts(2), pt(1), mn(1), nm(1), dn(2), kd(2), b‫(ݶ‬2), sf(1), kb(2), gb(1), pk(1), tk(1), dp(2), fș(1), sș(1) 5.2 The English CC clusters that have more than two different distinctive features (cf. Table 2-2) and can together be in the onset of the following syllable (108) nt(3), mp(3), ks(3), ns(5), ƾk(3), nj(4), sk(3), lj(3), kj(5), nv(5), lk(4), lt(4), mj(4), gm(3), sp(3), tr(5), kn(4), zm(5), tj(5), lm(3), nz(4), nf(6), zl(4), gn(3), lb(3), ls(5), gj(4), pj(5), lv(5), lf(6), bj(4), br(4), fr(4), kr(5), mz(5), mf5), dl(3), ft(3), sj(4), fl(5), k‫(ݕ‬3), ml(3), ms(6), gl(3), pr(5), ps(3), tl(4), vr(3), lg(3), lp(4), nș(5), fj(4), sm(6), kw(5), km(4), kl(4), zj(3), zb(3), zn(4), gz(3), gr(4), pn(4), tn(3), tf(3), vj(3), lw(4), ln(3), l‫(ݶ‬3), n‫(ݹ‬5), nh(4), n‫(ݕ‬6), bl(4), bs(4), dm(3), dj(4), mk(4), mr(4), mt(4), m‫(ݕ‬6), md(3), z‫(ݶ‬3), gw(4), pm(3), pl(4), tw(5), tm(4), vl(5), l‫(ݹ‬4), lr(3), l‫(ݕ‬6), ‫ݕ‬m(6), 4

See note 3. The number in brackets after the cluster indicates the number of differences between the individual segments in the cluster.

5

The Generative and the Structuralist Approach to the Syllable

211

‫ݕ‬n(6), ‫ݕ‬r(3), ‫ݕ‬t(3), nl(3), șm(6), șl(5), șj(4), bn(3), dr(4), dv(3), ðm(5), fg(4), fn(6), sw(4), sl(5), sn(5), sb(4) 6. The English word-medial CCC clusters (45) The number of differences between the neighbouring segments is in brackets after the clusters. For example, ‘spr (3-5) – s.pr’ means that there are three different distinctive features between ‘s’ and ‘p’ and five differences between ‘p’ and ‘r’ (see Table 2-2). The syllable boundary should be between ‘p’ and ‘r’: njh (4-2) – nj.h ndr (2-4) – n.dr ntr (3-5) – n.tr str (2-5) – s.tr mbr (2-4) – m.br ƾkw (3-5) – ƾ.kw kst (3-2) – ks.t mpr (3-5) – m.pr mpl (3-4) – m.pl ƾkt (3-1) – ƾk.t nfr (6-4) – nf.r ƾkr (3-5) – ƾ.kr nst (5-2) – ns.t mpt (3-1) – mp.t lst (5-2) – ls.t ltr (4-5) – l.tr ƾgw (2-4) – ƾ.gw ƾgr (2-4) – ƾ.gr ƾkl (3-4) – ƾ.kl ksw (3-4) – k.sw ksp (3-3) – k.sp or ks.p ktr (1-5) – k.tr mdr (3-4) – m.dr

mkw (4-5) – m.kw mbl (2-4) – m.bl mfl (5-5) – m.fl or mf.l6 ƾgh (2-3) – ƾ.gh ƾgt (2-2) – ƾ.gt or ƾg.t ƾgs (2-4) – ƾ.gs mpk (3-1) – mp.k mps (3-3) – m.ps or mp.s pst (3-2) – ps.t n‫ݹ‬b (5-3) - n‫ݹ‬.b ldr (3-4) – l.dr lfr (6-4) – lf.r lgr (3-4) – l.gr lpr (3-5) – l.pr nsl (5-5) – n.sl or ns.l ntl (3-4) – n.tl nsk (5-3) – ns.k nșr (5-4) – nș.r ntm (3-4) – n.tm dzw (2-3) – d.zw spr (3-5) – s.pr mfr (5-4) – mf.r

7. The English word-medial CCCC clusters (6) The number of differences between the neighbouring segments is in brackets after the clusters. For example, ‘nstr (5-2-5) – ns.tr’ means that 6

See note 3.

212

Appendix 8

there are five distinctive features between ‘n’ and ‘s’, two different features between ‘s’ and ‘t’, and five differences between ‘t’ and ‘r ’ (see Table 2-2). The syllable boundary should be between ‘s’ and ‘t’: nstr (5-2-5) – ns.tr mstr (6-2-5) – ms.tr kstr (3-2-5) – ks.tr

7

See note 3.

ƾkst (3-3-2) – ƾks.t nskr (5-3-5) – ns.kr kskl (3-3-4) – k.skl or ks.kl7

BIBLIOGRAPHY

Albrecht, Jörn. “European Structuralism.” In The Languages and Linguistics of Europe: A Comprehensive Guide, edited by Bernd Kortmann and Johan van der Auwera, 821–44. Berlin, New York: Mouton, 2011. Abercrombie, David. Elements of General Phonetics. Edinburgh: Edinburgh University Press, 1967. Akamatsu, Tsutomu. The Theory of Neutralization and the Archiphoneme in Functional Phonology. Amsterdam, Philadelphia: John Benjamins, 1988. Anderson, Stephen R. Phonology in the Twentieth Century: Theories of Rules and Theories of Representations. Chicago: The University of Chicago, 1985. Baertsch, Karen. “Sonority and Sonority-Based Relationships within American English Monosyllabic Words.” In The Sonority Controversy, edited by Steven Parker, 3–38. Berlin, Boston: Walter de Gruyter, 2012. Blevins, Juliette. “The Syllable in Phonological Theory.” In Handbook of Phonological Theory, edited by John A. Goldsmith, 206–244. Oxford: Basil Blackwell, 1995. Brown, Keith, and Sarah S. Ogilvie (eds.). Concise Encyclopaedia of Languages of the World. Oxford: Elsevier, 2009. Cairns, Charles E., and Eric Raimy (eds.). Handbook of the Syllable. Leiden: Brill, 2011. Clements, George N. “The Role of the Sonority Scale in Core Syllabification.” In Papers in Laboratory Phonology I. Between the Grammar and the Physics of Speech, edited by John C. Kingston and Mary E. Beckman, 283–333. Cambridge: Cambridge University Press, 1990. —. “Does Sonority Have a Phonetic Basis? Comments on the Chapter by Bert Vaux.” Ms. CNRS/SorbonneǦNouvelle, Paris, 2006. Clements, George N., and Elizabeth Hume. “The Internal Organization of Speech Sounds.” In Handbook of Phonological Theory, edited by John A. Goldsmith, 245–306. Oxford: Basil Blackwell, 1995. Clements, George N., and Samuel J. Keyser. CV Phonology. A Generative Theory of the Syllable. Cambridge, Massachusetts: MIT Press, 1983.

