Principles of Radical CV Phonology Is the Culmination of Many Years of Work and Principles of Reflection on the Nature of Phonological Representations

Edinburgh Studies in Theoretical Linguistics CV Phonology Principles of Radical

Books in the series address the core sub-disciplines of linguistics − phonology, Edinburgh Studies in Theoretical Linguistics 4 morphology, syntax, semantics and pragmatics − and their interfaces, with a particular focus on novel data from various sources and their challenges to linguistic theorising. ‘Principles of Radical CV Phonology is the culmination of many years of work and Principles of reflection on the nature of phonological representations. The theory, presented in a clear and thorough manner, builds on and synthesises ideas from a number of different theoretical and descriptive approaches. Its comprehensive treatment of Radical CV every major issue relating to the basic units of phonology makes this book a valuable resource for anyone with an interest in the cognitive organisation of speech sounds and its relation to phonetics.’ Elan Dresher, University of Toronto Phonology A new theory of the structure of phonological representations for segments and syllables A Theory of Based on 30 years of research, this book presents a theory about the ‘speech sounds’ that occur contrastively in human spoken and signed languages. Identifying the ultimate Segmental and Syllabic

elements of which speech sounds consist, Harry van der Hulst proposes a radical theory that recognises only two elements: |C| and |V|. Based on a small set of first principles, Structure the book explains what a possible speech sound is and provides explicit structures for all speech sounds that occur contrastively in the world’s languages. With numerous examples from hundreds of languages, including Dutch, Czech, Japanese, Kabardian, Hungarian, Korean and Zulu, the book also provides insight into current theories of segmental structure, commonly used feature systems and recurrent controversies. Key Features • Presents detailed analysis of phonological contrast • Introduces the central role of dependency relations in phonological structure

• Offers comparison to other models including dependency phonology and Harry van der Hulst government phonology • Covers both European and North-American theories • Contains introductions to many other theories including feature geometry models, the Toronto model, the parallel structure model and q-theory, which can be used for phonology courses • Relates theoretical proposals to a wide range of data as reported in several extensive typological surveys Harry van der Hulst is Professor of Linguistics at the University of Connecticut. He is the Editor-in-Chief of The Linguistic Review and co-editor of the series Studies in Generative Grammar.

Cover design: Stuart Dalziel Harry van der Hulst

2373 eup_van der Huls_PPC.indd 1 19/06/2020 08:32 Principles of Radical CV Phonology Edinburgh Studies in Theoretical Linguistics

Series Editors: Nikolas Gisborne, University of Edinburgh and Andrew Hippisley, Wichita State University

Books in the series address the core sub-disciplines of linguistics – phonology, morphology, syntax, semantics and pragmatics – and their interfaces, with a particular focus on novel data from various sources and their challenges to linguistic theorising.

Series Editors Nikolas Gisborne is Professor of Linguistics at the University of Edinburgh.

Andrew Hippisley is Dean of Arts and Sciences at Wichita State University.

Editorial Board Umberto Ansaldo, University of Sydney Balthasar Bickel, Universität Zürich Olivier Bonami, Université Paris Diderot Heinz Giegerich, University of Edinburgh Jen Hay, University of Canterbury Stefan Müller, Humboldt-Universität zu Berlin Mitsuhiko Ota, University of Edinburgh Robert Truswell, University of Edinburgh David Willis, University of Cambridge Alan Yu, University of Chicago

Titles available in the series: 1 Lexical Structures: Compounding and the Modules of Grammar Heinz J. Giegerich 2 A Substance-free Framework for Phonology: An Analysis of the Breton Dialect of Bothoa Pavel Iosad 3 Principles of Radical CV Phonology: A Theory of Segmental and Syllabic Structure Harry van der Hulst

Visit the Edinburgh Studies in Theoretical Linguistics website at www. edinburghuniversitypress.com/series-edinburgh-studies-in-theoretical- linguistics.html Principles of Radical CV Phonology A Theory of Segmental and Syllabic Structure

Harry van der Hulst

With the editorial assistance of Jeroen van de Weijer To Nancy

Edinburgh University Press is one of the leading university presses in the UK. We publish academic books and journals in our selected subject areas across the humanities and social sciences, combining cutting-edge scholarship with high editorial and production values to produce academic works of lasting importance. For more information visit our website: edinburghuniversitypress.com

Edinburgh University Press Ltd The Tun – Holyrood Road, 12(2f) Jackson’s Entry, Edinburgh EH8 8PJ

Typeset in Sabon by Servis Filmsetting Ltd, Stockport, Cheshire, and printed and bound in Great Britain.

