The
Sounds of
Korean
Tis introduction to the sounds o Korean is designed or English-speaking s tudents with no prior knowledge o the language, and includes online sound �les, which demonstrate the sounds and pronunciation pronunciation described. It will be an invaluable resource or students o Korean wanting wanting to understand the basis o the current state o Korean phonetics phonetics and phonology, phonology, as well as or those studying Korean linguistics. • Provides a complete and authoritative authoritative description and explanation explanation o the current state state o Korean phonetics and phonology • Gives clear comparisons comparisons with English, and provides practical practical advice on pronunciation pronunciation • Provides a wealth o authentic Korean Korean examples • Each chapter contains exercises and Did you know? sections sections to help students put their t heir knowledge into practice. ������� ���� is a Proessor Proessor in the Department o Korean Language and Literature at Korea University, Seoul. She has published widely in Korean phonetics/phonology and is the author o Understanding Speech Sounds (2000), Te Sound Pattern of Korean (2003, with Jaeeun Cha) and Articulatory Articulatory and Phonologic Phonological al Disorders Disorders (2007, with Sujin Kim). ����� ����� is Young Bin Min-Korea Foundation University Lecturer in Korean Language and Linguistics in the Oriental Institute o the University University o Oxord. Her current research is on the pragmatic aspects o syntactic architecture and the role o prosody in syntax. ������ ������ ��� is an Associate Proessor in the Department o Korean Language and Literature at Kyonggi University. She specialises in Korean phonology and the history o the Korean language and is the t he author o Middle Korean (1999) and Te Sound Korean Phonology Phonology (1999) Pattern of Korean (2003, with Jiyoung Jiyoung Shin).
The Sounds of
Korean Jiyoung Shin Jieun Kiaer Jaeeun Cha
��������� ���������� ����� Cambridge, New York, Melbourne, Madrid, Cape own, Singapore, São Paulo, Delhi, Mexico City Cambridge University Press Te Edinburgh Building, Cambridge CB2 8RU, UK Published in the United States o America by Cambridge University Press, New York www.cambridge.org Inormation on this title: www.cambridge.org/9781107672680 © Jiyoung Shin, Jieun Kiaer and Jaeeun Cha 2013 Tis publication is in copyright. Subject to statutory exception and to the provisions o relevant collective licensing agreements, no reproduction o any part may take place without the written permission o Cambridge University Press. First published 2013 Printed and Bound in the United Kingdom by the MPG Books Group A catalogue record for this publication is available from the British Library Library of Congress Cataloguing in Publication data
Sin, Chi-yong, 1966 Sept.Te sounds o Korean / Jiyoung Shin, Jieun Kiaer, and Jaeeun Cha. p. cm. Includes bibliographical reerences and index. ISBN 978-1-107-03005-3 (hardcopy) 1. Korean language – Phonology. 2. Korean language – Phonetics. 3. Korean language – Study and teaching – English speakers. I. Kiaer, Jieun. II. Ch’a, Chae-un. III. itle. PL915.S5525 2013 495.7 83421 – dc23 2012017770 ′
ISBN 978-1-107-03005-3 Hardback ISBN 978-1-107-67268-0 Paperback Tis book is published with the generous support o the International Center or Korean Studies, Korea University’s Research Institute o Korean Studies. Te International Center or Korean Studies was established in 2003 to support scholarship and exploration o Korea in the humanities and social sciences and to promote new research in Korean Studies to a wide international audience.
Additional resources or this publication at www.cambridge.org/shin Cambridge University Press has no responsibility or the persistence or accuracy o URLs or external or third-party internet websites reerred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.
Contents
Figures ables Preface Notational conventions
Characteristics of the Korean language 1.1 Origin, history and distribution 1
1.1.1 Te Korean language in East Asian history 1.1.2 Where is the Korean language rom? 1.1.3 Korean vs. Chinese and Korean vs. Japanese: are they related, and i so, how? 1.1.4 Korean as a global language: is the Korean language or the Korean peninsula only?
1.2 Korean alphabet and romanisation 1.2.1 Hangeul: the Korean alphabet 1.2.2 Romanisation: how many styles and how different is each option?
1.3 Lexicon 1.3.1 What is the proportional ratio between Sino-Korean and pure Korean words? 1.3.2 Increase o English loanwords 1.3.3 New words in the twenty-�rst century 1.3.4 Motion- and sound-symbolic words 1.3.5 Can South Koreans and North Koreans understand each other?
1.4 Structural properties o Korean 1.4.1 How are words and sentences composed? 1.4.2 Is the word order rigid or ree? 1.4.3 I it is not through word order, how are grammatical roles and relations expressed? 1.4.4 How is the speaker’s attitude expressed in Korean? 1.4.5 How is the mood o a sentence realised?
1.5 Socio-pragmatic characteristics o Korean 1.6 Dialectal variation
V
pag e x i xiv xvii xix
1 1 1 2 2 3
4 4 6
6 6 6 6 7 8
8 8 10 10 12 13
13 15
Contents
VI
Production of sounds 2.1 Vocal organs 2
2.1.1 2.1.2 2.1.3 2.1.4
Overview Larynx ongue Palate
2.2 Major cavities 2.3 Processes o speech production 2.3.1 Initiation 2.3.2 Phonation 2.3.3 Articulation
2.4 Classi�cation 2.4.1 Sonorants vs. obstruents 2.4.2 Consonants 2.4.2.1 Place o articulation 2.4.2.2 Manner o articulation 2.4.2.3 Phonation (or voice) types 2.4.3 Vowels 2.4.3.1 Height 2.4.3.2 Backness 2.4.3.3 Roundedness (protrusion o the lips)
16 16 16 16 17 18
18 20 20 23 25
26 26 27 27 28 31 35 35 36 37
2.5 Summary Exercises
37 38
Basic concepts of phonology 3.1 Speech sounds, phones and phonemes 3.2 Phonemes and allophones 3.3 Criteria or phoneme identi�cation
40 41 43 44
3
3.3.1 Minimal pairs 3.3.2 Complementary distribution, phonetic similarity and ree variation
3.4 Position o sounds
44 44
47
3.4.1 Sounds within a syllable: onset, nucleus and coda 3.4.2 Sounds in word-initial, word-medial and word-�nal position
48
3.5 Underlying orm, surace orm and phonological rules 3.6 Phonological eatures
48 50
3.6.1 Why are phonological eatures needed? 3.6.2 Major class eatures
47
51 52
3.7 Summary Exercises
54 54
Consonants 4.1 Consonant chart 4.2 Obstruents
56 56 57
4
4.2.1 Stops 4.2.1.1 Word-initial stops 4.2.1.2 Word-medial (intervocalic) stops
57 60 63
Contents
VII
4.2.1.3 Word-�nal stops 4.2.1.4 Allophones and phonological identity o lax stops 4.2.1.5 Major allophones o stops 4.2.2 Fricatives 4.2.2.1 Word-initial ricatives 4.2.2.2 Word-medial (intervocalic) ricatives 4.2.2.3 Allophones and phonological identity o the lax ricative 4.2.2.4 Major allophones o ricatives 4.2.3 Affricates 4.2.3.1 Word-initial affricates 4.2.3.2 Word-medial (intervocalic) affricates 4.2.3.3 Major allophones o affricates
4.3 Sonorants 4.3.1 Nasals 4.3.1.1 4.3.1.2 4.3.1.3 4.3.1.4 4.3.2 Liquids 4.3.2.1 4.3.2.2
65 65 68 68 71 73 74 75 76 78 79 80
80 Word-initial nasals Word-medial (intervocalic) nasals Word-�nal nasals Major allophones o nasals Phonetic realisation o liquids Phonemic realisation o /l/
4.4 Features or consonants 4.4.1 Features or manner o articulation 4.4.2 Features or place o articulation 4.4.3 Features or phonation types
81 82 83 83 83 83 83 85
88 88 90 92
4.5 Summary Exercises
93 94
Vowels 5.1 Phonological analysis o vowels
95 95
5
5.1.1 Monophthongs and diphthongs 5.1.2 Phonological status o the glide: consonant or vowel?
5.2 Monophthongs 5.2.1 How many are there? 5.2.1.1 Are there ront rounded vowels in Korean? 5.2.1.2 Are /e/ and /æ/ distinctive? 5.2.2 Te simple vowel system o Standard Korean 5.2.3 Te phonetic realisation o monophthongs
5.3 Diphthongs 5.3.1 Te diphthong // 5.3.1.1 Is // an on-glide or off-glide? 5.3.1.2 Te phonetic realisation o // 5.3.2 Te diphthongal system o Standard Korean
5.4 Features or vowels 5.5 Summary Exercises
95 95
97 97 98 99 101 102
109 110 110 111 112
115 116 117
Contents
VIII
Frequency trends of Korean sounds 6.1 Database 6
6.1.1 Dictionary 6.1.2 Spontaneous speech
6.2 Phoneme requency 6.2.1 Overall phoneme requency 6.2.2 Consonant requency 6.2.2.1 Overall consonant requency 6.2.2.2 Onset requency 6.2.2.3 Coda requency 6.2.3 Vowel requency
6.3 Syllable requency 6.3.1 High-requency syllables 6.3.2 Syllable type requency
6.4 Comparison with phoneme requency in English 6.4.1 Consonant requency in English 6.4.2 Vowel requency in English
118 119 119 119
120 120 121 121 127 130 133
137 137 140
142 142 145
6.5 Summary Exercises
150 150
Prosody 7.1 Linguistic unction o prosody 7.2 Prosodic structure o Korean 7.3 Syllable
151 151 155 156
7
7.3.1 7.3.2 7.3.3 7.3.4
7.4 7.5 7.6 7.7
Syllable structure Syllable types Building a syllable structure Syllabi�cation and phonological processes
Phonological words Phonological phrases Intonational phrases Summary Exercises
Phonological rules of Korean (I) 8.1 Phonological rules related to syllable structure constraints 8
8.1.1 Neutralisation 8.1.2 Consonant-cluster simpli�cation
8.2 Phonological rules via surace phonetic constraints 8.2.1 8.2.2 8.2.3 8.2.4 8.2.5 8.2.6 8.2.7
Post-obstruent ensi�cation Obstruent Nasalisation Liquid Nasalisation Lateralisation Non-coronalisation (Bilabialisation and Velarisation) Aspiration Similar-place Obstruent Deletion
157 160 161 161
164 165 170 176 176 178 181 181 183
187 188 189 190 192 192 193 194
Contents
IX
8.2.8 /j/ Deletion 8.2.9 /h/ Deletion
195 196
8.3 Summary Exercises
197 198
Phonological rules of Korean (II) 9.1 Background 9.2 Rules where two lexical morphemes meet
199 199 201
9
9.2.1 9.2.2 9.2.3 9.2.4
Lateralisation and Nasalisation o Sino-Korean words ensi�cation o Sino-Korean words /t/ Insertion /n/ Insertion
9.3 Rules between lexical morphemes and grammatical morphemes 9.3.1 9.3.2 9.3.3 9.3.4 9.3.5 9.3.6
Palatalisation Verbal Suffix ensi�cation: afer a nasal-ending stem Glide Formation in verbal in�ection /h/ Deletion in verbal in�ection /ɯ/ Deletion in verbal in�ection /ɑ/ or /ʌ/ Deletion in verbal in�ection
9.4 Rules relating to the word-initial liquid or nasal in Sino-Korean words 9.4.1 /l/ into /n/ in Sino-Korean words 9.4.2 /n/ Deletion in Sino-Korean words
9.5 Summary Exercises Loanword phonology 10.1 Origin and orthography o loanwords 10.2 Consonant correspondence between English and Korean 10
10.2.1 Correspondence between /p, t, k/ in English and /ph, th, kh/ in Korean 10.2.2 Correspondence between /b, d, g/ in English and /p, p*/, /t, t*/, /k, k*/ in Korean 10.2.3 Correspondence between //, /s/, /θ/, /∫/ in English and /ph/, /s, s*/, /s, s*/, /s+j, s*+j/ in Korean 10.2.4 Correspondence between /v/, /z/, /ᴣ/ and /ð/ in English and /p/, /ʨ/, /ʨ/, /t/ in Korean 10.2.5 Correspondence between /t∫/ and /dᴣ/ in English and /ʨh/ and /ʨ/ in Korean 10.2.6 Correspondence between /l/ and /r/ in English and /l/ in Korean 10.2.7 Correspondence between /j/ and /w/ in English and /j/ and /w/ in Korean
10.3 Vowel correspondence between English and Korean 10.4 Syllable structure and loanword phonology 10.5 Application o Korean phonological rules 10.5.1 Why is out pronounced as [ɑut] not [ɑus]? 10.5.2 Why is rugby pronounced as [lkp*i]? 10.5.3 Why is good morning pronounced as [kun moniŋ]?
201 203 203 206
207 207 208 209 210 210 211
211 211 211
215 215 217 218 219 220 221 221 222 223 223 224
224 225 228 228 228 229
X
Contents
10.5.4 Why is Hamlet pronounced as [hεmnit]? 10.5.5 Why is Finland pronounced as [phillantɯ]?
10.6 Summary Exercises Notes Bibliography Index
229 229
230 230 232 238 242
Figures
Figure 1.1 Figure 1.2 Figure 1.3 Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4 Figure 2.5 Figure 2.6 Figure 2.7 Figure 2.8 Figure 2.9 Figure 2.10 Figure 2.11 Figure 2.12 Figure 2.13 Figure 2.14 Figure 2.15 Figure 2.16 Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 XI
Te Korean language in East Asian history Distribution o the Korean language outside Korea Map o Korean dialects Te vocal organs View o the larynx rom above Subdivisions o the tongue Subdivisions o the roo o the mouth Te three major cavities involved in speech production wo passages o air�ow rom the glottis Schematic view o laryngeal con�gurations Schematic view o the vocal tract or stops and nasals Differences in the articulation o stops, taps and trills Differences in the articulation o stops, ricatives and approximants Difference between median and lateral articulations Schematic view o the differences between the articulation o stops and affricates Phonation types o stops during closure and release phases ongue height ongue backness Roundedness Spectrogram and waveorm o lax, tense and aspirated alveolar stops in word-initial position Spectrogram and waveorm o lax, tense and aspirated alveolar stops in word-medial (intervocalic) position Spectrogram and waveorm o the lax alveolar stop in word-medial (intervocalic) position Spectrogram and waveorm o the lax alveolar stop in word-�nal position Spectrogram and waveorm o the word /k ɑk ɑk/ ‘corner o the road’
pag e 2
4 14 17 17 18 19 19 20 24 29 30 30 31 31 33 36 37 37 61 64 65 66 66
Figures
XII
Figure 4.6 Figure 4.7 Figure 4.8 Figure 4.9 Figure 4.10 Figure 4.11 Figure 4.12 Figure 4.13 Figure 4.14 Figure 4.15 Figure 4.16 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 5.7 Figure 6.1 Figure 6.2 Figure 6.3 Figure 6.4 Figure 6.5 Figure 6.6 Figure 6.7 Figure 6.8 Figure 6.9 Figure 6.10 Figure 6.11 Figure 6.12
Spectrogram and waveorm o ricatives in word-initial position Spectrogram and waveorm o a tense alveolar ricative ollowed by the vowels / ɑ/ and /i/ in word-initial position Spectrogram and waveorm o a glottal ricative ollowed by the vowels /ɑ/, /i/, and /u/ in word-initial position Spectrogram and waveorm o ricatives in word-medial (intervocalic) position Spectrogram and waveorm o lax, tense and aspirated affricates in word-initial position Spectrogram and waveorm o lax, tense and aspirated affricates in word-medial (intervocalic) position Spectrogram and waveorm o bilabial and alveolar nasals in word-initial position Spectrogram and waveorm o bilabial and alveolar nasals in word-medial (intervocalic) position Spectrogram and waveorm o bilabial, alveolar and velar nasals in word-�nal position Spectrograms and waveorms o liquids in word-initial, word-medial (intervocalic) and word-�nal positions, and afer /l/ Place o articulation and eature value o the [anterior] eature Allophonic realisation o /l/ ollowed by glides in Korean words /pʰil + jo/ ‘necessity’ and /mil + w ʌl / ‘honeymoon’ Spectrogram o ‘outside’ and ‘stomach’ Formant plot o and (ten male speakers) Formant plot o and (ten emale speakers) Formant chart (ten male speakers) Formant chart (ten emale speakers) Comparison o ormant charts or English (RP) and Standard Korean Frequency trends by rank (all phonemes, dictionary vs. speech) Consonant requency (dictionary) Consonant requency (speech) Frequency trends by ranking (consonants, dictionary vs. speech) Consonant requency differences (dictionary vs. speech) Onset requency (dictionary) Onset requency (speech) Onset requency differences (dictionary vs. speech) Coda requency (dictionary) Coda requency (speech) Coda requency differences (dictionary vs. speech) Vowel requency (dictionary)
72 72 73 74 78 79 82 83 84 85 91 96 98 100 101 103 104 105 123 123 124 124 125 127 128 128 131 131 132 133
XIII
Figure 6.13 Figure 6.14 Figure 6.15 Figure 6.16 Figure 6.17 Figure 6.18 Figure 6.19 Figure 6.20 Figure 6.21 Figure 7.1 Figure 7.2 Figure 7.3 Figure 7.4 Figure 7.5 Figure 7.6 Figure 7.7 Figure 7.8
Figure 7.9 Figure 7.10 Figure 7.11 Figure 7.12 Figure 7.13 Figure 10.1
Figures
Vowel requency (speech) Vowel requency trends by ranking (dictionary vs. speech) Vowel requency (dictionary vs. speech) Consonant requency o English (dictionary) Consonant requency o English (speech) Differences in English consonant requency (dictionary vs. speech) Vowel requency in English (dictionary) Vowel requency in English (speech) Vowel requency in English (dictionary vs. speech) Stylised pitch movements in (5a–c) Prosodic structure o Korean Syllable structure o Korean Syllable structures or /h ɑlɑpʌʨi/, /hɑlmʌni/ and /ʨoŋihɑk/ Syllabi�cation o /k *oʨʰ/ and /k*oʨʰ + -i/ Syllabi�cation o /k ɑps/ and /k ɑps + -i/ wo pitch curves o /(mij ʌninɛ)ω (ʌmʌninɯn)ω (nɑ jʌnilɯl)ω (miw ʌhɑni)ω?/ spoken by a emale speaker o Standard Korean A pitch curve o /(k *ɑtɑloun)ω| (ʌmʌninɯn)ω| (hɑnɯn)ω (ili)ω| (nʌmunʌmu)ω| (mɑnthɑkonɯn)ω| (hɑsitʌnjɑ)ω?/ spoken by a emale speaker o Standard Korean ypical pitch patterns o phonological phrases with our syllables in Standard Korean ypical pitch patterns o phonological phrases with ewer than our syllables in Standard Korean ypical pitch patterns o phonological phrases with more than our syllables in Standard Korean Number o syllables per phonological phrase in spontaneous and read speech Pitch curves o two types o intonation patterns o [mijʌninɯn| ʨʰinkutɯlɯl| miw ʌhɛ jo] ‘Miyoen hates her riends’ Syllabi�cation o strike
134 135 135 143 143 145 146 146 148 154 156 158 162 163 163 167
168 169 169 169 170 172 227
Tables
able 1.1 able 1.2 able 1.3 able 2.1 able 2.2 able 2.3 able 3.1 able 3.2 able 4.1 able 4.2 able 4.3 able 4.4 able 4.5 able 4.6 able 4.7 able 4.8 able 4.9 able 4.10 able 4.11 able 4.12 able 4.13 able 4.14 able 4.15 able 5.1 able 5.2 able 5.3 able 5.4 able 5.5 XIV
Comparison o three romanisation methods Lexical differences between South and North Korea Examples o compound and derivational words Patterns o initiation Summary o places o articulation Summary o manners o articulation Major class eatures I Major class eatures II Consonants Stops Voiceless alveolar stops in Korean and English Major allophones o stops Fricatives Major allophones o ricatives Affricates Major allophones o affricates Sonorants Major allophones o nasals Major allophones o the liquid Classi�cation o Korean consonants using a eature matrix Features or place o articulation Features or phonation type Features or consonants in Korean Te ten monophthongs o SKP Eight monophthongs o SKP (without /y/ and / ø/) Monophthongs in Standard Korean Mean and standard deviation o F1 and F2 o ten male and emale speakers’ phonetic realisations o seven different monophthongs Te eleven diphthongs o SKP
page 5
8 9 21 28 32 53 53 57 58 59 69 69 77 77 80 81 84 86 89 91 92 92 98 99 102 103 109
Tables
XV
able 5.6 able 5.7 able 5.8 able 5.9 able 6.1 able 6.2 able 6.3 able 6.4 able 6.5 able 6.6 able 6.7 able 6.8 able 6.9 able 6.10 able 6.11 able 6.12 able 6.13 able 6.14 able 6.15 able 6.16 able 6.17 able 6.18 able 6.19 able 6.20 able 6.21 able 6.22 able 6.23 able 6.24 able 6.25 able 6.26 able 6.27 able 6.28 able 6.29 able 6.30
Te ten diphthongs o Standard Korean (tentative) Diphthongs in Standard Korean Features or vowels Features or vowels and glides Sound distribution (dictionary) Sound distribution (speech) Phoneme requency (dictionary) Phoneme requency (speech) ype and token requency o consonants (sonorants vs. obstruents and phonation types) ype and token requency o consonants by place o articulation ype and token requency o consonants by manner o articulation ype and token requency o onsets (sonorants vs. obstruents and phonation types) ype and token requency o onsets by place o articulation ype and token requency o onsets by manner o articulation ype and token requency o codas (sonorants vs. obstruents) ype and token requency o codas by place o articulation ype and token requency o codas by manner o articulation ype and token requency o vowels (monophthongs vs. diphthongs) ype and token requency o vowels by vowel height ype and token requency o vowels by tongue backness ype and token requency o vowels by roundedness Syllable requency (dictionary) Syllable requency (speech) High-requency syllables (dictionary and speech) oken and type requency o syllable types (dictionary vs. speech) oken and type requency o syllable types with/without onsets oken and type requency o syllables with/without codas ype and token requency o English consonants (sonorants vs. obstruents) ype and token requency o English consonants by place o articulation ype and token requency o English consonants by manner o articulation ype and token requency o English vowels (monophthongs vs. diphthongs) ype and token requency o English vowels by vowel height ype and token requency o English vowels by backness ype and token requency o English vowels by roundedness
109 112 115 116 119 120 121 122 126 126 126 129 129 130 132 132 133 136 136 136 136 138 139 140 141 141 142 144 144 144 147 147 147 147
Tables
XVI
able 6.31 able 6.32 able 7.1 able 8.1 able 8.2 able 8.3 able 8.4 able 9.1 able 9.2 able 9.3 able 9.4 able 9.5 able 9.6 able 9.7 able 9.8 able 9.9 able 10.1 able 10.2
Comparison o phoneme requency in Korean and English Comparison o consonant requency in Korean and English Statistical inormation relating to prosodic units Pronunciation o syllable-�nal consonants Pronunciation o consonant clusters in syllable-�nal position Dialectal differences in consonant cluster realisation Phonological rules o Korean (I) ense vs. lax contrast or words with the same orthography Phonetic realisation o /nl/: Lateralisation vs. Nasalisation ensi�cation via /t/ Insertion variation among subordinated compounds Pronunciation o days o the week showing /n/ Insertion in Gyeongsang dialect Pronunciation o word-initial and medial /l/ Pronunciation o word-initial and medial /n/ Tree types o application o the rules governing word-initial /l/ and /n/ Sino-Korean vocabularies with word-initial /l/ or /n/: South vs. North Phonological rules o Korean (II) Consonant correspondence charts o English and Korean Vowel correspondence examples between English and Korean
149 149 173 182 184 186 197 200 202 205 206 212 212 212 213 214 220 225
Preface
In learning a oreign language, it is essential to amiliarise onesel with the sound properties o the language such as the inventory o sounds and phonological processes. Native speakers will acquire these naturally, but learners o the second language must learn them ormally. Tis book aims to provide an introduction to the phonetics and phonology o Korean or those who study the Korean language and/or linguistics. More speci�cally, the target audience o this book consists both o academics in Korean Linguistics (or in Korean Studies or in linguistics) and o learners o Korean. Among the learners o Korean, this book is targeted particularly at learners o Korean who are amiliar with English. Hence, comparison with English is also provided whenever necessary. Tis book is rare in terms o being well balanced or these two audience groups. We believe that it will become the �rst comprehensive book to provide a complete and authoritative description and explanation o the current state o Korean phonetics and phonology. Previous knowledge o the Korean language or linguistics will be bene�cial to readers. We have provided the Korean alphabet with the revised Korean government romanisation or the IPA (International Phonetic Alphabet) transcription. We have adopted IPA (International Phonetic Alphabet) description at the phonemic level or each word. Knowing the Korean alphabet will be a great help in ollowing this book. Te website (www.cambridge.org/shin) will o course be a very useul/ practical guide to the sound o Korean. Ample examples rom the real use o contemporary spoken Korean orm the key eature o this book – making it accessible to all those who are interested in the Korean language. Each chapter contains two kinds o exercises – �rstly, those which help the learner get used to practical pronunciation (i.e., experiment yoursel/perormance exercises); and secondly, exercises which encourage learners to evaluate XVII
XVIII
Preface
what they have learned about the sounds o Korean (i.e., content exercises). Not only this, each chapter has a very intriguing ‘Did you know?’ question, so that the learner can easily be exposed to how the Korean language is used on a day-to-day basis. In this book, we ocus on Standard Seoul Korean pronunciation. Yet, whenever necessary, we have also included the synchronic and diachronic realisation o the sounds. Te majority o Korean grammars and textbooks only provide a ‘normative’ or ‘written’ grammar, ignoring how the language is truly ‘spoken’. Our book ocuses on Seoul Korean speakers’ ‘real’ pronunciation unless stated otherwise, but whenever comparison is needed, we also provide inormation regarding differences between dialects. Finally, we would like to thank our husbands – Dr DoneSik Yoo, Dr Ian Kiaer and Dr Naehyun Kwon – or their ull support in the process o writing this book. We are also grateul to Sookyoung Kwak, Marshall Craig, Deborah Smith and Matthew Hunter or their various editorial help. We thank our copy-editor Anna Oxbury or her wonderul work, and Helen Barton or her patience and encouragement throughout the project.
Notational conventions
. | ∥ () + # ω
syllable boundary minor break boundary (phonological phrase boundary) major break boundary (intonational phrase boundary) eojeol ‘word phrase’ unit bound morpheme morpheme boundary word boundary phonological word boundary
Consonants
XIX
Korean alphabet
IPA
Korean alphabet
IPA
Korean alphabet
IPA
k
ŋ
k *
n
ʨ
t*
t
ʨʰ
p*
l
kʰ
s*
m
tʰ
ʨ*
p
pʰ
s
h
XX
Notational conventions
Vowels Korean alphabet
IPA
Korean alphabet
IPA
ɑ
u
jɑ
ʌ
Korean alphabet
IPA
wɛ
ju
wi
ɯ
wɑ
jʌ
i
wʌ
o
ɛ
ɰi
jo
jɛ
Except in special cases, all phonetic notation in this book uses broad transcription at the phonemic level. We ollow in this book the Revised ROK Romanisation Convention.
Characteristics of the Korean language
1
In this chapter, we will provide an overview o the Korean language and brie�y discuss its main characteristics. In 1.1, we will discuss the origin, history and distribution o the Korean language; in 1.2, the Korean alphabet and its romanisation will be discussed; 1.3 ocuses on the characteristics o the Korean lexicon; in 1.4, the structural characteristics o Korean will be explored; and in 1.5, the socio-pragmatic characteristics o Korean will be discussed.
1.1
Origin, history and distribution
1.1.1
The Korean language in East Asian history It is impossible to think about the history o Korea without considering the history o Northeast Asia. In the same vein, the history o the Korean language cannot be considered without reerence to the in�uence o Korea’s neighbours; namely, China, Japan and Mongolia. Figure 1.1 shows how the Korean language has evolved rom Old Korean into Contemporary Korean within the bigger picture o East Asian history. Te classi�cation is based on K.-M. Lee (1998). As seen in Figure 1.1, social and political changes at home and abroad became the crucial actor in shaping the Korean language. For instance: the uni�cation o the Tree Kingdoms (676) resulted in the Silla language, the �rst uni�ed language on the Korean peninsula; later on, the establishment o the Koryo dynasty (918) gave rise to the central dialect o Korean, which became the basis o modern Korean; the Imjin War (1592–8) marks the division between late Middle Korean and Modern Korean; and lastly, the Korean War (1950–3) yielded the language division between North and South Korea.
1
Characteristics of the Korean language
2
5000 BC
3 00 B C
0 A D 100 AD
200 A D
600 AD 700 A D 800 AD 90 0 AD 1 000 AD 11 00 A D 1 200 A D 1300 AD 1 400 A D 1 500 A D 160 0 AD 1 700 AD 18 00 A D 1 900 AD 2000 AD
Korea 5000 BC Proto-Altaic
0 AD Old Korean
1592 AD Modern Korean
935 AD Middle Korean
Contemporary Korean
18 AD Paekche 37 AD Koguryo 57 AD Silla 676 AD United Silla 918 AD Koryo
1392 AD Choson
China 206 BC Han Dynasty
618 AD Tang Dynasty
960 AD Song Dynasty
1368 AD Ming Dynasty
1644 AD Qing Dynasty
221 BC Qin Dynasty Japan 250 AD Yamamoto Period
710 AD Nara Period
794 AD Heian Period
1336 AD Muromachi Period
1603 AD Edo Period
1185 AD Kamakura Period
Figure 1.1
1.1.2
The Korean language in the East Asian history
Where is the Korean language from? Te origin o the Korean language is still not clearly known. Although there were some predecessors, it was the Finnish linguist Ramstedt (1873–1959) who �rst proposed the genetic affinity between Korean and Altaic languages such as Manchu, Mongolian, ungus and urkish, through a systematic comparison. Tese languages share grammatical properties with Korean such as agglutinative morphology; that is, grammatical relations such as a subject or an object are mainly realised by attaching (or ‘gluing’) particles to nominal expressions. Nevertheless, it is hard to prove this genetic affinity with Altaic languages due to the lack o reliable evidence.
1.1.3
Korean vs. Chinese and Korean vs. Japanese: are they related, and if so, how? Korea and Japan, under the umbrella o the Chinese cultural sphere, have not only shared socio-cultural heritages, but also a shared linguistic heritage. Tis is represented in the lexicons o the Korean and Japanese languages. As we will see in 1.3, roughly 57 per cent o the Korean vocabulary is Sino-Korean and derived rom Chinese. Yet structurally, the two languages are completely unrelated.
3
1.1
Origin, history and distribution
Chinese has a strict subject–verb–object word order and does not have grammatical particles like those ound in Korean. Korean and Japanese, however, share a great deal o structural similarity. For instance, Japanese and Korean share an almost identical particle system. Nevertheless, it is still debatable whether Korean and Japanese belong to the same language amily. Vovin (2008) recently argued that there is no genetic relation between the two languages. Once again, however, a lack o reliable evidence makes it difficult to prove any linguistic affinity. Korean as a global language: is the Korean language for the Korean peninsula only? 1.1.4
Korean is no longer simply the language o the Korean peninsula, nor simply the language o erra Incognita. Tis is due to the increase o the Korean ‘diaspora’, now consisting o roughly 7 million people. Tey include both descendants o early emigrants rom the Korean peninsula, as well as more recent emigrants. Most o them live in China (2.34 million), the United States (2.1 million) and Japan (0.9 million). As a result, Korean is increasingly more widely spoken. With 79 million people speaking it across the globe (48.6 million in South Korea, 23.8 million in North Korea, 7 million overseas according to Statistics Korea), Korean is now the seventeenth most widely used amongst all the world languages (Ethnologue, 2008, www.ethnologue.com/web/asp). Figure 1.2 shows the distribution o the Korean language outside Korea as o 2010. Education in the Korean language overseas has also been expanded over the last ew decades. For instance, the number o students in the USA choosing Korean as their SA (Scholastic Aptitude est) oreign language has been steadily on the rise in recent years. In 2007, the Korean Embassy in the USA reported that Korean was the ourth most popular oreign language chosen by SA students. Economic development, cultural exchange and trade also provide motives or oreigners to learn Korean. From the late 1990s, there has been a rapid increase in the in�ux o oreign labour into Korea, particularly rom South Asia, as well as an increase in international marriages between Korean men and South Asian women. According to a recent survey by Statistics Korea, the number o students (o primary to high-school level) with oreign mothers was about 18,778 in 2009. Korean pop culture, or Hallyu, has also played an important role in raising interest in the Korean language and culture, particularly in China, Japan and Southeast Asia.
Characteristics of the Korean language
4
Russia 222 thousand Europe 118 thousand
China 2,337 thousand
North America 2,325 thousand Japan 592 thousand
Central and West Asia 327 thousand
Southeast Asia 304 thousand
Central and South America 107 thousand Oceania 157 thousand
Figure 1.2
1.2
Distribution of the Korean language outside Korea
Korean alphabet and romanisation
1.2.1
Hangeul: the Korean alphabet Hangeul , invented in 1443, is the unique alphabet used to write the Korean language. Hangeul is a phonemic alphabet; in other words, there is one-to-one correspondence between a phoneme and a letter. Consonants and vowels are given in alphabetic order in (1). Tere are twenty-our basic letters and sixteen complex letters. Among the twenty-our basic letters, ourteen are consonants and ten are vowels. (1)
Korean alphabet a. Basic letters or consonants (14) , , , , , , , , , , , , , b. Basic letters or vowels (10) , , , , , , , , , c. Complex letters or consonants (5) , , , , d. Complex letters or vowels (11) , , , , , , , , , ,
Systematic correspondence is observed between letters. For instance, an addition o a stroke makes a lax sound into an aspirated sound (e.g., → , → , →), whilst consonant doubling results in the tensing o sounds
5
1.2
Korean alphabet and romanisation
Table 1.1 Comparison of three romanisation methods
Word
M-R
Revised ROK Yale
Meaning
p’yŏngga
pyeongga
phyengka valuation, rating
chilli
jilli
cinli
truth, act
haetpyŏt
haetbyeot
hayspyet
sunlight
ch’ŏ tsarang cheotsarang
chessalang a �rst love
hyŏphoe
hyeopheo
hyephoy
a society, association, league
hakhoe
hakheo
hakhoy
an institute, academy
pibimpap bibimbap
pipimpap common Korean dish (rice topped with vegetables, usually with an egg and mincemeat)
tŏuk
deouk
tewuk
still more
yŏdŏl
yeodeol
yetelp
eight
halta
halda
halthta
lick
modŭn
modeun
motun
all, whole, every, each
chaemi
jaemi
caymi
interest
yŏgi
yeogi
yeki
here; a hobby
Sŏul
Seoul
Sewul
Seoul (Korea’s capital)
segi
segi
seyki
a century
ppalli
ppalli
ppalli
ast, rapidly, quickly
chaju
jaju
cacwu
ofen, repeatedly
kotkam
gotgam
kockam
dry persimmons
(e.g., → , → , →). Lastly, the sounds that are pronounced in the same place o articulation show visual similarity (e.g. ( ), ( ), ( ), ( , )). In 1933, the Korean Language Council ( Joseoneohakhoe ) decided that words should be spelled as they sound, but should also conorm to grammatical principles. Tey also decided that there should be a space between each word in the sentence and the particles attached to the previous word. Tese two rules became the basis o orthography in Contemporary Korean. Tis means that although the noun saram ‘person’ plus a subject particle - -i is pronounced [s ɑ .ɾa.mi], the written orm remains , respecting the morphological combination o noun plus particle. Tis is in contrast to Middle Korean orthography, where Korean is written as it is pronounced.
Characteristics of the Korean language
6
1.2.2
Romanisation: how many styles and how different is each option? Tere are three ways to romanise the Korean language: the McCune-Reischauer (M-R) system, the Revised Republic o Korean (ROK) system, and the Yale system. Inside Korea, the Revised ROK system is invariably used. (See www. korean.go.kr/09_new/dic/rule/rule_roman_0101.jsp or the Revised ROK Romanisation system.) Outside Korea, the McCune-Reischauer system and Yale system are mainly used. Whilst the McCune-Reischauer system respects the actual pronunciation o Korean, the Yale system ollows the original morphological orm. Te ormer system is used by most Koreanologists and Korean studies authorities including the Library o Congress. Te latter system is mainly used by linguists. Te ollowing table shows how the three systems romanise each sample word.
1.3
Lexicon
1.3.1
What is the proportional ratio between Sino-Korean and pure Korean words? In terms o vocabulary, the Korean language has been heavily in�uenced by the Chinese language. (Tis is also true or Japanese.) Most conceptual or proessional terms are Sino-Korean. Te more basic terms, however, tend to be pure Korean. According to the Standard Korean Dictionary edited by the National Institute o Korean Language (NIKL, 2000), and containing some 440,000 words, the ratio o (i) pure Korean (PK) words; (ii) Sino-Korean (SK) words; (iii) other oreign loanwords is 25.28 : 57.12 : 17.6.
1.3.2
Increase of English loanwords Te proportion o loanwords is closely related to a nation’s socio-cultural and political situation. As western in�uence grows rapidly in South Korea, the number o English loanwords has risen dramatically in the last �fy years. Te number o loanwords became one o the main causes o discrepancy between the North and South Korean languages. Whilst the South has adopted English loanwords, the North has replaced them with PK words wherever possible.
1.3.3
New words in the twenty-first century A language’s lexicon vividly re�ects the socio-cultural change o a particular society. Tis is also the case in the Korean lexicon. Korea University’s Korean Language Dictionary , published in 2009, contains words such as those presented in (2)–(4). Te words in (2) are inormation technology-related terms that have appeared in the last ten years. Te vocabulary in (3) re�ects the
7
1.3
Lexicon
socio-cultural aspects o present-day Korean society. For instance, gireogiappa ‘Wild Goose Father’ in (3) means a ather who sends his wie and children abroad to urther the children’s early oreign-language education, whilst he himsel remains in Korea to earn money to send to the amily. As shown in (4), some terms are English in origin, but are only used in Korea. (2) ak-peul ‘internet bullying’, net-maeng ‘internet-illiterate’, spam-mail ‘e-mail spam or junk email’, akseong-kodeu ‘malignant-code’, pro-gamer ‘proessional gamer’ (3) gireogiappa ‘wild goose ather’, kkonminam ‘pretty-boy’, saengeol ‘makeupree ace’, bihogam ‘muddied reputation’ (4) one-shot ‘bottoms up!’, skin-ship ‘physical contact’, cunning ‘cheating’ How many blues and blacks? Colour terms are well developed in Korean. Consider (5)–(6). All words in (5) reer to the colour blue and (6) reer to the colour black. (5) Blue , , , , , , , , , , (6) Black , , , , , , , , , , , , , ,
1.3.4
Motion- and sound-symbolic words Mimetic words (i.e., motion-symbolic words) and onomatopoeic words (sound-symbolic words) are also well developed in Korean. According to NIKL, the Korean language contains some 2,900 motion- or sound-symbolic words. Consider (7) and (8). (7) . (sound-symbolic) Jina-���� spit-��� kkulkkeok swallowed. ‘Jina swallowed with a gulp.’