214

Bibliography

Cole, Jennifer, and José I. Hualde. “Underlying Representations.” In The Blackwell Companion to Phonology: Segmental and Subsegmental Phonology, Vol. 1, edited by Marc van Oostendorp, Colin J. Ewen, Elizabeth Hume and Karen Rice, 1–27. Chichester: Blackwell Publishing Ltd., 2011. Crystal, David. A Dictionary of Linguistics and Phonetics. Oxford: Blackwell Publishing, 2008. ýerný, JiĜí. DČjiny lingvistiky. Praha: Votobia, 1996. Everaert, Martin, and Eric Reuland. “Generative Grammar in Europe: Some Historical Impressions.” In The Languages and Linguistics of Europe: A Comprehensive Guide, edited by Bernd Kortmann and Johan van der Auwera, 867–86. Berlin, New York: Mouton, 2011. Chafcouloff, Michel, and Alain Marchal. “Velopharyngeal Coarticulation.” In Coarticulation: Theory, Data and Techniques, edited by William J. Hardcastle and Nigel Hewlett, 69–79. Cambridge: Cambridge University Press, 1999. Chomsky, Noam, and Morris Halle. The Sound Pattern of English Studies in Language. Cambridge, MA and London: MIT Press, 1991. Cho, Young-mee Yo, and Tracy Holloway King. “Semisyllables and Universal Syllabification.” In The Syllable in Optimality Theory, edited by Caroline Féry and Ruben van de Vijver, 183–212. Cambridge: Cambridge University Press, 2003. de Saussure, Ferdinand. Cours de linguistique générale. Paris: Payot, 1922. —. Course in General Linguistics. Translated by Wade Baskin. Edited by Perry Meisel and Haun Saussy. New York: The University Library, 1959. Available at: https://archive.org/stream/courseingenerall00saus/courseingenerall00sa us_djvu.txt Donegan, Patricia, and David Stampe. “Hypotheses of Natural Phonology.” PoznaĔ Studies in Contemporary Linguistics 45, no. 1 (2009): 1–31. Duanmu, San. Syllable Structure. The Limits of Variation. Oxford: Oxford University Press, 2009. —. “The CVX Theory of Syllable Structure.” 2010. Accessed April 29, 2014. http://www-personal.umich.edu/~duanmu/CUNY-CVX2010Oct.pdf DubČda, Tomáš. Jazyky a jejich zvuky. Univerzalie a typologie ve fonetice a fonologii. Praha: Karolinum, 2005. Dvonþová, Jana. Fyziologická fonetika. Bratislava: Slovenské pedagogické nakladateĐstvo, 1980.

The Generative and the Structuralist Approach to the Syllable

215

Dvonþová, Jana, Gejza Jenþa, and Ábel KráĐ. Atlas slovenských hlások. Bratislava: VydavateĐstvo SAV, 1969. Evans, Nicholas, and Stephen Levinson. “The Myth of Language Universals: Language Diversity and its Importance for Cognitive Science.” Behavioral and Brain Sciences 32 (2009): 429– 92. Giegerich, Heinz J. English Phonology. An Introduction. Edinburgh: Edinburgh University Press, 1992. —. Lexical Strata in English. Morphological Causes, Phonological Effects. Cambridge: Cambridge University Press, 1999. Gibson, Eleanor, J, Anne Pisk, Harry Osser, and Marcia Hammond. “The Role of Grapheme-Phoneme Correspondence in the Perception of Words.” The American Journal of Psychology 75 (1962): 554–70. Glowacka, Dorota. “Unstressed Vowel Deletion and New Consonant Clusters in English.” PoznaĔ Studies in Contemporary Linguistics 3 (2001): 71–94. Goldsmith, John A. Autosegmental Phonology. [PhD dissertation]. Indiana University, 1976. —. “The Syllable.” In The Handbook of Phonological Theory. 2nd ed., edited by John A. Goldsmith, Jason Riggle and Alan C. L. Yu. Oxford: Blackwell Publishing, 2011. Goldsmith, John A., and Bernard Laks. “Generative Phonology: Its origins, Its Principles, and Its Successors.” In The Cambridge History of Linguistics, edited by Linda R. Waugh, Monique Monville-Burston and John E. Joseph. Cambridge: Cambridge University Press, 2011. Available at: http://hum.uchicago.edu/~jagoldsm/Papers/GenerativePhonology.pdf. Gorman, Kyle. Generative Phonotactics. [PhD dissertation], 2013. Gouskova, Maria. “Relational Hierarchies in Optimality Theory: The Case of Syllable Contact.” Phonology 21, no. 2 (2004): 201–50. Green, Antony D. “Extrasyllabic Consonants and Onset WellFormedness.” In The Syllable in Optimality Theory, edited by Caroline Féry and Ruben van de Vijver, 238–53. Cambridge: Cambridge University Press, 2003. Gregová, Renáta. Zvuková štruktúra slabiky. [MA Thesis]. Prešov: Filozofická fakulta Prešovskej Univerzity, 1998. —. Kvantita v štruktúre slovenskej a anglickej slabiky (v textoch masmediálnej komunikácie). [PhD Thesis]. Prešov: Filozofická fakulta Prešovskej Univerzity, 2004. —. “Quantity in Slovak and in British English.” SKASE Journal of Theoretical Linguistics 5, no. 1 (2008a): 17–41.

216

Bibliography

—. “Lexical Phonology and the Structure of the Slovak Syllable.” Suþasni doslidžennja z inozemnoji filolohiji. Vypusk 6: zbirnyk naukovych prac (2008b): 179–85. —. “A Comparative Analysis of Consonant Clusters in English and in Slovak.” Bulletin of the Transilvania University of Braúov 3, no. 52 (2010): 79–84. —. “The CVX Theory of Syllable. A Single-Slot Analysis of the Initial Consonant Clusters in English and in Slovak.” Bulletin of the Transilvania University of Brasov 4 (53), no. 1 (2011a): 205–12. —. “The CVX Theory of Syllable: The Analysis of Word-Final Rhymes in English and in Slovak.” Belgrade English Language and Literature Studies (Belgrade BELLS) III (2011b): 111–26. —. “Sonority-Based Analysis vs. Complex-Sound Analysis of the WordInitial Consonant Clusters in English and in Slovak.” Ostrava Journal of English Philology 5, no. 2 (2012a): 7–22. —. Kvantita v slovenþine a v angliþtine. Košice: Univerzita Pavla Jozefa Šafárika, 2012b. —. “On the Possibility of the English and the Slovak Consonant Clusters to Form Complex Segments.” In Within Language. Beyond Theories. Studies in Theoretical Linguistics, Vol. 1, edited by Anna Bondaruk and Anna PraĪmowska, 274–93. Newcastle: Cambridge Scholars Publishing, 2015. Hála, Bohuslav. Slabika, její podstata a vývoj. Praha: Nakladatelství ýSAV, 1956. Halle, Morris, and Jean-Roger Vergnaud. “Three-Dimensional Phonology.” Journal of Linguistic Research 1 (1980): 83–105. Halle, Morris, and Karuvannur Puthanveettil Mohanan. “Segmental Phonology of Modern English.” Linguistic Inquiry 16 (1985): 57–116. Hall, T. Allan. “Segmental Features.” In A Cambridge Handbook of Phonology, edited by Paul de Lacy, 311–34. Cambridge: Cambridge University Press, 2006. Halle, Morris and George Nick Clements. Problem Book in Phonology: A Workbook for Courses in Introductory Linguistics and Modern Phonology. Cambridge, Massachusetts and London: MIT Press, 1983. Hammond, Michael. The Phonology of English. A Prosodic-Optimality Theory Approach. Oxford: Oxford University Press, 1999. Hardcastle, William J. and Nigel Hewlett (eds.). Coarticulation: Theory, Data and Techniques. Cambridge: Cambridge University Press, 1999. Hasplemath, Martin. “Pre-Established Categories Don't Exist— Consequences for Language Description and Typology.” Linguistic Typology 11 (2007): 119–32.