A CIP record for this book is available from the British Library

ISBN 978 1 4744 5466 7 (hardback) ISBN 978 1 4744 5468 1 (webready PDF) ISBN 978 1 4744 5469 8 (epub)

The right of Harry van der Hulst to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988, and the Copyright and Related Rights Regulations 2003 (SI No. 2498). Contents

Preface xi List of abbreviations xiv

Introduction: contents of this book 1 1 Basic assumptions about phonology 2 2 Background: Dependency and Government Phonology 3 3 Radical CV Phonology 3 4 Manner 3 5 Place 3 6 Laryngeal: phonation and tone 4 7 Special structures 4 8 Predictability and preference 4 9 Minimal specification 5 10 Radical CV Phonology applied to sign phonology 5 11 Comparison to other models 5 12 Conclusions 6

Chapter 1 Basic assumptions about phonology 7 1.1 Introduction 7 1.2 What is phonology? 7 1.3 Six theses concerning phonological primes 17 1.3.1 Are features based on perception or articulation? 18 1.3.2 Are features innate? 18 1.3.3 Are features, or is phonology in general, substance-free? 19 1.3.4 Are phonological representations fully specified? 20 1.3.5 Is there such a thing as a segment inventory? 21 1.3.6 Are there still phonemes? 22 1.4 Is phonology different? 26 1.5 Alternations and processes 28 1.6 Summary and concluding remarks 34 vi Contents

Chapter 2 Background: Dependency and Government Phonology 35 2.1 Introduction 35 2.2 Monovalency, grouping, dependency and contrastivity 35 2.2.1 Monovalency 36 2.2.2 The triangular set 40 2.2.3 Grouping and elements in Dependency Phonology 45 2.2.4 Developments in Dependency Phonology 56 2.2.5 Minimal specification and polysystematicity 58 2.3 Government Phonology 60 2.4 Towards Radical CV Phonology 69 2.5 Summary and concluding remarks 71

Chapter 3 Radical CV Phonology 73 3.1 Introduction 73 3.2 An outline of Radical CV Phonology 73 3.2.1 The segmental model 73 3.2.2 Syllable structure 89 3.2.3 Empirical issues 96 3.2.4 The segment–syllable connection 98 3.2.5 Recursivity in syllables or foot structure in Radical CV Phonology 100 3.3 Summary and concluding remarks 105

Chapter 4 Manner 107 4.1 Introduction 107 4.2 Onset 109 4.2.1 Edge (onset head): obstruents 110 4.2.1.1 Edge (onset head): head class 110 4.2.1.2 Edge (onset head): dependent class 120 4.2.2 Bridge (onset dependent): sonorants 131 4.2.3 Sonorant consonants as onset heads (including taps/flaps) 138 4.2.4 Laryngeal consonants 147 4.3 Rhyme 150 4.3.1 Nucleus (rhyme head): vowels 150 4.3.1.1 Nucleus (rhyme head): head class 150 4.3.1.2 Nucleus (rhyme head): dependent class 154 4.3.2 Coda (rhyme dependent): sonorants 164 4.3.3 Coda conditions 165 4.4 Syllabic consonants (sonorants) 173 4.5 Long vowels, diphthongs and geminates 174 4.6 Summary and concluding remarks 176 Contents vii

Chapter 5 Place 179 5.1 Introduction 179 5.2 Edge (onset head): obstruents 179 5.2.1 Edge (onset head): head class 180 5.2.2 Edge (onset head): dependent class 187 5.2.3 Post-velar consonants: pharyngeals and laryngeals 192 5.2.4 Place distinctions for sonorant consonants in the edge 197 5.3 Nucleus (rhyme head): vowels 198 5.3.1 Nucleus (rhyme head): head class 199 5.3.2 Nucleus (rhyme head): dependent class 201 5.3.3 Syllabic consonants (sonorants) 202 5.4 Bridge and coda 204 5.5 Summary and concluding remarks 204