Characteristics of the Korean language
8
Table 1.2 Lexical differences between South and North Korea
Sino-Korean vocabulary
English loanwords South (EL) North (PK)
English gloss
South (SK) North (PK)
English gloss
knock
joint (bone)
record
a tall tree
spray
ability
syrup
sterilisation (pasteurisation)
jelly
sterilisation
sponge cake
wintering
curtain
labour orce
corner kick
harvest
hook
deluge, �ood
(8) . (motion-symbolic) Jina-���� nageunnageut (tenderly) smile-��� made. ‘Jina smiled tenderly.’
1.3.5
Can South Koreans and North Koreans understand each other? Te answer to this question is yes. However, it is not easy or South and North Koreans to understand each other one hundred per cent. Tis is due to the difference in vocabulary. As mentioned above, whilst the South has adopted English loanwords (EL) as they are, the North has invented corresponding PK words instead. In the North, not only EL, but also many SK words were replaced with PK words as shown in able 1.2.
1.4
Structural properties of Korean
1.4.1
How are words and sentences composed? Te smallest meaning-bearing unit in Korean is called a morpheme. Morphemes are divided into free and bound morphemes according to whether the morpheme can be used independently o any other host category or not. At the same time, a morpheme is classi�ed according to whether it has lexical meaning, or whether it only represents grammatical meaning. For instance, the verbal suffix -- cannot be used on its own (i.e., it is a bound morpheme),
1.4
9
Structural properties of Korean
Table 1.3 Examples of compound and derivational words
Derived words
Simple words
Pure Korean ‘hand’ ‘love’
Pre�x
Suffix
Compound words
- + → ‘bare hands’ - + → ‘�rst love’
+ - → ‘scissoring’ + - → ‘dandy’
+ → ‘chestnut tree’ + → ‘love quarrel’
() + -() → ‘the world o education’ () + -() → ‘home use’
() + () → ‘primary school’ () + ( )→ ‘political structure’
Sino() ()- + () Korean ‘river’ → ‘great senior’ () ‘mountain’ ()- + () → ‘incompletion’
but it contributes to the meaning o the past tense (i.e., it is a grammatical morpheme). Verbal stems contribute to the meaning o a word (i.e., they are lexical morphemes), but cannot be used on their own (i.e., they are bound morphemes). Te smallest ree or independent grammatical unit is a word. Korean words can have one o the ollowing structures: (9)
ypes o words a. simple word: root (e.g., ‘tree’) b. derived word: pre�x + root (e.g., -+ ‘�rst-love’) c. root + suffix (e.g., +- ‘scissoring’) d. compound word: root + root (e.g., + ‘chestnut tree’)
able 1.3 shows examples o compound and derivational words. Te next grammatical unit afer a word is the eojeol ‘word-phrase’. Word phrases are separated by a space in Korean orthography, and each is composed o a lexical morpheme (e.g., a noun) and a grammatical morpheme (e.g., a particle). Simply speaking, a word phrase is a basic grammatical unit that can unction as a subject or object within a sentence. Note that in (10) the word itsel does not have any grammatical role, but with the case particle – attached, it becomes the subject. (10) - - - . Jina-���� morning-at milk-��� drank ‘Jina drank milk in the morning.’
10
Characteristics of the Korean language
A nominal expression with a particle attached orms one word phrase, and a verb plus verbal suffixes also orms one word phrase, regardless o the number o suffixes attached. In Korean orthography, each word phrase is individually spaced. Consider (11). L stands or lexical morpheme, F stands or ree morpheme, G stands or grammatical morpheme and B stands or bound morpheme. Tereore (11) consists o three word phrases and seven morphemes. (11) Word phrase {} Subject Morpheme L/F-G/B
{} Object L/F-G/B
{} ‘Jina is having a meal.’ Verb L/B-G/B G/B
It is still an unsettled issue whether to regard a particle as an independent word or not. In South Korean grammar it is considered an independent word, but in North Korean grammar it is not. However, we will not dwell on this issue in this book. Word phrases whose host categories are nouns take particles, whereas word phrases whose host categories are verbs take in�ectional suffixes. We will return shortly to discussion o particles and suffixes. Word phrases orm a longer phrase such as a noun phrase or verb phrase, and these longer phrases then constitute a clause and �nally a sentence. 1.4.2
Is the word order rigid or free? Te word order in Korean is reer than in English, the only general rule being that the verb tends to come at the end o a sentence. Yet there are cases where the word order is more rigid; or instance, a modiying expression will always precede the noun being modi�ed. Consider (12). * denotes an ungrammatical sentence. (12) a. - - . Yuna-���� new shoes-��� put-on ‘Yuna put on new shoes.’ b. *- - . Yuna-���� shoes new-��� put-on (modi�ee precedes modiying expression)
If it is not through word order, how are grammatical roles and relations expressed? 1.4.3
Grammatical unctions are realised by ‘attaching’ or ‘gluing’ particles to the content words. For instance, regardless o the location o -, the agent
11
1.4
Structural properties of Korean
or subject o the sentence is the same; that is, in (13) is clearly the subject because o the subject particle -. (13) a. - - - . Jina-���� morning-at milk-��� drank ‘Jina drank milk in the morning.’ b. - - - . milk-��� Jina-���� morning-at drank ‘Jina drank milk in the morning.’ c. - - - . Morning-at milk-��� Jina-���� drank ‘Jina drank milk in the morning.’ Tis grammatical characteristic is shared by the so-called Altaic languages such as Mongolian, urkish, ungus and Manchurian. In the ollowing section, we will brie�y discuss the inventory o particles (or nominal expressions) and in�ectional suffixes (or verbal expressions).
Particles: What are they and how are they used? Particles that are attached to nouns can be classi�ed into two groups: (i) case particles; (ii) additional particles. Tere are no words in English that directly correspond to the subject/object particles ound in Korean. Prepositional/ad verbial particles, however, can be easily translated into in/on/at or to/from/ towards in English. Additional particles in Korean either add some additional meaning to the host noun or connect two nouns. A list o the most requently used particles according to B.-M. Kang and H.-G. Kim (2009)’s Sejong Corpus search is given below: (14) a. Case particles (i) Subject/object particles: -, -, -/ (ii) Prepositional/adverbial particles: -, -, -, -, -(), -, -, -, - b. Additional particles (i) Additional meaning particles: -/-, -, - (ii) Connective particles: -/, -(), -, - A word phrase whose stem is a verb can have a series o in�ectional suffixes. (In�ectional suffixes are also reerred to as endings in this book.) Tese in�ectional suffi xes can be classi�ed as ollows:
12
Characteristics of the Korean language
(15) Classi�cation o in�ectional suffixes (endings) I. Pre-�nal II. Final A. Sentence-�nal B. Non-sentence �nal (i) Conjunctive (ii) Function-converting a. Nominalising b. Adnominalising In the ollowing section, we will ocus on pre-�nal and sentence-�nal endings. 1.4.4
How is the speaker’s attitude expressed in Korean? In English, the speaker’s attitude is expressed through auxiliary verbs such as would , could , may , might , should etc. at the early stage o a sentence, as in (16). In Korean, the speaker’s attitude is expressed by a pre-�nal ending or an auxiliary verb that is attached by a connective to a main verb, as in (17). (16) a. Would you give me that book? b. You should come to the class. -. (17) a. - Birthday party -at come-please ‘Please, come to the birthday party.’ b. --? book please give -��� /��� -� ‘Will you give me the book, please?’ (18) shows the different kinds o pre-�nal sentence ending suffi xes and their primary unctions. Tese suffixes come right afer the verb stem but beore the sentence-�nal ending. (18) a. Honori�cation suffixes: -E.g. (to come) (to come; subject honori�cation) b. emporal/Aspectual suffixes: --/-- (past tense), --/-- (past perect) E.g. (to eat) (to eat; past tense) c. Modal suffixes: –- (volition) E.g. (to block) (to block; volition)
1.5
13
Socio-pragmatic characteristics of Korean
Where more than one suffi x is required, there is a speci�c order determining which suffix goes where in relation to the other suffi xes. Hence, the example in (19) is ungrammatical. (19) *---- (instead o ----) 1.4.5
How is the mood of a sentence realised? Sentence-�nal endings decide the mood o a sentence (e.g., declarative, imperative, question interrogative, etc.). (20) is a list o requently used sentence�nal endings according to B.-M. Kang and H.-G. Kim (2009)’s Sejong Corpus search. Even where sentence endings are morphologically identical, different moods can be realised by changing one’s intonation. A rising tone implies a question, whereas a alling tone normally implies a statement. In Chapter 7 , we will discuss the role o prosody in realising mood. (20) a. Declarative: -, -/, -/, -/, -/, -, -, -, -, - b. Exclamatory: -, -/, -/, -, -, -/ c. Interrogative: -/, -/, -, -, -/, -/, - d. Imperative: -/, -, - e. Propositive: -, -/, -/
1.5
Socio-pragmatic characteristics of Korean Unlike English, in Korean it is impossible to ‘�nish’ a sentence without knowing who is speaking. For instance, (21) shows a conversation between a boy and his neighbour, who is much older than him. Shifing the verb endings between the two parties would make the conversation sound rather odd. Although Korean society is becoming less hierarchical, misusing honori�cation can still insult the listener. Tereore (22) sounds inappropriate. (21) Old man (in the neighbourhood): ? ‘Have you eaten?’ Boy: , . ‘Not yet.’ (22) Boy: ? ‘Have you eaten?’ Old man (in the neighbourhood): ??? Tere are three kinds o honori�cation in Korean: subject, hearer and object honori�cation. Subject honori�cation is realised by the use o the pre-�nal ending -- as in (23).
14
Characteristics of the Korean language
Hamgyeong dialect Pyeongan dialect
Central District dialect
Standard Korean
Chungcheong dialect
Gyeongsang dialect
Jeolla dialect
Jeju dialect
Figure 1.3
Map of Korean dialects
(23) Subject honori�cation - - ---. Guests-���� home-at go-���-����-���� ‘Guests went home.’ Hearer honori�cation is now the most commonly used and is realised by different endings. (24a) shows high respect o the speaker towards the hearer, while (24b) reveals the hearer being lower than the speaker. (24) Hearer honori�cation a. -. ‘Please diligently study.’ b. -. ‘Please diligently study.’
15
1.6
Dialectal variation
Object honori�cation is rarely used; only its trace can be ound in the use o case particles such as -, or in certain words as in (25). (25) Object honori�cation a. Adverbial particle: - (← – ) b. Noun: (← ) ‘meal’, (← ) ‘house’ c. Verb: (← ) ‘to see’, (← ) ‘to eat’, (← ) ‘to ask’, (← ) ‘to give’ d. - - - --. Jina-���� grandather-to letter-��� give(���)-����-���� ‘Jina gave a letter to her grandather.’ Factors such as age, power and amiliarity determine whether honori�cation is required, and what degree o honori�cation is appropriate. 1.6
Dialectal variation We saw earlier that North and South Korean speakers can understand each other, although there is some difference between their vocabularies. How about the speakers within the South or within the North? Can people rom Seoul understand people rom Jeju or Pusan without diffi culty? Te answer is no. Tis is mostly due to the difference in prosody between dialects. Consider the dialectal map o Korea in Figure 1.3. As shown above, the sounds o Korean have many dialectal variants. Tis book, however, will concentrate on the Standard Seoul Korean dialect. In Chapter 7, we will discuss the prosodic difference between dialects urther. Now, let’s start our journey into the world o Korean sounds.
2
Production of sounds
Some o you may not be amiliar with Korean and may never have even heard the Korean language spoken. Fortunately or you, the vocal organs used to pronounce both Korean and English are almost the same. Moreover, the process o sound production is known to be universal and does not differ rom one language to another. We believe that understanding the general process o sound production will help greatly the reader’s understanding o the sounds o Korean and make you eel more amiliar with the sounds o the Korean language. Hence, in this chapter, we will discuss the vocal organs that are used in speech production in 2.1, the major cavities in 2.2, the processes o speech production in 2.3 and ways o classiying speech sounds in 2.4.
2.1
Vocal organs
2.1.1
Overview In this section, we will examine the various vocal organs involved in the production o speech. Figure 2.1 lists the vocal organs involved in speech production.
2.1.2
Larynx Te larynx is a common name or part o the vocal organs which extends vertically rom the inerior border o the cricoid cartilage to the hyoid bone. Te larynx is made up o �ve important cartilages: the thyroid cartilage, the cricoid cartilage, the epiglottis and the paired arytenoid cartilages. Figure 2.2 shows the larynx as viewed rom above. Te vocal olds are composed o ligaments, muscles and mucous membrane. One end o the vocal olds is connected to the arytenoid cartilage, and the other end to the thyroid cartilage. Te vocal olds are opened and closed through the adduction and 16
2.1
17
Vocal organs
1
2
7 10 4
8
9
6 5
14
11
3
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Nasal cavity Upper lip Lower lip Upper teeth Lower teeth Alveolar ridge Hard palate Soft palate (velum) Uvula Oral cavity Pharyngeal cavity Epiglottis Vocal folds Tongue
12 13
Figure 2.1
The vocal organs
Cricoid cartilage
Arytenoid cartilage Back
Glottis
Thyroid cartilage Vocal ligament Front
Figure 2.2
View of the larynx from above
abduction o the arytenoid cartilage. Te gap produced by the abduction o the vocal olds is called the glottis. Te vocal olds are the principal vocal organs involved in the process o phonation. 2.1.3
Tongue In the production o sound, the next most active organ afer the vocal olds is probably the tongue. In order to understand the production o sound, it
Production of sounds
18
Tongue body
Tongue tip
Front
Centre
Back
Tongue root
Tongue blade
Figure 2.3
Subdivisions of the tongue
is important to know the names o the various parts o the tongue. Tese are shown in Figure 2.3. Te oremost part o the tongue is called the tongue tip , and the part above it, which lies beneath the alveolar ridge when the tongue is in a relaxed state, is called the tongue blade, and measures about 10–15 millimetres. Te rest o the tongue is known as the tongue body, and can be urther divided into three parts, ront, central and back, or sometimes into just two parts, ront and back. I the body o the tongue is divided into two parts, the ront o the tongue reers to the part which makes contact with the hard palate during articulation, and the back o the tongue reers to the part which makes contact with the velum during articulation. Beyond the back o the tongue is the tongue root, which orms the ront wall o the pharynx. 2.1.4
Palate Figure 2.4 shows the various parts o the palate. As is shown, the palate is di vided into our areas: the protruding area behind the upper ront teeth, known as the alveolar ridge; the part supported by bone, known as the hard palate; the part which is not supported by bone, known as the sof palate or velum ; and �nally, the end part o the velum, known as the uvula .
2.2
Major cavities As shown in Figure 2.5, the upper part o the vocal olds is divided into three sections around the tongue, the pharyngeal wall and the palate, reerred to as the pharyngeal cavity, oral cavity and nasal cavity. Air rom the vocal olds is modi�ed in this upper space as it passes through. Te particular sound the air
19
2.2
Major cavities
Hard palate Alveolar ridge
Figure 2.4
Soft palate (velum) Uvula
Subdivisions of the roof of the mouth
Nasal cavity
Oral cavity Pharyngeal cavity
Figure 2.5
The three major cavities involved in speech production
makes when it is emitted as depends on the cavity in which it resonates as it leaves the larynx. Sound production is mostly done in the pharyngeal, oral and nasal cavities. Te varying movements o the vocal organs transorm the space inside these cavities, giving the air leaving the vocal olds a very different sound rom the uniorm sound as which it began. When we inhale, we usually breathe in and out o our noses with our mouths closed. Generally, when we exhale, air is released rom our lungs and leaves through the nose. But we can exhale through our mouths as well as through our noses. Figure 2.6 shows the passage o air�ow through the nasal and oral cavities when we exhale. In contrast to when we breathe, when we speak we use our mouths rather than our noses, and air rom the lungs only passes through the mouth. In order to prevent air rom entering the nasal cavity, the levator veli palatini, located at the back o the velum, contracts. When this muscle contracts, the
Production of sounds
20
2 1
Figure 2.6
Two passages of airflow from the glottis
palate rises and lies �ush with the pharyngeal wall. Tus, as the nasal passage is blocked, the air leaves through the oral cavity on its way out. In speech, sounds that are produced with the nasal passage obstructed are called oral sounds, and sounds produced with the nasal passage open are called nasal sounds. Most speech sounds are oral sounds, and may be assumed such unless otherwise speci�ed. In addition, as most pronounced sounds are oral – nasals are the exception – the velum is raised �at against the pharynx most o the time. However, the velum is in this position during breathing. 2.3
Processes of speech production Generally speaking, speech sounds are produced as air is exhaled rom the lungs through the vocal olds and the vocal tract. Tis process can be divided into three main stages. First, the supply o air needed or sound ormation; second, modulation o that air supply; and �nally, transormation o the regulated air. Te linguistic terms or these stages are initiation, phonation and articulation, respectively (Catord, 1988). In the ollowing, we will discuss each stage in detail.
2.3.1
Initiation Let’s take a closer look at initiation, the �rst stage. Initiation is the stage in which air is supplied or sound production. We can classiy initiation processes by initiator , the place rom which air is supplied, and also by direction
21
2.3
Processes of speech production
Table 2.1 Patterns of initiation
Airstream
Initiator
Direction
Names of stop consonants
Pulmonic egressive Lungs
Outwards
Stops
Glottalic egressive
Outwards
Ejectives
Glottalic ingressive Vocal olds
Inwards
Implosives
Velaric ingressive
Inwards
Clicks
Vocal olds Velum (tongue)
o air�ow. Te initiators o most sounds known to exist are the lungs, and the direction o air�ow is outward. We call sounds which are initiated rom the lungs pulmonic sounds. Sounds with an outward air�ow are called egressive sounds. In other words, egressive sounds are sounds that require exhalation in order to produce them. All o the sounds in the Korean language are pulmonic egressive sounds, sounds that are produced by air exhaled rom the lungs. Not only the sounds o Korean, but also the sounds o English, Japanese, Chinese, French and German are all pulmonic egressive sounds. In act, as noted above, most sounds known to exist are pulmonic egressive sounds. However, not all speech sounds are pulmonic egressive sounds. Tough rare, the vocal olds and the velum can also be the initiators o some speech sounds, and the direction o air�ow can also be rom the outside in rather than rom the inside out. Tus, theoretically, there are six possible patterns o initiation, according to type o initiator and direction o airstream, through which speech sounds can be produced: (i) pulmonic egressives; (ii) pulmonic ingressives; (iii) glottalic egressives; (iv) glottalic ingressives; (v) velaric egressives; and (vi) velaric ingressives. However, in reality, only (i) pulmonic egressives, (ii) glottalic egressives, (iii) glottalic ingressives and (iv) velaric ingressives are observed in actual speech sound production. For instance, ejectives are glottalic egressive sounds, implosives glottalic ingressives, and clicks velaric ingressives. Pulmonic ingressives and velaric egressives are not observed in the speech production o any o the world’s known languages. Patterns o initiation are listed in able 2.1.
DID YOU KNOW . . .?
Two kinds of ingressives When something surprising happens, or when you eel sympathy or somebody, what kinds o sound do you make? Let’s have a closer look at those sounds.
22
Production of sounds
Sounds made when something surprising happens
When something unexpected happens, what sound do you make? ry it now. Did you inhale or exhale to produce the sound? Native Korean speakers breathe in to produce this sound. In this case, initiation takes place in the lungs, and the direction o air�ow is ingressive. Tis sound is a pulmonic ingressive sound. Pulmonic ingressives are easy to make and amiliar to us. Interestingly, however, no pulmonic ingressive speech sounds are known to exist. Why might this be? o answer this question, try saying the words, “I am reading a book entitled Te Sounds of Korean” while breathing in – that is, with pulmonic ingressive sounds. You may well be out o breath beore �nishing the sentence. Pulmonic ingressives are ar too inefficient and unnatural to be used as speech sounds. Tis is likely to be the reason why no pulmonic ingressive speech sounds exist. Sounds made to express sympathy or disapproval
Tis time, make the “tut” sound that you would make when you eel sympathy or someone or to express disapproval. ry to think about the direction o air�ow. You will discover that air rushes into your mouth as the sound is being produced. You may be able to eel the direction o the air�ow more slowly by making a clicking noise. Next, click while holding your nose. When you click, you will notice that you are not breathing, although your mouth is open. You know that you cannot breathe while you are holding your nose, but you may wonder why you are not breathing through your mouth. Tis is because although the tip o the tongue is resting on the bottom o the mouth, the body o the tongue is obstructing the sof palate. When this clicking sound is produced, the body o the tongue touches the sof palate. At the same time, the blade o the tongue touches the alveolar ridge, and thus air is trapped in the space in between. Tis is the air used to produce the sound. Ultimately, according to its process o initiation, this clicking sound is a velaric ingressive, since the velum supplies the air, and the direction o air�ow is ingressive. Clicking sounds are easy to produce, but rarely ound as speech sounds, except in some Arican languages. What is the difference between these two ingressive sounds?
Although the sounds we produce when we are surprised or eel sympathy are both ingressive, there are important distinctions between the two.
23
2.3
Processes of speech production
First, as previously mentioned, when you produce a surprised sound, you can hold your nose and still breathe naturally. However, you cannot breathe i you hold your nose while making a clicking noise with your tongue. Te reason or this is that the mouth is blocked in two places, only one o which is released during the production o sound. Furthermore, even afer the sound is made, one end o the oral cavity remains obstructed, thus preventing air rom escaping. Second, reproducing the sound you make when surprised will require you to consciously inhale air. However, when producing clicking sounds, you inhale naturally without needing to consciously inhale. Te reason why clicking your tongue does not require you to consciously inhale air is because this particular airstream mechanism naturally generates inward air�ow. When you click with your tongue, the back and ront parts o the tongue cause an obstruction inside the mouth, trapping air between them. Beore the click sound is produced, the tongue pulls down this middle area that traps air, expanding the space a little. When this space is expanded, the pressure inside it drops, and the ront part o the tongue is released. Air rushes in because the pressure inside is lower than the pressure within the oral cavity.
2.3.2
Phonation Afer air is provided through initiation, it passes through the vocal olds and is modulated. Tis process is called phonation. Te vocal olds control the air exhaled rom the lungs by adduction and abduction. Tat is to say, the vocal olds unction as a valve which controls air�ow. For instance, when the vocal olds are closed, air cannot be released, but when the vocal olds are open, air can �ow reely. Te vocal olds can be opened and closed repeatedly. In everyday lie, when we strain our vocal olds are closed, when we breathe they are opened, and when we cry or shout they vibrate. We produce speech sounds by opening and closing our vocal olds. Some sounds are produced by a rapid and repeated opening and closing o the vocal olds. When the vocal olds rapidly open and close repeatedly, they vibrate. On the basis o their phonation, speech sounds can be divided into voiced sounds and voiceless sounds. Voiced sounds are accompanied by vibrations in the vocal olds, and voiceless sounds are not. A summary o voiced and voiceless sounds appears in Figure 2.7.
24
Production of sounds
(a) Breath
(b) Voiceless
(c) Normal voice
Arytenoid cartilage Glottis Thyroid cartilage
Figure 2.7
Schematic view of laryngeal configurations
When voiceless sounds are produced, the vocal olds either separate, causing the glottis to open, or come together, causing the glottis to close. When voiced sounds are produced, the vocal olds vibrate either completely or only partially.
DID YOU KNOW . . .?
Making vibrations in the vocal folds Tree conditions are required in order to produce vibrations in the vocal olds. When all three conditions are ul�lled, the vocal olds vibrate naturally. I any one o these conditions is not met, vibration is unlikely to occur. First, the vocal olds must be as close together as possible. Second, the vocal olds must be relaxed. Tird, subglottal air pressure must be sufficiently greater than supraglottal air pressure. Vowel sounds are produced with the mouth open, so subglottal air pressure is more likely to be greater than supraglottal air pressure. Tereore, all vowels are voiced sounds. However, it is easy or supraglottal air pressure to increase during the phonation o consonantal sounds such as stops or fricatives. Tis is because these sounds are produced either with both the oral and nasal cavities obstructed (i.e., stops), or with the nasal tract obstructed and the oral cavity severely narrowed (i.e., ricatives). Tereore, in these cases, additional effort is needed in order to keep the subglottal air pressure sufficiently higher than the supraglottal air pressure. Tis is why it is common or stops and ricatives to be classi�ed as voiceless. However, some stops and ricatives can also be produced as voiced sounds. Sounds that do not cause vibrations in the vocal olds are called voiceless sounds, and sounds that are accompanied by vibrations are called voiced sounds.
25
2.3
Processes of speech production
DID YOU KNOW . . .? How to quieten your voice You have bumped into a riend in one o the aisles in the library. You naturally lower your voice to greet your riend so as not to disturb other students. “Hi, long time no see. How are you?” Now, bring out your acting skills and talk as you would i you ound yoursel in such a situation. Is your phonation not different to what it would normally be? You will �nd that your voice is breathier than normal. Tis is because instead o the vocal olds vibrating in their entirety, only the ront part is now vibrating, leaving the part urther back open, in order to produce quieter sounds. Tis kind o phonation is called breathy voice. Tis time, imagine you are whispering to your riend something top secret, and say, “Tis is top secret. Please don’t tell anyone.” What about this time? Are you still using a breathy voice? Because breathy sounds are produced by vibrating vocal olds, you can eel vibrations in your throat. What about when you are whispering to your riend? Tis time, the vocal olds do not vibrate, and you will not be able to eel any vibrations in your throat. In act, these two phonation types are not effective methods o phonation. o begin with, the tone is quiet, and you need to breathe more ofen than normal in order to continue speaking in this way. O course, these sounds are inevitably quieter than normal voiced sounds, as the vocal olds are vibrating only partially (breathy sounds) or are not vibrating at all (voiceless sounds). Furthermore, as the vocal olds are entirely or partially open, more air will escape compared to when you produce normal voiced sounds, and it will thus be harder to speak or longer using one breath. Tough these phonations are ineffi cient, we still resort to them when we need to speak quietly or have a bad throat. Tis will help us to understand why, cross-linguistically, the voiced period o an utterance is much longer than the voiceless period. Voiced phonation is much more effective in producing louder sounds as it is accompanied by vibration. Also, air is released bit by bit, because the vocal olds are rapidly opening and closing, and so more sounds can be pronounced over a single breath.
2.3.3
Articulation As the air that is supplied through initiation enters the stage o phonation, it is modi�ed and transormed as it passes through the vocal tract. Te
Production of sounds
26
vocal tract reers to the oral and nasal tracts through which air rom the vocal olds passes beore being released. Tese cavities can take on many shapes to produce different kind o speech sounds. Tis process, called articulation, can produce two different types o sound: consonant sounds, which are produced by obstruction in the central part o the oral cavity, and vowel sounds, which are produced without such obstruction. 2.4
Classification
2.4.1
Sonorants vs. obstruents Speech sounds can be divided into two main categories: sonorants and obstruents. When sonorants are produced, air�ow is unhindered as it passes through the vocal tract. When obstruents are produced, however, air�ow leaving the vocal tract is obstructed. As a result, obstruents are accompanied by noise caused by disturbance in the air�ow, whereas sonorants are not. Tere are two general ways in which air�ow is obstructed. Te �rst is by constriction o the passageway through which air passes to create sound through turbulent air�ow. Te second is a complete obstruction and then reopening o the air passage to produce a momentary noise. Fricatives are o the ormer kind, and stops are o the latter kind. Both o these sounds are accompanied by noise, but the duration o that noise is different or each.
DID YOU KNOW . . .?
Sonorants and obstruents vs. consonants and vowels Sonorants and obstruents are distinguished on the basis o whether or not they are accompanied by noise, whereas consonants and vowels are distinguished on the basis o whether or not the central part o the oral cavity is obstructed. As noted above, consonantal sounds are produced with obstruction o the oral cavity, and vowel sounds are produced without any such obstruction. Tis obstruction is not made differently rom the way in which obstruents are created; both cause narrowing o the airway. Why, then, are obstruents and consonants categorised separately? Why are all vowels sonorants and all obstruents consonants? Furthermore, why are some sonorants consonants? Te answer lies in their very de�nitions. Although the nasal cavity is obstructed in the pronunciation o vowels, the central part o the oral
27
2.4
Classification
cavity remains open. Tus, there is no obstruction in the midsagittal region o the vocal tract. Furthermore, vowels are produced without any concomitant noise; hence they are sonorant. On the other hand, consonants can be either sonorant or obstruent. Fricatives, stops and affricates are obstruents, but nasal sounds and approximants are sonorants. Fricatives, stops and affricates are all sounds produced with the nasal cavity blocked off and by an obstruction to the centre o the oral cavity. As the air�ow meets this obstruction, noise is created. On the one hand, nasal sounds and approximants are grouped as consonants because they are produced by a ar narrower central part o the oral cavity than required when producing vowels, but they are classi�ed as sonorants as they are not accompanied by noise. First o all, like stops, nasal sounds are produced with complete obstruction o the oral cavity. However, unlike in the case o stops, air is allowed to escape through an unobstructed nasal cavity. Furthermore, approximants block air passing into the nasal cavity and constrict the central part o the vocal tract ar more than is the case in the production o vowels, but this constriction is not accompanied by noise, as the passageway is not narrow enough to produce turbulence.
2.4.2
Consonants As mentioned beore, consonants are sounds produced by obstruction along the centre line o the oral cavity during articulation. Te various types o consonants may differ in several ways depending on: (i) the place o obstruction within the oral cavity; (ii) the particular kind o obstruction o air�ow; and (iii) the state o the vocal olds. Tese three variables are linguistically known as place o articulation, manner o articulation and phonation type or voice type respectively. 2.4.2.1
Place of articulation
Consonants may be classed as bilabial, labiodental, dental, alveolar, palatoalveolar (or postalveolar), alveolo-palatal (or prepalatal), retro�ex, palatal, velar, uvular, pharyngeal or glottal according to their place o articulation. Places o articulation get their names rom the lower and upper articulators involved in the articulation process. For example, when the lower lip is the lower articulator and the upper lip is the upper articulator, this is reerred to as labiolabial. When the lower articulator is the lower lip and
28
Production of sounds
Table 2.2 Summary of places of articulation
Place of articulation
Lower articulator
Upper articulator
Bilabial
Lower lip
Upper lip
Labiodental
Lower lip
Upper teeth
Dental
ongue (tip or blade)
Upper teeth
Alveolar
ongue (tip or blade)
Alveolar ridge
Palato-alveolar (postalveolar) ongue (tip or blade)
Rear o alveolar ridge
Alveolo-palatal (prepalatal)
ongue (tip or blade)
Front o hard palate
Retro�ex
ongue (tip)
Hard palate
Palatal
ongue (body: back)
Hard palate
Velar
ongue (body: back)
Sof palate (velum)
Uvular
ongue (body: back)
Uvula
Pharyngeal
ongue (Root)
Pharyngeal wall
Glottal
Vocal old*
Vocal old
*
In the case o the glottal sounds, the vocal olds play the role o articulator. Tereore, unlike or other places o articulation, the concept o upper and lower articulators is not applicable here. Instead, just the two vocal olds are involved in articulation.
the upper articulator is the upper teeth, this is reerred to as labiodental. As labiolabial sounds are articulated with both lips, they are commonly reerred to as bilabial sounds. When the lower articulator is the tongue, the name o a place o articulation is determined by its upper articulator. Tus, dental sounds are produced when the lower articulator is the tongue and the upper articulator is the teeth, and alveolar sounds are produced when the lower articulator is the tongue and the upper articulator is the alveolar ridge. An overview o the places o articulation with their various articulators is given in able 2.2. 2.4.2.2
Manner of articulation
Consonants are also classed according to the manner o their articulation : stops/plosives, nasals, trills, taps/�aps, ricatives, lateral ricatives, approximants, lateral approximants and affricates, among others. Tese sounds are divided on the basis o whether or not they are produced with complete obstruction o the oral cavity. Whereas stops/plosives, nasals, trills and taps/�aps are produced by completely blocking off the oral cavity, ricatives and approximants are produced by allowing some air to escape through the oral cavity. Sounds produced by total obstruction o the oral cavity are known as sounds
29
2.4
Classification
Stop
Figure 2.8
Nasal
Schematic view of the vocal tract for stops and nasals
with complete oral closure, and sounds produced without total obstruction are known as sounds without complete oral closure. (1)
ypes o consonants according to the extent o oral closure • Oral cavity blocked: stops/plosives, nasals, trills, taps/�aps • Oral cavity open: ricatives, approximants
Among sounds with complete oral closure, nasal sounds differ considerably rom stops, trills and taps/�aps in that the velum is not raised and thus allows air to escape into the nasal cavity. Consequently, sounds with complete oral closure may be urther divided on the basis o whether the nasal cavity is open or closed. Tat is, nasals, which are produced with the oral cavity closed but the nasal cavity open are classed as one group, and stops, trills and taps/�aps, which are produced with both the oral and nasal cavities closed, orm another. Te differences in the articulation o stops and nasal sounds are shown in Figure 2.8. (2)
Sounds with complete oral closure divided according to the state o the nasal cavity • Closed nasal cavity: stops, trills, taps/�aps • Open nasal cavity: nasal sounds
Stops, trills and taps/�aps can be urther divided into two categories depending on the number o times the oral cavity is blocked. Te �rst category contains stops and taps/�aps, where closure occurs once, and the second contains trills, where closure occurs more than twice. How, then, are we to distinguish stops and taps/�aps rom each other? Te difference lies in the duration o their respective closure. Te closure that occurs in the production
30
Production of sounds
Stop
Figure 2.9
Tap
Trill
Differences in the articulation of stops, taps and trills
Turbulent airflow
Stop
Figure 2.10
Fricative
Approximant
Differences in the articulation of stops, fricatives and approximants
o stops lasts relatively longer than the obstruction that occurs in t he production o taps/�aps. Figure 2.9 shows the differences between stops, taps/�aps and trills. (3)
Consonantal sounds produced by a complete closure o the oral and nasal cavities • Number o closure: once (stops, taps/�aps), more than twice (trills) • Duration o closure: long (stops), short (taps/�aps, trills)
At the moment o emission, ricatives and approximants are both produced without complete closure o the oral tract, but they differ in whether or not they are accompanied by noise. A ricative is a sound produced by turbulent air�ow caused by a narrowed airway. By contrast, an approximant is a sound created with a wider airway, and hence no turbulent air�ow. Fricatives and approximants can be urther distinguished according to whether air escapes through the middle o the oral cavity or down the sides. Tese are called median ricatives and median approximants, and lateral ricatives and lateral approximants respectively. As ricatives and approximants are usually produced by constriction in the central part o the airway, they are reerred to simply as ricatives and approximants. However, the word ‘lateral’ must be attached to the names o those sounds that are produced by air escaping around the sides o the mouth when the centre is blocked, i.e. lateral ricative, lateral approximant. Te differences between stops, ricatives and approximants are shown in Figure 2.10 and the difference between median and lateral articulations is shown in Figure 2.11.
31
2.4
Classification
Median
Figure 2.11
Lateral
Difference between median and lateral articulations
Closing
Closed
Release (instant)
Stop
Figure 2.12
Closing
Closed
Release (delayed)
Affricate
Schematic view of the differences between the articulation of stops and affricates
(4)
ypes o consonants produced with an open oral cavity • Sounds produced with(out) noise produced by turbulence: with noise (ricative), without noise (approximant) • Sounds produced with(out) using the lateral passage: lateral (lateral ricative, lateral approximant), non-lateral (median ricative, median approximant)
Affricates are produced by successive complete and incomplete oral closure. Tus, like stops, affricates are initially articulated with complete oral closure, but, like ricatives, are disrupted by turbulent air created by narrowing o the airway. Figure 2.12 shows the differences in the articulation o stops and affricates. As shown in the diagram, whereas stop sounds are created by closure and sudden release, affricates are different as they are created by closure and delayed release (i.e., riction). able 2.3 summarises the manners o articulation examined so ar. 2.4.2.3
Phonation (or voice) types
Consonants are classi�ed not only by place and manner o articulation, but also by phonation (or voice) type. Te two most general phonation types are voiced and voiceless sounds. As all vowels are voiced sounds, the division between voiced and voiceless is irrelevant to them. However, some consonantal sounds are produced by vibrations in the vocal olds and some are not. Moreover, some consonantal sounds share the same place and manner o articulation and can only be distinguished according to whether or not the vocal olds vibrate. Why is this?
32
Production of sounds
Table 2.3 Summary of manners of articulation
Manner of articulation
Complete Nasal Closure Longer oral closure closure only once closure
Friction Lateral noise passage
Stop
°
°
°
°
–
–
Nasal
°
X
°
°
–
–
rill
°
°
X
X
–
–
ap
°
°
°
X
–
–
Fricative
X
°
–
–
°
X
Lateral ricative X
°
–
–
°
°
Approximant
X
°
–
–
–
X
Lateral approximant
X
°
–
–
–
°
Affricate
°
° = yes,
X
°
°
°
–
°
–
–
°
–
X
X = no, – = not applicable.
In order to answer this question, we need �rst to understand some points about vibrations in the vocal olds. Tree conditions must be met in order or the vocal olds to vibrate. O course, i any one o these conditions is not met, vibration will not occur, and additional effort will be required to cause vibrations. Te conditions are as ollows: �rst, the proximity between the vocal olds must be as close as possible; second, the vocal olds must be relaxed; third, air pressure rom below the vocal olds (subglottal air pressure) must be sufficiently greater than air pressure rom above the vocal olds (supraglottal air pressure). O these three conditions, the �rst two are controlled by the vocal olds, and the last condition depends on the airstream mechanism and is related to con�guration o the vocal tract. In other words, when the vocal tract is constricted excessively, air is delayed in escaping rom the vocal olds and the supraglottal pressure increases. As a result, the subglottal pressure is less likely to be su�ciently greater than the supraglottal pressure. On the other hand, when there is no such constriction and the airstream passage is wide, supraglottal pressure is more likely to be suffi ciently lower than subglottal pressure. Because o the aerodynamic actors noted above, vowels and sonorant consonants are by their nature voiced sounds, and obstruents are by their nature voiceless. Tis helps to explain why some languages have only voiceless obstruents, but no languages have only voiced obstruents. Apart rom vibrations in the vocal olds, phonation types are also distinguished by the presence o aspiration and the amount o tension in the vocal olds. Sounds produced by glottal riction are called aspirates, and sounds
33
2.4
Classification
Closed phase
Voiced
Voiceless unaspirated
Voiceless aspirated
Release phase
Voiced
Figure 2.13
Voiceless unaspirated
Glottal friction Voiceless aspirated
Phonation types of stops during closure and release phases
produced with tension in the vocal olds are called tense sounds. Aspiration is created when sounds are articulated with the vocal olds apart, and as air passing through the narrow gap between them becomes turbulent. ension is created by complete tension o articulators such as the vocal olds. Sounds are classed as ‘aspirated’ or ‘unaspirated’ depending on whether or not they are produced with aspiration, and as ‘ortis’ or ‘lenis’ depending on whether or not they are accompanied by tension in the vocal olds. Consonants can be basically divided, on the basis o voicing and aspiration, into the ollowing: voiced aspirated, voiceless aspirated, voiced unaspirated and voiceless unaspirated. Voiced sounds require the vocal olds to be close enough together to allow them to vibrate. Tereore aspirates, which are produced with the vocal olds parted, are naturally voiceless. Furthermore, voiced sounds are unaspirated, or the same reason. Tis is the reason why voiced unaspirated sounds are rarely ound in any language. Figure 2.13 shows the closed and release phases o three kinds o phonological stop sounds: voiced, voiceless unaspirated and voiceless aspirated stops. Although sounds are generally divided on the basis o voicing and aspiration, the Korean language cannot be subdivided in such a way. 1 Tis is because although it contains stops and affricates that share the same place and manner o articulation, these are all phonemically voiceless. Tereore, because the three types o stops and affricates in Korean are all voiceless, they cannot be distinguished by the presence o voicing, as shown in (5), Korean bilabial stops categorised by the presence o voicing.