The Generative and the Structuralist Approach to the Syllable

217

Hoad, T. F. (ed). The Concise Oxford Dictionary of English Etymology. Oxford: Oxford University Press, 1986. Holland, Norman N. The Critical I. New York: Columbia University Press, 1992. Hooper, Joan B. “The Syllable in Phonological Theory.”“ Language 48, no. 3 (1972): 525– 40. Horecký, Ján. “Generatívny opis fonologického systému spisovnej slovenþiny.” Acta Facultatis Philosophicae Universitatis Šafarikanae. Jazykovedný zborník 4 (1975): 5–41. Hyman, Larry. “Non-exhaustive syllabification: Evidence from Nigeria and Cameroon.” Proceedings of the Chicago Linguistics Society 26, no. 2 (1990): 175–95. —. “Universals in Phonology.” In The Linguistic Review, edited by Harry van der Hulst, 345–90, 2008. Jakobson, Roman, and Morris Halle. Fundamentals of Language. Gravenhage: Mouton & Co., 1956. Jakobson, Roman, Gunnar Fant, and Morris Halle. Preliminaries to Speech Analysis: The Distinctive Features and their Correlates. Cambridge, Massachusetts: MIT Press, 1961. Jakobson, Roman. “Typological Studies.” In Selected Writings 1: Phonological Studies, edited by Linda R. Waugh and Monique Monville-Burston, 523–32. Hague: Mouton, 1962. Johnson, Keith. Acoustic and Auditory Phonetics. Oxford: Blackwell Publishing, 2008. Jones, Charles. “Some Constraints on Medial Consonant Clusters.“ Language 52, no. 1 (1976): 121–30. Katamba, Francis. An Introduction to Phonology. London and New York: Longman, 1989. Kahn, Daniel. Syllable-Based Generalizations in English Phonology. [PhD Thesis]. MIT, 1976. Kaye, Jonathan. Phonology: A Cognitive View. Hillsdale, New Jersey: Lawrence Erlbaum, 1989. Kenstowicz, Michael, and Jerzy Rubach. “The Phonology of Syllabic Nuclei in Slovak.” Language 63, no. 3 (1987): 463–97. Kenstowicz, Michael. Phonology in Generative Grammar. Oxford: Blackwell Publishers, 1994. Koster, Jan. “Saussure Meets the Brain.” In Language and Cognition 5, edited by Roel Jonkers, Edith Kaan and Anko Wiegel, 115–20. Groningen: University of Groningen, 1996. KráĐ, Ábel. Pravidlá slovenskej výslovnosti. Systematika a ortoepický slovník. Martin: Matica slovenská, 2005.

218

Bibliography

KráĐ, Ábel, and Ján Sabol. Fonetika a fonológia. Bratislava: Slovenské pedagogické nakladateĐstvo, 1989. Krupa, Viktor. “O forme a obsahu v jazyku.” Slovenská reþ, 44 (1979): 267–70. Kuryáowicz, Jerzy. “Contribution à la thèorie de la syllable.” Biuletyn Polskiego towarzystwa jĊzykoznawczego VIII (1948): 80–114. Ladefoged, Peter, and Ian Maddieson. The Sounds of the World's Languages. Oxford: Blackwell Publishers, Ltd., 1996. Laeufer, Christiane. “Morphology and Syllabification Domains.” Lingua 97, no. 2–3 (1995): 101–21. Laks, Bernard. “Saussure’s Phonology.” In Take Danish, For Instance. Linguistic Studies in Honour of Hans Basboll, edited by Henrik Galberg Jacobsen, Dorthe Bleses, Thomas O. Madsen and Pia Thomsen, 199–211. Odense: University Press of Southern Denmark, 2003. Available at: bernardlaks.info/wp-content/uploads/.../Saussuresphonology-final.pdf. Liberman, Mark Y. The Intonational System of English. [PhD dissertation]. MIT 1975. Lyons, John. Language and Linguistics. An Introduction. Cambridge: Cambridge University Press, 1999. Mannell, Robert. Distinctive Features. Macquarie University, 2008. Marlo, Michael R. “CVX Theory in CCCCCCVX Languages: Implications for Universal Grammar.” Journal of Universal Language 5 (2004): 75– 99. Available at: www.unish.org/unish/DOWN/PDF/Michael%20Marlo.pdf Matthews, Peter H. Concise Dictionary of Linguistics. Oxford: Oxford University Press, 2007. McCarthy, John J. “Nonlinear Phonology: An Overview.” GLOW Newsletter 8 (1982): 63–77. McCarthy, John J., and Alan Prince. “Faithfulness and Reduplicative Identity.” In Papers in Optimality Theory, edited by Jill Beckman, Laura Walsh-Dickey and Suzanne Urbanczyk, 249–384. University of Massachusetts, 1995. McMahon, April. Lexical Phonology and the History of English. Cambridge: Cambridge University Press, 2000. —. An Introduction to English Phonology. Oxford: Oxford University Press, 2002. Mistrík, Jozef (ed.). Jazyk a štýl publicistiky. Bratislava: Slovenské pedagogické nakladateĐstvo, 1982. Mistrík, Jozef. Štylistika. Bratislava: Slovenské pedagogické nakladateĐstvo, 1997.

The Generative and the Structuralist Approach to the Syllable

219

Mohanan, Karuvannur Puthanveettil. Lexical Phonology. [PhD dissertation]. MIT, 1982. —. “The Organization of the Grammar.” In Handbook of Phonological Theory, edited by John A. Goldsmith, 24–69. Oxford: Basil Blackwell, 1995. Newman, John. Coursebook in Feature Geometry. LINCOM Coursebooks in Linguistics 02. Newcastle: Lincom Europa, 1997. Newman, Paul. “The reality of morphophonemes.” Language 44, no. 3 (1968): 507–15. Odden, David. Introducing Phonology. Cambridge: Cambridge University Press, 2005. Ondruš, Šimon, and Ján Sabol. Úvod do štúdia jazykov. Bratislava: Slovenské pedagogické nakladateĐstvo, 1987. Pauliny, Eugen. Slovenská fonológia. Bratislava: Slovenské pedagogické nakladateĐstvo, 1979. Pauliny, Eugen, Jozef Ružiþka, and Jozef Štolc. Slovenská gramatika. Bratislava: Slovenské pedagogické nakladateĐstvo, 1968. Petrišþáková, Katarína. PohĐad na problematiku dvojþlenných konsonantických skupín v slovenþine. [MA Thesis]. Masarykova univerzita, 2006. Pinker, Steven. The Language Instinct: How the Mind Creates Language. New York: W. Morrow, 1994. Považaj, Matej (ed.). Pravidlá slovenského pravopisu 3. Bratislava: Veda, 2000. Prince, Alan, and Paul Smolensky. Optimality Theory. Constraint Interaction in Generative Grammar. Malden, Oxford, Carlton: Blackwell Publishing, 1993. Pulgram, Ernst. Syllable, Word, Nexus, Cursus. Hague: Mouton, 1970. Recasens, Daniel. “Lingual Coarticulation.” In Coarticulation: Theory, Data and Techniques, edited by William J. Hardcastle and Nigel Hewlett, 80–104. Cambridge: Cambridge University Press, 1999. Roach, Peter. English Phonetics and Phonology. A practical course. Cambridge: Cambridge University Press, 2000. —. Personal Web Page. Accessed 14 September 2014. http://www.personal.rdg.ac.uk/~llsroach/phon2/asscoareli-into.htm Rocca, Iggy. Generative Phonology (Linguistic Theory Guides). London: Routledge, 2013. Romportl, Milan. “K analýze zvukového systému ruštiny.” Slovo a slovesnost 23, no. 4 (1962): 282–86. —. Studies in Phonetics. Prague: Academia, 1973. —. Základy fonetiky. Praha: Státní pedagogické nakladatelství, 1985.