Chapter 6 Laryngeal: phonation and tone 207 6.1 Introduction 207 6.2 Edge (onset head): consonants (phonation) 207 6.3 Nucleus (rhyme head): vowels (tone) 217 6.4 Four issues 224 6.4.1 Phonation oppositions in obstruents and sonorants 225 6.4.2 Laryngeal realism 230 6.4.3 Phonation in the nucleus 235 6.4.4 The correlation between tone and phonation 238 6.5 Bridge and coda 240 6.6 Summary and concluding remarks 240

Chapter 7 Special structures 242 7.1 Introduction 242 7.2 Incomplete structures 243 7.2.1 No content at all 243 7.2.1.1 Vowel/zero alternations 244 7.2.1.2 Consonant clusters 249 7.2.1.3 Initial geminates 257 7.2.1.4 Schwa 258 7.2.1.5 Consonant/zero alternations 258 7.2.1.6 Ghost consonants 259 7.2.1.7 Morphological templates 260 7.2.2 Partial content 261 7.2.2.1 No ‘no manner’ 261 7.2.2.2 Manner only 261 viii Contents

7.3 Overcomplete structures 263 7.3.1 Complex consonants 263 7.3.1.1 Affricates 264 7.3.1.2 Consonants with secondary manner 265 7.3.1.3 Consonants with secondary place 265 7.3.2 Consonants with two major places (clicks and multiply-articulated consonants) 266 7.3.3 Complex vowels 279 7.3.3.1 Short diphthongs 279 7.3.3.2 Vowels with contour tones 279 7.3.3.3 Vowels with special phonation 280 7.3.3.4 Vowels with special manner 280 7.3.4 Branching syllabic constituents or ‘two-root structures’ 281 7.4 Summary and concluding remarks 282

Chapter 8 Predictability and preference 283 8.1 Introduction 283 8.2 Harmony 284 8.2.1 Paradigmatic and cross-class harmony 284 8.2.2 Disharmony 300 8.3 Preference rankings of segments per syllabic position 300 8.3.1 Two determining principles: harmony and dispersion 301 8.3.1.1 Manner preferences 303 8.3.1.1.1 Manner preferences: edge 303 8.3.1.1.2 Manner preferences: nucleus 305 8.3.1.1.3 Manner preferences: bridge and coda 306 8.3.1.2 Place preferences 313 8.3.1.2.1 Place preferences: edge 313 8.3.1.2.2 Place preferences: nucleus 314 8.3.1.3 Laryngeal preferences 315 8.3.1.3.1 Laryngeal preferences: edge (phonation) 315 8.3.1.3.2 Laryngeal preferences: nucleus (tone) 315 8.3.1.4 Concluding remarks 316 8.4 Preferred segmental systems 318 8.4.1 The overall structure of segmental systems 318 8.4.2 Polysystematicity 322 8.4.3 Conclusions 323 8.5 Summary and concluding remarks 323 Contents ix

Chapter 9 Minimal specification 324 9.1 Introduction 324 9.2 A typology of redundant properties 325 9.3 Radical underspecification 328 9.4 Contrastive and radical underspecification in a unary framework 331 9.5 Markedness, complexity and salience 339 9.6 Examples of minimal specification 340 9.7 Can redundant elements become active? 346 9.8 Constraints and learnability 346 9.9 Summary and concluding remarks 352

Chapter 10 Radical CV Phonology applied to sign phonology 353 10.1 Introduction 353 10.2 The macrostructure of signs 357 10.3 The microstructure of signs 363 10.3.1 The articulator 363 10.3.1.1 FingerSelection 364 10.3.1.2 FingerConfiguration 366 10.3.2 Orientation 366 10.3.3 Place 368 10.3.4 Manner (‘movement’) 370 10.4 Two-handed signs 372 10.5 What about syllable structure? 374 10.6 Summary and concluding remarks 377

Chapter 11 Comparison to other models 378 11.1 Introduction 378 11.2 Feature Geometry models 378 11.3 Other models 389 11.3.1 Dependency models 389 11.3.2 The nested subregister model 391 11.3.3 The Toronto model 392 11.3.4 The parallel structure model 393 11.3.5 The channel-neutral model 398 11.3.6 The Duanmu model 403 11.3.7 Government Phonology 2.0 404 11.3.8 Q-theory 407 11.4 Summary and concluding remarks 409