34
Production of sounds
(5)
Korean bilabial stops with(out) voicing p () p* () ph () (vo (voicel iceles ess) s) (vo (voicel iceles ess) s) (voiceless)
On the other hand, they t hey may be distinguished by the presence o aspiration, aspiration, as shown in (6). As aspiration only provides two distinct categories o sound, it is insuffi cient or distinguishing disti nguishing the t he three types typ es o Korean stops and affricates affr icates rom one another. (6)
Korean bilabial stops with(out) aspiration p () p* () ph () (voi (voicel celess ess unaspi unaspira rated ted)) (voice (voiceles lesss unasp unaspira irated ted)) (voice (voiceles lesss aspira aspirated ted))
Tereore another class eature is needed here to distinguish between phonation types. Tis new class eature is tension. Sounds that are tense are called ortis, and sounds that are not are called lenis. Sounds that are tense tend to be strong sounds with lengthy articulation. With aspiration aspiration and tense as class eatures, eatures, we are now able to distinguish between the types o stops and affricates in the Korean language. (7) shows bilabial stops categorised according to the three phonation types o stops and affricates. (7)
Korean bilabial stops with(out) tension p () p* () (voiceless unaspirated (voiceless unaspirated lenis) ortis)
ph () (voiceless aspirated ortis)
In Korean, the voiceless unaspirated lenis is reerred to as lax, the voiceless unaspirated ortis as tense, and the voiceless aspirated ortis as aspirated. Korean linguists commonly reer to these three kinds o phonation types as lax, tense and aspirated. a spirated. However, some scholars may use the terms ‘plain’ or ‘lenis’ instead o lax, and ‘reinorced’ or ‘ortis’ instead o tense. For example, Peter Ladeoged calls the three Korean phonation types unaspirated, ortis and aspirated (www.phonetics.ucla.edu/appendix/languages/korean/korean.html (www.phonetics.ucla.edu/appendix/languages/korean/korean.html ). DID YOU KNOW . . .?
Notating Korean tense sounds Many Korean language scholars denote that a consonant is tense by placing an apostrophe afer it, or example, p’, t’, k’. However, the IPA uses the apostrophe to denote ejectives. Tis, o course, does not mean that
2.4
35
Classification
both these usages denote the same thing. During its initiation, a Korean Korean tense sound is an ordinary pulmonic egressive, but ejectives are glottalic egressives during initiation. O course, it is misleading that the sign used to denote an ejective should coincide with the sign used to denote Korean tense sounds. In order to get around this problem, Korean tense sounds are ofen denoted using a capital letter P, , K or an asterisk p *, t*, k*, although this differs rom scholar to scholar. Ladeoged’s method o using an asterisk to signiy tense sounds and using h to signiy aspiration has proved useul, and will be adopted in this book. Although Ladeoged’ Ladeoged’s method will be used to signiy tense sounds here, please bear in mind that terminologies other than Ladeoged’s have been used. Tis is because using the terms lax and tense is more appropriate than Ladeoged’s unaspirated and ortis or demonstrating the phonological eatures or differences in the phonation types o the three variants o o Korean Korean consonants. consonants.
2.4.3
Vowels Vowels Vowels are sounds that are emitted through the oral cavity without any obstruction during articulation. For this reason, vowel sounds cannot be grouped according to the place or manner o obstruction as can consonants. Furthermore, as previously mentioned, the production o vowel sounds ul�ls all the conditions necessary or the vocal cords to vibrate, and thus all vowels are voiced sounds. Hence it is not meaningul to distinguish vowels on the basis o phonation type. Vowels are produced by changes in the orm o the pharyngeal and oral cavities, caused by movements o the tongue and lips. Tereore, vowels are categorised according to the elements that comprise such changes. Te tongue and lips are the articulators that alter the orm o the two cavities. Vowels can be categorised using the ollowing class eatures: tongue height, tongue backness and roundedness o the lips. 2.4.3.1
Height
ongue height can be divided into our levels: high, mid-high, mid-low and low. ongue height is proportional to the openness o the mouth; hence high vowels vowels are reerred to as close vowels, vowels, and low vowels are reerred to as open
36
Production of sounds
High (close)
High mid (close-mid) Height Low mid (open-mid)
Low (open)
Figure 2.14
Tongue height
vowels. vowels. Tereore, Tereore, high, mid-high, mid-low, mid-low, and low vowels are also reerred to as close, close-mid (hal-close), open-mid (hal-open), and open vowels respectively (see Figure 2.14). 2.14 ). When only three levels o tongue height are speci�ed, the mid-high or close-mid, or sometimes the mid-low or open-mid position does not eature as part o the categorisation. Instead, a mid vowel is added between the high vowel and low vowel, vowel, thus leaving us with three levels: high, mid and low. Te reason why tongue height affects vowel quality is because the orm o the vocal tract changes according to tongue height. Even i all other conditions remained unchanged, the orm o the oral and pharyngeal cavities when the tongue was high would be different rom when the tongue was low, and such a difference would cause a difference in resonance requency, which in turn would produce a vowel sound with a different quality. 2.4.3.2
Backness
Vowels are ofen distinguished according to tongue backness, backness , the part o the vocal tract that is most constricted during their articulation, as shown in Figure 2.15, 2.15, using the ollowing designations: ront vowel; central vowel; and back vowel. Te orm o the vocal tract also changes according to which parts o the tongue and palate are used during the articulation o a vowel. I we consider the vocal tract simply as a single ‘tube’, then stricture at any point along that tube will divide it into two parts, ront and back. I the constriction takes place in the ront, the ront part will be shorter and the back part be b e longer; i the constriction takes place at the back, that is, at the base o the tongue, the ront part will be longer and the back part shorter. Variations in the length o this ‘tube’ will produce different vowel qualities caused by different resonance requencies.
2.5
37
Summary
Backness
Front
Figure 2.15
Central
Back
Tongue backness
Lip protrusion
Rounded
Figure 2.16
Unrounded
Roundedness
2.4.3.3
Roundedness (protrusion of the lips)
Te shape o the lips in vowel articulation is either termed ‘rounded’ ‘rounded ’ or ‘unrounded’ ‘unrounded’ depending on whether or not the lips protrude outwards. Again, i we consider the vocal tract as simply a single ‘tube’, protruding the lips will affect the length o the whole tract, as shown in Figure 2.16 . When the lips protrude, the vocal tract is longer than normal. Tis change produces differing vowel qualities caused caused by a change change in the resonance resonance requency requency.. 2.5
Summary Te production o sounds is divided into three main stages: initiation, phonation and articulation. Initiation involves supply o air�ow, phonation
Production of sounds
38
involves modulation o that air�ow and articulation involves transormation o the air�ow. Initiation, the stage in which air is supplied, may be classi�ed according to the initiator o the sound (lungs, vocal olds or velum), and whether the sound is ingressive or egressive. Among six possible combinations, our exist in spoken language: pulmonic egressive, glottalic ingressive (ejectives), glottalic ingressive (implosives) and velaric ingressive (clicks). Phonation, the stage in which air is modulated, may be divided into two main types, voiced and voiceless, depending on whether or not the sound is accompanied by vibration o the vocal olds. However, phonation also encompasses aspirated and unaspirated, and tense and lax sounds. During articulation, the stage at which air is transormed into sound, consonants are produced by obstruction along the centre line o the oral cavity, and vowels are produced without any such obstruction. Consonantal sounds are distinguished according to the place and manner o their articulation and their phonation type, while vowels are distinguished by tongue height, tongue backness and roundedness o the lips. EXERCISES 1
2
State whether the following sounds are easy to produce with(out) voicing, and explain why. a. Stops b. Fricatives c. Nasal sounds d. Vowels e. Aspirates What kind of sounds would the following forms of the vocal tract produce? (a)
(b)
Exercises
39
3
Identify the various parts of the vocal organs on the diagram as indicated.
1 2 3 4 5 7
3
4
6 5 6 2
7
1
4 5
Explain the similarities and differences of stops and taps/flaps. Explain the differences between the articulation of fricatives and that of affricates.
3
Basic concepts of phonology
In the previous chapter we discussed the general aspects o sound production, and learned that there are three steps in the production o sound: (i) initiation, (ii) phonation and (iii) articulation. Whereas the purpose o the last chapter was to help readers amiliarise themselves with the physical basis o sound production, this chapter will ocus on the abstract reality o sounds, and will introduce the background to phonology. Once speech sounds are produced by a speaker, they are transerred to a listener as a sequence o acoustic signals. Tese acoustic signals will then be understood by the listener according to his/ her knowledge o the relevant language. Even i the listener has perect hearing, without knowledge o the language being spoken, he/she will �nd it difficult to segment sounds or to distinguish between different sounds. Hence, in acquiring the sound system o our mother tongue we learn not only an inventory o sounds, but also how to determine which sound distinctions are signi�cant and which not. Understanding the phonological basis o a sound system is also crucial or second language speakers. ogether with Chapter 2, this chapter will provide a useul background or understanding the sounds o Korean. DID YOU KNOW…?
Do we actually listen with our ears? Why do Korean pigs say ‘ /k*ulk*ul/’ and English pigs ‘oink’?
Why is it that Korean people think that pigs say /k*ulk*ul/, but English people think that they say ‘oink’? Is it because onomatopoeic sounds are different in every language? Do English dogs ‘bow-wow’, and Korean dogs /mŋmŋ/, because the two make different noises? Do oreign languages exist or animals as well?
40
41
3.1
Speech sounds, phones and phonemes
Let’s think about the sound o a computer mouse click. Following the same logic, the reason Koreans hear /t* ɑlk*ɑk/ and English people hear ‘click’ when using a mouse would be that the two make different noises. However, the mouse noise is the same; it is just that Korean and English speakers hear it differently. Almost all onomatopoeic sounds differ rom language to language, not just animal sounds or mouse-clicking. Onomatopoeic words break up a sequence o continuous sounds into discrete segments. /k*ulk*ul/ and ‘oink’ are the result o the conversion o the continuous, indiscrete sound made by a pig into linguistically discrete segments. In other words, although onomatopoeia are expressions o natural sounds, they cannot but be in�uenced by the sounds o a particular language. Tereore, while the sound a pig makes is neither Korean nor English, the onomatopoeias /k*ulk*ul/ and ‘oink’ are ormed rom the sounds o those languages.
3.1
Speech sounds, phones and phonemes
Imagine a boy and a girl, saying /ɑnnjŋhɑsεjo/ ‘hello’ to each other. Although they have said the same word, is what they have said to each other truly the same? Are their speech sounds the same? Perhaps not, as they may have different voices and different accents. But, at the same time, it would be diffi cult to argue that their speech sounds are different. Te reason why the above question is diffi cult to answer is that the term ‘speech sounds’ can mean two different things. So ar we have used the term ‘speech sounds’ to reer to sound as both a physical and an abstract entity. But, strictly speaking, these terms need to be urther classi�ed. Sound as a physical entity is called a ‘phone’, whereas sound as an abstract and psychological entity is called a ‘phoneme’. Ofen, a phone is written inside square brackets [ ], whereas a phoneme is written between slashes / /. Tis is the convention adopted in this book. Te study o phones is called ‘phonetics’ and the study o phonemes is called ‘phonology’. Hence, there is a undamental difference between the oci o the two disciplines. In phonetics, a phone is understood as a physical, continuous, time-limited and quanti�able unit. It is continuous because it is difficult to distinguish the beginning and end o a sound; it is quanti�able because the various physical properties o a phone can be measured; it is time-limited because the length o a phone can be measured and is signi�cant. On the other hand, in phonology, a phoneme is understood as a psychological, discrete and
42
Basic concepts of phonology
unquanti�able unit which has no time constraints. It is discrete because it is a psychological entity, and has no time constraints because the physical length o a phoneme is o no signi�cance.
DID YOU KNOW…?
Speech Sounds are like Baron Ashura!
You may have heard o Baron Ashura, a Japanese manga animation character. Baron Ashura is a villain rom a Japanese made-or-V anime �lm called ‘Mazinkaiser Z’. However, or the bene�t o those who have not seen ‘Mazinkaiser Z’ (quite possibly many o you), it may be necessary to explain who the Baron is beore you can understand why he is being likened to the speech sound. Baron Ashura works or an evil boss, Dr Hell, who dreams o ruling the world. Te Baron is actually an arti�cial creation, which Dr Hell made by combining mummies rom Mycenae. As a result, the Baron is hal woman, and hal man. Te image in the �gure is the ace o the Baron rom the animation. As you can see, his right hal is emale, and his lef hal is male.
Is Baron Ashura a woman, or a man? You could quite rightly say, ‘the Baron is a man’, or ‘the Baron is a woman’, depending on which part o the ace you are looking at. You could also rightly say, ‘he seems to be both a man and a woman’, i you were looking at the whole ace. Te speech sound, like Baron Ashura, also has two ‘aces’. Depending on which part o its ‘ace’ you are looking at, the greetings /ɑnnjŋhɑsεjo/ spoken by a male and a emale student, as discussed in the example rom the previous section, can be said to sound both alike and different. Baron Ashura has only one name, but speech sound has a name or each o its aces: phone and phoneme.
43
3.2
3.2
Phonemes and allophones
Phonemes and allophones
A phoneme is ofen de�ned as the smallest sound unit which can alter the meaning o a word. A single phoneme may have different phonetic orms according to the phonetic environment in which it occurs. Te term ‘phoneme’ describes an abstract reality o the core properties which those different phonetic orms share.1 ry to pronounce the ollowing English words: (1)
pin
spin
nip
Afer that, put your hand in ront o your mouth and repeat. Do you notice any difference in the pronunciation o the three words? Although the three words in (1) all have /p/ as a phoneme, their phonetic realisations are different. At �rst, when you pronounce the /p/ in ‘pin’, you should be able to eel your breath strongly on your hand. However, you should not be able to eel any breath when you pronounce the /p/ in ‘spin’. When pronouncing the /p/ in ‘pin’ and ‘spin’, your lips should open afer you put them together. However, when you pronounce the /p/ in ‘nip’, your lips remain closed.2 In other words, the /p/ in ‘pin’ is aspirated, while the other two are unaspirated, and the /p/ in ‘nip’ is not released afer the closure o lips, whereas the other two are. Te phonetic realisations o /p/ in the words listed in (1) are shown in (2). (2)
pin [pʰn]
spin [spn]
nip [np ]
A superscript h reers to aspiration; hence ‘[ph]’ denotes an aspirated ‘p’. A ‘ ’ placed to the top-right o ‘p’ denotes that the stop sound ‘p’ is unreleased. Native speakers o English are ofen unable to tell the difference between these three sounds because they systematically occur in certain positions, as shown in (3). Tough these three sounds share a common abstract and psychological identity, they are still regarded as variants o the same phoneme, known as ‘allophones’. (3)
phoneme
/p/
h [p] [p¬ ] allophones [p ] word-initial word-initial following /s/ word-finial
However, i the words listed in (3) are heard by native speakers o Korean, they will invariably say that the /p/ sounds in ‘pin’ and ‘spin’ are different. Moreover, they would denote the /p/ in ‘pin’ and the /p/ in ‘spin’ in written orm with the Hangeul and respectively.
Basic concepts of phonology
44
Te reason why speakers o English and Korean perceive the same sounds differently is because the two languages have different inventories o phonemes. In other words, whether or not a consonant is aspirated can alter the meaning o a word in Korean, hence aspiration as a sound eature is signi�cant or Korean speakers; however, this is not the case or English speakers. In order to determine which sound properties are signi�cant in a language, we need to be able to distinguish the phonemes o that language. Guidelines or distinguishing phonemes are discussed below. 3.3
Criteria for phoneme identification
Suppose that we were asked to transcribe a sequence o speech sounds without any knowledge o the language in question and then to distinguish the allophones and phonemes within that sequence.3 How might we do this? In this section we will discuss criteria that can be used to distinguish phonemes and allophones. 3.3.1
Minimal pairs
Minimal pairs are pairs o words which differ in only one phonological element and have distinct meanings. Tat is to say, apart rom this one differing phonological element, such word pairs are phonetically identical, including having the same number o segments. For instance, ‘meal’ [mil] and ‘kneel’ [nil] orm a minimal pair, but ‘meal’ [mil] and ‘eel’ [il] do not because they have a different number o segments; ‘meal’ [mil] has three segments and ‘eel’ [il] has two. On the other hand, ‘meal’ [mil] and ‘mood’ [mud] do not orm a minimal pair because although they have the same number o segments, they differ in more than one segment. As shown by the minimal pair ‘meal’ and ‘kneel’, the phonetic difference between [m] and [n] has an effect on meaning in English. Te same phonetic dierence is also meaningul in Korean, as shown by the minimal pair [mɑl] ‘horse’ and [nɑl] ‘day’. Setting up minimal pairs is the key to determining whether or not a speech sound is a phoneme in any given language. 3.3.2
Complementary distribution, phonetic similarity and free variation
Looking at the distribution o speech sounds also provides insights useul or establishing an inventory o phonemes. I two speech sounds show complementary distribution, that is, they are mutually exclusive (when A appears B
45
3.3
Criteria for phoneme identification
does not and vice versa), it is highly unlikely that each sound acts independently as a phoneme. I A and B do not appear in the same phonetic environments, it will be impossible to �nd any minimal pairs or them. Tereore, i A and B are variants o the same phoneme, and thereore show complementary distribution, it should be impossible or them to orm minimal pairs. Most speech sounds that show complementary distribution are allophones o a single phoneme. Although each allophone is phonetically different, since they systematically occur in different environments, they are abstracted into one phoneme. Te three phonetic realisations o /p/ in English serve as one example o this phenomenon. Each sound shows complementary distribution: [ph] systematically appears in word-initial position; [p] in consonant clusters afer /s/; and [p ] in word-�nal position. Hence each becomes an allophone o the phoneme /p/. I this is the case, then can we say that all sounds that show complementary distribution are allophones o one phoneme? Let’s use the two criteria examined above, i.e. minimal pairs and complementary distribution, to see whether the sounds [h] and [ŋ] in English and Korean are two distinct phonemes, or allophones o one phoneme. In both Korean and English, [h] can be ound in syllable-initial position, as in [h ɑnɑ] and high [ha], but not in syllable-�nal positions. Conversely, [ŋ] can be ound in syllable-�nal position, as in [w ɑŋ] and king [k ŋ], but not in syllable-initial position. [h] and [ŋ] thereore show complementary distribution. However, no minimal pairs are available or these speech sounds in either language. In this case, can we say that the two sounds are allophones o one phoneme? Te answer is no, because there is no phonetic similarity between the two sounds. Phonetic similarity is the basis or grouping allophones and phonemes. Note that the three phonetic realisations o /p/ share a great deal o phonetic similarity. [p ] in word-initial position is a voiceless aspirated bilabial stop, [p] in a consonant cluster afer /s/ is a voiceless unaspirated bilabial stop and [p ] in word-�nal position is an unreleased voiceless bilabial stop. Phonetically speaking, all these sounds are voiceless bilabial stops. However, [h] and [ŋ] do not share any phonetic similarity. First o all, their respective places o articulation are different; [h] is glottal, whereas [ŋ] is velar. Moreover, they have different phonation types; [h] is voiceless, but [ŋ] is voiced. Te manner o articulation differs too, as [h] is ricative, but [ŋ] is nasal. Another important concept to consider is ree variation. Free variation reers to the phenomenon o two (or more) sounds or orms appearing in the same environment, without resulting in a change in meaning. I two sounds are realised in the same environment, but do not result in any change o meaning,
46
Basic concepts of phonology
then the phonetic difference between the two sounds is not signi�cant or native speakers o the language, and they cannot be regarded as two independent phonemes. For instance, in Korean, no distinction is made between /r/ and /l/. Hence, the English words rice and lice sound the same to a native speaker o Korean.4 In English, /r/ and /l/, which have minimal pairs such as rice and lice, are phonemes. In Korean, whether /r/ is pronounced as [ ɹ] or [l] its abstract orm is /l/. [ ɹ] and [l] in word-initial position can thereore be regarded as ree variants in Korean. Another example can be ound in released and unreleased stops in English. As discussed earlier, the /p/ in nip is pronounced [p ], where /p/ is unreleased. However, the /p/ in nip could also be pronounced [p], where /p/ is released, without any change in meaning rom the usual [p ]. Hence, sounds which show ree variation can be regarded as a single phoneme.
DID YOU KNOW . . .?
Superman and Clark Kent
We can understand principles o complementary distribution in phonemes and phonetic similarities more easily using the analogy o Superman and Clark Kent, or the analogy o a couple who cannot both attend a social unction because they have a newborn baby at home. First, let us look at the case o Superman and Clark. Clark disappears when Superman arrives on the scene, and when Clark comes back, Superman is gone. I can see that Superman is present only in emergency situations, while Clark is present in all other non-emergency situations. Tus, Superman and Clark show complementary distribution. Second, let us look at the case o the couple with a newborn baby. I see both o them individually at social gatherings, but I have never seen t he two o them together at a gathering. When Mrs K is present, Mr K is not, and vice versa. I can see that Mrs K attends gatherings which involve children and that Mr K attends all other social unctions. One can thereore say that their attendance o meetings also shows complementary distribution. Both Superman and Clark Kent, and Mr and Mrs K show complementary distribution, according to my observations. But why is it that I suspect that Superman and Clark might be the same person, but I think that Mr and Mrs K are two different people?
3.4
47
Position of sounds
Tis is due to similarities in the physical attributes o the �rst pair. Superman and Clark have the same height, the same build, and even the same acial eatures. However, Mr and Mrs K do not look alike at all. Mr K is 185 centimetres tall with broad shoulders, but Mrs K is 155 centimetres tall and very skinny. Teir acial eatures are also extremely different. Tus we might suspect that Superman and Clark are two different images o the same person, but we would never think the same or Mr and Mrs K. It is the same with sounds. Just because sounds show complementary distribution does not mean that they are allophones o the same phoneme. Tere must be phonetic similarities in addition to complementary distribution in order to conclude that two sounds are allophones. Because the voiceless glottal ricative /h/ and the voiced velar nasal /ŋ/ in both Korean and English show complementary distribution, but share no phonemic similarities, they are two different phonemes rather than allophones o the same phoneme. Tis corresponds to the analogy o Mr and Mrs K, who are just so physically different that even though they show a complementary distribution o sorts, they must be seen as two distinct individuals. Going off topic slightly, we may �nd it diffi cult to understand how Lois Lane doesn’t realise that Superman and Clark are the same person. Perhaps, the best way to distinguish who is who would be to ask both o them to turn up at the same place at the same time. Essentially, this is the logic behind the principle o complementary distribution.
3.4
Position of sounds
3.4.1
Sounds within a syllable: onset, nucleus and coda
A syllable, like a phoneme, is an abstract unit. Phonemes do not exist independently, but are grouped together with other phonemes according to a set o principles. Te unit which is ormed by a sequence o phonemes is termed a prosodic unit. A syllable is the smallest prosodic unit ormed by phonemes. 5 A syllable has at least one phoneme and an internal structure with three elements: onset, nucleus and coda, as shown in (4). O these, the nucleus is the most important part, being absolutely necessary or the ormation o a syllable. (4)
Syllable
Onset
Nucleus
Coda
Basic concepts of phonology
48
In Korean, only vowels can act as the nucleus o a syllable, whereas consonants can only be the onset or coda o a syllable. Only one consonant is allowed in either the onset or coda o a Korean syllable. Tis is not the case in English. As demonstrated by the words listen [ lsn] and shuttle [ʃtl], both vowels and consonants can be used as the nucleus o a syllable.6 Consonant clusters can also appear in the syllable-initial or syllable�nal position, as in the case o strip [strp] and texts [teksts]. A maximum o three consonants can appear in syllable-initial position and our in syllable�nal position. Te concepts o onset, nucleus and coda are universal to all languages, but the ways in which these come together to orm syllable structures vary rom language to language. 7 3.4.2
Sounds in word-initial, word-medial and word-final position
Te phonetic realisation o sounds is in�uenced by where they occur in a word. A consonant in word-initial position is always an onset, and a consonant in word-�nal position is always a coda. However, a consonant in word-medial position can be either an onset or a coda depending on the environment in which it occurs. A word-medial consonant can also appear either afer another consonant or between vowels. For instance, [m] and [p] in sympathy both occur in word-medial position, but [m] is a coda, whereas [p] is an onset. Te [p] in sympathy [s ´mpəɵ ] comes afer another consonant [m], but the same word-medial [p] also appears in the word topic [tɔ´pk] between vowels. In Korean, a coda sounds the same regardless o its position in a word. However, onsets are realised differently according to their position in a word. Hence, in comparison to word-medial and word-�nal onsets, word-medial and word-�nal codas are rarely discussed in Korean linguistics. In this book, ‘word-initial’ is taken to mean a word-initial onset and ‘word-medial’ is taken to mean a word-medial onset. However, ‘word-�nal’ reers here only to a word-�nal coda. With regard to the word-initial onset, we have mostly discussed only those cases where it occurs between vowels – i.e. in ‘intervocalic position’ – although, phonetically speaking, a word-medial onset occurring afer a consonant is not vastly different rom a word-medial onset that occurs between vowels. 3.5
Underlying form, surface form and phonological rules
Te underlying orm (UF) or underlying representation (UR) o a morpheme reers to its phonemic inormation. Tis is encoded in the mental lexicon o native speakers. On the other hand, surace orm (SF) or surace representation
49
3.5
Underlying form, surface form and phonological rules
(SR) o a morpheme reers to its phonetic orm – what is actually heard when the morpheme is spoken. In generative phonology, SF is achieved by applying a phonological rule (PR) to UF. Tis process is called derivation. (5)
PR UF
SF
o understand the notion o UF, let’s consider how the morpheme { k*oʨʰ} ‘�ower’ is realised. (6) a. . [k*oʨʰi| pʰinta] b. . [k*onman| t*anta] c. ! [k*ot] {k*oʨʰ} can have three different phonetic realisations; [k *oʨ], [k*on], [k*ot]. Tough they sound different, they all mean the same thing, ‘�ower’. In this case, or simplicity’s sake, it is better to consider one o these orms as the underlying orm, and the others as being derived rom that orm via different phonological rules. In this way, by positing the concept o underlying orm, we achieve simpler and more convincing explanations o phonological phenomena. Useul criteria or de�ning underlying orm include predictability, economy and plausibility. Let’s see how predictability is used in determining underlying orms. First o all, in positing that /k*oʨ/ is the underlying orm o {k *oʨʰ} we predict that /k*oʨʰ/ becomes [k*ot] at the end o a syllable, but [k *on] in ront o a nasal. o validate this underlying orm, we need to �nd examples o / ʨʰ/ becoming [t] in syllable-�nal position, as in [k*ot], but [n] in ront o nasals, as in [k *onman]. On the other hand, i /k *on/ is chosen as the underlying orm or {k *oʨʰ}, we must predict that the abstract sound /k*on/ becomes [k *ot] at the end o a syllable, but becomes [k*oʨʰ] between vowels. Likewise, or /k*ot/ to be the underlying orm or {k *oʨʰ}, we have to predict that /t/ in /k*ot/ becomes nasalised in ront o a nasal, but becomes [ ʨʰ] between vowels. Again, or this prediction to be true, we need to �nd examples where /t/ becomes [n] in ront o a nasal but [ ʨʰ] between vowels in Korean. o evaluate the second and third predictions, consider the word {k ɑn} ‘liver’. /k ɑn/ is realised as [k ɑn] in syllable-�nal position, but as [k ɑn] in intervocalic position. Tis proves that the second prediction is wrong. Consider another example {kot-} ‘straight’. Te underlying orm o [kot], /kot-/ is realised as [kotɑ] or [kotɯni] between vowels. Hence, the third prediction, that /t/ becomes [ʨʰ] in intervocalic position, is also wrong.
50
Basic concepts of phonology
Finally, let’s evaluate the �rst prediction. Te underlying orm o {n ɑʨʰ} ‘ace’ is /nɑʨʰ/. As in the case o /k*oʨʰ/, /nɑʨʰ/ is realised as [nɑt] at the end o a syllable, and as [nɑn] in ront o a nasal. Tereore, the �rst prediction, that / ʨʰ/ becomes [t] in syllable-�nal position, but becomes [n] in ront o a nasal, is valid. Next, let’s look at economy. Tis is related to the number o phonological rules and the efficiency o grammar. Since underlying orm is an abstract concept, the underlying orm or {k *oʨʰ} could thereore be either /k *on/ or even /k*o&/, which contains the very abstract phoneme /&/. Nevertheless, i we take /k*on/ as the underlying orm, then in order to derive [k *oʨʰ] rom /k*on/, we need a rule whereby /n/ changes into [ ʨʰ] between vowels. However, when we �rst selected /k*oʨʰ/ as the underlying orm, no such rule was required. In other words, i we select /k *on/ as the underlying orm we require one rule more than we did or /k *oʨʰ/ as the underlying orm, hence it can be said to be less economic. Moreover, this rule does not state that any /n/ appearing between vowels will turn into [ ʨʰ], just the /n/ in {k*oʨʰ}, ‘�ower’. Tereore, when /k*on/ is the underlying orm o the morpheme {k *oʨʰ}, there is a urther problem, in that the application o this rule is subject to special conditions. Furthermore, i the abstract orm /k *o&/ is set as the underlying orm or {k *oʨʰ} ‘�ower’, even more rules are required to link this underlying orm to its various surace representations. Needless to say, this is ar less economic. Finally, let us consider plausibility. Te rules by which ‘plausibility’ is derived or, one might say, the methods by which it is resolved, are natural parts o human language. In other words, plausibility is linked to what is natural in a language. In the example above, i /k *on/ is taken as the underlying orm or {k*oʨʰ} ‘�ower’, a rule is required whereby /n/ becomes [ ʨʰ] between vowels. However, or the nasal /n/, which is a sonorant consonant, to change into a voiceless aspirated obstruent such as [ʨʰ] between voiced sounds is always an unnatural process in phonetics. Tereore, it is difficult to consider this rule plausible. Underlying orms are determined by taking the above points, among others, into consideration, and in order to avoid excessive abstraction, an underlying orm is usually chosen rom one o its surace orms. In cases like the one above, where an abstract underlying orm must be determined in the absence o a suitable surace orm, an alternative rule or deriving the underlying orm is necessary.
3.6
Phonological features
So ar, we have assumed that a phoneme is the smallest unit o speech sounds. However, a phoneme can be analysed urther. In Chapter 2, we discussed how speech sounds can be classi�ed as either consonants or vowels. Consonants
51
3.6
Phonological features
can be urther classi�ed according to place o articulation, manner o articulation and phonation type. Vowels on the other hand can be classi�ed according to the height o the tongue (high or low), the backness o the tongue (e.g. back or ront) and roundedness o the lips (e.g. rounded or unrounded). With the advent o eature theory in phonology, the phoneme has come to be understood as a set o phonological eatures whose value can be represented either as positive (+) or negative (−). Phonological eatures are also called distinctive eatures. In this section, we will discuss the necessity o postulating phonological eatures and major class eatures. Features or consonants and eatures or vowels will be discussed in Chapters 4 and 5 respectively.
FURTHER STUDY
Chomsky and Halle (1968, abbreviated hereafer as SPE) criticised Jakobson, Fant and Halle (1952, abbreviated hereafer as JFH) or ocusing only on acoustic eatures. Tey then proposed to look at the articulatory eatures o sounds. Tough at times the eature system has been criticised, it nevertheless orms the basis o discussions and analyses in phonology. We will ollow the basic eature systems proposed in SPE, though we will only ocus on what is relevant to discussions o Korean phonology. Although the eature system in SPE applies universally to all languages, it contains certain eatures that are unnecessary in establishing a system or Korean. In addition, we have replaced some o the eatures rom SPE in order to provide a better explanation o Korean phonology. Tese newly introduced eatures include [+/−tense] and [+/−aspirated]. In SPE, phonological eatures are divided into the ollowing our major categories: major class eatures, cavity eatures, manner o articulation eatures and source eatures. Tis book, however, does not ollow SPE’s categorisation method. Instead, phonological eatures are divided into (i) major class eatures, (ii) eatures or consonants and (iii) eatures or vowels. Features or consonants are urther divided into the necessary eatures or manner o articulation, place o articulation and phonation types.
3.6.1
Why are phonological features needed?
Consider the Korean examples in (7). (7) /kuk + mul/ → [kuŋmul]
k → ŋ/ _ m8
‘the liquid part o a dish’
52
Basic concepts of phonology
/pɑp + mul/ → [pɑmmul]
p → m/ _ m
/ʨʰʌs + mʌli/ → [ʨʰʌnmʌli] s → m/ _ m /sok + nɛ- + -ɰi/ → [soŋnɛi ] k → ŋ/ _ n /thop + nɑl/ → [thomnɑl] p → m/ _ n /ʨʰʌs + nun/ → [ʨʰʌnnun] s → n/ _ n
‘water or boiling rice’ ‘the beginning’ ‘inner wear’ ‘a saw tooth’ ‘a �rst look’
(+ = morpheme boundary) In (7), we can see that the phonemes that undergo change, /k/, /p/, and /s/, the resulting phonemes, /ŋ/, /m/, and /n/, and the phonemes /m/ and /n/ that appear in the environment where change occurs all seem to have phonological similarity.9 Yet, this similarity is difficult to de�ne. Furthermore, it is di�cult to de�ne a clear reason as to why this sequence o obstruents is nasalised. However, by considering their phonological eatures, we discover a clue to this question. In (7), the input consonants are all obstruents and all o them are ollowed by nasals. Now this problem can be explained by stating that obstruents became nasalised when ollowed by nasal consonants. By positing eatures, it is possible not only to explain such phenomena but also to make predictions using relevant data. In the ollowing section, we will discuss major class eatures.
3.6.2
Major class features
Phonemes can be classi�ed into consonants, vowels and glides according to their phonetic and unctional characteristics. Consonants can be urther divided into obstruents and sonorants. A eature may be classed as one o two values, either positive (+) or negative (−). Tereore, in order to de�ne the three categories above, that is, consonant, vowel and glide, at least two eatures are needed. Tey are [consonantal] and [syllabic]. De�nitions o these eatures are given in (8). (8)
Major class eature (I) a. [+/−consonantal]: [+consonantal] sounds produced with radical obstruction o the midsaggital region o the vocal tract. b. [+/−syllabic]: [+syllabic] sounds are able to orm a syllable peak, and can thereore stand alone as a syllable.
According to the de�nition given in (8), the air�ow o a sound with the value [+consonantal] is obstructed while passing through the central part o the oral cavity, unlike a sound with the value [−consonantal]. Hence, consonants are [+consonantal], whereas vowels and glides are [−consonantal].
53
3.6
Phonological features
Table 3.1 Major class features I Consonant
Vowel
Glide
[consonantal]
+
−
−
[syllabic]
−
+
−
In order to have the value [+syllabic], a sound must be able to constitute a syllable on its own. Vowels can have the value [+syllabic], whereas consonants and glides can only have the value [ −syllabic]. Based on this observation, we can classiy consonants, vowels and glides in the manner shown in able 3.1. In order to distinguish obstruents rom sonorant consonants, another eature, [sonorant], is required. A sonorant can be de�ned as shown in (9). (9)
Major class eature (II) a. [+/−sonorant]: [+sonorant] sounds are produced with a vocal tract cavity con�guration in which spontaneous voicing is possible.
In the above de�nition, “a vocal tract cavity con�guration in which spontaneous voicing is possible” reers to a vocal tract con�guration which satis�es the aerodynamic conditions necessary or voicing. In other words, sonorants can be understood as a group o sounds which have the necessary conditions or voiced sounds to be produced with spontaneous voicing as the channels that they pass through are enlarged, causing air pressure in the vocal tract to drop. Tereore, all sonorant consonants have the value [+sonorant], as do all glide vowels. As shown in able 3.2, phonemes can be classi�ed into the major classes: obstruent, sonorant, consonant, vowel and glide, according to the three major class eatures [consonant], [syllabic] and [sonorant]. In this way it is possible to clariy the similarities and differences between the major categories o sounds. Table 3.2 Major class features II Consonant Obstruent
[consonantal] [syllabic] [sonorant]
Sonorant consonant
Vowel Glide
+
+
−
−
−
−
+
−
−
+
+
+
Basic concepts of phonology
54
3.7
Summary
In this chapter, we have discussed how to identiy the most basic units o sounds (Sections 3.1–3.3) and their structural properties (Section 3.4). Te same segment sequences can be interpreted differently by speakers o dierent languages due to the phonological differences between languages. We have also discussed the basics o major phonological concepts. First, in 3.1 we saw that there are two types o speech sounds, namely phones and phonemes. Whereas phonemes represent the psychological aspects o a sound, phones represent sounds as physical entities. Not all phones correspond to phonemes, and the inventory o phonemes differs in each language. For instance, in English, the phoneme /p/ is realised as [p h], [p] or [p ] according to the phonetic environment in which it appears. Tough each phonetic realisation is different, English native speakers are unable to distinguish between them. Hence, these sounds are allophones o one phoneme, /p/. In 3.3, we discussed criteria or identiying phonemes. Setting up minimal pairs, with two words whose meaning differs through only one sound in the same position in each word being different, plays a key role in establishing the phoneme inventory o a language. I two sounds show complementary distribution, then no minimal pairs are available or them; hence they cannot be allophones o one phoneme. However, phonetic similarity between sounds is essential in identiying the relationships between phoneme and allophones. wo (or more) sounds can appear in the same environment without resulting in a change in meaning, as in the case o [ɹ] and [l] in Korean. Tey are regarded as ree variants o the phoneme /l/. For a better explanation o phonological processes, however, we need to think about the sound properties which constitute each phoneme. Tese are called phonological eatures or distinctive eatures. We discussed major class eatures in 3.6. Features or consonants and vowels are discussed in Chapters 4 and 5 respectively. EXERCISES 1
2
Please state whether or not the following pairs of words form minimal pairs in English and explain why. a. pin, spin b. lice, r ice c. mince, nice d. nice, mice Explain the difference between a phone and a phoneme with examples.
Exercises
55
3 4
5 6
Can length or accent also be regarded as phonemes? Explain why or why not. Explain the benefits of using features in phonological explanation. Provide examples. What is a “phonological (distinctive) feature”? Define the features [sonorant] and [syllabic] with examples.