220

Bibliography

Rubach, Jerzy. The Lexical Phonology of Slovak. New York: Oxford University Press, 1993. Rubach, Jerzy, and Geert Booij. “Syllable Structure Assignment in Polish.” Phonology 7, no.1 (1990): 121–58. Sabol, Ján. “Frekvencia trojþlenných konsonantických skupín v spisovnej slovenþine.” Jazykovedný zborník 2 (1969a): 129–49. —. “Štvorþlenné konsonantické skupiny v slovenþine.” Slovenská reþ 34 (1969b): 30–33. —. “Spoluhláskové skupiny a fonologické protiklady.” Kultúra slova 9, no. 6 (1975): 204–8. —. Syntetická fonologická teória. Bratislava: Jazykovedný ústav ďudovíta Štúra Slovenskej akadémie vied, 1989. —. “Slovenská slabika (Náþrt problematiky).” In Studia Academica Slovaca 23. Prednášky XXX. letného seminára slovenského jazyka a kultúry, edited by Jozef Mlacek, 214–24. Bratislava: STIMIL– centrum informatiky a vzdelávania FF UK, 1994. —. “K typologickej charakteristike slovenskej slabiky.” In Zborník Filozofickej fakulty Univerzity Komenského. Philologica XLV, edited by Jozef Mlacek, 27–32. Bratislava: Univerzita Komenského, 1997. —. “Zvuk ako estetický fenomén detskej literatúry v akustických médiách.” In Médiá v umení a literatúre pre deti a mládež. Zborník z literárnovednej konferencie venovanej životnému jubileu prof. P. Petrusa, edited by Ondrej Sliacky and Bibiána Hlebová, 32–40. Prešov: Náuka, 2003. —. “Symetria a asymetria v jazyku.” In Vidy jazyka jazykovedy, edited by Martin Ološtiak, Martina Ivanová, and Daniela Slanþová, 32–40. Prešov: Prešovská univerzita v Prešove, 2011. Sabol, Ján and Július Zimmermann. Štatistika. Exaktné metódy v jazykovede a v literárnej vede. Košice: Rektorát UPJŠ, 1986. —. Komunikaþný štatút prízvuku v spisovnej slovenþine. Prešov: FF UPJŠ, 1994. Sagey, Elizabeth C. The Representation of Features and Relations in Nonlinear Phonology. [PhD dissertation]. MIT, 1986. Sampson, Geoffrey. Schools of Linguistics. Stanford: Stanford University Press, 1980. Scheer, Tobias. “Invariant syllable skeleton, complex segments and word edges.” Journal of Linguistics 48 (2012): 685–726. Sokolová, Miloslava, and Gustáv Moško, František Šimon, and Vladimír Benko. Morfematický slovník slovenþiny. Prešov: Náuka, 1999. Stanislav, Ján. Slowakische grammatik. Bratislava: Slovenské pedagogické nakladateĐstvo, 1977.

The Generative and the Structuralist Approach to the Syllable

221

Štekauer, Pavol. An Encyclopaedia of English Linguistics. Prešov: Slovacontact, 1995. Trubetzkoy, Nikolai S. “Travaux du cercle linguistique de Prague 7, 1939.” In Grundzüge der Phonologie [Principles of Phonology]. Translated by Christiane Baltaxe. Berkeley: University of California Press, 1969. Turner, Mark. Death is the Mother of Beauty: Mind, Metaphor, Criticism. University of Chicago Press, 1987. Vachek, Josef. Written Language Revisited. John Benjamins Publishing Company, 1989. —. Dictionary of the Prague School of Linguistics, edited by Libuše Dušková. John Benjamin Publishing Company, 2003. Wells, John C. Longman Pronunciation Dictionary. Harlow: Longman, 1993. —. “Whatever Happened to Received Pronunciation?” In II Jornadas de Estudios Ingleses, edited by Carmelo Medina Casado and Soto Palomo, 19–28. Universidad de Jaén, 1997. Zec, Draga. “The Syllable.” In A Cambridge Handbook of Phonology, edited by Paul de Lacy, 161–94. Cambridge: Cambridge University Press, 2006. Zimmermann, Július. Spektrografická a škálografická analýza akustického reþového signálu. Prešov: Náuka, 2002.

Online Sources: http://www.btinternet.com/~ted.power/clustersindex.html http://economics-files.pomona.edu/StatSite/ssp.html Krátky slovník slovenského jazyka [online dictionary]. http://kssj.juls.savba.sk/ http://manual.audacityteam.org/man/Background_Information_for_the_M edia http://www.softpedia.com/get/Multimedia/Audio/Other-AUDIOTools/Speech-Analyzer.shtml http://korpus.juls.savba.sk/dicts.html http://slovniky.korpus.sk/

AUTHOR INDEX

A Abercrombie, D., 10, 167, 213 Akamatsu, T., 213 Albrecht, J., xv, 155, 213 Anderson, S. R., 213 B Baertsch, K., 70, 213 Benko, V., 220 Blevins, J., xvi, 2, 4, 5, 7, 10, 11, 20, 22, 97, 98, 127, 151, 153, 213 Booij, G. E., 5, 220 Brown, K., xviii, 213 C Cairns, Ch. E., xvi, 2, 4, 100, 101, 213 Clements, G. N., 2, 4, 14, 17, 18, 60, 147, 148, 164, 213, 216 Cole, J., 214 Crystal, D., 6, 8, 153, 214 ý ýerný, J., xv, 99, 214

153, 155, 157, 165, 166, 167, 169, 170, 171, 172, 192, 214 DubČda, T., 10, 214 Dvonþová, J., 79, 115, 116, 214, 215 E Evans, N., 10, 152, 215 Everaert, M., xvi, 214 F Fant, G., 217 G Gibson, E., 8, 22, 215 Giegerich, H. J., 4, 8, 22, 27, 45, 62, 69, 128, 135, 164, 215 Glowacka, D., 196, 215 Goldsmith, J., xv, xvi, xviii, 1, 2, 11, 15, 97, 100, 155, 213, 215, 219 Gorman, K., 195, 196, 215 Gouskova, M., 87, 215 Green, A. D., 4, 215 Gregová, R., xvii, 25, 26, 36, 37, 40, 49, 107, 115, 155, 215, 216 Gunnar, C., 217