Chapter 12 Conclusions 411 12.1 Introduction 411 12.2 Goals and basic principles 411 12.3 X-bar structure everywhere 414 x Contents

12.4 Strengths and weaknesses 415 12.5 Some unresolved issues 419 12.6 What’s next? 424

Appendix 425 References 431 Subject index 479 Language index 494 Preface

This book has been a long time in the making. My work on features and segmental structure started in the early 1980s when I began explor- ing the new wave of autosegmental and metrical theories, applying the former to vowel harmony cases (leading to another book that was long in the making: van der Hulst (2018)) and the latter to my analysis of syllable structure and stress in Dutch (van der Hulst (1984)). Around that same time, learning about Dependency Phonology from my col- league Colin Ewen, I started considering the use of single-valued features and dependency. In 1990, I finished a manuscript that was entitled ‘The book of segments’, which I distributed on a small scale. This manuscript (essentially a forerunner of the present book), which contained an ambi- tiously ‘complete’ account of segmental representation in terms of unary features and dependency, has been the backbone of a lot of my work in this area since then. For each paper or talk on this subject, I would update the theory, which, in my view at least, each time made it better, giving it wider empirical scope and greater theoretical simplicity and elegance. In this endeavour, I collaborated with various colleagues such as Colin Ewen, Marcel den Dikken, Helga Humbert, Maarten Mous and Norval Smith. The model underwent many changes, slowly moving to an approach that uses a minimal number of phonological primes. Along the way, it became clearer to me what I was trying to achieve with my attempt to develop a structure that would account for all phonological ‘features’ and their interrelationships. In the early days of phonology, the basic units that form the perceptual side of language were thought to be ‘speech sounds’, or more technically phonemes, meaningless mental units of sound that in linear sequences would form meaningful units like morphemes and words. Phonemes were taken to be the ‘atoms’ of language. A new development introduced units that are smaller than phonemes, called (distinctive) features, which stand for properties of speech sounds, or could be seen as building blocks of phonemes. Early proposals for feature sets (such as Jakobson, Fant & Halle (1952)) presented a list of features arranged in certain groups, but these groups did not find a formal acknowledgement in theories of phonological structure

xi xii Preface

(such as Chomsky & Halle (1968)). Later, formal grouping structures were proposed (such as in Anderson & Ewen (1980); Clements (1985)). My dissatisfaction with these proposals was due to the fact that the sets of features being proposed essentially formed a list, or a group of lists (which could have been longer or shorter and structured in different ways). The lists and structures were inductively derived from observed processes and mechanism of articulation. In my proposal, the set of primes is no longer an arbitrary list, organised into an arbitrary geometry, but instead a set to which no prime could simply be added and from which no prime could be removed without implications elsewhere in the structure. Essentially, I was developing a metatheory of phonological ‘features’, which provides a principled, explanatory account of the structure of the set of ‘features’, based on cognitive principles of categorisation. Of course, the phonetic substance and processes have a say in the matter, but the emphasis in Radical CV Phonology is on cognitive principles that structure the phonetic substance into categories that then correlate with bits and pieces of this substance. To appreciate this theory, one has to share the idea that it makes sense to derive the set of primes that is necessary to express all possible phone- mic contrasts from a few general principles which determine the categorisation of acoustic percepts and their articulatory correlates. In my view, it is much more interesting to derive primes from general principles than to enumerate an essentially random list on a need-to-be basis. Phonological primes are basic mental concepts that as such (like most or perhaps all mental concepts) are created through mental categorisation processes, based on percepts of a pre-given substance. In the case of phonology, basic concepts are created from acoustic percepts and proprioceptive percepts of articulation for the specific purpose of being effective contrastive units whose function is to optimally differentiate meaningful entities (like morphemes and words). We thus expect the resulting categorisation to reflect properties of the perception of phonetic acoustic and articulatory substance, as well as properties that follow from whatever principles guide categorisation. My assumption is that these principles lead to the creation of discrete unary primes that are organised in a hierarchical dependency structure. This hierarchical structure, firstly, accommodates the set of primes and the paradigmatic relationships (i.e. affinities) that hold between them. Secondly, the structure also provides a basis for how the primes relate to the syntagmatic (that is, syllabic) structure. Thirdly, the structure accounts for the behaviour of primes in phonological alternations, that is phonological (as opposed to suppletive) allomorphy. Given that my proposal adopts the principles of Dependency Phonology, we will see that the head–dependent relationship pervades all phonological constructions. Crucially, in a dependency approach constructions are not constituents as understood in constituent formalisms. Preface xiii