4
Consonants
In Chapters 2 and 3, general aspects o sound systems were discussed. In Chapter 4, consonants in Korean will be discussed, and in particular, what consonants exist in Korean and what their phonetic and phonological characteristics are. With this aim, in 4.1, the inventory o Korean consonants will be discussed; in 4.2, the phonetic and phonological characteristics o obstruents; and in 4.3, the phonetic and phonological characteristics o sonorants. In 4.4, we will discuss phonological eatures needed to discuss consonants in Korean, and in 4.5 we conclude. 4.1
Consonant chart
able 4.1 shows the consonants o Korean classi�ed according to the place (across the top o the table) and manner (down the lef side o the table) o their articulation. Te nineteen consonants o Korean can be classi�ed as either (i) bilabial, (ii) alveolar, (iii) alveolo-palatal, (iv) velar or (v) glottal according to the place o their articulation. Tey can also be classi�ed as (i) stop, (ii) ricative, (iii) affricate, (iv) nasal or (v) liquid (lateral approximant) according to the manner o their articulation.1 Furthermore, each consonant class has two or three sub-classi�cations. For instance, stop sounds have lax, tense or aspirated sounds, and ricative sounds can be classi�ed as either lax or tense. DID YOU KNOW . . . ?
Dialectal differences in the list of consonants
Tere are ew differences in the inventory o consonants between regional dialects. In certain dialects the phonetic realisation o consonants is
56
4.2
57
Obstruents
different, but the actual list o consonants is usually the same as that o Standard Korean. However, in certain regional dialects o Gyeongsang province, no contrast between /s/:/s */ : is observed. Tus, to the speakers o these regional dialects, there is no audible difference between /sɑl/ ‘�esh’ and /s*ɑl/ ‘rice’.
Table 4.1 Consonants
Stop (plosive) Lax ense Aspirated Fricative Lax ense Affricate Lax
4.2
Bilabial
Alveolar
Alveolo-palatal Velar
p () p* ()
t () t*()
k () k* ()
pʰ ()
tʰ ()
k ʰ ()
s () s* ()
Glottal
h ()
ʨ ()
ense
ʨ* ()
Aspirated Nasal Liquid (lateral approximant)
ʨʰ () m ()
n () l ()
ŋ ()
Obstruents
An obstruent is a sound made as air�ow is obstructed by narrowing o the vocal tract. Korean has a rich inventory o these sounds; �feen o its nineteen consonants are obstruents. Te main characteristic o Korean obstruents is that they are, without exception, voiceless. Te phonetic and phonological characteristics o Korean obstruents will be discussed individually below. 4.2.1
Stops
As shown in able 4.2, Korean stop consonants, that is, bilabial, alveolar and velar sounds, all have lax, tense and aspirated types. First o all, let’s discuss the
58
Consonants
Table 4.2 Stops
Bilabial
Alveolar
Alveolo-palatal
Velar
Lax ense
p () p* ()
t () t*()
k () k* ()
Aspirated
pʰ ()
tʰ ()
k ʰ ()
Glottal
places o articulation or Korean stop sounds. Bilabial sounds are produced using the lips. Alveolar sounds are pronounced with the tip o the tongue (or the tip and the blade) touching the teeth and alveolar ridge. Velar sounds are pronounced with the body o the tongue touching the velum. As the tongue is not involved in their articulation, bilabial sounds are not in�uenced by surrounding vowels. However, alveolar and, to a greater extent, velar sounds are affected by surrounding vowels. Let’s look at some veri�able minimal sets or Korean stop sounds. (1)
Minimal sets /p/ : /p*/ : /p/
/t/ : /t*/ : /t/
/k/ : /k*/ : /k /
/pul/ ‘�re’ : /p*ul/ ‘horn’ : /pul/ ‘grass’ /pɑn/ ‘class’ : /p*ɑn/ ‘to suck (adnominal)’ : /pɑn/ ‘board’ /pε/ ‘ship’ : /p*ε/ ‘to extract (imperative)’ : /pε/ ‘tag’ /pitɑ/ ‘be empty’ : /p*itɑ/ ‘to sprain’ : /pitɑ/ ‘to bloom’ /tɑl/ ‘moon’ : /t*ɑl/ ‘daughter’ : /tɑl/ ‘mask’ /tɑ/ ‘all’ : /t*ɑ/ ‘to pick (imperative)’ : /tɑ/ ‘to ride (imperative)’ /tɑn/ ‘column’ : /t*ɑn/ ‘to pick (adnominal)’ : /tɑn/ ‘to ride (adnominal)’ /tε/ ‘bamboo’ : /t*ε/ ‘dirt’ : /tε/ ‘crack’ /tltɑ/ ‘deduct’ : /t*ltɑ/ ‘tremble’ : /tltɑ/ ‘shake off’ /k ɑn/ ‘liver’ : /k*ɑn/ ‘estimation’ : /k ɑn/
‘blank’ /ki/ ‘energy’ : /k*i/ ‘talent’ : /k i/ ‘one’s height’ /kεtɑ/ ‘to old’ : /k*εtɑ/ ‘to break’ : /k εtɑ/ ‘to dig’ In the case o velar consonants, where the body o the tongue is involved in articulation, the point at which air�ow is blocked varies according to the ollowing
59
4.2
Obstruents
Table 4.3 Voiceless alveolar stops in Korean and English
Korean
/tʰɑ/ /tʰi/
English
‘to ride (imperative)’ ‘a speck’
/tɑ/ /ti/
‘ta’ ‘tea’
vowel. For instance, when /k ɑ/ and /ki/ are pronounced by a native speaker, the /k/ in /ki/, which is ollowed by a ront vowel, is pronounced more at the ront than the /k/ in /k ɑ/, which is ollowed by a back vowel. Phonetically speaking, the /k/ in /ki/ is actually closer to a palatal sound than to a velar sound. Velar sounds that appear in onset position are more in�uenced by the vowel that ollows. Alveolar sounds in Korean are articulated with the blade o the tongue (or the tip and the blade) touching the teeth and alveolar ridge, as air�ow is blocked and then immediately released. Alveolar sounds in Korean are slightly different rom English alveolar sounds, in that English alveolars are articulated urther back than in Korean; Korean alveolar sounds are articulated with the blade or the tip o the tongue touching the teeth and alveolar ridge, whereas with English alveolars, the tongue only touches the alveolar ridge. For this reason, it sounds awkward when the Korean sounds /t ɑ/ and /ti/ are pronounced like the English ‘ta’ and ‘tea’, or conversely i the English words ta and tea are pronounced as /tɑ/ and /ti/ . (See able 4.3). Let’s look at the three classi�cations o Korean stop sounds. In Chapter 2 , we de�ned the three pronunciation orms or Korean stop consonants as lax, tense and aspirated, and characterised them as possessing the phonetic eatures voiceless unaspirated lenis, voiceless unaspirated ortis and voiceless aspirated ortis respectively. Lax sounds can appear in both onset and coda positions, and when realised as onsets they can appear either in word-initial or word-medial position. By contrast, tense and aspirated sounds may only appear in onset position, and cannot be realised in coda position. 2 Nevertheless, all o these sounds may be realised as word-initial and word-medial onsets. DID YOU KNOW . . . ?
Which Korean food do you like? (rib) or (noodle)?
Harry is an Englishman who came to Korea to study the Korean language. He ofen met up with his Korean riend to practise speaking Korean, and one day, they talked about their avourite Korean dishes. When the Korean
60
Consonants
riend asked him what he liked best, Harry immediately replied “” [k ɑlpi]. Ten, the Korean riend taught him that he had to pronounce it “” [k ɑlpi], not “” [k ɑlpi]. Soon afer, the Englishman went to a Korean restaurant with his Korean riend. At the restaurant, his riend asked him what he would like to eat. Tis time, the Englishman replied very con�dently, “” [k ɑlkuks*u]. Unortunately or him, this time it should have been /k /, not /k/.
4.2.1.1
Word-initial stops
Figure 4.1 shows a wide-band spectrogram and waveorm o a native Korean speaker, o the standard Seoul dialect, a woman in her orties, 3 saying the syllables /tɑ/, /t*ɑ/ and /tɑ/. Te spectrogram displays a two-dimensional plane rom a three-dimensional image o changes in the intensity and requency o the sound signal as time passes. Te horizontal axis denotes time and the vertical axis denotes requency. Intensity is represented using grey level mapping: when intensity is high in relation to time and requency, the spectrogram will be darker, and when intensity is low in relation to time and requency, it will be ainter. Te waveorm is a two-dimensional image which displays changes in the intensity o a sound signal over time. ime is displayed along the horizontal axis, and intensity is displayed along the vertical axis. Te letters (a), (b) and (c) denote /tɑ/, /t*ɑ/ and /tɑ/ respectively,4 and the arrow shows the point o release or each. As can be seen in Figure 4.1, all three types o stop sound are released afer closure. Te momentary expulsion o air due to the release o the stop sound can be observed on the spectrogram as a vertical spike in the position indicated by the arrow. Te white space which appears on the spectrogram beore the vertical spike is a eature that appears when the level o sound energy is close to zero. Let’s look more closely at the three kinds o Korean stop consonant on the spectrogram and waveorm. First o all, let’s look at closure, that is, the point beore the vertical spike created by the release o air. It can be seen that the three kinds o Korean stop consonant are all voiceless, that is, the stop sound is not accompanied by any vibration. I they were voiced sounds, then a periodic wave with small oscillations during the period o closure would be visible on the waveorm, and on the spectrogram, an area o low requency would be visible in the orm o bars at regular intervals, known as voice bars, in the lower part o the diagram.
61
4.2
Obstruents
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a)
(b) t
ɑ
t
(c) ɑ
0
th
ɑ 2.0
Time (s)
Figure 4.1
Spectrogram and waveform of lax, tense and aspirated alveolar stops in word-initial position
Now let us turn our attention to the sound characteristics displayed by the three kinds o stop sounds afer their release. Te three show wide variance in phonation type afer the vertical spike (which is common to all three). First o all, in the lax and aspirated sounds, a blurred mark can be seen on the right o the picture just afer the vertical spike appears on the spectrogram, and an aperiodic wave can be seen on the waveorm; 5 this is not the case or tense sounds, or which it would be diffi cult to discern either a blurred mark on the spectrogram or an aperiodic wave on the waveorm. 6 Similar eatures can be observed on the spectrograms and waveorms o the three kinds o stop sounds, and show that all stop sounds are accompanied by a degree o riction in the glottis afer their release. Tat is to say, they are produced with a degree o aspiration. Based on this, we can see that although tense sounds in Korean are produced with almost no aspiration, lax and aspirated sounds in word-initial position are articulated with aspiration. It can also be seen that compared to lax sounds, aspirated sounds have a ar greater degree o aspiration. Lax sounds in Korean have been de�ned above as lenis voiceless unaspirated sounds. However, having observed their real phonetic orms, it can be seen that lax sounds are aspirated, not unaspirated in word-initial position. I this is the case, then why must Korean lax sounds be viewed as unaspirated sounds phonologically? In order to answer this question we must take a more comprehensive view o the characteristics o Korean stop sounds, including those realised not only in word-initial position but in other environments too.
62
Consonants
TRY YOURSELF!
Ask a Korean riend to make the ollowing sounds in sequence – /pɑ/, /p*ɑ/, /pɑ/ – and place your hand in ront o your riend’s mouth. Ten notice the difference in the amount o air�ow that can be elt on your palm. When she/he makes the /pɑ/ sound, which is lax, you should eel a slight �ow o air on your palm, but when she/he makes the /p *ɑ/ sound, which is tense, you should hardly eel any air�ow. When she/he makes the aspirated /pɑ/ sound, however, you should eel a strong current on your palm. Te amount o air�ow that can be elt on your palm thereore correlates to the degree o aspiration; whether it is lax, tense or aspirated. I your native language is English, try to pronounce pie and bye with your palm in ront o your mouth. When you make the pie sound, you ought to eel a strong rush o air as with aspirated sounds in Korean, but when you make the bye sound, you should eel almost nothing, as with tensed sounds in Korean.
DID YOU KNOW . . . ?
Distinguishing the difference between three types of stops
As the story o and shows, the three-way contrast in word-initial consonants is particularly difficult or learners o Korean to distinguish. Stop sounds in Korean in word-initial position are generally divided into two categories; those with aspiration, i.e. lax and aspirated stops, and those without aspiration, i.e. tense stops. In English, stops can be phonologically divided into voiced and voiceless stops. However, all stops are voiceless in word-initial position. Te only difference observed among stops in word-initial position is thereore the presence/absence o aspiration, as in Korean. Hence, Korean and English native speakers can both easily distinguish unaspirated stops in word-initial position. Ofen, Koreans regard them as tense consonants, while English speakers regard them as voiced consonants. However, it is not easy or English speakers to distinguish lax and aspirated stops in Korean, since both sounds are aspirated. Te only difference between them lies in the degree o aspiration. Lax stops are aspirated but the degree o aspiration is very weak compared to that o
63
4.2
Obstruents
aspirated stops. More importantly, the pitch o the vowel that ollows a lax stop is very different to the pitch o the vowel that ollows an aspirated stop. When you listen careully to the pronunciation o [p ɑ] and [pɑ] by Korean speakers, you will be able to tell that the [ ɑ] that ollows the aspirated stop is much higher in pitch than the [ ɑ] that ollows the lax stop. Te diagram below shows the pitch difference in the pronunciations o [tɑlilɑs] ‘because (it is) a moon’, where the �rst consonant is a lax stop (a) and [tɑlilɑs] ‘because (it is) a mask’, where the �rst consonant is an aspirated stop (b). a.
b.
[tɑlilɑsʌ]
[tʰɑlilɑsʌ]
Pitch patterns in word-initial position are important in understanding the prosody o Korean. Tis will be discussed urther in 7.5.
4.2.1.2
Word-medial (intervocalic) stops
Figure 4.2 shows the spectrogram and waveorm o three alveolar stops pronounced between two /ɑ/ vowels. (a), (b) and (c) represent [ ɑtɑ],7 [ ɑt*ɑ] and [ɑtʰɑ] respectively. One directional arrow indicates the time o the release and the other, double-ended arrow indicates the closure period. 8 All three alveolar stops show the release, but beore and afer the point o release they differ signi�cantly. Most signi�cantly, the duration o closure differs; the tense stop displays the longest closure duration ollowed by the aspirated stop, while the lax stop has the shortest closure duration.9 Te second noticeable difference is seen in the presence, or absence, o aspiration afer the release. Afer the release o lax and tense sounds, the vocal olds begin to vibrate immediately or the ollowing vowel, as can be observed. However, in aspirated sounds, the point when the vocal olds begin to vibrate is relatively late due to the wide glottal opening. Tis results in longer VO and strong glottal riction beore the beginning o vibration o the vocal olds or the ollowing vowel. In other words, in word-medial position, lax and
64
Consonants
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a) ɑ
(b) t
ɑ
ɑ
(c) *
t
0
ɑ
ɑ
th
ɑ 2.0
Time (s)
Figure 4.2
Spectrogram and waveform of lax, tense and aspirated alveolar stops in word-medial (intervocalic) position
tense sounds are articulated as unaspirated sounds, whereas only aspirated sounds are articulated with aspiration. It should be noted that the presence o aspiration in the lax sound here is different rom when it appears as a wordinitial onset. Lax sounds in word-initial position are realised phonetically as aspirates, whereas they are realised as unaspirated sounds in word-medial position. However, tense sounds and aspirates are pronounced the same in both word-initial and word-medial position, as unaspirated and aspirated sounds respectively.10 Te third difference observed is whether or not there is any vibration o the vocal cords during the closure period. Figure 4.3, which is a magni�cation o (a) in Figure 4.2, shows that the vocal olds vibrate when a lax stop appears between vowels. Te period o voicing or the lax sound can be seen on the spectrogram as a voice bar and on the waveorm as a periodic wave, which occurs or the period o closure marked by a double-headed arrow. In other words, this means that unlike tense or aspirated stops, lax stops are voiced between vowels. In addition, a difference can be seen in the length o the vowel that precedes the closure period, which varies according to the phonation type o the stop sound. As shown in Figure 4.2, the vowel preceding the lax stop is much longer than the vowels that precede the tense or aspirated stops. 11 However, the most important thing to note is that lax stops in word-medial position are realised as voiced unaspirated sounds, which is different rom word-initial position.12 Unlike lax stops, tense and aspirated stops do not change rom word-initial position, and are realised as voiceless unaspirated sounds and voiceless aspirated sounds respectively.
65
4.2
Obstruents
Spectrogram Hz 8000 6000 4000 2000
Waveform
ɑ
t
ɑ
0
0.5 Time (s)
Figure 4.3
Spectrogram and waveform of the lax alveolar stop in word-medial (intervocalic) position
4.2.1.3
Word-final stops
Finally, we will discuss the characteristics o word-�nal stops. Because stop consonants in word-�nal position should not be released afer closure, only lax stops can occur in word-�nal position. Figure 4.4 shows the spectrogram and waveorm o a word-�nal lax stop. (a) indicates the duration o the vowel sound and (b) indicates the duration o the word-�nal lax stop. 13 No vertical spike indicates that the air has not been released afer the closure. Hence, these sounds are realised as unreleased consonants. 4.2.1.4
Allophones and phonological identity of lax stops
So ar, we have discussed how, unlike tense or aspirated stops, lax stops vary according to where they occur. Tat is, in word-initial position lax stops are aspirated, and in the word-medial position they are released as voiced unaspirated sounds. Finally, in word-�nal position, they are released as voiceless unreleased stops. o demonstrate the various allophones o lax stops, we have provided a spectrogram and waveorm o the word /k ɑk ɑk/.14 In Figure 4.5, (a) indicates the word-initial /k/; (b) indicates the word-medial /k/ and (c) indicates the word-�nal /k/. Te slanted arrow again indicates the point o release afer closure. Figure 4.5 clearly shows how the same phoneme /k/ can be realised differently according to the phonetic environment in which it occurs. Aspiration due to glottal riction can be seen in word-initial /k/ shown at (a). Conversely, word-medial, intervocalic /k/ at (b) has no such riction, but does show
66
Consonants
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a)
(b) ɑ
t
0
0.5 Time (s)
Figure 4.4
Spectrogram and waveform of the lax alveolar stop in word-final position Spectrogram Hz 8000 6000 4000 2000
Waveform
(a)
(b) k
ɑ
(c) k
ɑ
0
k 0.7
Time (s)
Figure 4.5
Spectrogram and waveform of the word /kɑkɑk/ ‘corner of the road’
vibration o the vocal olds. In other words, it is a voiced unaspirated sound. In word-�nal position at (c), /k/ is realised as a voiceless unreleased sound. One question remains: i a lax stop is realised differently according to where it occurs, then why is the lax stop regarded phonologically as a voiceless unaspirated sound? o answer this question, we need to observe how native Korean speakers interpret and use lax stops phonologically. First, let’s see why native speakers o Korean regard lax stops not as aspirated, but as unaspirated sounds. In Korean, when a lax stop is preceded or ollowed by a glottal /h/ at a morphological boundary, it becomes aspirated through the process o phonological contraction. Examples are given in (2).
67
4.2
Obstruents
(2) a. /noh + ko/ → [nok ʰo] ‘to put (connective)’ b. /pʌp + hɑk/ → [pʌpʰɑk] ‘law’ (+ = morpheme boundary) Te aspiration processes shown in (2) demonstrate that native speakers o Korean regard lax stops as unaspirated sounds. I we assumed that lax stops in Korean were aspirated sounds, then the examples in (2) would appear to show an aspirated sound becoming another aspirated sound when it meets a glottal /h/, which is itsel an aspirated sound.15 It is difficult to �nd a good reason to explain why an aspirated sound should change into another aspirated sound in this case, hence why lax stops are regarded as unaspirated sounds. Furthermore, native speakers o Korean perceive lax stops as being voiceless. Te phonological behaviour o lax stops is very different rom that o voiced consonants such as nasals or liquids. I lax stops were voiced consonants, they would show some similarity with other voiced consonants. However there are no phonological processes in Korean that can be applied exclusively to voiced consonants and lax stops. Te process o Neutralisation, which will be discussed in 8.1.1, also supports the view that lax stops are voiceless. Due to Neutralisation, aspirated or tense consonants cannot appear in syllable-�nal position in Korean. Instead, those consonants are replaced with their lax counterparts as shown in (3). (3) a. /pɑk*/ → [pɑk] ‘outside’ c.. /pɑk* + i/ → [pɑ.k*i] ‘outside (subjective marker)’ b. /puʌk ʰ/ → [pu.ʌk] ‘kitchen’ c.. /puʌk ʰ + i/ → [pu.ʌ.k ʰi] ‘kitchen (subjective marker)’ c. /pɑk/ → [pɑk] ‘gourd’ c.. /pɑk + i/ → [pɑ.ki] ‘gourd (subjective marker)’ (. = syllable boundary) In other words, the distinction between tense and aspirated sounds in syllable-�nal position is neutralised. I lax stops were voiced, this rule would suggest that voiceless unaspirated sounds and voiceless aspirated sounds b oth become voiced in syllable-�nal position. However, cross-linguistically, syllable�nal position tends to be where voiced sounds become voiceless rather than the other way round. o say that Korean is an exception to this rule would not be a satisactory explanation. In terms o distribution, lax sounds are much more prevalent than tense or aspirated ones. As will be discussed urther in Chapter 6, out o lax, aspirated and tensed consonants, lax sounds are used the most, constituting 68.6 per cent and 71.9 per cent o dictionary usage and spontaneous speech respectively.
Consonants
68
Te �nal thing to add is that Korean lax stops are not always realised as voiced sounds in word-medial position. In the above examples, we saw how Korean lax stops are realised as voiced sounds in intervocalic position; this is the case in all the examples in Figure 4.3 and Figure 4.5. However, by analysing the data o the actual utterances, it is possible to con�rm whether or not lax stops appearing between two voiced sounds are necessarily realised as voiced sounds themselves. It has been observed that the rate at which voiced sounds are realised varies massively according to the speaker and the speed o the utterance (S.-A. Jun, 1995; J.-Y. Shin, 2000b). Afer J.-Y. Shin (2000b) observed the realisation angle o lax stops in various environments, she noted that the most regularly observed phenomenon in the phonetic realisation o lax sounds in Korean was not lax sounds in word-medial position realised as voiced sounds, but lax sounds realised with aspiration in word-initial position. Tereore, i sounds realised with aspiration in word-initial position are reerred to as ‘weak-aspirated’ sounds, then the standard realisation o word-initial lax sounds may be termed ‘weak aspiration’. Te results o this research and the conclusion it has produced means many contemporary Korean phonologists �nd it difficult to view the process by which Korean lax sounds become word-medial voiced sounds as a phonological or allophonic process. Hence it is more appropriate to reer to the voicing o lax sounds in word-medial position as a phonetic implementation which varies according to the speaker, the speed o utterance and phonetic environment. For the reasons stated above, it is better to regard lax consonants in Korean as voiceless. Neutralisation in syllable-�nal position also supports the view that lax sounds are the most unmarked. Frequency data also con�rm this observation. 4.2.1.5
Major allophones of stops
able 4.4 summarises the discussion o the major allophones o Korean stops thus ar. In the table, the ‘˺’ sign which appears above and the ‘ ˚’ sign which appears above or below them are the signs or unreleased and unvoiced realisations o the phonetic sign respectively in IPA. 4.2.2
Fricatives
As can be seen in able 4.5, the index o Korean ricatives is very simple indeed, especially considering the nine stops which exist in Korean. Fricatives in Korean are articulated in only two places: glottal and alveolar positions. 16 However, ricatives articulated at the alveolar ridge can be subdivided into lax or tense ricatives, whereas glottal ricatives have no such distinction.
69
4.2
Obstruents
Table 4.4 Major allophones of stops
Category
Phoneme
Major allophone
Environment
Lax
p
p ̥ b/b
Word-initial Word-medial onset
p̚
Coda
t ̥ d/d
Word-initial Word-medial onset
t̚
Coda
k ̊ g/g
Word-initial Word-medial onset
k̚
Coda
p* t* k* pʰ tʰ k ʰ
Onset Onset
t
k
p* t* k* pʰ tʰ k ʰ
Tense Aspirated
Let’s look at some minimal sets that can be used to determine the inventory o ricatives in Korean. (4)
Minimal sets /s/ : /s*/ : /h/
/sɑl/ ‘�esh’ : /s*ɑl/ ‘rice’: /hɑl/ ‘percentage’ /sʌltɑ/ ‘to be unamiliar’ : /s*ʌltɑ/ ‘to cut’ : /hʌltɑ/ ‘to get old’ /sɑli/ ‘reason’ : /s*ɑli/ ‘bush clover’ :
/hɑli/ ‘low offi cial’ /sotɑ/ ‘soda’ : /s*otɑ/ ‘to shoot’ : /hotɑ / ‘to sew together’ /sɑtɑ/ ‘to buy’ : /s*ɑtɑ/ ‘to be cheap’ : /hɑtɑ / ‘to do’ /sɑk/ ‘month’ : /s*ɑk/ ‘sprout’ : /hɑk/ ‘crane’
Table 4.5 Fricatives
Bilabial
Lax ense
Alveolar
s () s* ()
Alveolo-palatal
Velar
Glottal
h ()
70
Consonants
/sɑm/ ‘�ax’ : /s*ɑm/ ‘rice wrapped leaves’ :
/hɑm/ ‘box’ /sɑmʨi/ ‘three �ngers’ : /s*ɑmʨi/ tobacco pouch : /hɑmʨi/ ‘wooden bowl’ /sɑŋ/ ‘table’ : /s*ɑŋ/ ‘pair’ : /hɑŋ/ ‘item’ /sεŋ/ ‘lie’ : /s*εŋ/ ‘ping’ : /hεŋ/ ‘line’ Let’s examine the places o articulation or Korean ricatives. Alveolar ricatives are articulated in the narrow passageway between the tongue blade, and the back o the upper teeth and alveolar ridge. Glottal ricatives are articulated in the narrow passageway between the vocal olds. Korean alveolar ricatives are articulated in the same place as Korean alveolar stops. A good way to identiy the location o a Korean alveolar ricative is to articulate it without enunciating the ollowing vowel, and immediately aferwards to breathe in. When breathing in, the place that eels coolest is the place that was most constricted, and it was in that place that the sound was articulated. Afer the articulation o a Korean alveolar ricative, the back part o the upper teeth and the ront part o the alveolar ridge will eel coolest. However, an alveolar ricative’s place o articulation differs when ollowed by the vowels /i/ or /j/ or the diphthong /wi/. Under the in�uence o these ollowing vowels, the alveolar ricative shifs its place o articulation, becoming an alveolo-palatal ricative [ ɕ. ɕ*]. Korean alveolo-palatal ricatives are articulated slightly urther back in the mouth than the English palatoalveolar ricatives /∫/ or / ʒ/, and the lips protrude slightly at the time o articulation. Protrusion o the lips is also characteristic during the articulation o palato-alveolar ricatives in English. Lip protrusion aside, let’s take a closer look at the differences between Korean alveolo-palatal ricatives, that is, the allophonic variations o alveolar ricatives, and English palatoalveolar ricatives. Te biggest difference between the two can be seen or onesel, using the tip o the tongue. English palato-alveolar ricatives are usually articulated with the edge o the tongue touching only the lower gums. By contrast, in Korean, alveolo-palatal ricatives are articulated with the tip o the tongue touching the lower teeth as well as lower gums. I we pronounce the two sounds without ollowing vowels and then inhale, we can �nd the place o articulation more precisely. Pronounce Korean /s/ beore /ɯ/, and English / ʃ / and Korean /s/ beore /i/, without ollowing vowels, and then inhale each time. You will eel that the place where you can eel cool air shifs urther backwards.
71
4.2
Obstruents
Finally, in English, the place o articulation o alveolar ricatives does not change under the in�uence o the vowel that ollows them, whereas this type o allophonic variation is observed in Korean and Japanese. It is or this reason that Korean and Japanese native speakers ofen mispronounce the English word ‘see’, as they apply the allophonic rules o their native language to the pronunciation o the English word. Glottal ricatives are articulated in the glottis; indeed, there is no other place within the oral cavity where they could be pronounced. Te glottal ricative’s place o articulation within the oral cavity is determined by the ollowing vowel. vowel. o o con�rm this, try saying /hɑ/ and /hi/. At what point does the tongue assume the position necessary or pronouncing / ɑ/ and /i/? You will notice that when pronouncing pronouncing the series s eries /h ɑ/ /hi/, the tongue is already in the correct position to pronounce the ollowing vowel as you pronounce /h/. Te same is the case in English. Unlike other obstruents, alveolar ricatives do not possess three subdivisions; they can only be urther urt her classi�ed as lax or tense. Te lack o an aspirated variant is an important important eature eature o ricatives. ricatives. What What ollows ollows is an examination examination o ricatives similar to our earlier examination o stop sounds, bearing in mind that ricatives only have two subclassi�cations. In the syllable structure o Korean, ricatives can only exist at the onset o a syllable, as there is a general principle in Korean that all syllable-�nal sounds should not be released afer central closure. Korean ricatives thereore can only be observed in word-initial and word-medial positions. 4.2.2.1
Word-initial fricatives
Figure 4.6 shows the realisation o the Korean word-initial alveolar ricatives [s. s*] and also the glottal ricative [h] as represented on a spectrogram. In the picture, (a), (b) and (c) indicate the lax alveolar ricative, tense alveolar ricative and glottal ricative ric ative respectively. respectively. Te difference in energy distribution requency between alveolar and glottal ricatives is clearly visible. On the one hand, it is possible to observe energy distribution in the high requency area o the alveolar ricative, 17 and on the other hand, strong energy energy can be seen in the ormant o the vowel ollowing the glottal ricative. 18 Just by comparing comparing (a) and (b) we can see differences in the articulation o the two alveolar ricatives. It is characteristic o both the lax alveolar ricative /s/ and also o the tense /s */ that afer starting st arting rom a low requency with no energy, articulation is only accompanied by high requency noise energy. However, a difference in phonation type can be seen at the offset o the ricative. Rather
72
Consonants
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a)
(b)
s
*
ɑ
s
(c) ɑ
h
ɑ
0
1.7 Time (s)
Figure 4.6
Spectrogram and waveform of fricatives in word-initial position Spectrogram Hz 8000 6000 4000 2000
Waveform
(a)
(b) *
s
s
ɑ
*
i
0
1.5 Time (s)
Figure 4.7
Spectrogram and waveform of a tense alveolar fricative followed by the vowels / ɑ/ and /i/ in word-initial position
like obstruents, the lax ricative in Korean is articulated with aspiration. In the diagram, the circled area indicates the aspiration that accompani accompanies es the articulation o the lax alveolar ricative. 19 However, such aspiration cannot be seen on the spectrogram o the articulation or /s */. As we have seen above, in Korean the vowel ollowing a ricative can cause allophonic allophonic variation. In Figure 4.7 we can see s ee the tense sound s ound /s */ ollowed by /ɑ/ and /i/. Te areas on the diagram marked (a) and (b) indicate the riction duration o the two ricatives. Te energy distribution o the two sounds differs greatly. Te energy distribution o the riction noise o the sound has been circled. As can be seen, there is a big difference in the distribution o noise energy. When /s */ is ollowed by /i/, requency o noise energy is relatively
73
4.2
Obstruents
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a) h
(b) ɑ
h
(c) i
h
u
0
2.0 Time (s)
Figure 4.8
Spectrogram and waveform of a glottal fricative followed by the vowels /ɑ/, /i/, and /u/ in word-initial position
lower than when it is ollowed by / ɑ/. Tis is because alveolar ricatives are realised as alveolo-palatal alveolo-palatal ricatives when ollowed by /i/, as mentioned earlier. earlier. Figure 4.8 shows /ɑ/, /i/ and /u/ with a preceding /h/ sound and the relevant relevant waveorms. waveorms. As can be seen, the distribution o noise energy o /h/ ollows ollows the pattern o the articulation o the ollowing vowel sound. 4.2.2.2
Word-medial (intervocalic) fricatives
Figure 4.9 shows the three different Korean Korean ricatives between two t wo / ɑ/ sounds and the relevant spectrograms and waveorms. Te �gure shows patterns o energy distribution or alveolar and glottal sounds similar to those observed in word-initial position. When alveolar ricatives in intervocalic position are compared to those in word-initial position, there is hardly any aspiration o the lax alveolar ricative.20 Also, riction duration or the tense ricative is longer than or the lax rictive in word-medial position. 21 A difference can also be seen in the duration o the preceding vowel. 22 Similar phenomena have been observed in Korean obstruents. Looking at (c) in Figure 4.9 it is very clear that /h/ becomes voiced when located between two vowels. Looking at the spectrogram o the intervocalic glottal ricative in Korean, we can see the voice bar at a lower requency during the articulation o the consonant, as well as riction noise near the ormant requency o the neighbouring vowels. Similar allophonic variation due to vowel sounds has been seen in wordinitial position. Te alveolar ricatives, both lax and tense, become alveolo-palatal
74
Consonants
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a) ɑ
s
(b) ɑ
ɑ
(c) *
s
ɑ
ɑ
h
ɑ
0
2.0 Time (s)
Figure 4.9
Spectrogram and waveform of fricatives in word-medial (intervocalic) position
ricatives when ollowed by /i/. Te preceding vowel, however, has no effect on the place o articulation o the ollowing ricatives. Glottal ricatives also tend to be in�uenced more by the vowel that ollows them than by the one that precedes them. 4.2.2.3
Allophones and phonological identity of the lax fricative
As we move away rom our previous discussion o lax and tense types o alveolar ricatives, it is necessary to question whether such a distinction is natural, given that Korean ricatives differ rom other obstruents in having only two subclassi�cations. Kagaya (1974 (1974)) and S.-J. Moon (1997 (1997)) both point out that the distinction here is not between lax and tense, but between tense and aspirated. Kagaya (1974 (1974), ), the originator o this idea, used analyses o laryngeal gestures with a �brescope to argue or an aspirated/tense distinction rather than a lax/tense distinction in Korean ricatives. Tis is because similarities were noticed between Korean Korean lax alveolar ricatives and the aspirated types o other obstruents. In addition, S.-J. Moon (1997 (1997)) posited that lax ricatives do not possess the same voicing quality as is observed obser ved in other lax obstruents and urthermore also have an aspirated quality. However, such a discussion has two inherent problems. Te �rst is that, as has been noted, phonological categories should be based on phonological behaviours rather than phonetic properties. Tereore, the ocus o this discussion is whether lax ricatives share common phonological behaviours with other Korean lax obstruents. It is relatively easy to see that this is the case by looking at the process by which consonants become tense. As can be seen in (5a), /s/ is realised as [s *] afer another obstruent, like the other lax obstruents
75
4.2
Obstruents
shown in (5b). I it were truly an aspirated consonant, consonant, then we would expect it to undergo no change, like those seen in (5c).
ɑʨuk + sin/ → [k ɑʨ ɑʨuks*in] (5) a. /k ɑʨ /pɑp + sotʰ/ → [pɑps*ot] b. /ɑk + po/ → [ɑkp*o] /ɑpʰ + twi/ → [ɑpt*wi] /ɑpʰ + kil/ → [ɑpk*il] /t*ʌk + ʨip/ → [t*ʌk ʨ*ip] c. /mɑk + pʰɑn/ → [mɑkpʰɑn] /ʨip + tʰʌ/ → [ʨiptʰʌ] ʰɑn/ → [ʨipk ʰɑ ʰɑn] /ʨip + k ʰɑ ʨʰɑ] /mɑk + ʨʰɑ/ → [mɑk ʨʰɑ
‘leather shoes’ ‘rice pot’ ‘musical score’ ‘the ront and the rear’ ‘the road ahead’ ‘rice cake shop’ ‘the last round’ ‘housing site’ ‘house unit’ ‘the last train’
Te second problem is that, compar c ompared ed to changes seen in some phonetic eatures, the differences between the ricatives under discussion and other lax obstruents do not seem as signi�cant. As in both the discussion discuss ion above o stops and also the discussion below o affricates, the lax variant either shows considerable aspiration in word-initial position or does not show any aspiration at all in wordmedial position. Add to this the act that only the ricatives under discussion do not undergo the voicing process in word-medial position and it becomes di�cult to reach a de�nitive conclusion. conclusion.23 o sum up, ricatives do appear to share some eatures with other lax obstruents. 24 Tereore it is necessary to conclude that it is possible to see a distinction between lax and tense in Korean ricatives, but difficult to make a case cas e or a distinction distincti on between aspirated aspirate d and tense. 4.2.2.4
Major allophones of fricatives
On the basis o the preceding discussion it is possible to draw up a chart like that in able 4.6. 4.6. It ollows rom rom the table t able that the main actor in determining the allophonic variation o a ricative is the vowel sound that ollows it.
DID YOU KNOW . . . ?
Why are fricative sounds so difficult for Korean speakers?
In Korean, Korean, there are only three ricative sounds; the alveolar ricatives /s/, /s*/ and the glottal ricative /h/. In English, however, there are as many as nine ricative phonemes, as indicated in the t he table. Tey are the labiodental
Consonants
76
ricatives //, /v/, the dental ricatives /θ/, /ð/, the alveolar ricatives /s/, /z/, the palato-alveolar ricatives / ʃ /, / ʒ/, and the glottal ricative /h/.
Voiceless Voiced
Labiodental
Dental
Alveolar
v
θ ð
s z
Palato-alveolar Glottal
ʃ
h
ʒ
Consequently, it is not easy or native Korean speakers to distinguish between the ricative sounds in English and to pronounce them correctly. Te �rst difficulty arises rom the act that, in English, ricative sounds are made in �ve different places, whereas in Korean, all ricatives are pronounced either in the alveolar or glottal positions. Te second difficulty arises rom the act that, except or the glottal ricative, all other ricatives in English have a voiced and voiceless contrast. Because there are no voiced obstruents among Korean phonemes, they are diffi cult or Korean speakers to pronounce. In addition, ricatives show the longest duration among obstruents, which means that there must be a prolonged vibration o the vocal olds in order to make a voiced ricative sound. Due to these diffi culties, in Korean // is replaced by /pʰ/ , /v/ by /p/ , /θ/ by /s */ , /ð/ by /t/ , /s/ by /s/ or /s*/ , / ʃ / by /swi/ , and / ʒ/ by /ʨwi/ . We will come back to this issue when we discuss English loanwords in Chapter 10.