D

H

de Saussure, F., xv, 99, 100, 214, Donegan, P., 1, 214 Duanmu, S., xviii, 4, 11, 12, 13, 14, 15, 18, 19, 20, 21, 22, 23, 25, 27, 30, 31, 35, 36, 39, 40, 42, 44, 45, 46, 47, 48, 49, 50, 51, 53, 55, 57, 58, 59, 69, 70, 87, 90, 96, 97, 111, 116, 146, 147,

Hála, B., 101, 106, 118, 119, 216 Hall, T. A., 15, 42, 165, 166, 216 Halle, M., xv, 1, 2, 7, 8, 28, 32, 60, 96, 101, 147, 151, 164, 214, 216, 217 Hammond, M., 22, 215, 216 Hardcastle, W. J., 116, 118, 214, 216, 219 Haspelmath, M., 160

The Generative and the Structuralist Approach to the Syllable Hewlett, N., 116, 118, 214, 216, 219 Hoad, T. F., 111, 217 Holland, N. N., 155, 217 Hooper, J. B., 2, 147, 151, 155, 217 Horecký, J., 155, 217 Hualde, J. I., 214 Hume, E., 14, 15, 16, 17, 18, 213, 214 Hyman, L., 152, 153, 217 CH Chafcouloff, M., 116, 117, 214 Cho, Y. Y., 4, 5, 23, 214 Chomsky, N., xv, 1, 7, 8, 9, 10, 96, 151, 166, 214 J Jakobson, R., 99, 101, 217 Jenþa, K., 217 Johnson, K., 112, 217 Jones, Ch., 5, 22, 128, 167, 217 K Kahn, D., 2, 217 Katamba, F., 1, 217 Kaye, J., 217 Kenstowicz, M., 61, 70, 87, 217 Keyser, S. J., 2, 4, 60, 213 King, T. H., 4, 5, 23, 214 Koster, J., 155, 160, 217 KráĐ, Á., 31, 42, 53, 79, 81, 108, 116, 215, 217, 218 Krupa, V., 218 Kuryáowicz, J., 106, 127, 128, 140, 141, 143, 144, 218 L Ladefoged, P., 41, 42, 116, 218 Laeufer, Ch., 5, 218 Laks, B., xv, xvi, xviii, 1, 11, 97, 99, 155, 215, 218 Levinson, S., 10, 215 Lyons, J., 7, 218

223

M Maddieson, I., 41, 42, 116, 218 Mannell, R., 6, 218 Marchal, A., 116, 117, 214 Marlo, M., 15, 17, 18, 218 Matthews, P. H., 6, 218 McCarthy, J. J., 4, 14, 218 McMahon, A., 28, 43, 45, 69, 90, 128, 143, 218 Mistrík, J., 37, 218 Mohanan, K. P., xviii, 2, 8, 9, 98, 147, 216, 219 Moško, G., 220 N Newman, J., 14, 15, 16, 17, 35 O Odden, P., 6, 219 Ogilvie, S., xviii, 213 Ondruš, Š, 107, 110, 219 Osser, H., 215 P Pauliny, E., xvi, 101, 106, 108, 109, 135, 136, 137, 139, 147, 159, 219 Petrišþáková, K., 24, 219 Pinker, S., 10, 219 Pisk, A., 215 Prince, A., 2, 4, 218, 219 Pulgram, E., 10, 219 R Raimy, E., xvi, 2, 4, 100, 101, 213 Recasens, D., 116, 219 Reuland, E., xvi, 214 Roach, P., 10, 22, 23, 27, 43, 45, 63, 96, 116, 128, 143, 195, 219 Rocca, I., 2, 219 Romportl, M., 10, 36, 111, 116, 219 Rubach, J., xviii, 4, 5, 61, 62, 63, 65, 66, 67, 70, 74, 76, 77, 78,

Author Index

224

79, 80, 81, 82, 85, 86, 155, 158, 217, 220 Ružiþka, J., 159, 219 S Sabol, J., xv, xvi, xviii, 19, 22, 24, 31, 37, 39, 42, 53, 56, 79, 80, 81, 101, 102, 103, 104, 105, 106, 107, 110, 116, 128, 140, 141, 144, 145, 147, 154, 156, 158, 197, 218, 219, 220 Sagey, E. C., xix, 14, 15, 16, 17, 18, 19, 20, 35, 39, 42, 43, 116, 177, 220 Sampson, G., xv, xvii, 1, 7, 8, 9, 10, 42, 76, 151, 220 Scheer, T., xix, 14, 15, 20, 42, 96, 220 Smolensky, P., 2, 219 Sokolová, M., 77, 107, 220 Stampe, D., 1, 214 Stanislav, J., 77, 220

Š Šimon, F., 219, 220 Štekauer, P., 104, 221 Štolc, J., 159, 219 T Trubetzkoy, N. S., xv, 99, 100, 221 Turner, M., xvii, 156, 160, 221 V Vachek, J., 100, 112, 221 Vergnaud, J. R., 2, 60, 147, 216 W Wells, J. C., 37, 221 Z Zec, D., 70, 83, 89, 221 Zimmermann, J., 37, 39, 220, 221

SUBJECT INDEX

A absolute universals, see universals abstract, 102, 103, 104, 142, see also universal affix, 13, 26, 48, 50, 108, 109, 146 potential, 44, 45, 51 real, 48 Affix Rule, 13, 14, 44, 45, 46, 50, 51, 52, 54, 55, 58, 59, 95, 157 affricate, 19, 20, 25, 28, 31, 32, 36, 70, 84, 120, 167, 177, 198, 199 alternation, 4, 104, 108, 144, 145 analytic universals, see universals Anti-Allomorphy, 13, 14, 46, 52, 58, 95, 157 aperture, 105, 119 architectural universals, see universals articulation, 16, 17, 26, 27, 28, 31, 32, 41, 42, 43, 60, 106, 115, 119, 128, 136, 164 double, 41 multiple, 14, 19, 39, 41, 42, 95, 157 secondary, 41 simultaneous, 41 Articulator-based Feature Geometry, 18, 27, 29, 31, 42, 44, 53 articulator-bound feature, see feature articulator-free feature, see feature articulatory phonology, see phonology asymmetry, 107

autosegmental phonology, see phonology B boundary, 1, 2, 5, 11, 37, 49, 77, 107, 108, 112, 128, 133, 146, 159 morpheme, 77, 105, 107, 108, 109, 110, 141, 146 syllable, 47, 77, 81, 86, 105, 106, 107, 109, 110, 118, 119, 121, 122, 123, 124, 128, 133, 134, 135, 136, 137, 138, 141, 144, 146, 150, 206, 209, 211, 212 brackets, 9, 24, 52, 201, 202, 205, 206, 209, 210, 211 slanted, 6, 70, 145 square, 6, 13, 17, 70, 145 C C, see consonant class node, see node cluster, 14, 20, 21, 22, 23, 24, 25, 26, 30, 31, 35, 37, 38, 39, 42, 43, 44, 45, 46, 47, 48, 50, 51, 52, 53, 54, 55, 56, 57, 70, 71, 72, 73, 74, 75, 76, 78, 81, 82, 83, 84, 85, 87, 88, 89, 90, 91, 92, 109, 114, 117, 118, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 147, 148, 153, 157, 159, 165, 167, 119, 170, 171, 172, 175, 178, 191,