The basic tenet of my approach was much inspired by John Anderson’s Dependency Phonology model. The developments of my own version of Dependency Phonology turned out to converge with certain aspects of Jonathan Kaye’s Government Phonology model, which undoubtedly had its own influence on my thinking processes. Having published bits and pieces of the model during the 1990s and early 2000s, I formulated a complete proposal in van der Hulst (2005a) in a Festschrift for John Anderson. After this, another period followed of fine tuning and modify- ing which has led to the present monograph, which, no doubt, does not contain the final version of the theory. At some point, after I had proposed that the minimal set of elements could just contain two elements, |C| and |V|, someone (I forget who, sorry), suggested that I call my theory Radical CV Phonology (RCVP). I am first and foremost indebted to John Anderson and Jonathan Kaye, leading thinkers in the field of modern phonology. Over the years, I have collaborated with many other phonologists and I owe them for the development of RCVP: Norval Smith, Colin Ewen, Jeroen van de Weijer, Marcel den Dikken, Rob Goedemans, Helga Humbert, Nancy Ritter, Maarten Mous and Chris Golston. I am also grateful to several people who have commented on a synopsis of the RCVP model that I made for the specific purpose of getting feedback without burdening them with the entire book manuscript or who have answered specific queries: John Anderson, Ksenia Bogomolets, Cor van Bree, Marcel den Dikken, Matthew Gordon, Alex Vaxman, Song Zhenjun and Jeroen van de Weijer. Of course my thinking about features has also been inspired by the work of numerous other phonologists, through their work or during personal conversations. I single out Nick Clements, whose original ideas, enthusi- asm and encouragement have always been an inspiration. Finally, I would like to thank Fiona Sewell for her careful editing of the final manuscript. List of abbreviations

AC arytenoid cartilages AE Anderson & Ewen (1987) ASL American Sign Language ATR Advanced Tongue Root C, V the two basic elements of Radical CV Phonology CG consonant–glide CL consonant–liquid CN consonant–nasal DP Dependency Phonology FG Feature Geometry GP Government Phonology h, l high, low (tonal register) H, L, M high, low, mid (tone) IPA International Phonetic Alphabet KLV85 Kaye, Lowenstamm & Vergnaud (1985) KLV90 Kaye, Lowenstamm & Vergnaud (1990) LM Ladefoged & Maddieson (1996) MAC multiply-articulated consonant MD Maddieson (1984) OT Optimality Theory PP Particle Phonology RcvP Radical CV Phonology RTR Retracted Tongue Root SAA Structural Analogy Assumption SCL Syllable Contact Law SDA Successive Division Algorithm SPE The Sound Pattern of English S Strong TR Tongue Root UHR Universal Harmony Rule VOT Voice Onset Time W Weak

xiv Introduction: contents of this book

This book presents a theory of phonological structures with roots in the framework of Dependency Phonology (henceforth DP), but proposes a rather different ‘geometry’, which reduces the set of unary elements to just two: |C| and |V| (which explains the name of the theory). The structure that is proposed accommodates all and only phonological distinctions that have been found to function contrastively in (at least one of) the world’s languages. As such, the theory explains the set of contrastive distinctions, rather than (as is common) presenting it as a ‘random’ list (with or without a ‘random’ geometrical organisation). The book also provides an account of the relationship between syllable structure and segmental structure, but not in detail of phonological alternations (allomorphy). Although the theory mainly deals with spoken language phonology, it is also shown how it provides an account of segmental and syllabic structure in sign languages. The proposal made here concurs with the view that the building blocks of segmental structures are unary elements, a view shared with models such as DP and Government Phonology (henceforth GP) as well as several other models. In agreement with DP and the approach called ‘Feature Geometry’ (henceforth FG), my proposal organises the elements into a segment-internal structure. Both this structure and the ‘syntax’ of element combinations are fundamentally based on the notion of dependency, with a striking recurrence of an ‘X-bar’-like structure (that is, a head with two levels of dependents). My proposal can be regarded as a metatheory of phonological features. It will be shown that many prior proposals for feature systems (for manner, place or laryngeal distinctions) find a ‘home’ in this model. The detailed analysis of phonological contrast provides a typological window on segmental and syllabic inventories in the world’s languages, as well as on the relationship between these two levels. The book introduces the reader to the central role of dependency relations in phonological structure, while advocating the idea that this structure can be derived from a small set of basic principles. Through comparison to other models, this work also provides a window on