4.2.3
Affricates
As can be seen rom able 4.7, in Korean there is only one place o articulation or affricates, though there are three different phonation types. Tat this place o articulation is not shared by other obstruents is a particular eature o Korean. Let’s look at some minimal sets that can be used to determine the inventory o Korean affricates. (6)
Minimal sets /ʨ/ : /ʨ*/ : /ʨʰ/
/ʨɑtɑ/ ‘to sleep’ : /ʨ*ɑtɑ/ ‘to squeeze’ : /ʨʰɑtɑ/ ‘to kick’ /ʨitɑ/ ‘to be deeated’ : /ʨ*itɑ/ ‘to steam’ : /ʨʰitɑ/ ‘to hit’ /ʨεtɑ/ ‘to measure’ : /ʨ*εtɑ/ ‘to cut open’ : /ʨʰεtɑ/ ‘to be kicked’
77
4.2
Obstruents
Table 4.6 Major allophones of fricatives
Category
Phoneme
Major allophone
Environment
ɕ s ɕ* s* ç x
beore /j, i, wi/ beore all other vowels beore /j, i/ beore all other vowels beore /j, i/ in word-initial position beore /ɯ/ in word-initial position
ɸw h
beore /u, o/ in word-initial position beore all other vowels in word-initial position voiced counterparts in word-medial position
Alveolar Lax
s
Tense
s*
Glottal
h
ʝ/ɣ/β/ɦ/
Table 4.7 Affricates
Bilabial Lax Tense Aspirated
Alveolar
Alveolo-palatal
Velar
Glottal
ʨ () ʨ*() ʨʰ()
/ʨok/ ‘trotter’ : /ʨ*ok/ ‘indigo plant’ : /ʨʰok/ ‘candle-power’
Te place o articulation or affricates is the alveolo-palatal region, sometimes also reerred to as the prepalatal region, that is, where the palate meets the alveolar ridge. Korean alveolo-palatal affricates are articulated with the tip o the tongue either touching the lower teeth or both the lower teeth and the lower gums; the tongue blade brie�y touches the ront palate beore air is released. Afer momentarily blocking air�ow, the tongue blade is slightly lowered and the air passage slightly narrowed, creating turbulent air�ow. Te place o articulation or affricates is more or less the same as that or the allophonic variation o alveolar ricatives when articulated beore the vowel /i/.
78
Consonants
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a) t
(b) ɑ
(c) *
t
ɑ
th
ɑ
0
1.7 Time (s)
Figure 4.10
Spectrogram and waveform of lax, tense and aspirated affricates in word-initial position
Te place o articulation or Korean affricates is different rom that o English affricates, in the same way that the allophones o the Korean alveolar ricative [ɕ] and the English / ʃ / differ, as described above. Te English affricates / ʧ / and /ʤ/ are articulated slightly urther orward in the palato-alveolar region – the back o the alveolar ridge – than their Korean counterparts, and are always accompanied by protrusion o the lips. English affricates are produced in the same place as the English ricatives / ʃ , ʒ/, also with the lips protruding. English affricates are pronounced with the tip o the tongue touching only the lower gums. It is worth noting that English affricates are not articulated with t he tip o the tongue touching the lower teeth; this is the biggest difference rom their Korean counterparts. Affricates can be classi�ed as either lax, tense or aspirated sounds according to phonation type. All are voiceless. Just like ricatives, affricates can exist only in syllable-initial position. Tis is due to the act that syllable�nal sounds should not be released afer central closure. In the ollowing, we will discuss the phonetic characteristics o affricates in the word-initial and -medial positions. 4.2.3.1
Word-initial affricates
Figure 4.10 shows the word-initial realisation o / ʨ, ʨ*, ʨʰ/. (a), (b) and (c) indicate the realisations o the three different phonation types o affricates. Just as with other obstruents, where air�ow is obstructed in sound production, the spectrogram o affricates shows vertical lines indicating the release o closure, indicated by arrows.25 Afer the point indicated by the arrow, the kind o
79
4.2
Obstruents
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a) ɑ
t
(b) ɑ
ɑ
(c) *
t
ɑ
ɑ
th
ɑ
0
2.0 Time (s)
Figure 4.11
Spectrogram and waveform of lax, tense and aspirated affricates in word-medial (intervocalic) position
energy distribution associated with high-requency riction noise can be identi�ed. Aspiration o the lax and aspirated variants o the affricates is observed as with other obstruents, but the tense variant does not possess the same aspiration, as can be seen through the absence o a circled area on the diagram. 4.2.3.2
Word-medial (intervocalic) affricates
Figure 4.11 shows the three types o aricates in intervocalic position. In the diagram, the three types o aricates are labelled (a), (b) and (c), with a double-headed arrow indicating the duration o closure, and a diagonal arrow indicating the release o the closure. In word-medial position, riction is observed at the point o release, similar to what was ound in wordinitial position. However, in the case o tense and lax aricates the onset o the ollowing vowel occurs immediately aterwards, whereas the release o the aspirated aricate is ollowed irst by aspiration and then by the onset o the ollowing vowel. Like other obstruents, the lax aricate in wordmedial position is articulated with no aspiration, unlike in word-initial position. It can be seen that the closure period, riction period and length o preceding vowel in affricates all vary according to phonation type. 26 Tis is a particular eature o word-medial affricates, and is also seen in other obstruents. Another point to note is that in the examples in Figure 4.11 , the lax affricates in intervocalic position are not realised as voiced sounds. Te voice bars on the spectrogram and the periodic wave in the waveorm which are usually
Consonants
80
Table 4.8 Major allophones of affricates
Category
Phoneme
Allophone
Environment
Lax
ʨ
ʨ
Word-initial
ʥ/ʥ̥
Word-medial
ense
ʨ*
ʨ*
Onset
Aspirated
ʨʰ
ʨʰ
Onset
observed with voiced sounds are not present. However, as noted above, lax aricates in Korean are not necessarily realised by all speakers as voiced sounds in word-medial onset position. More importantly, in this environment, lax sounds are not articulated with aspiration. 4.2.3.3
Major allophones of affricates
able 4.8 shows the major allophones o Korean affricates. As can be seen, the major allophones o affricates are similar to those o stop sounds. 4.3
Sonorants
Sonorants are sounds produced without any obstruction as air�ow passes through the vocal tract. able 4.9 shows a list o sonorants in Korean. As can be seen, sonorants in Korean can be urther classi�ed into (i) nasal and (ii) liquid. Tere are three nasals in Korean, but only one liquid. All sonorants are voiced consonants. Let’s look at some minimal sets which can help us to de�ne the inventory o sonorants in Korean. Velar nasals can only appear in syllable-�nal position, and not syllable-initial position, whereas lateral sounds cannot appear in word-initial position, except in the case o loanwords. Tis means that minimal sets or sonorants need only ocus on syllable-�nal position. Te ollowing examples show minimal sets which differ in syllable-�nal position. (7)
Minimal sets /m/ : /n/ : /ŋ/ : /l/ /tɑm/ ‘ence’ : /tɑn/ ‘column’ : /tɑŋ/ ‘party’ : /tɑl/ ‘moon’ /mɑm/ ‘heart’ : /mɑn/ ‘ten-thousand’ : /mɑŋ/ ‘watch’ : /mɑl/ ‘language’ /pɑm/ ‘night’ : /pɑn/ ‘class’ : /pɑŋ/ ‘room’ : /pɑl/ ‘oot’
4.3
81
Sonorants
Table 4.9 Sonorants
Nasal
Bilabial
Alveolar
m ()
n ( ) l ()
Liquid
Alveolo-palatal
Velar
Glottal
ŋ ()
/sɑm/ ‘�ax’ : /sɑn/ ‘mountain’ : /sɑŋ/ ‘table’ : /sɑl/ ‘�esh’ /s*ɑm/ ‘lettuce-wrapped rice’ : /s*ɑn/ ‘to be cheap (adnominal)’ : /s*ɑŋ/ pair : /s*ɑl/ ‘rice’ /ʨɑm/ ‘sleep’ : /ʨɑn/ ‘glass’ : /ʨɑŋ/ ‘market’ : /ʨɑl/ ‘well’ /tʰɑm/ ‘greed’ : /tʰɑn/ ‘coal’ : /tʰɑŋ/ ‘soup’ : /tʰɑl/ ‘mask’ /hɑm/ ‘box’ : /hɑn/ ‘grudge’ : /hɑŋ/ ‘item’ : /hɑl/
‘percentage’ 4.3.1
Nasals
Nasals in Korean can be pronounced in the bilabial (i.e. /m/), alveolar (i.e. /n/) and velar positions (i.e. /ŋ/). Bilabial and alveolar nasals can occur either in word-initial or word-�nal position, but the velar nasal can only appear in word�nal position. Major phonetic characteristics o Korean nasals are weak nasality in word-initial position and weak nasalisation o preceding vowels in syllable�nal position. For this reason, native speakers o French, Italian or Japanese ofen recognise word-initial nasals as voiced stops instead o nasals. 27 As discussed earlier, in order to produce nasals, the velum needs to be lowered, so that air can �ow through the nasal cavity without any riction. When a vowel precedes a nasal, native speakers automatically lower the velum. Te point at which the velum is lowered varies cross-linguistically. In Korean, the velum is lowered very late, so that the articulation o the preceding vowel is not greatly affected. However, in English and Japanese, the velum is lowered early, so that the preceding vowels are strongly nasalised. 28 DID YOU KNOW . . . ?
Korean [mun] vs. English [mun]
As mentioned earlier, word-initial nasals in Korean are different rom those in English, French, Italian or Japanese, all o which have voiced stops. Compared to those languages, Korean word-initial nasals are weak and short. Vowels
82
Consonants
preceding word-�nal nasals are nasalised, but the degree o nasalisation is weaker in Korean in comparison with English and Japanese. (5)
Korean a. /mun/ ‘door’ b. /nun/ ‘eye’ c. /mɑm/ ‘heart’
English /mun/ ‘moon’ /nun/ ‘noon’ /mɑm/ ‘mom’
As noted above, the words listed are represented by the same symbols in IPA, but when one actually listens to the way they are pronounced, one can hear that the degree o nasality differs greatly between Korean and English. We can easily see that word-initial nasals in Korean are shorter and weaker than in English. At the same time, the nasality o the preceding vowels is also weaker in Korean than in English, as they are barely nasalised. Tis shows that even the same sequence o segments can be pronounced differently according to the coordination timing between vocal organs. Moreover, differences in coordination timing give rise to different coarticulation characteristics across languages.
4.3.1.1
Word-initial nasals
Figure 4.12 shows the spectrograms and waveorms o the bilabial nasal /m/ and alveolar nasal /n/ in word-initial position. Te word-initial nasal is relatively short as noted earlier, and thereore has a low degree o nasality.
Spectrogram Hz 8000 6000 4000 2000
Waveform
(a) m
(b) ɑ
n
ɑ
0
1.5 Time (s)
Figure 4.12
Spectrogram and waveform of bilabial and alveolar nasals in word-initial position
4.3
83
Sonorants
Spectrogram 8000
Hz
6000 4000 2000
Waveform
(a) ɑ
m
(b) ɑ
ɑ
n
ɑ
0
1.5 Time (s)
Figure 4.13
Spectrogram and waveform of bilabial and alveolar nasals in word-medial (intervocalic) position 4.3.1.2
Word-medial (intervocalic) nasals
Now, let’s see the spectrograms and waveorms o the bilabial nasal /m/ and the alveolar nasal /n/ in word-medial intervocalic position. Figure 4.13 shows how bilabial nasal /m/ and alveolar nasal /n/ are realised between / ɑ/ and /ɑ/. Figure 4.13 shows that word-medial (intervocalic) nasals are longer than word-initial nasals, and have a high degree o nasality. 4.3.1.3
Word-final nasals
Te bilabial /m/, the alveolar /n/ and the velar /ŋ/ may all appear in word�nal position. Figure 4.14 shows the spectrograms and waveorms o the three nasals in word-�nal position. Tey are pronounced much longer than in wordinitial or word-medial position. As in the case o other consonants, nasals in word-�nal position are pronounced unreleased. 4.3.1.4
Major allophones of nasals
able 4.10 shows the major allophones o Korean nasals. As can be seen, the major allophones are realised within a syllable structure. 4.3.2
Liquids 4.3.2.1
Phonetic realisation of liquids
In the consonant inventory o Korean, there is only one liquid , /l/. However, phonetically speaking, two allophones o /l/ exist. Tat is, /l/ can be
84
Consonants
Table 4.10 Major allophones of nasals
Phoneme
Major allophones
Environment
mnŋ
mn m¬ n¬ ŋ¬
Onset Coda
Spectrogram 8000
Hz
6000 4000 2000
Waveform
(a) ɑ
m
(b) ɑ
(c) n
ɑ
ŋ
0
1.7 Time (s)
Figure 4.14
Spectrogram and waveform of bilabial, alveolar and velar nasals in word-final position
realised either as the alveolar lateral approximant [l], or as the alveolar tap [ɾ]. Figure 4.15 shows the spectrogram and waveorms o the liquid /l/ in our different phonetic environments: (a) word-initial, as in /l ɑ/, (b) inter vocalic, as in / ɑlɑ/, (c) word-�nal, as in /ɑl/, and (d) afer /l/, as in / ɑll ɑ/. /l/ is pronounced [ɾ] in word-initial and word-medial (intervocalic) position as shown in (a) and (b). In the case o (c) and (d), /l/ is pronounced [l] in word-�nal position, and when it occurs afer another /l/ (i.e. in gemination). However, although [l] and [ ɾ] are very different phonetically, Korean native speakers do not regard them as two distinctive phonemes, but as allophones o /l/. Te use o /l/ in word-initial position is subject to restrictions. Tis is ofen called the Law o Initials. No pure Korean words begin with /l/, only loanwords. A syllable-initial /l/ can however occur in word-medial (intervocalic) position. Furthermore, /l/ cannot be realised afer another consonant, unless that consonant is also /l/.29
85
4.3
Sonorants
Spectrogram 8000 Hz 6000 4000 2000
Waveform
(a) l
(b) ɑ
ɑ
l
(c) ɑ
ɑ
l
(d) ɑ
ll
ɑ
0
2.2 Time (s)
Figure 4.15
Spectrograms and waveforms of liquids in word-initial, word-medial (intervocalic), and word-final positions, and after /l/
DID YOU KNOW . . . ?
Two ‘lice’ burgers please
Yuna went to a hamburger place in London with Henry. She saw a new burger made using rice instead o bread. She thought it would be a good idea to order it, since she wanted to eat rice. So, she said “two lice burgers please”. Te person in the shop looked puzzled and Henry could not help but laugh at her. Tis is one o the most common mistakes that native speakers o Korean make when learning English. Tey ail to distinguish between /r/ and /l/ sounds, which leads them to pronounce them the same. We will come back to this issue in Chapter 10.
4.3.2.2
Phonemic realisation of /l/
Previously, we have observed that /l/ is realised either as [l] or [ ɾ] depending on the phonetic environment in which it is ound. able 4.11 shows the phonetic realisations o /l/. Here a question arises as to why [l] is regarded as a phoneme, but not [ ɾ]. What is the basis or this? o answer this question, we need to consider the phonological patterns o the liquid in Korean. Consider the ollowing. (8) a. /tɑl + nɑlɑ/ → [tɑllɑlɑ] /tɯl + nɑmul/ → [tɯllɑmul]
‘lunar world’ ‘�eld herbs’
86
Consonants
Table 4.11 Major allophones of the liquid
Phoneme
Major allophone
l
l ɾ
Environment
Coda and afer /l/ (in gemination) Word-initial and intervocalic
b. /nɑn + lo/ → [nɑllo] /kw ɑn + ljo/ → [kw ɑlljo] c. /pi + kw ɑn + lon/ → [pikw ɑnnon] /hɯp + in + ljʌk/ → [hɯpinnjʌk]
‘a stove’ ‘a bureaucrat’ ‘pessimism’ ‘suction power or personal magnetism’
When /n/ and /l/ appear in a sequence either as /l/ + /n/ (8a) or /n/ + /l/ (8b), one o the sounds is assimilated into the other, as a sequence o /n/ and /l/ is not allowed in Korean phonology. 30 I [ ɾ] were a phoneme and not [l], it would be hard to explain why the assimilation observed in (8a–c) occurs at all; why should the tap [ɾ] become a nasal? Phonetically, the most prominent difference between a tap and a nasal sound is the openness o the nasal cavity. However, this cannot be used as the sole criterion or distinguishing a tap rom a nasal. For instance, this can also be used to distinguish a stop sound rom a nasal. Hence, we are still unable to explain why [ ɾ] is assimilated into the nasal [n].31 However, i we assume [l] as a phoneme, we can more easily explain this assimilation, since although [l] and [n] are both sonorants, only [l] has a lateral eature. I the lateral sound assimilates the nasal, then the sequences [ln] or [nl] will become [ll] as in (8a–b), and i the lateral sound loses the lateral eature, then the sequences will become [nn] as in (8c).32
DID YOU KNOW . . . ?
What the sound of farting can tell us about the sounds of Korean?
Te Korean language has a highly developed set o onomatopoeias. Te list o onomatopoeic words that describe the sound o a art alone is very extensive: /p*oŋ/ , /p*uŋ/ , /pʰik/ , /pʰisisik/ , /p*ulɯlɯk/ , /pulɯlɯk/ . O course, not all o these words are listed in the dictionary, but they are ound in everyday use. By examining the
87
4.4
Features for consonants
onomatopoeic words used to describe the sound o arts, one discovers a very interesting phenomenon. First, these onomatopoeic words can be divided into three general categories o sound: /p *oŋ/, and /p*uŋ/ denote the sound o loud, single arts (let’s call these /p *oŋ/ type); /pʰisisik/ the kind o arts that are more prolonged in length (let’s call these /p ʰisisik/ type); and /p*ulɯlɯk/ and /pulɯlɯk/ the kind o /p*oŋ/ type that is repeated at least twice (let’s call these /p *ulɯlɯk/ type). Different categories o onomatopoeia exist or the sound o arts because there are different ways in which the gas produced in the intestine exits through the anal passage. /p *oŋ/ type occurs when the gas that was trapped inside the intestine exits all at once, /p ʰisisik/ type when the gas inside the intestine exits through small crevices o the anus, creating air turbulence and /p*ulɯlɯk/ type when the /p *oŋ/ type is repeated more than twice. Tese onomatopoeic sounds re�ect patterns ound in the articulation o consonants. Te /p*oŋ/ type can be seen as stop sounds, the /p ʰisisik/ type as ricative sounds and the /p*ulɯlɯk/ type as trill sounds. Te consonants in these onomatopoeic words, interestingly, re�ect the acoustic characteristics o the arts themselves. Te /p *oŋ/ type starts with the stop sound, /p */, which has no aspiration, and has only one syllable, expressing a momentary popping sound. Te /p ʰisisik/ type starts with the stop sound /pʰ/ – which has strong aspiration – and the ricative /s/ that is repeated twice to create three syllables, expressing a continuing noise. In addition, the only vowel is the high vowel, /i/, which clearly emphasises the riction o the preceding consonants caused by vowel devoicing. Te /p*ulɯlɯk/ type starts with an unaspirated stop sound, like the /p*oŋ/ type, but the liquid sound between the vowels is realised as a tap sound. Tis means that in two out o three syllables the tap sound [ ɾ] is produced, which is the closest sound to a trill sound that can be pronounced in Korean. I there were trill sounds in Korean, the second syllable in /p *ulɯlɯk/ type would have been replaced by one. Te act that the Korean onomatopoeic words or arts re�ect the acoustic characteristics o the arts through the choice o consonants and the number o syllables is very interesting.
Consonants
88
4.4
Features for consonants
Consider the Korean consonant charts again. As shown in able 4.1 , there are �feen obstruents and our sonorants in Korean. Now we shall see how these nineteen consonants can be urther classi�ed by using phonological eatures. 4.4.1
Features for manner of articulation
Let’s �rst consider the eatures related to manner o articulation. Tere are �ve ways o classiying consonants in Korean according to their manner o articulation. o classiy the �ve groups (or categories), we need three eatures, as shown in (9). (9)
Features or manner o articulation a. [+/−continuant]: [+continuant] sounds are produced without air�ow being blocked where the primary constriction takes place, but continuing throughout the process o sound production. b. [+/−delayed release]: Tis eature is or sounds which are made with complete closure o the vocal tract. [+delayed release] sounds are produced without instantaneous release, as is the case with stops. c. [+/−lateral]: Tis eature is restricted to coronal consonants. 33 [+lateral] sounds are produced by lowering the mid section o the tongue at both sides or at only one side, thereby allowing air to �ow out o the mouth in the vicinity o the molar teeth.
Tese three eatures can be used to distinguish �ve types o Korean consonants as shown in able 4.12. Stops and affricates share [−continuant] and [−lateral] eatures, but they can be distinguished using the [delayed release] eature, as stops have the [−delayed release] eature, whereas affricates have the [+delayed release] eature. FURTHER STUDY
[lateral] vs. [nasal]
Sonorant consonants in Korean can be divided into two categories, either nasal or lateral sounds. In this book, we have classi�ed sonorants using the [lateral] eature. However, it is also possible to classiy them by using the [nasal] eature. Te result is that nasal sounds are classi�ed as
89
4.4
Features for consonants
Table 4.12 Classification of Korean consonants using a feature matrix
Stop
Fricative
Affricate
Nasal
Lateral
[continuant]
−
+
−
−
−
[delayed release]
−
+
−
[lateral]c
−
−
−
b −
a
+
a
In some languages, lateral s ounds have [+continuant] eatures, while in others, they have [ −continuant] eatures. his i s because although the central oral cavity, where airlow is blocked, is completely closed, the sides o the cavity are open. Given the closure o the central oral cavity, sounds may be des cribed as having [−continuant] eatures, but i one considers that the sides o the cavity remain open, then the sound may be described as having [+continuant] eatures. hereore, the phonological behaviours o lateral s ounds need to considered urther in order to determine whether they possess [+continuant] or [−continuant] eatures. For instance, one must consider how they behave when they occur with sounds which have [+continuant] or [ −continuant] eatures. As we shall see in 8.1.1, in Korean, only seven consonants /p, t, k, m, n, ŋ, l/ can occur in the syllable-inal position. Six o those sounds, /p, t, k, m, n, ŋ/ , have the eature [−continuant]. It would be unnatural to assume that o t hese seven sounds, only the lateral has the eature [+continuant], and we thereore assume the lateral as having the eature [−continuant] in Korean. b Tis is a eature limited to sounds created by closure o the vocal tract. Nasals, as sounds created by closure and immediate release o the vocal tract, possess the eature [−delayed release]. However this eature is not applicable to ricatives and laterals, which are not created by closure o the vocal tract, and it i s not used with reerence to them. c As in the de�nition, the eature [lateral] is limited to consonants which are [+coronal]. Tereore, the eature [−lateral] is not accorded to all stops, ricatives, and nasals, but only those which are [+coronal]. As place o articulation is not shown clearly in this table, the three kinds o sound here are given the eature [−lateral].
[+nasal] and lateral sounds as [ −nasal]; On the other hand, i sonorants are classi�ed using the [lateral] eature, then lateral sounds come to be classi�ed as [+lateral] and nasal sounds as [ −lateral]. Te reason why we have chosen the [lateral] over the [nasal] eature in this book is due to certain phonological behaviours that have been observed, as shown below. Consider the ollowing examples: (8)
/kjʌul+nɑl/ /kw ɑil+nɑmu / /tɑl+nim/ /ʨul+nʌm+ki/ /k ʰɑl+nɑl/ /hɑnɯl+nɑlɑ/
[kjʌullɑl] [kw ɑillɑmu] [tɑllim] [ʨullʌmk*i] [k ʰɑllɑl] [hɑnɯllɑlɑ]
‘winter days’ ‘ruit tree’ ‘moon’ ‘skipping’ ‘blade o a knie’ ‘heaven’
90
Consonants
Tese examples display a phonological process known as lateralisation that we will discuss in Chapter 8. In the examples above, alveolar nasals become laterals due to the preceding lateral sound. I we were to replace the eature [lateral] with [nasal], this process could be understood as ollows: lateral sounds with the [ −nasal] eatures change sounds that have the [+nasal] eature into lateral sounds with the [−nasal] eature. Tis way o understanding Lateralisation may seem a little unnatural. However, i we understand this process as lateral sounds with [+lateral] eatures changing ollowing alveolar nasals with [ −lateral] eatures into lateral sounds with [+lateral] eatures, the explanation seems more plausible. Moreover, when we explain Lateralisation using the [lateral] eature, we can understand why this happens only to alveolar nasals, since bilabial or velar nasals are not subject to the [lateral] eatures. As discussed earlier, only sounds with [+coronal] eatures are subject to the [lateral] eatures. However, bilabial and velar nasals are not coronal sounds. In other words, they are [ −coronal].
4.4.2
Features for place of articulation
Consonants in Korean can also be classi�ed according to their place o articulation. o distinguish each group, we need to postulate a different set o eatures, which is shown in (10). (10) Features or place o articulation a. [+/−coronal]: [+coronal] sounds are produced with the blade o the tongue raised rom the neutral position. b. [+/−anterior]: [+anterior] sounds are produced by an obstruction in the area orward o the palato-alveolar region o the mouth. In order to understand the [coronal] eature, we must �rst understand what is meant by ‘neutral position’ o the tongue. Te de�nition given or neutral position in SPE is shown in (11). (11) Neutral position It can be observed that just prior to speaking, the subject positions his vocal tract in a certain characteristic manner. Te velum is raised and the air�ow through the nose is shut off. Te body o the tongue, which in quiet breathing lies in a relaxed state on the �oor o the mouth, is
91
4.4
Features for consonants
Table 4.13 Features for place of articulation
Bilabial [coronal]
−
[anterior]
+
Alveolar
Alveolo-palatal
Velar
Glottal
+ +
+
−
−
−
−
−
Palato–alveolar [+anterior] Bilabial
Alveolar
[−anterior]
[ ∫ ]
Alveolo–palatal
Velar
Lips
Figure 4.16
Glottal
Vocal folds
Place of articulation and feature value of the [anterior] feature
raised in the neutral position to about the level that it occupies in the articulation o the English vowel [e] in the word bed ; but the blade o the tongue remains in about the same position as in quiet breathing. able 4.13 sums up how consonants can be urther classi�ed by eatures or place o articulation. As shown, bilabial and velar sounds are distinguished using the [+/−anterior] eature and bilabial and alveolar sounds are distinguished using the [+/−coronal] eature. Only alveolar and alveolo-palatal consonants have the [+coronal] eature since when these sounds are produced, the blade o the tongue is lifed up. Tis is not observed in any other consonants, which thereore have the [ −coronal] eature. Bilabial and alveolar sounds share the [+anterior] eature since they are pronounced at the ront o the mouth. When alveolar and alveolo-palatal sounds are produced, the blade o the tongue is higher than when in the neutral position. Hence, they have the [+coronal] eature. However, when bilabial, velar or glottal sounds are produced, the blade o the tongue is lower than when in the neutral position. Hence, they have the eature [ −coronal]. On the other hand, bilabial and alveolar sounds are produced closer to the ront o mouth than palato-alveolar sounds, such as [ ʃ ]. Hence, they have the eature [+anterior], whereas alveolopalatal, velar and glottal sounds, which are produced urther back in the mouth, possess the eature [−anterior].
Consonants
92
Table 4.14 Features for phonation type
[tense] [aspirated]
4.4.3
Lax
Tense
Aspirated
– –
+ –
+ +
Features for phonation types
able 4.14 shows the three-way contrast in stops and affricates in Korean. Al veolar ricatives also show a distinction between lax and tense types. o distinguish these consonants, we need two eatures, [+/ −tense] and [+/−aspirated]. (12) Features or phonation type a. [+/−tense]:34 [+tense] sounds are produced with tension in the vocal olds. b. [+/−aspirated]:35 [+aspirated] sounds are produced with the two vocal olds ar apart, and with subsequent opening o the glottis. Table 4.15 Features for consonants in Korean
p p* pʰ
t
t* tʰ
k k* k ʰ s
s* ʨ ʨ* ʨʰ m n
ŋ
l
h
Major class features consonantal
+ + + + + + + + + + + + + + + + + + +
syllabic
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
sonorant
–
–
–
–
–
–
–
–
–
–
–
–
–
–
+
+
+
+
–
–
–
–
–
–
–
–
–
–
+
+
–
–
–
–
–
–
–
+
–
–
–
–
–
–
–
–
–
+ + + –
–
–
–
–
–
–
–
Manner continuant delayed release lateral
–
–
–
–
+
Place coronal
– – – + + + – – – +
+ + + + – + – + –
anterior
+ + + + + + – – – + + – – – + + – + –
Phonation types tense
– + + – + + – + + –
+ – + + – – – – –
aspirated
–
–
–
+
–
–
+
–
–
+
–
–
–
+
–
–
–
–
+
4.5
93
Summary
Using these two eatures, phonation types can be classi�ed as in able 4.14. Aspirated sounds and tense sounds shared the [+tense] eature, distinguishing them rom lax sounds which have the [ −tense] eature. Aspirated and tense sounds are distinguished rom one another due to the act that tense sounds do not possess the [+aspirated] eature. able 4.15 summarises the eatures relating to consonants in Korean.
DID YOU KNOW . . . ?
Why is it unnecessary to posit a [voiced] feature in Korean?
Suppose you were in a mixed school. A [emale] eature would be necessary to distinguish emale students rom male students. However, in a girls’ school, this [emale] eature would be redundant, because all the students are emale. For the same reason, a [voiced] eature would be redundant in Korean, unlike in English or in Japanese. Tis is because all three kinds o obstruent in Korean are voiceless sounds that do not entail vibration o the vocal olds, and all the sounds that can be classi�ed as sonorant are voiced sounds, i.e. they are accompanied by vibration o the vocal olds. Tereore, in Korean consonants, all sonorant sounds are voiced, and all obstruents are voiceless.
4.5
Summary
In this chapter, we have discussed the consonants o Korean. Altogether, there are nineteen consonants in Korean. Four o them are sonorants and the rest are obstruents. Consonants in Korean can be urther classi�ed according to (i) place o articulation, (ii) manner o articulation and (iii) phonation types. Te main characteristics o Korean consonants can be summarised as ollows. First, Korean has a rich inventory o obstruents: �feen out o nineteen consonants are obstruents. Second, Korean obstruents are all voiceless. Unlike English, Korean has no voiced obstruent in the inventory. Tird, Korean shows a three-way contrast among stops and affricates. Cross-linguistically, such a three-way contrast is relatively rare. Te pronunciation o stops differs according to where they occur. Te threeway contrast is neutralised in word-�nal or syllable-�nal position. We will discuss the reason why this is the case in more detail in Chapter 8 . In contrast to stops, Korean has relatively ew ricatives in comparison with English. In addition,
Consonants
94
only one liquid phoneme exists. Word-initial nasals in Korean are short and weak compared with English. In order to provide more natural explanations or consonant-related phonological processes, we need to classiy each consonant as a set o eatures. Major eatures or consonants can either be those or manner o articulation (e.g. [+/–continuant], [+/–delayed release], [+/–lateral]) or place o articulation (e.g. [+/–coronal], [+/–anterior]). Consonants can also be classi�ed according to phonation types (e.g. [+/ –tense], [+/–aspirated]). EXERCISES 1
2
Why is [pʰ] a phoneme in Korean but not in English? Explain with examples from Korean and English. Circle the words which start with consonants articulated in the region specified.
3 1 2
3
4
5 6
7
8 9 10
11
a. Region 1: /pɑl/ ‘foot’, /t*ɑŋ/ ‘earth’, /hɑnɯl/ ‘sky’, /mɑɯm/ ‘heart’, /pʰoto/ ‘grape’ b. Region 2: /kʌmi/ ‘spider’, /tɑl/ ‘moon’, /kɑŋ/ ‘river’, /sɑlɑm/ ‘human’, /p ʰɑl/ ‘arm’ c. Region 3: /ʨɑŋkɑp/ ‘gloves’, /kɑŋɑʨi/ ‘puppy’, /ʨʰɑ/ ‘car’, /tʰɑʨo/ ‘ostrich’ List examples which illustrate the three-way contrast (lax, tense and aspirated) among stops in Korean. List examples which illustrate the three-way contrast (lax, tense and aspirated) among affricates in Korean. Of lax, tense and aspirated stops, which are most commonly used in Korean? Compare word-initial and word-final stops. What is the difference between them? Compare word-initial and word-medial (intervocalic) stops. What is the difference between them? How many liquids exist in Korean? Explain the [lateral] feature and provide examples with the feature [+lateral]. What is the difference between bilabial and velar sounds? Explain using relevant features. Discuss dialectal variations in Korean consonant pronunciation.
5
Vowels
In the previous chapter, we discussed consonants in Korean. In this chapter, we will discuss vowels in Korean. In particular, we will explore which vowels are phonemes in Korean and what their phonetic and phonological characteristics are. In 5.1, we will introduce the basic properties o vowels. In 5.2 and 5.3, we will examine in detail the phonetic and phonological characteristics o monophthongs (simple vowels) and diphthongs in Korean and 5.4 covers eatures or vowels. Te chapter summary can be ound in 5.5. 5.1
Phonological analysis of vowels
5.1.1
Monophthongs and diphthongs A monophthong (or simple vowel) is a vowel consisting o one articulation rom beginning to end. Likewise, a diphthong is a vowel made up o two articulations and a triphthong is a vowel made up o three articulations. Diphthongs can be urther analysed into their constituent parts, glides and monophthongs. Te articulation o a glide is similar to that o a vowel, but as it is pronounced much aster it does not remain stable during articulation, unlike vowels. Moreover, unlike a vowel, a glide cannot orm a syllable by itsel. Diphthongs can be categorised as on-glide diphthongs or off-glide diphthongs according to the order in which the glide and monophthong are combined. An on-glide diphthong is one in which the glide precedes the monophthong, and an off-glide is one where the glide ollows the monophthong. In Korean, only on-glide diphthongs exist.
5.1.2
Phonological status of the glide: consonant or vowel? In the previous section, we said that a diphthong is seen as a combination o glides and monophthongs. A glide is also called a semivowel or an 95
96
Vowels
/phil + jo/
/mil + w^ l/
[ɾ]
[ɾ]
Hz 7000 6500 6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000 500
Figure 5.1
Allophonic realisation of /l/ followed by glides in Korean words /p ʰil + jo/ ‘necessity’ and /mil + wʌl/ ‘honeymoon’
approximant and has the properties o vowels as well as consonants. Because o this, glides are regarded as vowels in some languages but as consonants in others. In Korean, glides are classi�ed as a type o vowel, yet in English they jɔt are regarded as consonants. For instance, the �rst sound /j/ in yacht [ ] in English and /j/ in /jo/ ‘underquilt’ in Korean is not so very different phonetically, but /j/ is a vowel in Korean and a consonant in English. In English, words starting with a glide such as window or yacht take the inde�nite article a instead o an. Tis could be a piece o evidence which supports the analysis that an English glide is regarded as a consonant not a vowel. In the case o Korean, it is difficult to use similar methods because grammatical morphemes are always attached at the end o a word and grammatical morphemes in Korean are only sensitive to the phonological category o the word-�nal sound. Also, as mentioned earlier, only on-glide diphthongs exist in Korean. Tis means that there are no words in Korean which end with a glide. However, we can con�rm that a glide is a vowel by observing different phonetic realisations o /l/. As discussed in Chapter 4, /l/ is realised as [l] beore a consonant or afer /l/ but as [ɾ] beore a vowel. I a glide is a vowel, /l/ beore a glide should be realised as / ɾ/. But otherwise, /l/ should be realised as /l/. Te words in (1) show that /l/ beore a glide becomes [ ɾ] as shown in Figure 5.1. Tis supports the case or glides being vowels. (1) a. /pʰil + jo/ ‘necessity’ b. /mil + w ʌl/ ‘honeymoon’ DID YOU KNOW . . . ?
Why do [twin] in Korean and [twin] in English sound different? In Korean, [twin] is an adnominal o the verb {twi-} - ‘to spring’, and means ‘springing’. Although [twin] is represented identically in IPA in both
5.2
97
Monophthongs
Korean and English as [twin], the English word twin [twin] and the Korean word [twin] sound different. Why is this the case? Te contrast in their pronunciation re�ects two different methods o pronouncing the same combination o consonant and glide. When the Korean glide [w] ollows a consonant, it makes the preceding consonant labialised, rather than being pronounced as an independent sound o its own. In English, on the other hand, although the succeeding glide labialises the preceding consonant to a certain extent, the glide, [w], is pronounced ‘independently’ with a noticeable duration o its own. Hence, [twin] in English is actually pronounced as [t w win]. Tis helps to explain why native Korean speakers ofen make mistakes when pronouncing the sound /kw/, as ound in such words as queen and question. Korean speakers have a particular tendency to pronounce /kw/ as [kw ] rather than [kw]. Tey thereore pronounce ‘queen’ and ‘question’ as [kw in] and [kw esʧən] rather than [kwin] and [kwesʧən], and this phenomenon arises rom the act that a combination o a consonant and a glide is pronounced differently in the two languages.
5.2
Monophthongs In this section, we will discuss monophthongs in Korean. Unlike consonants, there is some noticeable difference between dialects with regard to vowels. In this book, we have ocused on what is known as Standard Korean, which is based on the Seoul dialect. Nevertheless, de�ning Standard Korean is not easy and in act, what is reerred to as ‘Standard Korean Pronunciation’ (SKP) is slightly different rom Seoul speakers’ real speech (or pronunciation), as we shall see in this section. For instance, it is known that according to SKP, Standard Korean has ten monophthongs, yet as we shall see in 5.2.2, whether Korean has ten monophthongs or not is still debatable and in reality it is more likely that Korean has seven monophthongs instead o ten. In the ollowing, we will �rst discuss the simple vowel system based on SKP and some problems o this system based on analysis o spontaneous speech.
5.2.1
How many are there? According to SKP, Standard Korean has ten simple vowels as listed in able 5.1.
98
Vowels
Table 5.1 The ten monophthongs of SKP
Front
Back
Unrounded
Rounded
Unrounded
Rounded
High
i ()
y ()
u ()
Mid
e ()
ø ()
ɯ () ʌ ()
Low
æ ()
o ()
ɑ ()
Formant transition
Hz 7000 6500 6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000 500
[wε]
Figure 5.2
[wi]
Spectrogram of ‘outside’ and ‘stomach’
5.2.1.1
Are there front rounded vowels in Korean?
As we can see, SKP postulates ten vowels, among which �ve are ront vowels and �ve are back vowels. Both ront and back vowels have roundedness/nonroundedness contrast and also can be classi�ed into high, mid and low vowels according to the height o the tongue during articulation. Tis appears to be a systematic and stable vowel system, but perhaps doesn’t quite re�ect the reality o Korean vowels. In the ollowing, we will discuss two aspects o the SKP system which reveal these problems. Firstly, the two ront rounded vowels /y/ and /ø/ are phonemes in SKP. However, in Seoul speakers’ real speech, we do not observe the monophthongs /y/ and /ø/. Native speakers o Seoul Korean pronounce words with and spelling as diphthongs instead o monophthongs. Te best way o �nding out whether // and // are simple vowels or diphthongs is to pronounce words with or spelling such as ‘outside, except’ and ‘stomach’. I the shape o mouth, that is, roundedness, is the same rom the beginning to the end o articulation, then the vowel is a simple vowel. On the other hand, i the roundedness o the mouth disappears during the course o pronunciation, then it is a diphthong. Figure 5.2 shows the spectrogram or the two words ‘outside’ and ‘stomach’ by a emale native Korean speaker. We can see
99
5.2
Monophthongs
Table 5.2 Eight monophthongs of SKP (without /y/ and /ø/)
Front
Back
Unrounded
Unrounded
Rounded
High
i
u
Mid
e
Low
æ
ɯ ʌ ɑ
o
ormant transition in both �gures. Tis shows that / / and // are diphthongs not monophthongs. In act, SKP also notices the dual nature o / / and // and mentions that // and // ‘can’ be pronounced as diphthongs. However, in both natural and read speech, // and // are never pronounced as simple vowels but as diphthongs without exception. Based on this observation, we propose an eight simple-vowel system as in able 5.2. 5.2.1.2
Are /e/ and /æ/ distinctive?