226

Subject Index

192, 195, 196, 197, 198, 200, 201, 202, 203, 204, 205, 206, 209, 210, 211 coarticulation, 106, 115, 116, 118, 143, 144, 148 nasal, 116, 117 velopharyngeal, 116 coda, 4, 13, 21, 22, 25, 26, 44, 45, 47, 48, 49, 50, 61, 62, 66, 73, 74, 78, 79, 80, 81, 85, 87, 91, 92, 94, 96, 98, 107, 112, 127, 128, 129, 131, 146 Coda Rule, 62, 63, 64, 65, 66, 67, 72, 74, 75, 82 Complex Coda Rule, 62, 65, 67, 72, 74, 75, 82 Complex Onset Rule, 63 complex segment, xix, 5, 14, 15, 18, 19, 20, 21, 27, 31, 35, 36, 38, 39, 40, 41, 42, 43, 44, 48, 52, 56, 57, 95, 96, 105, 116, 140, 145, 147, 148, 157, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 186, 188, 192, 193, 194, see also complex sound complex sound, 12, 13, 14, 18, 19, 20, 21, 26, 27, 30, 31, 35, 36, 38, 40, 41, 42, 43, 44, 50, 51, 52, 53, 54, 55, 57, 58, 95, 96, 146, 157, 165, 169, 171, see also complex segment complex-segment analysis, 21 concrete, 102, 103, 104, 142, 158, 159, see also individual consonant, 4, 5, 7, 12, 13, 14, 15, 20, 21, 22, 23, 24, 25, 26, 27, 28, 31, 35, 36, 37, 38, 39, 40, 41, 43, 44, 45, 46, 47, 48, 49, 50, 51, 54, 55, 56, 57, 58, 59, 60, 61, 66, 67, 69, 70, 72, 73, 74, 76, 78, 79, 80, 81, 82, 83, 84, 85, 87, 88, 89, 90, 91, 93, 94, 95, 98, 99, 100, 105, 106, 107, 108, 111, 112, 115, 116,

117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 131, 133, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 147, 149, 151, 155, 156, 158, 159, 161, 163, 164, 166, 169, 172, 195, 198, 201 extra, 4, 12, 13, 43, 46, 67, 146 extrasyllabic, 4, 14, 53, 56, 66, 67, 85, 86 final, 13, 23, 45, 101 initial, 101 intervocalic, 74 medial, 22, 85, 133, 195 nasal, 48, 54, 95, 96, 111, 115, 116, 117, 118 stray, 4 unsyllabified, 4, 44, 50, 98, 148, 157 word-final, 26, 44, 106 word-initial, 23, 24, 49, 51, 57, 87, 88, 101 word-medial, 22, 43, 47, 56, 71, 86, 87, 95, 106, 118, 133, 136, 138, 195, 196, 198, 201, 203 consonant cluster, 14, 15, 20, 21, 22, 23, 24, 25, 26, 35, 37, 38, 39, 42, 43, 44, 45, 46, 47, 50, 51, 52, 53, 54, 56, 57, 58, 71, 72, 73, 74, 75, 76, 82, 83, 84, 85, 88, 89, 90, 91, 93, 94, 95, 96, 105, 109, 114, 115, 116, 117, 118, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 147, 149, 155, 157, 159, 165, 166, 168, 169, 170, 171, 175, 178, 191, 192, 195, 196, 197, 198, 200, 201, 202, 203, 204, 205, 206, 209, 210, 211 final, see word-final initial, see word-initial medial, see word-medial word-final, 23, 26, 44, 106

The Generative and the Structuralist Approach to the Syllable word-initial, 23, 24, 49, 57, 87, 88 word-medial, 22, 43, 47, 56, 71, 85, 86, 87, 95, 106, 118, 133, 136, 138, 195, 196, 198, 201, 203 content, 7, 104, 107, 140, 149, 152 contour node, see node contrast, 5, 101, 105, 106, 135, 136, 137, 138, 139, 141, 143, 144, 147, 156, 158, 160, 206 Copenhagen School, xv core syllable, see syllable Coronal, 19, 28, 29, 32, 33, 34, 35, 43, 45, 46, 55, 135, 136, 163, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194 Coronal node, see node CV-phonology, see phonology CV-Rule, 62, 63, 64, 66, 67, 72, 74, 75, 77, 82, 143 CVX theory of syllable, xvii, xviii, xix, 11, 12, 95, 97, 98, 147, 148, 149, 151, 152, 154, 155, 156, 159 D degree of stricture, see stricture derivational morpheme, see morpheme derivational prefix, see prefix descriptive universals, see universals diphthong, 44, 45, 48, 49, 52, 53, 54, 58, 59, 91, 123 distinctive feature, see feature Dorsal node, see node double articulation, see articulation E European structuralism, see structuralism

227

exhaustiveness, 4 explosion, 99, 100 extra consonant, see consonant extrasyllabic consonant, see consonant F feature, 2, 4, 6, 8, 10, 14, 15, 16, 17, 18, 19, 20, 27, 30, 39, 42, 55, 60, 61, 62, 66, 79, 81, 95, 100, 101, 104, 106, 111, 113, 115, 135, 137, 138, 141, 144, 147, 152, 154, 156, 158, 163, 165, 166, 177, 178, 179, 180, 181, 182, 183, 184, 185, 189, 190, 191, 193, 194, 209, 212 articulator-bound, 17 articulator-free, 17 distinctive, 7, 8, 14, 17, 28, 31, 32, 36, 88, 101, 103, 105, 135, 136, 137, 138, 139, 141, 146, 147, 158, 163, 205, 206, 209, 210, 211, 212 phonetic, 6, 7, 60, 61, 79 place, 17, 136 prosodic, 2, 100, 101, 158 simultaneous, 19, 35, 177, 178, 194 stricture, 17, 19, 136, 165 terminal, 16, 17, 27 Feature Geometry, 15, 16, 17, 18, 219, see also Articulator-based Feature Geometry FG, see Feature Geometry final consonant, see consonant flat syllable structure, see syllable structure form, 4, 5, 6, 7, 13, 19, 25, 37, 44, 48, 50, 52, 57, 65, 67, 74, 78, 81, 85, 96, 104, 107, 108, 109, 110, 112, 115, 117, 119, 121, 126, 128, 129, 131, 133, 134, 140, 141, 152, 155 form morpheme, see morpheme formal universals, see universals