1 2 Introduction current theories of segmental structure, commonly used feature systems and recurrent controversies. My approach is to first outline the background to my proposal in DP and GP (although mainly the former), followed by a systematic presen- tation of my model of segmental and syllabic structure (Chapters 1–3). I then develop this model in subsequent chapters for the element classes manner, place and laryngeal, motivating the details of the proposal on the basis of typological findings regarding segment inventories and contrast (and with reference to phonological rules, e.g. vowel harmony) (Chapters 4–6). Chapter 7 deals with ‘special structures’ which are either incomplete (missing, for example, place or manner) or overcomplete (such as different kinds of ‘complex segments’). Subsequently, I show how the model accounts for a number of central claims in phonology involving predictability (Chapter 8) and ‘minimal specification’ (Chapter 9). In Chapter 10, I show how the Radical CV Phonology (henceforth RCVP) model can be applied to sign language structure (based on my own work in this area), while Chapter 11 offers a comparison between my model and a selection of other, prevalent models. Finally, Chapter 12 reviews my goals and the basic principles of RCVP and mentions strengths, weaknesses and possible further developments. The RCVP model has been developed in a series of earlier articles over the last two decades. However, this work is not simply a collection of articles; far from this, it is a new, completely (re)written text, with many new proposals and coverage of content that was never addressed in earlier articles. Following, I offer a chapter-by-chapter summary so that the reader can decide whether to read the book in sequence or skip to certain preferred chapters. If the reader wants to tackle the RCVP model head-on, the first two chapters can be skipped at first to focus on Chapters 3–7.

1 Basic assumptions about phonology In this chapter, I outline my basic assumptions about the enterprise called phonology, starting with my understanding of the scope of phonology, which I break down into three assumptions: the need for a separation of grammatical phonology and phonetic implementation, the idea that ‘phoneme’-sized segments are pivotal units in phonological representations, and the idea that there is a syllabic organisation. I then situate RCVP within the field of current phonological approaches. Subsequently, I discuss six theses regarding phonological primes and representations. I conclude with addressing the question of structural analogy, that is, whether phonology is different in a fundamental way from syntax, answering this question in the negative. Introduction 3

2 Background: Dependency and Government Phonology In this chapter, I briefly discuss the background of the theory that is proposed in this book, which I see as a development of DP, with which it thus necessarily shares a number of design properties which are presented in some detail, both with respect to the DP proposal for elements, and with respect to its basic principles. My model also has certain properties in common with varieties of GP, which I also review in this chapter. Comparison to other phonological models is covered in Chapter 11.

3 Radical CV Phonology In this chapter, I present an introduction to the RCVP model, focus- ing on the ‘syntax’ of C/V combinations but without providing details of the segmental structure or empirical underpinning. In the chapters that follow, I go into the details for each class of phonological elements (manner, place, laryngeal) and provide typological empirical support. In this chapter, I also provide the RCVP model of the syllable (which accommodates four core positions) and discuss how segmental structure and syllable structure are connected. This chapter also discusses the question of whether major class distinctions can be expressed structurally, as opposed to adopting a separate major category class. I show how the two RCVP ‘C’ and ‘V’ elements are phonetically interpreted by providing a set of interpretation functions that form part of the phonetic implementation system.

4 Manner In this chapter, I focus on the manner class. I take manner to be the head class of the segmental structure because manner elements determine the distribution of segments in the phonotactic organisation of words. Also, manner elements are the most stable elements, often resisting assimila- tion. Following the notion of polysystematicity (the idea that sets of contrastive segments differ for different phonotactic positions) I discuss manner distinctions for each of the four core syllabic positions and, where relevant, for the head (primary) and dependent (secondary) class separately.

5 Place Place elements are mostly relevant for the syllabic head position (that is, the onset head and the rhyme head position), because dependent units (onset dependent and rhyme dependent or coda), as proposed in Chapter 3, have limited distinctive location properties. In this chapter, I 4 Introduction adopt the same structure as for the chapter on manner (which is also used for the laryngeal class in the next chapter).