Let’s move on to the next issue which is also controversial. ables 5.1 and 5.2 (above) show that there are high, mid and low vowels in Korean, and that the de�nition o high, mid and low depends on the height o the tongue in articulation. However, it is not so obvious whether /e/ (ront-mid vowel) and /æ/ (ront-low vowel) are distinctive enough to native speakers o Standard Korean. Spellings corresponding to /e/ and /æ/ are and in Korean. Orthographically, ‘crab’ and ‘dog’ are different and ofen it is expected that the two words should sound different, given that /e/ and /æ/ are distinctive. In spontaneous speech, however, native speakers can’t distinguish between the two sounds. H.-W. Choi (2002) shows that among 210 native speakers o Standard Korean between 20 and 80 years old, more than 80 per cent o the time participants pronounced the two sounds the same. Tat is, without any appropriate context, when (2) is heard, it can thereore reer to either (2a) or (2b). Not only in perception but also in production, native speakers can’t distinguish the two sounds. (2)
. I like crabs. I like dogs.
Minimal pairs which show and contrast are given in (3). Native speakers will pronounce all words in (3a) and (3b) the same.
100
Vowels
F2 (Hz) 3000 2500 2000
1500
1000
500 200
300
400
500
F1 (Hz) 600
700
e
800
A 900 1000 1100
Figure 5.3
Formant plot of and (ten male speakers)
(3)
Minimal pairs having and a. words with : ‘hemp cloth’, ‘disenrolment’, ‘your’ b. words with : ‘belly’, ‘enrolment’, ‘my’
Interestingly, because /e/ and /æ/ have become largely no longer distinctive, when there are two requently used words with only an /e/ and /æ/ distinction, there is a tendency to use an alternative word to avoid any ambiguity (i.e., by replacing either the /e/ or /æ/ vowel or preceding consonant). For instance, in speech, ‘my’ and ‘your’ can be ambiguous and cause diffi culty in the proper understanding o a sentence. o avoid this problem, native speakers tend to much preer /ni/ to or the word ‘your’. Also, except or words with and which are requently used ofen in transcribing names, or example, people need to con�rm the spelling because it is hard to tell which one is which. Figures 5.3 and 5.4 con�rm our observation. Both �gures show the results o a study in which ten male and ten emale native speakers took part. In this
5.2
101
Monophthongs
F2 (Hz) 3000 2500 2000
1500
1000
500 200
300
400
500
F1 (Hz) 600
700
e
800
A 900 1000 1100
Figure 5.4
Formant plot of and (ten female speakers)
study, participants were asked to take part in a quiz, in which the answers to all o the questions were words which, orthographically speaking, contained either or ollowed by the {-t ɑ} - ending. Te test was repeated. Figures 5.3 and 5.4 show the ormant charts o / / and // in terms o the ormant-measures F1 and F2. No signi�cant difference is observed between // and //. Tis indicates that native speakers o Korean cannot distinguish /e/ rom /æ/. 5.2.2
The simple vowel system of Standard Korean able 5.3 sums up our observations and presents an updated version o the range o monophthongs available in Standard Korean, a total o seven sounds. From this table, it can be seen that there are more back vowels than ront vowels in Korean. All back vowels except the low vowel /ɑ/ have a rounded and unrounded vowel pair and there is a symmetry between (high/mid) ront vowels and (high/mid) back vowels.
102
Vowels
Table 5.3 Monophthongs in Standard Korean
Front
Back
Unrounded
Unrounded
Rounded
High
i
u
Mid
ɛ
ɯ ʌ ɑ
Low
o
o con�rm the accuracy o the above table, let us now turn to look at minimal pairs in Korean. (4)
Minimal pairs Front vowels /i/ : /ɛ/ /kimi/ ‘reckles’ /mi/ ‘beauty’ /si/ ‘poetry’ /sitɑ/ ‘to be sour’ Back vowels /ɯ/ : /ʌ/ /k ɯl/ ‘sentence’
/u/ : /o/
/ɑ/ : /ʌ/
5.2.3
/tɯl/ ‘�eld’ /nɯl/ ‘always’ /s*ɯltɑ/ ‘to sweep’ /kuŋ/ ‘palace’ /sum/ ‘breath’ /sul/ ‘tassel/liquor’ /kuptɑ/ ‘to be bent’ /k ɑm/ ‘persimmon’ /k ɑmi/ ‘seasoning’ /sɑki/ ‘raud’ /k ɑmtɑ/ ‘to wind’
/k ɛmi/ ‘ant’ /mɛ/ ‘hawk’ /sɛ/ ‘bird’ /sɛtɑ/ ‘to leak’ /k ʌl/ ‘three points in a game o yut’ /tʌl/ ‘less’ /nʌl/ ‘board’ /s*ʌltɑ/ ‘to cut’ /koŋ/ ‘merit’ /som/ ‘cotton’ /sol/ ‘brush’ /koptɑ/ ‘to be beautiul’ /k ʌm/ ‘sword’ /k ʌmi/ ‘spider’ /sʌki/ ‘clerk’ /k ʌmtɑ/ ‘to be black’
The phonetic realisation of monophthongs So ar, we have observed the phonological aspects o the monophthong. In this section, we will discuss the phonetic realisation o the monophthong. able 5.4 shows the results o a case-study investigation, in which ten male and emale native Korean speakers were asked to pronounce the seven monophthongs outlined in able 5.3 above. Tey were asked to pronounce the target vowels in a sequence o /h_t ɑ/. able 5.4 shows the mean and standard deviation
103
5.2
Monophthongs
Table 5.4 Mean and standard deviation of F1 and F2 of ten male and female speakers’ phonetic realisations of seven different monophthongs
Male
Female
F1 (SD)
F2 (SD)
F1 (SD)
F2 (SD)
i
258.9 (40.0)
2065.9 (130.6)
291.2 (24.8)
2730.7 (133.8)
ɛ ɑ
489.5 (78.6)
1828.1 (121.2)
589.6 (108.5)
2309.2 (172.6)
788.1 (55.3)
1406.9 (66.4)
990.9 (67.9)
1716.2 (103.7)
560.4 (45.4)
1045.4 (94.3)
688.5 (107.6)
1293.4 (101.2)
333.5 (71.7)
1517.6 (169.9)
322.2 (62.9)
1666.7 (171.0)
o
356.2 (42.8)
795.0 (116.7)
398.1 (63.1)
739.7 (110.9)
u
280.1 (37.0)
858.4 (146.6)
321.1 (24.8)
800.7 (112.9)
F2 (Hz) 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 18001700 1600 1500 1400 13001200 11001000
900 800 700
600
500 200 250
u
i ɯ
o
300 350 400 450
ε
500
550
) z H 600 ( 1 650 F 700 750
ɑ
800 850 900 950 1000
Figure 5.5
Formant chart (ten male speakers)
(SD) o F1 and F2 or seven monophthongs. Except the vowel / ɛ/, which has orty values (10 speakers × 2 times × 2 vowels (// and //), all other vowels had twenty values (10 speakers × 2 times). In the case o // and //, native speakers were asked to pronounce words that are spelled and . Figures 5.5 and 5.6 show ormant charts drawn rom the same data that were summarised in able 5.4.
Vowels
104
F2 (Hz) 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 18001700 1600 1500 1400 13001200 1100 1000 900
800 700
600
500 200 250 300
m
i
u 350
o
400 450 500 550
) z H 600 ( 1 650 F
ε
700 750 800 850 900 950
ɑ
Figure 5.6
1000
Formant chart (ten female speakers)
Figure 5.7 compares the ormant charts o Standard Korean and English (RP).1 Te solid line represents Korean and the dotted line represents English. We can see that monophthongs in the two languages are quite different rom each other. Te number o monophthongs in Korean (seven) is smaller than in English (eleven) and the F2-value range o vowel pronunciation is smaller in Korean as a result. Interestingly, such a range difference is observed more signi�cantly in terms o the ront–back contrast than in terms o the high– mid–low contrast. Also, though / ɑ/ and /ʌ/ exist both in Korean and English, Figure 5.7 shows that the Korean / ɑ/ is lower than English / ɑ/ and also pronounced more at the ront, whereas Korean / ʌ/ is pronounced urther back than English /ʌ/ and the mouth is less open. Korean / ɯ/ is not ound in English and English-speaking learners o Korean may �nd this sound unamiliar. In act, not only English-speaking learners but learners o Korean in general �nd it very difficult to distinguish between /ɯ/ and /ʌ/, two back-unrounded vowels in Korean. In addition, /ɯ/ and /ʌ/ and /ʌ/ and /o/ are equally diffi cult to distinguish.
105
5.2
Monophthongs
F2 (Hz) 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 18001700 1600 1500 1400 13001200 1100 1000 900
i
800 700
6 00
250
u
iE
300
uE
ɯ
IE
350
o
Ω
400
E
ɔE
ε εE
ɒE
E
E
450 500
) z H 600 ( 1 650 F 550
V
E
500 200
ɑE
700 750
ɑ
800 850
E = English
900 950 1000
Figure 5.7
Comparison of formant charts for English (RP) and Standard Korean
DID YOU KNOW . . . ?
Are Korean taxi drivers still rip-offs? Eric studies Korean at university. Wanting to improve his Korean, he decided to go to Korea, and �nally arrived at Incheon International Airport. Tinking to himsel that he could now �nally practise his Korean as he wanted, he decided to take a cab to his riend’s house, and got in a cab. Eric:
, 2. ‘I would like to go to the Sinchon station, please.’
Driver:
, . ‘Okay. You speak very good Korean.’
Having been complimented or his Korean shortly afer his arrival, Eric elt very proud o himsel or having studied so hard. According to his riend, Sinchon was not very ar rom Incheon Airport. However, the taxi driver, despite having travelled quite ar, did not seem to be stopping at all. Eric was a bit bewildered. Was the driver trying to rip him off? While Eric was thinking that the driver was a bit suspicious, the taxi driver spoke.
106
Vowels
Driver:
, . 7 8. ‘We have arrived. Tat will be 78,000 won.’
Eric:
? . 5. ‘Really? I can’t believe that. My riend said it would cost about 50,000 won at most.’
Driver:
, 5 . ‘He must be wrong. It is impossible to get here or 50,000 won.’
Eric thought to himsel, “And they said there weren’t any rip-off taxi drivers in Korea anymore! Am I being completely ripped off by this guy? Well, I suppose the ares are very cheap compared to England.” Comorting himsel with this realisation, he looked or exit no. 4 rom the subway station. He had promised his riend to meet at a Burger King near exit no. 4. But no matter where he looked, he couldn’t �nd a Burger King next to exit no. 4. While he was panicking, Eric saw that the subway station read Sincheon not Sinchon! He had mispronounced [sinʨʰon], as [sinʨʰʌn], and so the taxi driver had dropped him off at Sincheon station instead o Sinchon station. “Distinguishing /ʌ/ rom /o/ in Korean is so hard! But I had better make sure o the difference rom now on”, thought Eric.
DID YOU KNOW . . . ?
The six-vowel system in the dialect of Gyeongsang Province As mentioned earlier, the vowel system shows more discrepancies across dialects than consonants do. Among the Korean dialects, the Gyeongsang dialect has the smallest number o single vowels. Unlike Standard Korean, the Gyeongsang dialect has a six-vowel system; this is because /ʌ/ and /ɯ/ are not distinct in this dialect. Figure A compares the monophthongs o the Daegu dialect – a branch o the Gyeongsang dialect – with those o the Seoul dialect. Te data or the Daegu dialect were drawn rom ten male speakers in their orties, while the data or the Seoul dialect were drawn rom ten male speakers in their twenties. Te average F1 and F2 values o the seven monophthongs in the Seoul dialect are indicated by the solid lines, while the average F1 and F2 values o the Daegu dialect are indicated by the dotted lines. Te
107
5.2
Monophthongs
F2 (Hz) 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 18001700 1600 1500 1400 13001200 11001000 900 800 700
600 500 200
iS uD
uS
iD ɯS
250 300
0S
350 400
ɯ / D
450
εD εS
0D S
500 550 600 650
) z H ( 1 F
700
ɑD
750
ɑS
800
D = Daegu S = Seoul
850 900 950 1000
Comparison between Seoul and Gyeongsang ( Daegu ) dialects (ten male speakers of the Daegu dialect in their forties) A:
Daegu dialect speakers in their orties did not show a signi�cant difference in their pronunciation o /ʌ/ and /ɯ/, and thereore the difference in the average F1 and F2 values o these two vowels is not statistically signi�cant. Te average value o the combined vowel is indicated in the chart as ɯ/ʌD. As the �gure shows, the combined vowel is situated acoustically about halway between /ʌ/ and /ɯ/ in the Seoul dialect. It is interesting to observe that the combination o these two vowels in�uences the way that the other surrounding vowels are pronounced. Tis effect can be seen more distinctly with back vowels, rather than with ront vowels. As the �gure shows, Daegu dialect speakers tend to distinguish /u/ rom /o/ ar more clearly than Seoul dialect speakers. In particular, there is a noticeable difference in the way that /o/ is pronounced in these two dialects; in the Daegu dialect /o/ is pronounced with the mouth opened more widely than in the Seoul dialect. On the other hand, the open vowel /ɑ/ is not pronounced with the mouth opened as widely in the Daegu
108
Vowels
dialect as it is in the Seoul dialect. Tis is because the back unrounded vowels that result rom the merging o the two unrounded vowels do not require the mouth to be open as wide as do back unrounded vowels in the Seoul dialect. However, this vowel combination phenomenon was not observed among speakers in their twenties. According to H.-J. Jang and J.-Y. Shin (2006), Daegu dialect speakers in their twenties, regardless o gender, distinguish between seven monophthongs in the same way as speakers o the Seoul dialect, because o the in�uence o Standard Korean. Figure B is a chart that combines the data that H.-J. Jang and J.-Y. Shin (2006) drew rom ten male Daegu dialect speakers in their twenties with the data drawn rom ten male Seoul dialect speakers mentioned earlier in this book. As the Daegu dialect speakers in their twenties acquired a seven-vowel system, their overall vowel pronunciation patterns came to resemble those o Seoul dialect speakers. F2 (Hz) 2800 2700 2600 2500 2400 2300 2200 2100 2000 1900 18001700 1600 1500 1400 13001200 1100 1000 900
800 700
6 00
500 200
iS
uS
iD
250
uD
ɯS
0S ɯD
0D
300
350
400
εS
450
εD D
S
500
550
) z H 600 ( 1 F 650
700
750
ɑD
ɑS 800
D = Daegu S = Seoul
850
900
950
1000
Comparison between Seoul and Gyeongsang ( Daegu ) dialects (ten male speakers of the Daegu dialect in their twenties) B:
5.3
109
Diphthongs
Table 5.5 The eleven diphthongs of SKP
i
e
æ
y
ø
ɯ
ʌ
ɑ
j
–
je ()
jæ ()
–
–
w
–
we ()
wæ () –
–
ɯ j () jʌ () jɑ () – w ʌ () w ɑ ()
u
o
Glide
ju () jo () –
–
Table 5.6 The ten diphthongs of Standard Korean (tentative)
Monophthongs
5.3
i
ɛ
ɯ
ʌ
ɑ
u
o
Glide j
–
jɛ
ɯ j ()
jʌ
jɑ
ju
jo
w
wi
w ɛ
–
w ʌ
w ɑ
–
–
Diphthongs In this section, we will discuss diphthongs in Korean. According to SKP, there are eleven diphthongs as listed in able 5.5. However, i we adopt the seven-monophthong system, which assumes no distinction between /e/ and /æ/, /je/ and /jæ/ as well as /we/ and /wæ/ will also have no distinction. In addition, /wi/ should be categorised as a diphthong. Tis will give us a revised inventory o Korean diphthongs as shown in able 5.6. According to able 5.6, we can observe one off-glide () along with nine on-glides in Korean. Is the diphthong / / the only off-glide in Korean? Tis will be discussed in the next section.
DID YOU KNOW . . . ?
Is it or ? People ofen conuse and and mistake the adnominal or in writing. For example, they ofen write as , or as . Alternatively, they make the opposite mistake, writing as . Tese mistakes in orthography are due to the act the two sounds are indistinguishable.
110
5.3.1
Vowels
The diphthong // 5.3.1.1
Is // an on-glide or off-glide?
Whether to treat // in Korean as an on-glide or off-glide is one o the di�cult problems in Korean phonology. Most researchers regard / / as an offglide yet it is still debatable. Te problem is this: given that // is a diphthong, one o the vowels should be the syllable nucleus and the other should be a glide. However, it is not easy to decide which one should be a syllable nucleus and which one should be a glide. I we analyse // as /ɯ j/, a /ɯ/ vowel plus a glide /j/, this yields a unique case o off-glide in Korean phonology. We can wonder i it is explanatorily adequate to postulate an off-glide sound in Korean to explain / /. On the other hand, i we analyse // as /ɰi/, where /i/ is a monophthong and / ɰ/ is attached to it as an on-glide, we won’t have a unique off-glide. Nevertheless, this means we have an extra glide /ɰ/ in the inventory o Korean. In this book, we adopt the latter view and argue that /ɰ/ should be added in the list o glides in Korea. Tere are two major reasons or this claim. First, i we assume that / / has an off-glide, we would have to change our assumption that diphthongs in Korean are alling. Moreover, we would need to explain why the glide /j/ combines in this way only with /ɯ/, but not with other vowels, as we can see rom able 5.6. On the other hand, i we assume that // has an on-glide, then we avoid the problems outlined above, and simply state that the reason / ɰ/ combines only with /i/ in this way is due to the constraints placed on glide and monophthong combinations. Second, Korean diphthongs have certain constraints placed on them with regard to preceding consonants. For instance, the glide /j/ cannot be combined with coronal consonants such as alveolar or alveolo-palatal sounds within a morpheme. As a result, in pure Korean words, coronal consonants can never be combined with /j/ and in Sino-Korean words, coronal obstruents can never be combined with /j/. On the other hand, /w/ cannot be combined with bilabial consonants. (5)
*
/sjo/ , */sju/ , */tɕ ju/ , */tɕ* ju/ , */mwa/
Tese constraints are only valid within a morpheme, though in cases where a sequence o morphemes is contracted, the constraints are not observed in the surace phonetic realisation. For instance, the phoneme /sj ʌ/ is never observed as an independent morpheme, but is observed as the pronunciation o the contracted orm o /hɑ- + -si- + -ʌ/, as shown in (6).
111
5.3
(6)
Diphthongs
/hɑ- + -si- + -ʌ/ → [hɑsjʌ]
Morphologically, ‘ ’ cannot be preceded by a consonant. Although // is preceded by a consonant in words such as (/nɰillili/), (/hɰi+n+sɛk/) and (/munɰi/), in the actual pronunciation, // is pronounced as /i/ in all o these words. O course, / / can be combined with consonants in its surace orm. Standard Seoul Korean speakers will thereore pronounce (/pʌp # ɰihɑk/) ‘legal medicine’ or (/hɯntɯl # ɰiʨɑ/) ‘rocking chair’ as [p ʌ.pɰi.hɑk] and [hɯn.tɯ.lɰi.ʨɑ]. In sum, i we assume / / to be a alling diphthong (i.e. a diphthong with an off-glide), we have to postulate a new combining constraint or alling diphthongs in general. On the other hand, i we assume / / to be a rising diphthong (i.e., a diphthong with an on-glide), we can say that the combining constraint observed above is not applicable to all rising diphthongs, but only the glide // is combined with certain consonants. Tis is more plausible, as this constraint is universal to all languages.
5.3.1.2
The phonetic realisation of //
In spontaneous speech, /ɰi/ is realised as [ɰi] only when it appears as the �rst syllable o a ree morpheme. Otherwise, / ɰi/ is realised as [i] in general as in (7) below. Te only exception is when the �rst segment o the �rst syllable is a consonant as in (7c) below. (7) a. /ɰisɑ/ [ɰisɑ] ‘doctor’, /ɰiʨɑ/ [ɰiʨɑ] ‘chair’ /jʌ # ɰisɑ/ [jʌɰisɑ] ‘emale doctor’, /hɯntɯl # ɰiʨɑ/ [hɯntɯlɰiʨɑ] ‘rocking chair’ b. /minʨuʨuɰi/ [minʨuʨui] ‘democracy’, /k ɑŋɰi/ [k ɑŋɰi] ‘lecture’ c. /hɰimɑŋ/ [himɑŋ] ‘hope’ (# = word boundary) I /minʨuʨuɰi / and /k ɑŋɰi/ in (7b) are pronounced as [minʨuʨuɰi] or [k ɑŋɰi] instead o [minʨuʨui] and [k ɑŋi], it sounds very unnatural. Interestingly, however, genitive {ɰi} is pronounced not as [ɰi] but as [ ɛ] in general. For instance, /(suni + - ɰi) (sin)/ ‘Suni’s shoes’ is more naturally pronounced as [suni ɛsin] instead o [suniɰisin].
112
Vowels
Table 5.7 Diphthongs in Standard Korean
i
ɛ
ɑ
ɯ
u
ʌ
o
j
–
jɛ
jɑ
–
ju
jʌ
jo
w
wi
w ɛ
w ɑ
–
–
w ʌ
–
ɰ
ɰi
–
–
–
–
–
–
Glide
DID YOU KNOW . . . ?
No courtesy in the Jeolla dialect? Tere are discrepancies across dialects even in the list o diphthongs. Te Jeolla dialect is a prime example o this. In the Jeolla dialect, unlike in Standard Korean, the glide /j/ cannot be combined with any ront vowel. Tus, neither /jε/ nor /ji/ exists in its list o diphthongs. Te diphthong /ɰi/ does not exist in the Jeolla dialect either. Because o this, the /j ɛ/ o standard Korean is replaced by / ɛ/ in the Jeolla dialect, and /ɰi/ by /ɯ/. Tereore, while /j ɛɰi/ ‘courtesy’ is pronounced as [jɛi] in Standard Korean, it is pronounced as [ɛɯ] in Jeolla dialect. In other words, this means that there is no [jɛɰi], or courtesy, in the Jeolla dialect.
5.3.2
The diphthongal system of Standard Korean able 5.7 provides a urther revised chart or diphthongs in Korean. Altogether it has ten vowels, which are indeed the combination o seven monophthongs and three glides. As we can see rom the gaps in the chart, not all glides can combine with all monophthongs and not all monophthongs can combine with all glides. Tere are some restrictions in diphthong ormation as in (8). (8) Some constraints in diphthong ormation in Korean a. All diphthongs in Korean have a structure o a glide plus a monophthong, where a glide precedes and does not ollow the monophthong (i.e. on-glides). No glides can occur afer a monophthong (i.e. no off-glides). b. No glide can be attached to /ɯ/.
113
5.3
Diphthongs
c. /j/ can’t be attached to /i/ and /w/ can’t be attached to /u/ and /o/. Tis shows that the possible combinations are restricted when the two sounds are too similar. d. /ɰ/ can be attached to /i/ only. We will conclude this section by discussing the bene�t o considering the diphthong in Korean as a combination o two phonemes instead o one independent phoneme. Economy and simplicity o the system is the �rst reason. I we regard each diphthong as an independent phoneme, we will have 19 consonants, 7 monophthongs and 10 diphthongs, which altogether are 36 phonemes. But, i we consider a diphthong as the combination o a monophthong and a glide, we will have 19 consonants, 7 monophthongs, and 3 glides, which is 29 altogether. Not only this, we can also provide a systematic explanation or why only certain diphthongs can exist as we discussed above.
DID YOU KNOW . . .?
Knowing how to pronounce things correctly can save your life! Some sparrows were sitting in a row on an electric line and chatting away. Ten, suddenly, another sparrow �ew over and said urgently to the perched sparrows, “Guys, we’re in big trouble! Tere’s a hunter in town! But, apparently, he asks you where your hometown is �rst, then, he kills all the sparrows rom Gyeongsang-do and spares all the rest!” Te sparrow rom Gyeongsang-do was overcome with ear at the news. However, he considered himsel lucky to have been warned in good time. He �gured that as long as he lied about his hometown, he would be spared. Ten one day, the hunter appeared beore the sparrow rom Gyeongsangdo. Just as it was rumoured, the hunter asked the sparrow where he was rom. Hunter: Sparrow:
(pointing the gun towards the sparrow) Where is your hometown? (speaking as calmly as possible) [k ɛŋkito] !
Needless to say, the hunter shot the sparrow straight away. Te story above can be very unny with just a little background knowledge about the Gyeongsang dialect. In the Gyeongsang dialect, the C + /jʌ/ o Standard Korean corresponds to C + / ɛ/; Gyeongsang dialect speakers thereore pronounce [kj ʌŋʨɛ] , [mjʌŋʨɛ] , and [pjʌŋ]
114
Vowels
in Standard Korean as [k ɛŋʨɛ], [mɛŋʨɛ], [pɛŋ] respectively. It was inevitable that the sparrow rom Gyeongsang-do would pronounce [kjʌŋkito] in Standard Korean as [k ɛŋkito]. In the end, the alse hometown that the sparrow had decided on revealed most clearly the phonemic characteristics o the dialect o his true hometown. I this sparrow had been more knowledgeable about the differences in pronunciation across dialects, he would have been spared his lie!
DID YOU KNOW . . . ?
Triphthongs found in Korean As noted above, the vowel system o Korean consists o seven monophthongs and ten diphthongs. However, i one examines the surace level phonetics careully, one �nds that there are not only diphthongs but also triphthongs. riphthongs are not ound within morphemes, but observed only in particular cases o the joining o morphemes. In particular, this sound can be heard only in the contraction that occurs when a stem that ends with /wi/ is combined with an ending that starts with /ʌ/. O course, this cannot be observed in the writing itsel, but only in speech. Tere is as yet no letter in Korean orthography that can represent this sound. Some examples are as ollows: /pɑk*wi (‘to change’) + ʌ/ → [pɑk*wjʌ]/[pɑk* jʌ] /sɑkwi (‘to make riends with’) + ʌ/ → [sɑkwjʌ]/[sɑkjʌ] /swi (‘to rest’) + ʌ/ → [swjʌ]/[sjʌ] /nwi (‘to lie down’) + ʌ / → [nwjʌ]/[njʌ] /t*wi (‘to run’) + ʌ/ → [t*wjʌ]/[t* jʌ]
As these examples show, when the diphthong, /wi/, in the word stem meets the /ʌ/ word ending, it sometimes results in the contraction, [wjʌ], or sometimes the [w] drops out to orm [jʌ]. Te omission o the semivowel [w] occurs most commonly when it is located in the second syllable or afer. Te question arises, thereore, o how the triphthong, which exists in Korean phonetics, ought to be represented in the Korean phonemic
115
5.4
Features for vowels
inventory. Whilst it is true that triphthongs exist in surace level phonetics, this sound cannot be observed within a morpheme, but only where contraction occurs between particular morphemes. Te triphthong thereore cannot acquire the status o a phoneme, since this sound only occurs as a by-product o common speech (and does not exist in the oundations o Korean phonology). Te above is similar to the phenomenon in which the alveolar affricative [ts], although absent rom the list o English phonemes, is heard in surace level phonetics, in words such as /sports/ or /texts/. Te alveolar affricative, /ʦ/, does not exist in the oundational phoneme system o English. However, when /t/ and /s/ are brought together by a combination o certain morphemes, the alveolar affricative, / ʦ/, can be observed in the surace level phonetics. o conclude, the alveolar affricative [ ʦ] in English, and the triphthong [wj ʌ] in Korean, are not necessarily included in the phoneme inventory o those languages; rather, they are observed as surace level sounds that occur through the bridging o the phonemes /t/ and /s/, and /wi/ and /ʌ/.
Table 5.8 Features for vowels
5.4
i
ɛ
ɯ
ʌ
u
o
ɑ
[high]
+
−
+
−
+
−
[low]
−
−
−
−
−
−
+
[back]
−
−
+
+
+
+
+
[round]
−
−
−
−
+
+
−
Features for vowels Tere are three actors that play an important role in the articulation o a vowel: (i) roundedness (lip protrusion), (ii) tongue backness and (iii) tongue height. Te �rst actor is related to the lips and the second and third are related to the body o the tongue. Vowels are articulated mostly by using the body o the tongue, but consonants are articulated in general by using the tip or blade o the tongue. Tis is why most eatures or vowels are related to the body o the tongue. o distinguish seven monophthongs in Korean, we need our eatures as given in able 5.8.
Vowels
116
Table 5.9 Features for vowels and glides
i
ɛ
ɯ
ʌ
u
o
ɑ
j
w
ɰ
[syllablic]
[high]
[low]
[back]
[round]
Let’s think brie�y about the our eatures used in the table. O these, three eatures relate to the body o the tongue, namely [high], [low] and [back], while the eature [round] relates to the lips. Each eature is explained in (9). (9) Features or vowels a. Features related to the body o the tongue (i) [+/−high]: [+high] sounds are produced by raising the body o the tongue above the level that it occupies in the neutral position.2 (ii) [+/−low]: [+low] sounds are produced by lowering the body o the tongue below the level that it occupies in the neutral position. (iii) [+/−back]: [+back] sounds are produced by retracting the body o the tongue rom the neutral position. b. Features related to the lips (i) [+/−round]: [+round] sounds are produced with the lips narrowed. Glides and vowels share the same eatures, except or the eature [syllabic]: vowels possess the eature [+syllabic], whereas glides are [–syllabic]. Tis means that /j, w, ɰ/ are exactly the same was /i, u, ɰ/ except or the value o their syllabic eatures, as shown in able 5.9. 5.5
Summary In this chapter, we have discussed vowels in Korean. Tere are three types o vowels in Korean: monophthongs (simple vowels), diphthongs and glides. Glides are consonants in English, but it is better to view them as vowels in Korean. We demonstrated that all diphthongs in Korean are on-glides, that is, made up o glides preceding monophthongs. Monophthongs can be classi�ed
Exercises
117
by the height and backness o the tongue and by the roundedness o the lips. Some vowel distinctions in Korean, such as the one between /e/ and /æ/, no longer seem to exist. aking into account the lack o true distinction between /e/ and /æ/, we have shown that there are seven monophthongs in Korean, and ten diphthongs, which are combinations o these seven monophthongs with the three glides. In order to provide more natural explanations or vowelrelated phonological processes, we need to classiy each vowel as a set o eatures. O the our major eatures or vowels, three are related to the position o the body o the tongue (e.g. [ +/−high], [+/−low], [+/−back]), and one is related to the �at or rounded shape o the lips (i.e. [+/ −round]). EXERCISES 1 2
3 4 5
6 7 8 9
Define monophthongs, diphthongs and glides. Is a glide a consonant or a vowel in Korean? Explain why with relevant examples. Are // and // monophthongs or diphthongs? Is there any difference between /e/ and /æ/ in contemporary Korean? State whether the following pairs of words form minimal pairs in Korean and explain why. Also, write down the Korean letter(s) as the relevant phoneme(s). a. ‘gold’, ‘bear’ b. ‘ant’, ‘spider’ c. ‘brain’, ‘cave’ d. ‘curve’, ‘to be pretty’ e. ‘I’, ‘you’ f. ‘soup’, ‘song’ Which monophthong(s) can be found in Korean, but not in English? What is the difference between English diphthongs and Korean diphthongs? Discuss dialectal differences in Korean vowels. Classify the seven monophthongs in Korean using the features [high], [low], [back] and [round].
6 Frequency trends of Korean sounds
In Chapters 4 and 5, we have discussed the phonetic and phonological characteristics o consonants and vowels in Korean. However, not all o the sounds which exist are ound requently in Korean, and some are used more ofen than others. In this chapter, we would like to explore the requency o sounds in Korean to �nd out which sounds are requently used and which are used less. We will try to �nd the answers by looking at dictionary data and spontaneous speech data. In addition, we will compare the requency o sounds in Korean with that o sounds in English. Tere are 19 consonants and 7 monophthongs in Korean, along with 10 diphthongs made rom 7 monophthongs and 3 glides. Tereore, 36 sounds exist in total in the Korean sounds inventory. In the previous chapter, we considered a diphthong as a combination o a monophthong and a glide. Tis view can provide a better explanation because it doesn’t increase the number o phonemes unnecessarily. In addition, this can capture the phonetic similarity between the corresponding monophthong and the glide and hence increase explanatory adequacy in accounting or the behaviours o speech sounds. However, in terms o explaining phoneme requency, it is better to analyse diphthongs as one discrete phoneme instead o a combination o two phonemes. By doing so, we can easily observe the requency difference between diphthongs as well as monophthongs. Tereore, in this chapter we will regard diphthongs as one phoneme. As a result, the number o phonemes in Korean increases rom 29 to 36 (19 consonants, 7 monophthongs, 10 diphthongs). Te organisation o this chapter is as ollows: in 6.1, we will look at how the data under discussion have been collected; in 6.2 and 6.3, we will
118
6.1 Database
119
Table 6.1 Sound distribution (dictionary)
Lexemes
47,401
Syllables
122,761
Sounds Vowels
122,761 (42.3%)
Consonants
Onset
110,015 (65.7%)
Coda
57,355 (34.3%)
Subtotal
Total
167,370 (57.7%) 290,131
discuss phoneme and syllable requency respectively. In 6.4, we will compare sound requency in Korean with that o English, and the conclusion ollows in 6.5. 6.1 Database 6.1.1
Dictionary We have chosen the Yonsei Korean Language Dictionary (henceorth ‘dictionary’) or our discussion in this se ction. Compared to other dictionaries, the Yonsei dictionary better re�ects ‘real’ usage in selecting word entries. Among the 49,553 entries in the Yonsei dictionary, we have excluded those entries that are not ound in the Standard Korean Language Dictionary .1 Tis leaves us with 47,401 entries. able 6.1 shows the statistics or those entries.
6.1.2
Spontaneous speech o investigate sound distribution in spontaneous speech, we have used a database called SLILC (Spoken Language Inormation Lab Corpus), established by one o the authors (J.-Y. Shin 2008). o create this database, 57 native speakers’ dialogue was recorded and subsequently transcribed. O those 57 speakers, 28 were male and 29 emale; their ages ranged rom 19 to 32; and all 57 spoke Seoul Korean. Recording took place in a phonetic booth, and three people participated in each conversation. Te total length o recorded material amounts to approximately 23 hours. able 6.2 shows the statistics or the sounds occurring in the database.
120
Frequency trends of Korean sounds
Table 6.2 Sound distribution (speech)
Speakers (M, F)
57 (28, 29)
Utterances
35,439
Syllables
403,605
Sounds Vowels
403,605 (47.0%)
Consonants
Onset
345,128 (75.9%)
Coda
109,779 (24.1%)
Subtotal
454,907 (53.0%)
Total
858,512
6.2 Phoneme frequency Based on the database established in 6.1, in this section, we will investigate phoneme requency. As mentioned, there are 36 sounds in Korean, consisting o 19 consonants, 7 monophthongs and 10 diphthongs. In the ollowing, while discussing phoneme requency, we will treat the diphthong as an independent phoneme instead o as a combination o a monophthong and a glide, or simplicity o discussion. In 6.2.1, we will discuss overall phoneme requency and in 6.2.2 and 6.2.3, we will discuss the distribution o consonants and vowels in the dictionary and in spontaneous speech, respectively. 6.2.1
Overall phoneme frequency able 6.3 shows the rankings o phoneme requency ound in the dictionary. Likewise, able 6.4 shows the rankings o phoneme requency as ound in spontaneous speech. We can see that in both databases, / ɑ/ is the most requently used phoneme and / ɰi/ is the least requently used phoneme. As can be seen in Figure 6.1, not all phonemes appeared with similar requency. ‘DIC’ indicates dictionary data and ‘SP’ indicates spontaneous speech. In other words, in both the dictionary and in speech, some phonemes were much more requently used than others. In the dictionary, the our most highly ranked phonemes constituted 30.90% o total distribution, and the eight most highly ranked phonemes constituted 51.86%. However, the bottom 16 phonemes constituted only 9.82%. Phoneme distribution becomes even more asymmetric in the case o spontaneous speech. Here, the top three phonemes constituted 30.43% o total distribution, and the top seven constituted 55.78% o total distribution, whereas the bottom 18 phonemes constituted only 9.95%.
6.2 Phoneme frequency
121
Table 6.3 Phoneme frequency (dictionary)
Rank Phoneme Frequency %
Cumulative (%)
Rank Phoneme Frequency
1
ɑ
28,844
9.9
9.90
19
2
k
22,284
7.7
17.62
3
n
19,385
6.7
24.30
Cumulative % (%)
5,290
1.8
89.04
20
ʨʰ pʰ
3,289
1.1
90.18
21
ʨ*
2,922
1.0
91.18
*
4
i
19,133
6.6
30.90
22
k
2,784
1.0
92.14
5
ŋ
16,726
5.8
36.66
23
2,772
1
93.10
6
l
16,090
5.5
42.21
24
w ɛ s*
2,726
0.9
94.04
7
ʌ
14,783
5.1
47.30
25
w ɑ
2,662
0.9
94.96
8
ʨ
13,205
4.6
51.86
26
th
2,609
0.9
95.85
9
s
12,818
4.4
56.27
27
2,361
0.8
96.67
10
o
12,817
4.4
60.69
28
k ʰ t*
1,821
0.6
97.30
11
u
12,557
4.3
65.02
29
jo
1,529
0.5
97.83
12
m
11,802
4.1
69.09
30
ju
1,329
0.5
98.28
13
h
11,224
3.9
72.96
1,232
0.4
98.71
14
p
10,161
3.5
76.46
31 jɑ * 32 p
1,216
0.4
99.12
15
ɛ
9,309
3.2
79.67
33
wi
998
0.3
99.47
16
t
8,657
3.0
82.65
34
w ʌ
997
0.3
99.81
17
ɯ
6,882
2.4
85.02
35 jɛ
314
0.1
99.92
18
jʌ
6,373
2.2
87.22
36
230
0.1
100.00
6.2.2
ɰi
Consonant frequency In this section, we will discuss phonemic distribution with respect to consonants. Section 6.2.2.1 will show overall distribution and then examine the distribution o consonants with regard to their position within the syllable; in 6.2.2.2 we will examine consonant distribution in syllable-initial, or onset, position, and in 6.2.2.3 we will look at consonant distribution in syllable-�nal, or coda, position. 6.2.2.1
Overall consonant frequency
Figure 6.2 shows overall consonant requency as observed in the dictionary. Te most requently used consonant was /k/, ollowed by /n/, /ŋ/, /l/ and / ʨ/.