Subject Index

228

fricative, 19, 20, 23, 25, 26, 27, 28, 29, 31, 32, 33, 35, 61, 69, 70, 72, 74, 79, 80, 81, 82, 84, 90, 120, 128, 145, 163, 165, 166, 167, 171, 173, 177, 178, 183, 184, 185, 193, 194, 198, 199, 200 G generative phonology, see phonology generativism, xvi, xvii, xix, 155, 156, 159 Geneva School, xv gesture, 18, 19, 20, 27, 30, 35, 42, 44, 46, 55, 165, 166, 168, 169, 170, 171, 172, 173, 175, 177, 178, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194 glide, 62, 67, 70, 73, 78, 79, 80, 81, 82, 158 grammatical morpheme, see morpheme graphic form, 6, 52, 76, see also graphic representation graphic representation, see representation H hierarchical syllable structure, see syllable structure hypermorphophoneme, 103, 105 hypophone, 103, 104, 105 I implosion, 99, 100 individual, 102, 103, 104, 106, 142, 144, 159, see also concrete inherence, 36 Initial Adjunction Rule, 85 initial consonant, see consonant intervocalic consonant, see consonant

initial consonant cluster, see consonant cluster invariant, 103, 104, see also universal K Kuryáowicz's rule, 106, 127, 128, 141, 144 L Labial, 16, 17, 19, 28, 29, 32, 34, 46, 163, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, Labial node, see node level, 8, 9, 10, 25, 66, 76, 80, 96, 97, 98, 101, 102, 103, 104, 105, 119, 144, 145, 146, 147, 151, 152, 153, 157, 158, 160 of abstraction, 102, 144, 145, 147, 152, 160 of the morphophoneme, 8, 104, 144, 145, 151, 160 of the phone, 104, 144, 145, 151, 160 of the phoneme, 104, 144, 145, 151, 160 Lexical Phonology, see phonology linguistic sign, 105, 107, 108, 140, 155 linguistic universals, see universals Liquid Syllabification, 82 London School, xv long vowel, see vowel LP, see Lexical Phonology M Maximal Onset Principle, 143 medial consonant, see consonant medial consonant cluster, see consonant cluster

The Generative and the Structuralist Approach to the Syllable melodic tier, see tier metrical phonology, see phonology minimal sonority distance, 70, 71, 72, 73, 83, 85, 87, 88, 89, 90, 98, 198, 199, 200, 201, 202 monophthong, 44, 45, 49, 52, 53, 55, 58, 59, 91 monosyllabic word, 4, 13, 25, 45, 51, 52, 66, 91, 96, 128 morph thematic, 108 morpheme, 1, 5, 6, 14, 50, 77, 97, 103, 104, 105, 107, 108, 109, 110, 111, 141, 142, 144, 145, 146, 195 derivational, 107, 110, 143 form, 107, 110 grammatical, 107, 108, 143 root, 51, 107, 108, 110, 142 morpheme boundary, see boundary morpheme structure, see structure morphology, 5, 6, 9, 12, 13, 21, 44, 46, 50, 51, 52, 54, 58, 97, 98, 146 morphophoneme, 8, 103, 104, 105, 144, 145, 146, 152, 160, 219 Moscow School, xv mot node, see node MSD, see minimal sonority distance multiple articulation, see articulation N nasal coarticulation, see coarticulation nasal consonant, see consonant nasalization, 54, 95, 116 natural phonology, see phonology neutralization, 25, 104, 108, 144, 145, 146 No Contour Principle, 19, 169, 171 node, 15, 16, 17, 19, 27, 60, 63, 66, 67, 85, 86 class, 16, 17, 18, 27, 29, 34 contour, 18, 27

229

Coronal, 16, 17 Dorsal, 17 Labial, 21 mot, 108 Place, 16, 17, 19 root, 14, 16, 18 Soft Palate, 16, 17 Supralaryngeal, 16 syllable, 2, 63, 85 non-linear phonology, see phonology N-Placement Rule, 62, 66, 67, 72, 74, 75, 79, 82 O obstruent, 69, 74, 82, 83, 100, 128, 133, 163 Obstruent Sequencing Principle, 74, 81 onset, 4, 13, 21, 22, 25, 26, 31, 43, 44, 45, 47, 48, 49, 51, 52, 53, 58, 61, 62, 66, 70, 71, 72, 74, 78, 79, 81, 83, 84, 85, 87, 89, 90, 91, 92, 94, 98, 105, 106, 112, 114, 117, 121, 126, 127, 128, 129, 130, 131, 133, 134, 136, 138, 139, 140, 144, 146, 147, 153, 157, 158, 160, 198, 205, 210 Onset Rule, 62, 63, 64, 65, 66, 67, 71, 72, 73, 74, 75, 82, 84, 85 Optimality Theory, 2, 22, 214, 215, 218, 219 orthographic form, 8, see also orthographic representation orthographic representation, see representation orthography, 7, 8 oscillogram, 38, 112, 113, 114 OSP, see Obstruent Sequencing Principle P phone, 22, 103, 104, 105, 119, 144, 145, 152, 160

230

Subject Index

phoneme, xv, xvi, 6, 8, 17, 19, 22, 24, 27, 28, 29, 31, 32, 34, 45, 52, 54, 80, 81, 83, 88, 100, 101, 103, 104, 105, 108, 127, 128, 134, 135, 136, 137, 138, 140, 144, 145, 146, 147, 151, 160, 161, 163, 177 phonetic feature, see feature phonetics, xv, xvii, 8, 41, 42, 97, 143, 146, 160 physical, 8 physiological, 42 systematic, 8 phonological representation, see representation phonological typology, see typology phonological universals, see universals phonology, xv, xvii, xviii, 1, 6, 8, 9, 11, 14, 42, 81, 97, 98, 99, 100, 101, 141, 143, 146, 147, 151, 155, 160 articulatory, 11 autosegmental, 2, 11, 15 CV-phonology, 2 generative, xv, xvii, xviii, 1, 2, 4, 6, 7, 8, 9, 11, 22, 66, 97, 144, 147, 148, 151, 155, 156, 157 Lexical Phonology, xvii, xviii, 2, 9, 11, 60, 74, 79, 80, 84, 85, 86, 91, 94, 97, 98, 108, 143, 148, 150, 151, 155, 156, 158, 159, 216, 218, 219, 220 metrical, 2, 11 natural, 1, 11, 97 non-linear phonology, 14, 20, 36 prosodic, 11 structuralist, xv, xvi, xviii, 11, 97, 152 three-dimensional, 2 phonotactics, 10, 81, 87, 99, 106, 127, 128, 133, 141, 143, 158 physical phonetics, see phonetics physiological phonetics, see phonetics

place feature, see feature Place node, see node polysyllabic word, 57, 105, 159 Postcyclic Adjunction Rule, 85, 86 potential affix, see affix potential suffix, see suffix Potential Vowel Rule, 13, 14, 44, 46, 50, 51, 52, 53, 54, 55, 58, 59, 95, 157 Prague Linguistic Circle, xv Prague School of Linguistics, xv, 99, 100, 221 prefix, 5, 13, 14, 44, 49, 50, 51, 77, 82, 98, 109, 110, 144 derivational, 107, 108, 109 pre-surface form, 152, 160 Progressive Gliding Rule, 62 prosodic feature, see feature prosodic phonology, see phonology Prosodification Constraint, 77, 78, 81, 82, 98, 146 R real affix, see affix real suffix, see suffix Regressive Gliding Rule, 62 representation, 2, 5, 6, 7, 8, 9, 14, 19, 22, 36, 37, 48, 54, 57, 60, 61, 62, 66, 79, 80, 81, 96, 97, 98, 101, 112, 147, 151, 157, 192 graphic, 7, 22, 37, 112, see also graphic form orthographic, 7, see also orthographic form phonological, 7, 9, 11, 17, 60, 97, 98 surface, 7, 8, 15, 95, 98, 157 three-tier phonological, 158 underlying, 6, 7, 8, 62, 98, 153, 157 rhyme, 3, 4, 44, 45, 46, 47, 48, 49, 52, 53, 54, 55, 57, 58, 59, 63 Right-hand Rule, 63, 129, 131, 143 root morpheme, see morpheme root node, see node