6 Laryngeal: phonation and tone In this chapter, I discuss the laryngeal class as needed for phonation types in consonants. For vowels, laryngeal distinctions cover tone proper and register. I discuss at length the position of laryngeal realism as it plays out in analyses of phonation types in different Germanic languages. I then turn to a number of specific issues in typologies of phonation types and phonation on vowels. Finally, I review correlations between tone and phonation.

7 Special structures In this chapter, I propose both incomplete and overcomplete structures for segment types that call for one or the other. Incomplete structures are structures that miss one of the element classes. In previous chapters, we have seen examples of this, in the sense that a non-tonal language does not use the laryngeal node for vowels. We have also seen that the place class can be missing, as in central vowels and pharyngeal and laryngeal consonants. What cannot be missing is the manner class because this, being the head class, is obligatory. I discuss one possible exception to this. I consider whether the manner class, being obligatory, can be ‘empty’, then provide a different approach that may not require this kind of ‘abstractness’. Subsequently, I turn to overcomplete structures which are necessary for various classes of so-called complex segments, such as clicks, multiply-articulated consonants (henceforth MACs), short diphthongs and some others.

8 Predictability and preference In this chapter I discuss how RCVP relates to some recurrent concepts in phonological theory. This discussion is framed by the following issues and questions:

• Predictability of elements: can some C/V choices be predicted from the syllable structure specification of segments or other elements within the segmental structure? • Preference of occurrence (lexical or token frequency): how does RCVP predict preference of occurrence of segment types, with reference to position in the syllable (or larger units)? • Preferred segmental systems (system typology, type frequency): how does RCVP predict the shape of preferred phoneme systems? Introduction 5

Throughout the preceding chapters, I have occasionally indicated how RCVP addresses these questions, but in this chapter I offer a systematic account.

9 Minimal specification In this chapter I address the following issues:

• Redundancy and default: are predictable (i.e. redundant) elements underspecified, and what about elements that are contrastive? In other words, is there a role for (contrastive and radical) underspecification in RCVP? • Markedness: how does RCVP express markedness? What is ‘markedness’ in RCVP? • Minimal specification: given that underspecification plays a role in RCVP, how are minimal (i.e. maximally underspecified) representations determined? • Phonological activity: can redundant or default properties be phonologically active? • Constraints: which types of constraints can be formulated to restrict the set of possible segments to precisely those that occur in a given language?

10 Radical CV Phonology applied to sign phonology As explained in Chapter 1, a central aspect of RCVP is that ‘features’ (elements) are not as such innate. Elements emerge from successive splitting during language acquisition, based on the occurrence of contrast. However, the RCVP splitting model is not specific to the phonetic modality. This implies that the model can also be applied to sign language phonology, for which I have developed an explicit model in previous work over the last two decades. In this previous work, I have not emphasised how an appropriate structure for signs can be derived from the principles of RCVP, instead using more ‘descriptive’ labels for contrastive specifica- tions. In this chapter I show how such a structure can be represented in terms of a binary ‘C/V’ choice within the relevant classes.

11 Comparison to other models In this chapter, I offer a comparison between my model and a selection of other models. While it is of course important to highlight how the RCVP model differs from other models, primarily in its radical proposal to reduce all distinctions to two elements which occur in multiple roles (head, dependent) in different element classes, I am more interested in 6 Introduction showing how the RCVP model is compatible with feature sets and FGs that have been proposed, and motivated, in other proposals. To the extent that these proposals have been made on solid empirical grounds, they also support the RCVP choices. What RCVP adds to these specific proposals is that it is shown how they can be derived from the basic RCVP principles, adding an element of explanatory adequacy to these proposals. I point out striking analogies between RCVP and other proposals which have often been developed independently of DP and RCVP, with no apparent knowledge of such historically earlier models, which only adds to their value as independent confirmation of the model proposed here.

12 Conclusions This chapter summarises the contribution of RCVP to our understanding of segmental and syllabic representations. I outline strengths as well as weaknesses, and point to future research which will contribute to the advancement of models in this domain. I also briefly consider the exten- sion of RCVP to higher levels of phonological/prosodic structure. In the Appendix, I provide a summary of all relevant structures and their interpretations.