Frequency trends of Korean sounds
122
Table 6.4 Phoneme frequency (speech)
Rank Phoneme Frequency
%
Cumulative (%)
%
Cumulative (%)
11,355
1.32
91.38
10,638
1.24
92.62
Rank Phoneme Frequency
1
ɑ
95,853
11.17
11.17
19
2
n
92,621
10.79
21.95
20
k* s*
3
k
72,756
8.47
30.43
21
t
8,873
1.03
93.65
4
i
60,318
7.03
37.45
22
7,480
0.87
94.52
5
l
54,253
6.32
43.77
23
6,640
0.77
95.29
6
52,381
6.11
49.88
24
6,349
0.74
96.03
50,704
5.91
55.78
25
6,229
0.73
96.76
8
ʌ ɯ ɛ
50,024
5.83
61.61
26
5,553
0.65
97.41
9
m
36,132
4.21
65.82
27
3,792
0.44
97.85
10
o
31,268
3.64
69.46
28
3,628
0.42
98.27
11
t
30,704
3.58
73.04
29
ʨ* jɑ ʨʰ tʰ k ʰ w ɛ w ɑ pʰ
3,399
0.4
98.67
12
ʨ
27,076
3.15
76.19
30
2,874
0.33
99.00
13
u
24,061
2.8
79.00
31
jo p*
2,457
0.29
99.29
14
ŋ
23,874
2.78
81.78
32
w ʌ
2,345
0.27
99.56
15
s
22,827
2.66
84.44
33
jɛ
1,614
0.19
99.75
16
h
16,494
1.92
86.36
34
ju
1,165
0.14
99.88
17
jʌ
15,915
1.85
88.21
35
wi
905
0.11
99.99
18
p
15,837
1.84
90.05
36
ɰi
85
0.01
100.00
7
*
Te above �ve consonants constituted 52.4% o total consonant distribution. On the other hand, the bottom seven consonants, shown in the righthand box, constituted just 9.8% o the total consonant distribution. It is worth noting that both /k/ and /ŋ/ are highly ranked in the distribution. Given that it occurs only in syllable-�nal position, the act that /ŋ/ is the third most requently used consonant is particularly signi�cant. Figure 6.3 shows overall consonant requency as observed in spontaneous speech. Te most requently used consonant here was /n/, ollowed by /k/, /l/, /m/, /t/, and /ʨ/. On the other hand, the least requently used consonant was /p*/. Te top three consonants constituted 48.3% o total distribution, whereas the bottom nine constituted only 13.7%. Te results show that in both the dictionary and in spontaneous speech, obstruent consonants were more requently used than sonorant consonants and among obstruents, lax obstruents were more requently used than their tense or aspirated counterparts.
123
6.2 Phoneme frequency
12
10 DIC
SP
) 8 % ( y c n 6 e u q e r F 4
2
0
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
35
Ranking
Figure 6.1
Frequency trends by rank (all phonemes, dictionary vs. speech) 24
20
) 16 % ( y c n 12 e u q e r F 8
52.4%
k n ŋ
l t
s
m
h
p
9.8%
t th
4
ph
*
*
*
t
k
s
1.7
1.7
1.6
th
kh
*
*
t
p
1.1
0.7
0 13.3 11.6 10.0 9.6
Figure 6.2
7.9
7.7
7.1
6.7
6.1
5.2
3.2
2.0
1.6
1.4
Consonant frequency (dictionary)
Figure 6.4 shows the requency ranking o consonants ound in the dictionary and in speech. As the �gure shows, the curve was smoother in the dictionary than in speech. In other words, in the dictionary, the distribution o consonants is even, rather than being concentrated on a limited set o sounds, as in speech. Figure 6.5 shows a comparison between the data rom the dictionary and rom spontaneous speech. Te boxed areas show cases where the distributional
124
Frequency trends of Korean sounds
24
n 20 48.3%
k
) 16 % ( y c n 12 e u q e r F 8
l m
t
t
13.7%
ŋ
s
4
h
p
k
*
s
*
t
*
*
t
th th
kh
ph p
1.4
1.2
0.7
*
0 20.4 16.0 11.9 7.9
Figure 6.3
6.7
6.0
5.2
5.0
3.6
3.5
2.5
2.3
2.0
1.6
1.4
0.5
Consonant frequency (speech) 24
20 DIC
SP
16
) % ( y c n 12 e u q e r F 8
4
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Ranking
Figure 6.4
Frequency trends by ranking (consonants, dictionary vs. speech)
difference between the dictionary and spontaneous speech is either greater than 5% (solid box) or between 3% and 5% (dashed box). O all the consonants, the distribution o /n/, which was used 8.8% more requently in spontaneous speech than in the dictionary was most asymmetric. On the other hand, /ŋ/ and /h/ were used more requently in the dictionary than in spontaneous speech: /ŋ/ was used 4.8% more in the dictionary than in speech and /h/ was used 3.1% more in the dictionary than in speech. In short, /n/ was ound requently in speech and /ŋ/ and /h/ were ound requently in the dictionary.
125
6.2 Phoneme frequency
24
20
) 16 % ( y c n 12 e u q e r F 8
4
0 k
n
ŋ
s
m
h
p
t
th
ph
t
k
7.9
7.7
7.1
6 .7
6.1
5.2
3.2
2.0
1 .7
11.9 6.0
5.0
7.9
3 .6
3.5
6.7
1.4
0.7
1.6
l t
DIC 13.3 11.6 10.0 9.6 SP 16.0 20.4 5.2
Figure 6.5
*
s
th
k h
t
1 .7
1.6
1.6
1 .4
1 .1
0.7
2.5
2.3
1.4
1.2
2.0
0.5
*
*
*
*
p
Consonant frequency differences (dictionary vs. speech)
Te act that /k/ is used less requently in speech while /n/ is used very requently is related to the use o grammatical morphemes. Indeed /n/ is the most requently used sound among grammatical morphemes and constituted o 29.8% o all consonants ound in them. For instance, grammatical morphemes such as –, a topic or adnominal particle, or the verbal ending – , are both used very requently, and both contain /n/. ables 6.5–6.7 show consonant distribution classi�ed according to obstruents vs. sonorants (able 6.5), place o articulation (able 6.6) and manner o articulation (able 6.7). able 6.5 shows that in both the dictionary and spontaneous speech, obstruents are used more requently than sonorants. In particular, in speech, obstruents occurred 7.2% more ofen than in the dictionary. Among obstruents, tense consonants were observed 5.2% more ofen in speech than in the dictionary, while aspirated consonants were observed 5.4% less requently in speech than in the dictionary. Consider able 6.6. When consonants were classi�ed according to place o articulation, the most signi�cant difference between speech and the dictionary was observed in alveolar sounds. Alveolar sounds were observed 11.4% more requently in speech than in the dictionary. Now consider able 6.7 or differences in manner o articulation. When consonants were classi�ed according to manner o articulation, the most signi�cant difference between speech and the
126
Frequency trends of Korean sounds
Table 6.5 Type and token Frequency of consonants (sonorants vs. obstruents and phonation types)
Token (%) Type (%)
DIC
SP
Δ
Sonorants
21.1
38.2
45.5
Obstruents
78.9
61.8
54.5
Lax
35.7
72.8
73.1
−
Tense
35.7
12.4
17.6
−
Aspirated
28.6
14.7
9.3
7.2
−
7.2 0.2 5.2 5.4
Table 6.6 Type and token frequency of consonants by place of articulation
Token (%) Type (%)
DIC
SP
Δ
Bilabial
21.1
15.8
12.7
3.1
Alveolar
36.8
38.3
49.7
Alveolo-palatal
15.8
12.8
9.0
3.8
Velar
21.1
26.4
25.0
1.4
Glottal
5.3
6.7
3.6
3.1
11.4
−
Table 6.7 Type and token frequency of consonants by manner of articulation
oken (%) ype (%)
DIC
SP
Δ
Stop
47.4
33.0
34.5
Fricative
15.8
16.0
11.0
5.0
Affricate
15.8
12.8
9.0
3.8
Nasal
15.8
28.6
33.6
−
Liquid
5.3
9.6
11.9
−
1.6
−
4.9 2.3
127
6.2 Phoneme frequency
24
20 46.2%
) 16 % ( y c n 12 e u q e r F 8
k
t
s 35.9%
h t
p
m
l
17.9%
n
t
h
4
ph
t
*
*
k
s
2.5
2.5
*
th
kh
*
t
p
1.7
1.1
*
0 12.3 12.0 11.7 10.2
Figure 6.6
7.0
6.8
6.3
6.2
4.9
4.8
3.0
2.7
2.4
2.1
Onset frequency (dictionary)
dictionary was observed in ricatives and nasals, which were observed 5.0% more and 4.9% less requently respectively in the dictionary than in speech. 6.2.2.2
Onset frequency
In this section, we will look at consonant distribution in syllable-initial position. As discussed in Chapter 4, only eighteen consonants can appear in this position. Figure 6.6 shows the data observed or dictionary usage. In the dictionary, /k/ was most requently used, while /p */ was least requently used. Figure 6.6 shows the consonants grouped according to their requency. Te three groups were divided at points where requency drops signi�cantly. Te lef-most group consists o high-requency consonants; the middle group consists o mid-requency consonants; and the right-most group consists o the consonants least requently used. Along with /k/, /ʨ, s, h/ appear in the high-requency consonant group. Except or /k/, all o these consonants are either ricatives or affricates. Te mid-requency consonants include /t, p, m, l, n, ʨh/. Except or / ʨh/, all were either lax stops or sonorants. In both the dictionary and in speech, /ph, ʨ*, k*, s*, t h, k h, t *, p */ appear in the low-requency group, and these consonants are all either tense or aspirated consonants. Next, we will discuss the data rom spontaneous speech. Figure 6.7 shows the data rom speech. In speech, the most requently used syllable-initial consonant was /k/ (19.2%). Te least requently used consonant was /p */ (0.7%). As shown in Figure 6.7, the top our consonants constituted 51.2% o the
128
Frequency trends of Korean sounds
24 51.2%
k
20
) 16 % ( y c n 12 e u q e r F 8
n l
t
t
m
39.7%
s
9.1%
h 4
p
*
k
*
s
*
t
*
t
th th
kh
1.8
1.8
1.6
1.0
s
th
kh
t
p
ph
*
p
0 19.2 13.4 9.8
Figure 6.7
8.8
7.8
7.8
6.6
t
p
m
4.8
3.7
3.3
3.1
2.6
2.2
l
n
th
ph
t
k
0.7
Onset frequency (speech) 24
20
) % ( 16 y c n e u 12 q e r F 8
4
0 k
Figure 6.8
*
*
*
*
s
DIC 12.3
12.0
11.7
10.2
7.0
6.8
6.3
6.2
4.9
4.8
3.0
2.7
2.7
2.5
2.4
2.1
1.7
1.1
SP
7.8
6.6
4.8
8.8
3.7
7.8
9.8
13.4
1.8
1.0
2.2
3.3
3.1
1.8
1.6
2.6
0.7
19.2
h
*
t
Onset frequency differences (dictionary vs. speech)
overall distribution o syllable-initial consonants, while the bottom six constituted just 9.1% altogether. Figure 6.8 shows a comparison between the data rom the dictionary and rom speech. Te boxed areas show cases where the distributional difference between the dictionary and spontaneous speech is either greater than 5% or between 3% and 5%. Consonants which show a difference greater than 5% in both speech and the dictionary include /k/ (6.9% more in speech), /s/ (5.1% more in the dictionary), /h/ (5.4% more in the dictionary), and /n/ (8.5% more in speech). Furthermore, consonants which show a difference between 3% and
129
6.2 Phoneme frequency
Table 6.8 Type and token frequency of onsets (sonorants vs. obstruents and phonation types)
Token (%) Type (%)
DIC
SP
Δ
Sonorants
16.7
17.3
31.0
Obstruents
83.3
82.7
69.0
13.7
Lax
35.7
68.6
71.9
−
Tense
35.7
14.4
18.4
−
Aspirated
28.6
17.0
9.7
13.7
−
3.3 4.0 7.3
Table 6.9 Type and token frequency of onsets by place of articulation
Token (%) Type (%)
DIC
SP
Δ
Bilabial
22.2
17.2
13.2
4.0
Alveolar
38.9
36.1
46.1
Alveolo-palatal
16.7
19.5
11.9
Velar
16.7
17.0
24.1
Glottal
5.6
10.2
4.8
9.9
−
7.6 7.1
−
5.4
5% include / ʨ/ (4.2% more in the dictionary), /p/ (3.1% more in dictionary), /l/ (3.6% more in speech), and / ʨh/ (3% more in the dictionary). Tis data shows that in syllable-initial position, /n/, /k/ and /l/ are used relatively more requently in speech than in the dictionary, while /h/, /s/, / ʨ/, /p/, /ʨh/ are used relatively more requently in the dictionary than in speech. ables 6.8–6.10 show consonant distribution classi�ed according to obstruents vs. sonorants (able 6.8), place o articulation (able 6.9) and manner o articulation (able 6.10). able 6.8 shows that in both the dictionary and spontaneous speech, obstruents are used more requently than sonorants. In particular, obstruents occurred 13.7% more ofen in the dictionary. Among obstruents, in both the dictionary and in spontaneous speech, lax consonants were much more requently used than their tense and aspirated counterparts. Aspirated consonants, which account or 17.0% o total distribution in the dictionary, only accounted or 9.7% in speech.
130
Frequency trends of Korean sounds
Table 6.10 Type and token frequency of onsets by manner of articulation
Token (%) Type (%)
DIC
SP
Δ
Stop
50.0
38.9
42.7
Fricative
16.7
24.3
14.5
9.9
Affricate
16.7
19.5
11.9
7.6
Nasal
11.1
11.2
21.1
Liquid
5.6
6.2
9.8
3.8
−
10.0
−
3.7
−
able 6.9 shows that alveolar sounds are the most requently used, both in the dictionary and in speech, but were much more requently used in speech than in the dictionary. able 6.10 shows that stops are the most requently used in both the dictionary and in speech. Fricatives and affricates were more requent in the dictionary than in speech, but on the other hand, nasals were more requent in speech than in the dictionary. 6.2.2.3
Coda frequency
In the ollowing section, we will discuss the distribution o syllable-�nal consonants, that is, codas. As discussed in Chapter 4, only seven consonants can appear in this position. Figure 6.9 shows the data rom the dictionary or syllable-�nal consonants. Among syllable-�nal consonants, /ŋ/ was the most requently used (29.1%), whereas the alveolar stop /t/ was the least requently used (1.8%). Along with /ŋ/, /n/ was requently used, and together the two consonants account or 53.6% o overall distribution. Te three consonants that occurred least requently, i.e. /m, p, t/, only account or 15.0% o overall distribution. Figure 6.10 shows the data rom speech. Again, as in the dictionary, /n/ was the most requently used syllable-�nal consonant (42.3%), while the al veolar stop /t/ was the least requently used (0.3%). Next to /n/, /ŋ/ was most requently used, and together the two consonants account or 64% o overall distribution. Te three consonants that occurred least requently, i.e. /k, p, t/, only account or 9% o overall distribution. Consider Figure 6.11. Te boxed areas show cases where the distributional difference between the dictionary and spontaneous speech is greater than 5%. Among these, /ŋ/ was observed 7.4% more requently in the dictionary than
131
6.2 Phoneme frequency
32
ŋ 53.6%
28
n 24
) % ( 20 y c n 16 e u q e r 12 F
31.4%
l
k 15.0%
m
8
p
4 0
Figure 6.9
t 29.1
24.5
16.2
15.2
8.5
4.7
1.8
Coda frequency (dictionary) 45
n
64.0%
40 35 30
) % ( 25 y c n 20 e u q 15 e r F
27.0%
ŋ l
9.0%
m
10
k p
5
t
0 42.3
Figure 6.10
21.7
18.5
8.5
5.9
2.8
0.3
Coda frequency (speech)
in speech, /n/ was observed 17.8% more requently in speech than in the dictionary, and /k/ was observed 9.3% more requently in the dictionary than in speech. ables 6.11–6.13 show consonant distribution in codas classi�ed according to obstruents vs. sonorants (6.11), place o articulation (6.12), and manner o articulation (6.13). able 6.11 shows that in both the dictionary and in spontaneous speech, sonorants were more requently used than obstruents in syllable-�nal (coda) position. In particular, in speech, sonorants accounted or 91.0% o overall distribution, 12.7% higher than in the dictionary, where sonorants accounted or 78.3% o consonants in syllable-�nal position.
132
Frequency trends of Korean sounds
Table 6.11 Type and token frequency of codas (sonorants vs. obstruents)
oken (%) ype (%)
DIC
SP
Sonorants
57.1
78.3
91.0
Obstruents
42.9
21.7
9.0
Δ 12.7
−
12.7
Table 6.12 Type and token frequency of codas (place of articulation)
Token (%) Type (%)
DIC
SP
Bilabial
28.6
13.2
11.3
Alveolar
42.9
42.5
61.1
Velar
28.6
44.3
27.6
Δ
1.9 18.6
−
16.7
45 40 35
) 30 % ( y 25 c n e u 20 q e r F 15 10 5 0
Figure 6.11
ŋ
n
l
k
m
p
t
DIC
29.1
24.5
16.2
15.2
8.5
4.7
1.8
SP
21.7
42.3
18.5
5.9
8.5
2.8
0.3
Coda frequency differences (dictionary vs. speech)
able 6.12 shows that alveolar and velar sounds appear requently, while the requency o bilabial sounds is low. In particular, in speech, alveolar sounds accounted or 61.1% o overall distribution. Tis is 18.6% higher than in the dictionary, where alveolar sounds accounted or 42.5% o consonants in syllable-�nal position.
6.2 Phoneme frequency
133
Table 6.13 Type and token frequency of codas (manner of articulation)
Token (%) Type (%)
DIC
SP
Δ
Stop
42.9
21.6
9.0
12.7
Nasal
42.9
62.1
72.5
Liquid
14.3
16.2
18.5
10.4
−
2.3
−
ɑ
24
51.1% 20
i
) 16 % ( y c n 12 e u q e r F 8
o
u ε ɯ
9.8%
j
4
wε wɑ
jo
ju
jɑ
wi w
jε
i
1.2
1.1
1.0
0.8
0.3
0.2
0 23.5 15.6 12.0 10.4 10.2 7.6
Figure 6.12
5.6
5.2
2.3
2.2
0.8
Vowel frequency (dictionary)
able 6.13 shows that in both the dictionary and in speech, nasals are dominant. In the dictionary, the proportion o stops was relatively higher than in speech, while the proportion o nasals was higher in speech than in the dictionary. No signi�cant difference is observed in the distribution o liquids between the dictionary and speech. 6.2.3
Vowel frequency In this section, we will discuss the phonemic distribution o vowels. Figure 6.12 shows vowels ranked according to the requency with which they appear in the dictionary. Te most requently observed vowel was / ɑ/ (23.5%), ollowed by, /i/, / ʌ/, and /o/. It is worth noting that the top seven vowels, which amount to 85.0% o all vowel occurrences, are all monophthongs. Tis is much higher than the total or diphthongs (15.0%). Te three most requently used
134
Frequency trends of Korean sounds
24
ɑ 51.7%
20
) 16 % ( y c n 12 e u q e r F 8
i
ɯ
ε
o u
9.6%
j 4
jɑ
wε wɑ
jo w jε ju
wi
i
0.2
0.0
0 23.8 14.9 13.0
Figure 6.13
12.6
12.4
7.7
6.0
3.9 1.7
0.9
0.9
0.7
0.6
0.4
0.3
Vowel frequency (speech)
vowels constituted 51.1% o overall distribution, whereas the nine least used vowels accounted or only 9.8% o the total. O monophthongs, / ɯ/ was the least requently observed and o diphthongs, /j ʌ/ was the most requently observed. Figure 6.13 shows vowels according to their requency in speech. Te most requently observed vowel was / ɑ/ (23.8%). Following this, /i/, /ʌ/, /ɯ/, /ε/ were most requent. Just as in the dictionary, the seven most requently used vowels are all monophthongs. Te three most requently used vowels constitute 51.7% o overall distribution, whereas the ten least used vowels constitute only 9.6% o the total. Among monophthongs, /u/ was least observed and among diphthongs, /jʌ/ was most requently observed. Consider Figure 6.14. Unlike in the case o consonants, distributional patterns or vowels were similar both in speech and in the dictionary. Tat is, it cannot be said that distributional asymmetry was more marked in speech than in the dictionary. Figure 6.15 shows a comparison between the data rom the dictionary and rom speech. Te only vowel which shows a requency difference between speech and dictionary greater than 5% is / ɯ/ (7% more in speech). Vowels which show a difference between 3% and 5% include /u/ (4.3% more in the dictionary) and /ε/ (4.8 % more in the dictionary). While /ɯ/ is requent in speech, /u/ and /ε/ are requent in the dictionary. Overall, diphthongs are even less requently observed in speech than in the dictionary.
135
6.2 Phoneme frequency
24
20 DIC
SP
16
) % ( y c n 12 e u q e r F 8
4
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Ranking
Figure 6.14
Vowel Vowel frequency trends by ranking (dictionary vs. speech) 24
20
) 16 % ( y c n 12 e u q e r F 8 4
0
Figure 6.15
ɑ
i
o
u
DIC 23.5
15.6
12.0
10.4
10.2
SP
14.9
13.0
7.7
6.0
23.8
ε
ɯ
j
wε wɑ
7.6
5.6
5.2
2.3
12.4
12.6
3.9
0.9
jo
ju
jɑ
wi
w
jε
i
2.2
1.2
1.1
1.0
0.8
0.8
0.3
0.2
0.9
0.7
0.3
1.7
0.2
0.6
0.4
0.0
Vowel Vowel frequency (dictionary vs. speech)
ables 6.14 6 .14–6.17 –6.17 show the classi�cation o vowels. Te results in able 6.14 show that monophthongs are much more requently used than diphthongs. In particular, monophthongs are more requently used than diphthongs in speech, where they accounted or 90.4% o total distribution. In the dictionary, monophthongs accounted or 85.0% o total distribution.
136
Frequency trends of Korean sounds
frequency of vowels vowels Table 6.14 Type and token frequency (monophthongs vs. diphthongs)
Token (%) Type (%)
DIC
SP
Monophthong
41.2
85.0
90.4
Diphthong
58.8
15.0
9.6
Δ
5.4
−
5.4
Table 6.15 Type and token frequency frequency of vowels vowels by vowel height
Token (%) Type (%)
DIC
SP
Δ
High
42.9
37.0
37.1
−
Mid
42.9
35.4
36.6
−
Low
14.3
27.6
26.3
0.1 1.2 1.3
Table 6.16 Type and token frequency frequency of vowels vowels by tongue backness
Token (%) Type (%)
DIC
SP
Front
28.6
27.3
30.2
Back
71.4
72.7
69.8
Δ
2.9
−
2.9
Table 6.17 Type and token frequency frequency of vowels vowels by roundedness
Token (%) Type (%)
DIC
SP
Δ
Rounded
28.6
24.3
15.2
9.1
Unrounded
71.4
75.7
84.8
9.1
−
able 6.15 shows that tongue height does not appear to have a major impact on vowel distribution. able 6.16 shows that in both the dictionary and in speech, back vowels were more requently used than ront vowels. able 6.17 shows that in both the dictionary and in speech, unrounded vowels were more requently used than rounded vowels.
137
6.3 Syllable frequency
6.3 Syllable frequency 6.3.1
High-frequency syllables In the dictionary, 1,283 syllable types were observed in total. Among these, the most requently used syllable was /h ɑ/, which was observed 5,414 times. Te syllables /li/ and / ʨi/ were also requently used. Te reason why /h ɑ/ was so requent is i s related to the t he widespread widesprea d use o the t he verbal suffi s uffix /-ha-/ -- ‘to do’ in Korean. able 6.18 shows 82 syllable types which constitute more than 50% o overall usage. It is signi�cant that the usage o these 82 syllable types, just 6.39% o all observed 1,283 syllable types, constitute 50.29% o the total distribution. High-requency syllable structures are in general either o the V type or CV type. Among seven possible V types, �ve are ound in the high-requency group. In addition, among 127 observed CV types, 45 are ound in the highrequency group. On the other hand, although approximately 778 structures o the CVC type have been observed, only 21 are ound in the high-requency group. In speech, 1,212 syllable types were observed in total. O these, the most requently-used syllable was /k ɯ/, which was observed 13,755 times, constituting 3.41% o the overall distribution. /k ɑ/ and /i/ were also requently used. Tere were 36 syllable types which together accounted or approximately hal o the total distribution. In other words, this means that just 2.97% o all observed structure types accounted or 50% o the syllables used in speech. able 6.19 shows a list o the 36 most requently used syllables. Most o those 36 were either V or CV types. O seven possible V types, our appear in the high-requency high-requency group. In addition, o 126 possible CV type syllable structures, 28 appear in the high-requency group. On the other hand, although there are 712 observed structures o the CVC type, only two, /nɯn/ and / mjʌn/, appear in the high-requency group. able 6.20 shows a list o the high-requency syllables that comprise more than 50 per cent o total distribution, as ound both in the dictionary and in speech. Tese 27 syllables were ound both in the 82 most requently used syllables in the dictionary as well as in the 36 most requently used syllables in speech. O the 27 syllables in able 6.20, 6.20 , /ʨi/ and /i/ ranked among the top ten both in the dictionary and in speech. Some o the syllables, such as /k ɯ/, /lʌ/, /ni/ occurred more requently in speech than in the dictionary, whereas others, such as /si/ and /s ɑ/, /to/ and /ʨɑ/ occurred more requently requently in the dictionary than in speech. Te distributions o /tε/, /i/, /l ɑ/ and /nε/ were similar both in the dictionary and in speech.
Frequency trends of Korean sounds
138
Table 6.18 Syllable frequency (dictionary)
Rank Syllable Frequency
%
Cumulative (%)
Rank Syllable Frequency
%
Cumulative (%)
1
hɑ
5,414
4.41
4.41
33
ʨo
635
0.52
31.24
2
li
1,808
1.47
5.88
34
toŋ
634
0.52
31.76
3
ʨi
1,642
1.34
7.22
35
596
0.49
32.25
4
i
1,629
1.33
8.55
36
588
0.48
32.73
5
ki
1,627
1.33
9.87
37
584
0.48
33.20
6
sɑ
1,389
1.13
11.00
38
581
0.47
33.67
7
1,367
1.11
12.12
39
579
0.47
34.15
1,209
0.98
13.10
40
563
0.46
24.60
9
ʨʌk tw ɛ ʨɑ
ʨ*ʌk pʰa sʌ nɑ sʌŋ tɑ
1,198
0.98
14.08
41
in
560
0.46
35.06
10
si
1,175
0.96
15.04
42
koŋ
550
0.45
35.51
11
su
1,154
0.94
15.98
43
mi
546
0.44
35.95
12
tɛ
1,095
0.89
16.87
44
nɛ
543
0.44
36.40
13
k ɑ
1,065
0.87
17.74
44
sɯ
543
0.44
36.84
14
ʨʌŋ
1,011
0.82
18.56
46
po
542
0.44
37.28
15
pu
1,006
0.82
19.38
47
ma
533
0.43
37.71
16
949
0.77
20.15
48
s in
519
0.42
38.14
17
ʨɛ ʰɑ k ʰɑ
946
0.77
20.92
49
sɛ
517
0.42
38.56
18
ko
904
0.74
21.66
50
ʌ
512
0.42
38.97
19
sɑŋ
881
0.72
22.38
51
mul
500
0.41
39.38
20
k ɛ
880
0.72
23.09
52
498
0.41
39.79
21
860
0.70
23.79
53
487
0.40
40.18
850
0.69
24.49
54
ni
481
0.39
40.58
23
ʨɑŋ hw ɑ ʨu
ɑ lɑ
847
0.68
25.18
55
mun
480
0.39
40.97
24
to
836
0.68
25.86
55
k ʰi
480
0.39
41.36
25
mu
827
0.67
26.53
57
sʌn
467
0.38
41.74
26
ʨʌn
815
0.66
27.19
58
mo
464
0.38
42.12
27
pi
784
0.64
27.83
59
ʨʰɛ
462
0.38
42.49
28
ku
781
0.64
28.47
60
il
458
0.37
42.87
29
so
750
0.61
29.08
61
w ʌn
450
0.37
43.23
30
ʨʰi
711
0.58
29.66
62
s*ʌŋ
449
0.37
43.60
31
ju
659
0.54
30.20
63
kjʌŋ
446
0.36
43.96
32
k ʌ
652
0.53
30.73
64
k ɯ
438
0.36
44.32
8
22
6.3 Syllable frequency
139
Table 6.18 (cont. ) )
Rank Syllable Frequency
%
Cumulative (%)
Rank Syllable Frequency
%
Cumulative (%)
65
mjʌŋ
438
0.36
44.68
74
lɛ
405
0.33
47.75
66
u
434
0.35
45.03
75
pul
399
0.33
48.08
67
sil
433
0.35
45.38
76
ʨʰɑ
395
0.32
48.40
68
jʌn
428
0.35
45.73
77
kjo
392
0.32
48.72
69
ʨʌ ʨuŋ pɑl pɑŋ lʌ
420
0.34
46.07
78
tʰɑ
388
0.32
49.03
419
0.34
46.41
79
kw ɑn
387
0.32
49.35
414
0.34
46.75
79
tɑn
387
0.32
49.66
414
0.34
47.09
81
sɑn
386
0.31
49.98
409
0.33
47.42
82
o
381
0.31
50.29
%
Cumulative (%)
70 71 71 73
Table 6.19 Syllable frequency (speech)
Rank Syllable Frequency
%
Cumulative (%)
Rank Syllable Frequency
1
k ɯ
13,755
3.41
3.41
19
ko
4,880
1.21
34.60
2
k ɑ
12,386
3.07
6.48
20
ku
4,632
1.15
35.75
3
i
10,920
2.71
9.18
21
s*ʌ
4,591
1.14
36.89
4
nɑ
9,609
2.38
11.56
22
lɑ
4,530
1.12
38.01
5
tɛ
9,362
2.32
13.88
23
lɛ
4,283
1.06
39.07
6
nɯn
8,606
2.13
16.02
24
jɑ
4,272
1.06
40.13
7
ni
8,143
2.02
18.03
25
mɑ
4,143
1.03
41.16
8
k ʌ
7,114
1.76
19.80
26
4,044
1.00
42.16
9
6,608
1.64
21.43
27
3,763
0.93
43.09
6,100
1.51
22.94
28
3,659
0.91
44.00
5,815
1.44
24.38
29
3,386
0.84
44.84
5,476
1.36
25.74
30
si
3,236
0.80
45.64
5,459
1.35
27.09
31
k *ɑ
3,229
0.80
46.44
5,279
1.31
28.40
32
hɛ
3,018
0.75
47.19
15
ɑ ʨi k ɛ hɑ lʌ sʌ tɑ
ʨɑ ɛ jʌ ʌ
5,146
1.28
29.68
33
2,909
0.72
47.91
16
ki
5,131
1.27
30.95
34
tɯ to
2,888
0.72
48.62
17
li
4,958
1.23
32.18
35
mjʌn
2,794
0.69
49.32
18
nɛ
4,914
1.22
33.39
36
sɑ
2,762
0.68
50.00
10 11 12 13 14
140
Frequency trends of Korean sounds
Table 6.20 High-frequency syllables (dictionary and speech)
6.3.2
Syllable
Ranking (DIC)
Ranking (SP)
Syllable
Ranking (DIC)
Ranking (SP)
hɑ
1
12
k ʌ
32
8
li
2
17
sʌ
37
14
ʨi
3
10
nɑ
38
4
i
4
3
tɑ
40
15
ki
5
16
nɛ
44
18
sɑ
6
36
mɑ
47
25
ʨɑ
9
26
50
29
si
10
30
52
9
tɛ
12
5
ʌ ɑ lɑ
53
22
k ɑ
13
2
ni
54
7
ko
18
19
k ɯ
64
1
k ɛ
20
11
lʌ
73
13
to
24
34
lɛ
74
23
ku
28
20
Syllable type frequency able 6.21 shows the distribution o syllable types. In the dictionary, o eight possible syllable structures, the CV type was the most requently used, comprising 42.8% o all occurring syllable types. Te GV type, on the other hand, was the least requently used, accounting or only 2.1% o all occurrences. CVC types comprise 60.6% o all observed syllable types, yet accounted or only 36.1% o the syllables in the dictionary. Tis contrasts with the CV type, which comprises 9.9% o possible syllable types, but accounts or 42.8% o all occurrences. In speech too, the CV type was the most requently used, comprising 59.1% o all syllable types. Unlike in the dictionary, the CGVC type was the least requently used, comprising only 2.3% o all occurrences. CVC types comprise 58.8% o all possible syllable types, yet were used in only 21% o occurrences in the dictionary. Tis again contrasts with the CV type, which comprises 10.4% o all possible syllable types, but is used 59.1% o the time. Te most signi�cant difference between the dictionary and speech was ound in the requency o CV and CVC types. Te CV type was used 16.3% more in speech than in the dictionary. On the other hand, the CVC type was
141
6.3 Syllable frequency
Table 6.21 Token and type frequency of syllable types (dictionary vs. speech; C consonant, G glide, V vowel) =
=
=
Token (%)
DIC
Type (%)
SP
Δ
DIC
SP
Δ
4.0
0.6
0.6
0.0
1.4
0.8
0.8
0.0
16.3
9.9
10.4
−
2.3
7.7
8.3
−
0.1
3.7
3.9
−
−
V
3.2
7.2
−
GV
2.1
3.4
−
CV
42.8
59.1
CGV
5.3
3.1
VC
2.9
3.0
GVC
2.3
0.9
1.4
2.6
3.3
CVC
36.1
21.0
15.1
60.6
58.8
1.9
5.4
2.3
3.1
14.2
14.0
0.2
100.0
100.0
0.0
100.0
100.0
0.0
CGVC otal
−
−
0.5 0.5 0.2 0.7
Table 6.22 Token and type frequency of syllable types with/without onsets
Token (%)
Type (%)
DIC
SP
Δ
DIC
SP
Δ
With onset
89.6
85.5
4.1
92.5
91.4
1.0
Without onset
10.4
14.5
4.1
7.6
8.6
−
1.0
−
ound to occur 15.1% more in the dictionary than in speech. Tis shows that the requency o the CVC type was relatively higher in the dictionary than in speech, whereas the requency o the CV type was relatively higher in speech than in the dictionary. No signi�cant difference was observed in terms o possible types o syllables between the dictionary and speech. In both, the requency o syllables including glides, i.e. GV, CGV, GVC, CGVC, was low both in terms o type and token requency. In the dictionary, the type and token requencies o syllables including a glide were 25.3% and 15.0% respectively. In speech, the type and token requencies o syllables including a glide were 26.4% and 9.7% respectively. able 6.22 shows the requencies o syllables with and without onsets. In both the dictionary and in speech, the majority o syllables (92.5% o syllables in the dictionary and 91.4% o those in speech) had onsets. As we can see in the distribution, syllables with onsets accounted or 89.6% in the dictionary and 85.5% o those in speech.
142
Frequency trends of Korean sounds
Table 6.23 Token and type frequency of syllables with/without codas
Token (%) DIC
SP
Without coda
53.4
72.8
With coda
46.6
27.2
Type (%) Δ
DIC
SP
19.4
19.0
20.0
19.4
81.0
80.0
−
Δ
1.0
−
1.0
able 6.23 shows the requency o syllables with and without codas. In both the dictionary and in speech, the majority o syllable types did not have codas – that is, they were open syllables – and this was mirrored in the actual distribution. Syllables without codas, – that is, open syllables – were observed ar more requently. Te type requency o open syllables was 19.0% in the dictionary and 20.0% in speech. On the other hand, the token requency o open syllables was 53.4% in the dictionary and 72.8% in speech. In terms o type requency, no signi�cant difference was observed between speech and the dictionary. However, in terms o token requency, open syllables occurred 19.4% more requently in speech than in the dictionary. 6.4 Comparison with phoneme frequency in English In this section, we will brie�y discuss phoneme requency in English in comparison with Korean. From language to language, it is not only the sound in ventory, but also the requency o occurrence o those sounds that differs. Tis section contains the requency o phonemes ound in Received Pronunciation (RP), based on the 70,646 entries ound in the Oxford Advanced Learner’s Dictionary by John Higgins. Te data is available at http://linguistlist.org/ issues/4/4–294.html#1. It also contains the requency o phonemes ound in speech, based on Crystal (1995)’s work which originates rom Fry (1947). 6.4.1
Consonant frequency in English Consider Figures 6.16 and 6.17. Tese �gures show consonant requency as observed in the dictionary (Figure 6.16) and in speech (Figure 6.17). In the dictionary, /t/ was the most requently used, ollowed by /s/, /n/, /l/, /r/, /k/, /d/, /z/. On the other hand, in speech, /n/ was most requently used, ollowed by /t/, /d/, /s/, l/, / ð /, r/. ables 6.24–6.26 show observed types o consonants and their actual realisation in the dictionary and in speech. As shown in able 6.24 , obstruents were
143
6.4 Comparison with phoneme frequency in English
15
t
s
) 10 % ( y c n e u q e r F 5
n l r
k
d
z p m b
ŋ
f
0
Figure 6.16
ʃ
v w d h
10.910.810.2 8.8 7.4 7.2 6.8 6.4 5.0 4.7 3.5 2.9 2.8 2.0 2.0 1.9 1.5 1.2
j
tʃ θ
ð
1.2 1.1 0.9 0.5 0.2 0.1
Consonant frequency of English (dictionary) 15
n t ) 10 % ( y c n e u q e r F 5
d
s l
ð
r
m k
w
z
v
b
f
p
h
ŋ
ʃ
d tʃ θ
0 12.510.6 8.5 7.9 6.0 5.9 5.8 5.3 5.1 4.6 4.0 3.3 3.2 2.9 2.9 2.4 1.9 1.7 1.6 1.4 1.0 0.7 0.6 0.2
Figure 6.17
Consonant frequency of English (speech)
more requently used than sonorants both in the dictionary and in speech. Among the obstruents, voiceless sounds were more requent than voiced sounds. Tis distinction is, however, less pronounced in speech than in the dictionary. able 6.25 shows that both in the dictionary and in speech, alveolar sounds are more requent than bilabials, and that bilabials are more requent than velars. Alveolar sounds, which account or 29.2% o the possible sound types, occurred most ofen, accounting or 61.3% o occurrences in the dictionary and 55.2% o occurrences in speech. able 6.26 shows that in terms o manner o articulation, both in the dictionary and in speech, stop sounds were most requent, accounting or 35.4%
144
Frequency trends of Korean sounds
Table 6.24 Type and token frequency of English consonants (sonorants vs. obstruents)
oken (%) ype (%)
DIC
SP
Sonorants
29.2
36.6
37.5
Obstruents
70.8
63.4
62.5
0.9
Voiceless
47.1
65.1
55.5
9.5
52.9
34.9
44.5
Voiced
Δ 0.9
−
9.5
−
Table 6.25 Type and token frequency of English consonants by place of articulation
oken (%) ype (%)
DIC
SP
Δ
Bilabial
12.5
13.2
11.5
Labial- velar
4.2
1.5
4.6
−
Labiodental
8.3
4.7
6.2
−
Dental
8.3
0.7
6.5
−
Alveolar
29.2
61.3
55.2
6.1
Palato-alveolar
16.7
4.2
3.4
0.8
Palatal
4.2
1.1
1.4
Velar
12.5
12.1
8.7
Glottal
4.2
1.2
2.4
1.7 3.2 1.5 5.8
0.3
−
3.4 1.2
−
Table 6.26 Type and token frequency of English consonants by manner of articulation
Token (%) Type (%)
DIC
SP
Stop
25.0
35.4
32.0
Fricative
37.5
25.9
28.8
Affricate
8.3
2.1
1.7
Nasal
12.5
17.9
19.7
Lateral
4.2
8.8
6.0
12.5
10.0
11.8
Approximant
Δ
3.4 2.9
−
0.4 1.8
−
2.7 1.9
−
6.4 Comparison with phoneme frequency in English
145
15
) % ( 10 y c n e u q e r F 5
0 t
Figure 6.18
s
n
l
r
k
d
z
p
m
b
ŋ
f
ʃ
v
w
d
h
j
tʃ
θ
ð
DIC 10.9 10.8 10.2 8.8 7.4 7.2 6.8 6.4 5.0 4.7 3.5 2.9 2.8 2.0 2.0 1.9 1.5
1.2 1.2 1.1 0.9 0.5 0.2 0.1
SP 1 0. 6 7 .9 1 2. 5 6.0 5. 8 5.1 8.5 4.0 2.9 5.3 3.2
1.0 2.4 1.4 0.7 0.6 5.9 0.2
1.9 2.9 1.7
1.6 3.3 4.6
Differences in English consonant frequency (dictionary vs. speech)
o occurrences in the dictionary and 32.0% in speech, ollowed by ricatives and nasals. Fricatives were ound only 25.9% o the time in the dictionary, but 28.8% o the time in speech. Nasals were ound 17.9% o the time in the dictionary, and 19.7% o the time in speech. Both in the dictionary and in speech, affricates occurred least requently, accounting or 2.1% o occurrences in the dictionary and 1.7% in speech. Finally, Figure 6.18 shows a comparison between consonant requency in the dictionary and in speech. Te boxed areas show cases where the difference between the dictionary and speech is greater than 5% (solid box) or between 3% and 5% (dashed box). We can see that /ð/ and /w/ show a difference greater than 3%. It is noticeable that /ð/, which is used only 0.2% o the time in the dictionary, is ound 5.9% o the time in speech. Along with /ð/, /w/ is also used only 1.5% o the time in the dictionary, but is observed 4.6% o the time in speech. 6.4.2
Vowel frequency in English Figures 6.19 and 6.20 show vowel requency in English as observed both in the dictionary and in speech. Te most requently used vowel in the dictionary was /ɪ/, ollowed by /ǝ/, /æ/, /e/, /eɪ/. On the other hand, / ǝ/ was used most requently in speech, ollowed by / ɪ/, /e/, /aɪ/, /ʌ/. Both in the dictionary and in speech, the two most requently used vowels accounted or the majority o the total vowel distribution (46.0% in the dictionary and 48.6% in speech).