The Generative and the Structuralist Approach to the Syllable rule-based algorithm, 11 rule-based approach, 11, 98, 158 S secondary articulation, see articulation segment, 1, 2, 4, 5, 7, 11, 13, 14, 15, 17, 18, 19, 20, 21, 23, 24, 26, 31, 35, 36, 38, 39, 40, 41, 42, 43, 44, 45, 46, 50, 51, 52, 57, 60, 61, 62, 63, 64, 65, 66, 69, 80, 85, 86, 87, 89, 90, 91, 95, 97, 98, 100, 103, 112, 113, 115, 117, 123, 124, 125, 135, 136, 144, 146, 147, 148, 150, 153, 157, 158, 160, 163, 169, 172, 177, 192, 206, 209, 210, 211 semisyllable, 5 semivowel, 66, 70, 79, 120, 123, 199, 200 short vowel, see vowel simultaneous articulation, see articulation simultaneous feature, see feature single-slot analysis, 31, 43, 51 skeletal tier, see tier slanted brackets, see brackets Soft Palate node, see node sonorant, 17, 66, 80, 82, 84, 100, 112, 113, 114, 119, 120, 128, 135, 144, 163, 171, 183, 184, 185, 191, 192, 193, 194 Sonorant Syllabification, 66, 67, 74, 85 sonority, 4, 44, 61, 67, 69, 70, 71, 72, 73, 74, 83, 84, 87, 88, 89, 90, 91, 98, 105, 119, 120, 123, 125, 146, 148, 158 s. degree, 25, 70, 71, 73, 83, 84, 89, 119 s. graph, 123, 125 s. hierarchy, 61, 70, 83, 87, 88, 123, 125 s. scale, 69

231

Sonority Sequencing Generalization, 4, 5, 61, 66, 67, 69, 70, 71, 73, 74, 80, 81, 83, 84, 85, 86, 87, 88, 89, 90, 91, 98, 158, 201, 202 sonority-based analysis, 71, 83, 87, 88, 90, 91, 198, 200, 201, 202 sound form, 7, 13, 23, 24, 50, 52, 53, 119 Sound Pattern of English, 1, 7, 8, 9, 152, 214 SPE, see Sound Pattern of English spectrogram, 112, 113 SPT, see synthetic phonological theory square brackets, see brackets SSA, see Syllable Structure Algorithm SSG, see Sonority Sequencing Generalization stop, 19, 20, 27, 28, 29, 31, 32, 33, 35, 36, 42, 43, 61, 69, 74, 82, 112, 119, 128, 135, 154, 163, 165, 166, 167, 168, 171, 177, 178, 184, 193, 194, 198, 199, 200 stray consonant, see consonant Stray Erasure, 5 stricture, 17, 19, 28, 32, 41, 105, 106, 119, 121, 122, 123, 124, 136, 143, 165, 166, 171, 173, 184, 191, 192, 193 degree of s., 17, 41, 106, 118, 119, 120, 121, 126, 144 stricture feature, see feature structuralism European, xv, 101, 147 structuralist phonology, see phonology structure morpheme, 5, 50, 77, 107, 108, 109, 110, 111, 141, 142, 144, 145, 147 syllable, 1, 2, 4, 5, 10, 11, 13, 21, 44, 45, 46, 48, 49, 50, 51, 52, 54, 55, 60, 61, 62, 66, 67,

232

Subject Index

74, 77, 84, 85, 91, 93, 94, 96, 97, 98, 99, 100, 101, 108, 109, 110, 111, 112, 114, 115, 117, 118, 119, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 136, 137, 138, 139, 140, 141, 142, 144, 145, 146, 151, 153, 155, 156, 158, 159 flat, 2 hierarchical, 2 substantive universals, see universals suffix, 5, 13, 24, 45, 48, 53, 54, 58, 107, 109, 110 potential, 45 real, 13, 45, 48 Supralaryngeal node, see node surface form, 1, 6, 8, 144, 152, 160, see also surface representation surface representation, see representation syllabification algorithm, 11 syllable, 1, 2, 4, 5, 10, 11, 12, 13, 14, 20, 21, 22, 23, 24, 25, 36, 43, 44, 45, 46, 47, 49, 50, 51, 52, 53, 54, 55, 57, 58, 60, 61, 62, 63, 65, 66, 67, 69, 71, 72, 74, 76, 77, 78, 79, 80, 81, 83, 84, 85, 86, 87, 90, 91, 95, 96, 97, 98, 99, 100, 101, 105, 106, 107, 108, 109, 110, 111, 112, 114, 115, 117, 118, 119, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 150, 151, 153, 155, 156, 158, 159, 160, 169, 198, 205, 206, 209, 210, 211, 212, 218, 220 core, 5 syllable boundary, see boundary syllable node, see node syllable structure, see structure

Syllable Structure Algorithm, 11, 62, 65, 66, 67, 69, 71, 72, 73, 74, 75, 77, 78, 79, 81, 82, 83, 84, 85, 86, 91, 92, 94, 97, 98, 146, 147, 148, 149, 150, 151, 155, 156, 158, 159 syllable tier, see tier syllable weld, 137 symmetry, 106, 107, 109, 141 synthetic phonological theory, 101, 102, 105, 142, 144, 146, 147, 148, 149, 150, 151, 156, 158, 160 systematic phonetics, see phonetics T template-matching algorithm, 11 terminal feature, see feature thematic morph, see morph theory-driven universals, see universals three-dimensional phonology, see phonology three-tier phonological representation, see representation tier, 19, 60, 61, 62, 79, 81, 147 melodic, 60, 61, 62, 66, 79, 81, 147 skeletal, 60, 66 syllable, 60 transgression, 115 typology, xviii, 101, 104 phonological, 152 U underlying form, 1, 6, 7, 8, 151, 153, 160, see also underlying representation underlying representation, see representation universal, 102, 103, 104, 106, 142, 144 , see also abstract universal grammar, 218

The Generative and the Structuralist Approach to the Syllable universality, xviii, 4, 10, 96, 97, 154, 155, 157, 158, 159, 160 universals, 10, 96, 154 absolute, 154 analytic, 154 architectural, 154 descriptive, 153 formal, 96 linguistic, 96 phonological, xvii substantive, 96 theory-driven, 154 unsyllabified consonant, see consonant V V, see vowel variant, 81, 103, 104, 145 velopharyngeal coarticulation, see coarticulation vocalic quality, 45, 48, 53, 58, 116, 117 vocalic quantity, 36, 37, 38, 39, 40, 41, 42, 45, 48, 53, 58, 96, 99, 119 vocalic sound, 53 vowel, 4, 7, 12, 13, 20, 22, 23, 24, 36, 43, 45, 47, 48, 49, 50, 51,

233

53, 54, 58, 60, 61, 62, 66, 69, 79, 81, 95, 96, 100, 101, 106, 111, 112, 113, 115, 116, 117, 119, 123, 125, 135, 139, 152 long, 44, 45, 48, 53, 54, 55, 58, 60, 66, 96, 104 short, 44, 45, 52, 53, 55, 60, 104 vowel epenthesis, 4 W Weight-Stress Principle, 47, 57 word-final consonant, see consonant word-final consonant cluster, see consonant cluster word-initial consonant, see consonant word-initial consonant cluster, see consonant cluster word-medial consonant, see consonant word-medial consonant cluster, see consonant cluster X X-slot, 14, 60, 157