146
Frequency trends of Korean sounds
30
25
) 20 % ( y c n 15 e u q e r F 10
ə
e
e
5
ɒ
a
əυ
u
ɔ
ɑ
ə
aυ
1.2 1 .1
υ υə eə
o
0 28.8 17.2 6.4
Figure 6.19
6.3
5.7
4.4
4.1
4.0
3.7
3.7
2.7
2.6
ɒ
ɔ
u
2.4
2.3
1.7
υ
ɑ
aυ
0.6
0.5
0.4
Vowel frequency in English (dictionary)
30
ə 25
) 20 % ( y c n 15 e u q e r F 10
5
e a
e
i
əυ
0
eə ə
o υə
27 .4 2 1. 2 7. 6 4. 7 4. 5 4. 4 4. 2 3 .9 3 .7 3 .5 3 .2 2. 9 2. 2 2. 0 1. 6 1. 3 0 .9 0 .5 0 .4 0 .2
Figure 6.20
Vowel frequency in English (speech)
ables 6.27–6.30 show the classi�cation o vowels. Te results shown in able 6.27 reveal that monophthongs are much more requently used than diphthongs. able 6.28 shows that mid and high vowels appeared more requently than low vowels. able 6.29 shows that ront vowels appeared more requently than back vowels. able 6.30 shows that unrounded vowels appeared more requently than rounded vowels. Figure 6.21 shows a comparison between vowel requency in the dictionary and in speech. Te boxed areas show cases where the difference between
147
6.4 Comparison with phoneme frequency in English
Table 6.27 Type and token frequency of English vowels (monophthongs vs. diphthongs)
Token (%) Type (%)
DIC
SP
Monophthong
60.0
81.4
83.7
Diphthong
40.0
18.6
16.3
Δ
2.3
−
2.3
Table 6.28 Type and token frequency of English vowels by vowel height
Token (%) Type (%)
DIC
SP
High
33.3
44.6
36.5
Mid
41.7
39.1
52.5
Low
25.0
16.3
11.0
Δ
8.1 13.4
−
5.3
Table 6.29 Type and token frequency of English vowels by backness
Token (%) Type (%)
DIC
SP
Front
33.3
55.6
43.9
Central
25.0
28.2
39.7
Back
41.7
16.2
16.4
Δ
11.7 11.5
−
0.2
−
Table 6.30 Type and token frequency of English vowels by roundedness
Token (%) Type (%)
DIC
SP
Rounded
33.3
13.3
14.0
Unrounded
66.7
86.7
86.0
Δ
0.7
−
0.7
148
Frequency trends of Korean sounds
30
25
) 20 % ( y c n 15 e u q e r F 10
5
0
Figure 6.21
ə
e
e
ɒ
a
i
əυ
u
ɔ
ɑ
ə
aυ
υ
DIC 28.8 17.2
6.4
6.3
5.7
4.4
4.1
4.0
3.7
3.7
2.7
2.6
2.4
2.3
1.7
1.2 1.1
SP
3.7
7.6
4.4
3.5
4.7
4.5
4.2
3.9
2.9
3.2
2.0
0.5
1.3
1.6 2.2
21.2 27.4
υə
eə
o
0.6
0.5
0.4
0.2
0.9
0.4
Vowel frequency in English (dictionary vs. speech)
the dictionary and speech is greater than 5%. /ɪ/ and /ǝ/ show a difference o greater than 5% between dictionary and speech usage. In the dictionary, / ɪ/ is the most requently used vowel, comprising 28.8% o the total distribution, whereas in speech, /ǝ/ is the most requently used vowel, comprising 27.4% o the total distribution. o sum up, we can observe the ollowing characteristics in our comparison o phoneme requency in English and Korean. First, let’s look at the requency o /t/, /k/, /s/. In Korean, /t/ occurred 5.2% o the time and ranked tenth in the dictionary, and in speech, it occurred 6.7% o the time and ranked �fh. However, in English, /t/ was ound to be the most requently used consonant in the dictionary (10.9%) and the second most requently used consonant in speech (10.6%). In Korean, on the other hand, /k/ was ound to be the most requently used consonant in the dictionary (13.3%) and the second most requently used consonant in speech (16.0%). In English, /k/ occurred 7.2% o the time and ranked sixth in the dictionary, and in speech, it occurred 5.1% o the time and ranked ninth. Te distribution o /s/ also differed. In Korean, /s/ occurred 7.7% o the time and ranked sixth in the dictionary, and in speech, it occurred 5.0% o the time and ranked eighth. However, in English, /s/ was ranked second in the dictionary (10.8%) and ranked ourth in speech (10.6%). Te results are summarised in able 6.31. In other words, in Korean velars occur more requently than alveolar sounds. In English, all o the �ve most requently used consonants are alveolars, whereas in Korean only two o the most requently used consonants in
149
6.4 Comparison with phoneme frequency in English
Table 6.31 Comparison of phoneme frequency in Korean and English
Korean (rank/%)
English (rank/%)
Dictionary
Speech
Dictionary
Speech
/t/
10th / 5.2%
5th / 6.7%
1st / 10.9%
2nd / 10.6%
/k/
1st / 13.3%
2nd / 16.0%
6th / 7.2%
9th / 5.1%
/s/
6th / 7.7%
8th / 5.0%
2nd / 10.8%
4th / 10.6%
Table 6.32 Comparison of consonant frequency in Korean and English
Korean
English
Dictionary
Speech
Dictionary
Speech
Fricatives
16.0%
11.0%
25.9%
28.8%
Affricates
12.8%
9.0%
2.1%
1.7%
Nasals
28.6%
33.6%
17.9%
19.7%
the dictionary and three in speech are alveolars. On the other hand, in English no velar sounds were requently used, whereas in Korean two o the �ve most requently used consonants in the dictionary were velars, and one in speech. In terms o manner o articulation, the distribution o ricatives, affricates and nasals is signi�cantly different in Korean and English. In short, ricatives were less requent in Korean than in English, while nasals and affricates were more requent in Korean than in English. able 6.32 summarises these results. THINK MORE! How would English sound to Koreans? And how would Korean sound to you? Although these words are not officially listed in the dictionary, how English sounds to Korean people who do not speak English is represented as (or ; although is more commonly ound in writing, it is more commonly pronounced as ) or (or ). Why do Korean people think that English sounds like , , , ? Might these onomatopoeic expressions be related to the sounds characteristic o the language in question, or its high-requency phonemes? Tis hypothesis is not totally ungrounded i we pay attention to the alveolar, the alveolar ricative, or the lateral sounds that are
Frequency trends of Korean sounds
150
characteristically ound requently in English pronunciation. And, conversely, what would people who heard Korean, but did not understand it, think that Korean sounded like?
6.5 Summary In this chapter, we have examined the requency o phonemes and syllables in Korean, based on dictionary and spontaneous speech data. Te result shows that overall, /ɑ/ was used most requently and / ɰi/ used least requently. Among consonants, /k/ and /n/ were most requently used. Overall requency o obstruents is higher than sonorants. In terms o place o articulation, alveolar sounds were most requent and in terms o manner o articulation, stops were most requent. O the vowels, monophthongs were much more requent than diphthongs. As or syllable requency, CV types were more requently ound than CVC types. Comparisons with English were also discussed. In terms o place o articulation, velars were requently observed in Korean but not in English. On the other hand, alveolars were requently observed in English than in Korean. In terms o manner o articulation, ricatives were requent in English, but not in Korean. Both affricates and nasals were more requent in Korean than in English. EXERCISES 1
What are the most and least frequently used consonants in Korean, both in the dictionary and in speech?
2
What are the most and least frequently used vowels in Korean, both in the dictionary and in speech?
3
Is there any difference, in terms of consonant distribution, between word-initial and word-final positions?
4
Which syllable structures are frequently used in Korean?
5
Compare phoneme frequency in Korean and English. What is the significant difference?
6
How does Korean sound to you? State which sounds are most frequently heard by you when you listen to the Korean language.
7
Prosody
So ar, we have observed Korean sounds at the segmental level. In particular, we have considered the phonetic characteristics o each sound. However, the same sound can have different phonetic realisations; or instance, / ɑ/ can be produced either with high pitch or low pitch, as a long vowel or a short vowel, and sometimes loudly or quietly. Tis is not only the case or individual sounds but also or sequences o sounds or segments. Hence, the same sequence o sounds (or segments) may be realised with a different pitch, loudness or length, and these are known as ‘supra-segmental eature s’ or ‘prosodic eatures’. Te term ‘supra-segmental eatures’ emphasises the sound ‘unit’ in which those eatures appear. By contrast, the term ‘prosodic eatures’ draws emphasis to the sound properties that are maniested within the sequence o segments. We will use the term ‘prosodic eatures’ throughout this chapter, since we are more interested in the nature o sound properties than the sound units bearing these properties. In this chapter, we will discuss the prosody o Korean: in 7.1, we will examine the linguistic unction o prosody; 7.2 will provide an overview o the prosodic structure o Korean; 7.3–7.6 look at each o the linguistic units which comprise the prosodic structure o Korean rom the syllable, the smallest u nit, to the phonological word, phonological phrase, and �nally the intonational phrase; and in 7.7, we conclude.
7.1
Linguistic function of prosody
Prosody includes such variables as pitch, loudness, length and pause. O course, pitch, loudness, length and pause all have their corresponding physical reality, such as undamental requency or pitch, intensity or loudness, duration or length and silence duration or pause; or instance, high-pitch sounds 151
152
Prosody
have a higher undamental requency than lower-pitch sounds. Likewise, loud sounds will be more intense than quiet sounds; long sounds will have a longer duration than short sounds; pauses naturally grow longer as silence duration becomes longer. However, native speakers are not typically conscious o properties such as pitch, loudness, length and pause, or they are naturally attuned to the music o their language. Across all languages, even when there might be no difference in sound at a segmental level, different meanings can be expressed through variation in prosody: or instance, Mandarin Chinese is a tonal language, and so even when a group o words all have exactly the same sound properties at a segmental level, the differences in pitch yield words with different meanings. An example o this phenomenon is presented below. (1) a. b. c. d.
[ma] with high level tone: ‘mother’ [ma] with high rising tone: ‘hemp’ [ma] with low alling-rising tone : ‘a horse’ [ma] with high alling tone: ‘to scold’
Let’s consider the case o Korean. (2) a. [k ɑ] with rising pitch b. [k ɑ] with alling pitch o Korean speakers, (2a) and (2b) sound different. Tough the sound sequence is exactly the same, when the sequence /kɑ/ is pronounced at the end o a sentence in a rising tone, it turns the sentence into a question, whereas the same sequence in a alling tone makes the sentence declarative or imperative. Does this mean that Korean is a tonal language like Chinese? Te answer is no. In the case o (2), the difference in pitch marks a difference in mood; that is, a rising pitch makes the sentence interrogative, whereas a alling pitch makes the sentence declarative or imperative. In terms o word meaning, however, /k a/ in (2) means the same thing, ‘to go’. In other words, in the case o Korean, unlike languages such as Chinese, variation in pitch contributes to the meaning o a sentence, but not to the meaning o a word .1 Now then, what about the duration o sound? According to SKP, the meanings o individual words can change according to the vowel length. Listed in (3) are minimal pairs which demonstrate this: or instance, /nu:n/ meaning ‘snow’ is pronounced with a longer vowel than /nun/ meaning ‘human eyes’, and /pɑ:m/ meaning ‘chestnut’ is pronounced with a longer vowel than /pɑm/ meaning ‘night’. Likewise, /mɑ:l/ meaning ‘language’ is pronounced with a longer vowel than /m ɑl/ meaning ‘horse’.
153
7.1
Linguistic function of prosody
(3) a. b. c. d.
/nu:n/ ‘snow’ /pɑ:m/ ‘chestnut’ /mɑ:l/ ‘language’ /pɑ:l/ ‘blind’
/nun/ ‘eye’ /pɑm/ ‘night’ /mɑl/ ‘horse’ /pɑl/ ‘oot’
However, in everyday language usage, duration o sound alone doesn’t play an important role in distinguishing the meaning o one word rom another. It is indeed hard to �nd any word that has relatively long vowels in contemporary spoken Korean. Neither does loudness or pause play a role in distinguishing between words. DID YOU KNOW . . . ?
Is length distinctive in Seoul Korean?
SKP assumes length as a distinctive eature in Korean. However, this is ar rom what is observed in ‘real’ contemporary spoken Korean. S.-N. Lee (1960) argued that rom his own observation (he is a native Seoul Korean speaker and a renowned Korean phonologist), length is becoming less and less distinctive and indeed has almost disappeared in his time. He also claimed that a view o length as a distinctive eature is not based on real spoken data but purely based on dictionary knowledge. About hal a century has passed since S.-N. Lee (1960), but normative pronunciation still assumes that length is a distinctive eature in Seoul Korean. In short, in the case o Korean, prosody doesn’t distinguish one word rom another. What, then, is the linguistic unction o prosody in Korean? o understand this, one must consider linguistic units higher than a word. Consider (4). (4) [sɑ.nɛ. k ɑ.jo] (. = syllable boundary) As we saw in (2), prosody determines the mood o a sentence. It is diffi cult to recognise the meaning o sentence (4) without an awareness o the prosody o the �nal syllables. For instance, a alling tone would make the sentence either declarative or imperative, whilst a rising tone would make it a yes/no question. In this way, prosody plays a crucial role in realising grammatical and pragmatic inormation. Listen careully to the utterances in (5). (5) a. [ɑ. pʌ. ʨi. k ɑ. pɑŋ. ɛ. tɯ. lʌ. k ɑ. sin. tɑ.] b. [ɑ. pʌ. ʨi. k ɑ. pɑŋ. ɛ. tɯ. lʌ. k ɑ. sin. tɑ.] c. [ɑ. pʌ. ʨi. k ɑ. pɑŋ. ɛ. tɯ. lʌ. k ɑ. sin. tɑ.] (. = syllable boundary)
154
Prosody
(5a) (
= syllable)
(5b)
(5c)
Figure 7.1
Stylised pitch movements in (5a–c)
(6) a. ()()() [ɑ. pʌ. ʨi. k ɑ | pɑŋ. ɛ| tɯ. lʌ. k ɑ. sin. tɑ.] b. ()()() [ɑ. pʌ. ʨi| k ɑ. pɑŋ. ɛ| tɯ. lʌ. k ɑ. sin. tɑ.] c. ()() [ɑ. pʌ. ʨi. k ɑ. pɑŋ. ɛ| tɯ. lʌ. k ɑ. sin. tɑ.] (. = syllable boundary, | = minor break boundary) Although all the examples in (5) consist o the same syllables, in the same order (and thereore o the same segments), the meaning o each sentence is different according to the way in which the syllables are grouped together. When pronounced, (6a) means ‘(my) ather goes into a room’, but (6b) means ‘(my) ather goes into a bag’, and (6c) means ‘(somebody who is older than me) goes into (my) ather’s bag’. Te different interpretations result rom different phrasing. We will come back to this issue in 7.5. Beore going urther, let’s look closely at the prosodic differences between the three utterances in (5). As noted, although the three utterances have the same sequences o segments, prosodically, they are all different. Te most prominent difference is observed in pitch pattern, as we can see in Figure 7.1. Consider the boxed areas in Figure 7.1. In (5a), the �rst syllable has a low pitch, the second has a high pitch, the third is a little lower again, and then the ourth syllable has a high pitch. Afer that, the pitch drops radically at the �fh syllable and in the sixth syllable, the pitch rises again, as in the ourth syllable. On the other hand, in (5b), the �rst syllable has a low pitch as in (5a), but the second syllable does not have a high pitch, unlike (5a). However, the pitch rises suddenly in the third syllable and then alls down again in the ourth
155
7.2
Prosodic structure of Korean
syllable. Te pitch rises a little in the �fh syllable, similar to the second syllable. Te pitch rises urther in the sixth syllable until it is as high as the third syllable. Finally in (5c), the �rst syllable has a low pitch, rising to the second syllable and then descending in the third, ourth and �fh syllables, beore rising again suddenly in the sixth syllable. Different pitch movements cause native speakers to group the same segments differently and subsequently cause a difference in meaning. In (5a) in Figure 7.1, different pitch patterns cause the �rst our syllables to be grouped together and the other two syllables to be grouped separately. Likewise, in (5b) in Figure 7.1, pitch patterns cause the �rst three syllables to be grouped as one and the ollowing three syllables to be grouped together. In (5c) in Figure 7.1 , the pitch pattern causes all six syllables to orm one group. While difference in pitch movement is one o the easiest things to notice, there are o course other prosodic differences to be ound in different utterances. 7.2
Prosodic structure of Korean
Listen careully to the ollowing sentence, spoken at a normal speed. (7) . (mijʌŋ + -i + -nɯn) (pus + -ɯl) (sɑ- + -nɯntɛ), (nɑ jʌŋ + -i + -nɯn) (mʌk + -ɯl) (sɑ- + -ntɑ). (+ = morpheme boundary, ( ) = eojeol unit) Te sentence in (7) has six word phrases (i.e. eojeols). Listen again and this time, try to �gure out where the longest and shortest pauses are located. A native speaker o Korean would be able to tell that the longest pause is between the third and ourth word phrases, and that there are breaks between the second and third word phrases, as well as between the �fh and sixth word phrases, as demonstrated in (8). In (8), one bar (|) reers to a shorter pause, compared to two bars ( ∥) denoting a longer pause. (8) | | [(mijʌŋ + -i + -nɯn)| (pus + -ɯl) (sɑ- + -nɯntɛ)∥ (nɑ jʌŋ + -i + -nɯn)| (mʌk + -ɯl) (sɑ- + -ntɑ)] (+ = morpeme boundary, ( ) = eojeol unit) Te bracketing in (8) shows how each segment orms a prosodic or phonological unit. Tese prosodic units then build up a larger prosodic/phonological structure. Te way in which the actual prosodic structure is built up
Prosody
156
Utterance Intonational phrase Phonological phrase
Phonological word
Syllable
Figure 7.2
(Intonational phrase) ...
(Phonological phrase) ...
(Phonological word) ...
(Syllable) ...
Prosodic structure of Korean
differs rom language to language. Figure 7.2 shows the prosodic structure o Korean. As shown in Figure 7.2, the smallest prosodic unit is a syllable and the largest possible unit is an utterance. In Korean, an utterance contains at least one intonational phrase (IP), and an intonational phrase contains at least one phonological phrase (PP).2 A phonological phrase contains at least one phonological word (PW), and a phonological word contains at least one syllable. In the ollowing sections, we will discuss each prosodic structure in more detail. 7.3
Syllable
A ‘syllable’ is the smallest unit within the prosodic structure o Korean.3 A syllable is ormed o phonemes, and syllables can be made up o different numbers o phonemes. For instance, the word /u ʨusʌn/ ‘spaceship’ has six segments: /u/, /ʨ/, /u/, /s/, /ʌ/, /n/. Tese six segments can be grouped into three syllables as in [u], [ ʨu], [sʌn]. However, each syllable has a different number o segments: the �rst syllable [u] has one segment, the second syllable [ʨu] has two segments, and the third syllable [s ʌn] has three segments. Any native Korean speaker will agree with this way o grouping syllables. A syllable is an abstract, psychological unit, not a concrete, physical unit, and it differs rom language to language. Hence, even when the same string o sounds is heard, native speakers o other languages will group the segments differently. For example, the English word strike is recognised as a one-syllable word by native English speakers, but as a �ve-syllable word by native Korean speakers.
157
7.3
Syllable
DID YOU KNOW . . . ?
How many syllables does milk have?
Syllable structures differ rom language to language. Hence, even the same sequence o segments will be grouped differently. Let’s try an experiment. Ask an English native speaker to pronounce milk. And then, ask English, Korean and Japanese speakers to clap their hands according to the number o syllables milk has. You will see that English speakers clap once, Koreans twice and Japanese three times. Tis is due to the different syllable structures o the three languages. In English, milk is syllabi�ed as one syllable because consonant clusters such as /lk/ are allowed in word-�nal position. On the other hand, in Korean, consonant clusters are not allowed in any position and thereore, /lk/ is pronounced as [lk ɯ], with an additional deault vowel /ɯ/ afer /k/. Tis makes Koreans regard milk as a two-syllable word. On the other hand, Japanese, which is a CV language, cannot have any consonant at all in word-�nal position, except or nasals. Tereore, /l/ and /k/ can only occur in word-initial position with the added deault vowel /u/. Hence, the English /milk/ is recognised as a three-syllable word, /miruku/. We will come back to this in Chapter 10.
7.3.1
Syllable structure A syllable consists o a syllable nucleus and a syllable margin. Just as a cell must have a cell nucleus, each syllable must have a syllable nucleus. Tis is the same in any language, but the inventory o the possible sounds that can act as a syllable nucleus or a syllable margin differs rom language to language, as does syllable structure. In the ollowing section, we will brie�y discuss the syllable nucleus. Ofen, it is simply called ‘nucleus’, and orms the core part o a syllable. Tese syllable nuclei can only be ormed o syllabic sounds. Tese are sounds that can orm a syllable on their own without a syllable margin. In the case o Korean, only vowels are syllabic, but in other languages there are syllabic consonants; in English, or instance, sonorant consonants such as /n/ and /l/ are syllabic and can orm a separate syllable as in button [bʌ-tn ̩] and jungle [d ʒʌŋ-gl]. Button and jungle are thereore considered two-syllable (disyllabic) words in English. A glide can be attached to the nucleus either beore or afer the vowel. In contemporary Korean, as we discussed in Chapter 5, a glide can only precede
158
Prosody
Syllable (σ)
Onset
(C)
Nucleus
(G)
Coda
V
(C)
V = vowel; C = consonant; G = glide; ( ) = optional
Figure 7.3
Syllable structure of Korean
the vowel in a syllable nucleus. In short, a syllable nucleus in Korean consists o an obligatory vowel and optional glide preceding the vowel. Unlike the syllable nucleus, a syllable margin is optional and can precede or ollow the syllable nucleus. Syllable margins are ormed o consonants. A consonant that precedes a nucleus is an onset and one that ollows is a coda. Which particular sounds can be onset or coda, and how many o them are allowed to orm one syllable, differs across languages. In contemporary Korean, only one consonant can appear beore or afer the syllable nucleus, meaning consonant clusters are not possible. Figure 7.3 summarises the syllable structure o Korean. As mentioned earlier, every language has a unique syllable structure. Tis is why English can have one-syllable (monosyllabic) words containing as many as three consonants in the onset position, but a maximum o our consonants in the coda position, such as strike [stɹɑɪk] and texts [tɛksts]. In the case o Japanese, only one consonant is allowed in the onset position and only nasal sounds at the coda. FURTHER STUDY
Cross-linguistic differences in syllabic structure
A language can be classi�ed as either a CV or CVC language, depending on whether or not its syllable structure allows consonants in the coda position. For example, Japanese, Italian and Spanish are CV languages, while English, German and Mandarin Chinese are CVC languages. Korean is a CVC language because we can easily �nd Korean words with a CVC syllable structure, such as // or //. Due to different syllable structures, the same loanwords may thereore be adopted differently. For instance, the English words tile and helmet have been adopted in Japanese as tairu and
159
7.3
Syllable
herumetto , but as thail and helmet in Korean. We will get back to this issue in Chapter 10. A different way of pronouncing coda sounds
In languages like French, coda sounds must be released afer central closure. Hence, the third person, eminine pronoun elle should be pronounced not as [ ɛl̚], but as [ɛl]. Tis is in contrast to Korean, where coda sounds must never be released afer central closure. On the other hand, this tendency does not exist in English. Tereore, English native speakers will recognise the pronunciations [kæt ̚] with closure, or [kæt] with release o the closure, as the same one-syllable word, cat . What would the syllable structure of fifteenth-century Korean have been like?
As a language changes, so too does its syllable structure. One complete character in Hangeul orms one syllable. Hence, as noted above, it is easy to count syllables in Korean. In act, when it was �rst invented, Hangeul orthography was more aithul to actual pronunciation than to morphological inormation. Tereore, it is easy to reconstruct the syllabic structure o �feenth-century Korean by looking at the documents in this period. Fifeenth-century Korean displays a different syllable structure rom that o contemporary Korean. Most signi�cantly, glides can not only precede, but also ollow vowels. In �feenth-century Korean and were regarded as a combination o /ɑ/ plus /j/ and /ʌ/ plus /j/, that is, /j/ ollowing simple vowels such as and . and can be then understood as the combination o a glide attached beore and afer the simple vowel. In addition, consonant clusters existed in both syllable-initial (onset) and syllable-�nal (coda) positions. Fifeenth-century documents allow consonant clusters o up to three consonants. For example, ound in Hunminjeongeum Haerye ‘Explanation and Examples o Hunminjeongeum’, has three consonants at the coda o the �rst syllable and three at the onset o the second syllable. Some scholars argue that the words like ‘chicken’ are clear evidence o the existence o an /lk/ sequence in �feenth-century Korean. Moreover, in �feenth-century Korean, ‘millet’ combines with ‘rice’ to produce ‘millet’. Te afer is the result o being at that time. Similar words are listed below. ‘newly-harvested rice’ ‘glutinous rice’
‘unglutinous rice’ ‘nonglutinous rice’
160
7.3.2
Prosody
Syllable types In this section, we will discuss the various syllable types possible in Korean. Based on the syllable structure o Korean given in Figure 7.3, we can think o about eight possible syllable types in Korean, as shown in (9). Here, C reers to a consonant, V reers to a vowel and G reers to a glide: (9) ype 1: A syllable with one vowel: V, e.g. /i/ ‘tooth’ ype 2: A syllable with a glide and a vowel: GV, e.g. /jo/ ‘underquilt’ ype 3: A syllable with a consonant and a vowel: CV, e.g. /no/ ‘oar’ ype 4: A syllable with a consonant, a glide and a vowel: CGV, e.g. /mw ʌ/ ‘what’ ype 5: A syllable with a vowel and a consonant: VC, e.g. /ok/ ‘jade’ ype 6: A syllable with a glide, a vowel and a consonant: GVC, e.g. /jok/ ‘abuse’ ype 7: A syllable with a consonant, a vowel and a consonant: CVC, e.g. /nok/ ‘rust’ ype 8: A syllable with a consonant, a glide, a vowel and a consonant: CGVC, e.g. /pjʌk/ ‘wall’ Tis con�rms that: (i) all syllable types have a vowel as a syllable nucleus; (ii) glides can only precede the vowel as in types 2, 4, 6 and 8 in (9); (iii) only one consonant can precede or ollow the vowel as in types 3, 4, 5, 6, 7 and 8 in (9).
FURTHER STUDY!
Body vs. rhyme structure
Te syllable structure given in Figure 7.3 is a �at structure, but two more syllable structures are available in which the onset and nucleus or nucleus and coda orm a sub-structure �rst. (a)
(b)
σ
B
R O σ
N
σ
C
O
N
C
= syllable; R = rhyme; B = body; O = onset; N = nucleus; C = coda
161
7.3
Syllable
Te ormer is called a rhyme structure, and languages such as English ollow this pattern o syllable structure ormation. Te latter is called a body structure. It is not easy to decide whether Korean has a body or rhyme structure. Te ollowing examples suggest that Korean has a body structure rather than a rhyme structure. a. b. c.
- /t*ɑk-t*ɑtɑk/, - /p*ɑŋ-p*ɑpɑŋ/, - /k ʰuŋ-k ʰukuŋ/, - /p*oŋ-p*opoŋ/, - /p*uŋ-p*upuŋ/ - /pʰɑtɑk-pʰɑtɑtɑk/, - /pʰutɑk-pʰutɑtɑk/, - /hutɑk-hutɑtɑk/ - /tu-tuŋsil/
On the other hand, the ollowing examples suggest that Korean has a rhyme structure. d.
- /ulthuŋ-pulthuŋ/, - /ollok-pollok/, - /ulluk-pulluk/
Recent studies in psycholinguistics suggest that Korean has a body structure, rather than a rhyme structure, but more research is needed to con�rm this (see Yoon and Derwing, 2001 or more discussion).
7.3.3
Building a syllable structure In this section, we will discuss how segments are built up into a syllable in Korean. Te most important principle in this is the onset-�rst principle . Tis principle states that as many o the margins (consonants) as possible should be included as onsets preceding the nucleus, and then any margins that remain should be included as the coda, afer the nucleus. Figure 7.4 shows how the syllable structures in /hɑlɑpʌʨi/ ‘grandather’, /h ɑlmʌni/ ‘grandmother’ and /ʨoŋihɑk/ ‘paper crane’ are built up. As shown in stage (a), the �rst thing to do is to �nd the syllable nucleus and build a skeleton. Ten aferwards, as in stage (b), as many o the consonants as possible should be incorporated into that nucleus ollowing the onset-�rst principle. Finally, as in stage (c), any remaining segments need to be linked to the coda.
7.3.4
Syllabification and phonological processes In Korean, those consonants that can appear in the onset position are different rom those that appear in the coda position. All consonants except /ŋ/, and
162
Prosody
(a) First step: find a nucleus and build the bone structure of a syllable h ɑ l ɑ p t i
N
N
σ
σ
N
σ
hɑ l m n i
N
N
N
N
σ
σ
σ
σ
t o ŋ i h ɑ k
N
σ
N
N
σ
σ
(b) Second step: link the onset by the onset-first principle h ɑ l ɑ p t i
h ɑ l m n i
t o ŋ i h ɑ k
O N O N O N O N
ON
ON
σ
σ
σ
σ
O N ON
σ
σ
σ
σ
N ON
σ
σ
(c) Third step: link the coda with the residual segment h ɑ l ɑ p t i
h ɑ l m n i
ONONON ON
ONC ONON
σ
σ
σ
σ
σ
σ
σ
t o ŋ i
h ɑ k
ONCN ONC
σ
σ
σ
––––––– = nucleus linking; ............= onset linking; –– . –– = coda linking Figure 7.4
Syllable structures for /hɑlɑpʌʨi/, /hɑlmʌni/ and /ʨoŋihɑk/
thereore eighteen out o nineteen consonants, can occur in the onset position. Tese sounds are pronounced with no change in their basic pronunciation. However, only seven o the nineteen consonants (i.e. /k, n, t, l, m, p, ŋ/) can appear in the coda position. All other sounds are thereore substituted by one o these seven sounds; in other words, the same consonant is pronounced dierently depending on the position in which it occurs within a syllable structure. Why is this the case? It is because all coda sounds in Korean must be a ‘stop’ sound involving the closure o the mouth, and cannot simply be released, as we discussed in
163
7.3
Syllable
*
(a) /k ot h / *
k
o
th
O N C
*
(b) /k ot h + −i/ *
k
O N O N
σ
Figure 7.5
σ
σ
Syllabification of /k*oʨʰ/ and /k*oʨʰ + -i/
(a) /k ɑ ps/ k ɑ p s
O N C
σ
Figure 7.6
o t h i
(b) /k ɑ ps + i/ k ɑ
p s
i
O N C O N
σ
σ
Syllabification of /kaps/ and /kaps + -i/
4.2.1.3. Due to this restriction, those sounds which end with a release o air instead o ‘closure’, are changed into sounds produced with complete closure. Hence, ricatives and affricatives are changed into the ‘stop’ sounds that are produced in the same manner o articulation. We will come back to this issue in Chapter 8. Put simply, all syllables in Korean should end with the central closure o the oral cavity. I these coda sounds were ricative, they would be regarded as the onset o the next syllable instead o the coda o the current syllable. For instance, / ʨh/ is pronounced differently in /k *oʨh/ and /k*oʨh + -i/ + : /ʨh/ is pronounced as [t] in /k*oʨh/, but as [ʨh] in /k*oʨh + -i/. Te reason behind this is the difference in syllabi�cation. As shown in Figure 7.5, /ʨh/ is in the coda position in /k*oʨh/ , whereas in /k*oʨh + -i/ + , it is in the onset position o the second syllable. Another phonological process related to syllable structure is consonantcluster simpli�cation. In Korean, although consonant clusters can be expressed orthographically, only one o these sounds is pronounced because only one consonant can appear in the coda position. Tis is why in /k ɑps/ only /p/ is pronounced, so it becomes [k ɑp] as in Figure 7.6a. However, in /k ɑps + -i/ + where a subject particle is attached, there are two syllable nuclei, so the second consonant /s/ can also be pronounced, as it becomes the onset o the
164
Prosody
second syllable, as shown in Figure 7.6b. Further discussions on consonantcluster simpli�cation will be given in Chapter 8. 7.4
Phonological words
Syllables are grouped together to orm a phonological word, which is different rom a morphological word . In morphology, a word is de�ned as the smallest independent, indivisible unit that contains no pause. In most cases, a morphological word and a phonological word are the same. However, we also �nd examples o morphological words that don’t coincide with phonological words. Consider (10): the syllables bracketed together are recognised as orming one phonological word within an utterance, yet this same unit can consist o more than one morphological word. (10) a. () ()ω. (k ɑl) (su)ω (ʌpsʌs*tɑ)ω ‘(I) could not go there’ () ()ω. (mʌk- + -ɯl) (k ʌs)ω (ʨhɑʨ- + -nɯn + -tɑ)ω ‘I am searching or something to eat’ () ()ω. (k ɯlʌ- + -l) (li)ω (ʌps- + ʌ)ω ‘I doubt it’ b. ()ω? (k ɯ) (sɑlɑm)ω (ʌtis*- + -ni)ω * ← () () (ʌti)ω (is - ± -ni)ω ‘where is he?’ , ()ω. (ʨɑ)ω (jʌkis*- + -tɑ)ω * ← () () (jʌki)ω (is - + -tɑ)ω ‘here it is’ c. (ωω) (pʌm-ω# minʨok ʨʌk ω) ‘cross-national’ (ωω) (ʨʰo-ω#hohw ɑpʰɑnω) ‘extra-luxurious’ (ωω) (pɑn-ω# minʨuʨuɰiω) ‘anti-democracy’ (ωω) (mɛŋ-ω# hw ɑljɑksɑŋω) ‘playing extremely well’ (ωω) (jʌk-ω# ʨʰɑpjʌlʨuɰiω) ‘counter-differentia’ (+ = morpheme boundary, ( ) = eojeol unit, ω = phonological word boundary # = word boundary)
165
7.5
Phonological phrases
For instance, in (10a), when the �rst two words are pronounced, they are always ound as one whole unit without any pause in between. Although morphologically each sound unit orms two individual words, phonologically they orm one word. Shown in (10b) are other cases where two morphological words have been contracted into one phonological word. In (10c), however, the words listed are morphologically recognised as one word, but phonologically regarded as two. Tis is because these words are ormed o a pre�x and a stem. Since there is a strong pause between the pre�x and stem, it is natural to regard them phonologically as two separate words.
7.5
Phonological phrases
Phonological words orm phonological phrases. An understanding o phonological phrases is crucial in explaining various phonological processes in Korean. Consider (11). (11) (mɑʨimɑk)ω (sɑlɑm + -tɯl + -i)ω (mok + soli + -nɯn)ω (ɑlɯmtɑw- + -ʌ jo)ω . ‘Te last group o people have beautiul voices.’ Te sentence in (11) has our phonological words and can thereore be uttered in our different ways, as in (12). Te pronunciation o these sentences is given in (13). (12) a. (mɑʨimɑk)ω| (sɑlɑm + tɯl + -i)ω| (mok + soli + -nɯn)ω| (ɑlɯmtɑw- + -ʌ jo)ω |||. b. (mɑʨimɑk)ω (sɑlɑm + -tɯl + -i)ω| (mok + soli + -nɯn)ω (ɑlɯmtɑw- + -ʌ jo)ω |. c. (mɑʨimɑk)ω| (sɑlɑm + -tɯl + -i)ω| (mok + soli + -nɯn)ω (ɑlɯmtɑw- + -ʌ jo)ω ||. d. (mɑʨimɑk)ω (sɑlɑm + -tɯl + -i)ω| (mok + soli + -nɯn)ω| (ɑlɯmtɑw- + -ʌ jo)ω ||. (13) a. |||. [mɑ.ʨi.mɑk| sɑ.lɑm.tɯ.li| mok.s*o.li.nɯn| ɑ.lɯm.tɑ.w ʌ.jo]
166
Prosody
b. |. [mɑ.ʨi.mɑk.s*ɑ.lɑm.tɯ.li| mok.s*o.li.nɯ.nɑ.lɯm.tɑ.w ʌ.jo] c. ||. [mɑ.ʨi.mɑk| sɑ.lɑm.tɯ.li| mok.s*o.li.nɯ.nɑ.lɯm.tɑ.w ʌ.jo] d. ||. [mɑ.ʨi.mɑk.s*ɑ.lɑm.tɯ.li| mok.s*o.li.nɯn| ɑ.lɯm.tɑ.w ʌ.jo] In (13), when /s/ in /s ɑlɑmtɯli/ is phrased together with the preceding word /mɑʨimɑk/ which ends with /k/, /s/ is tensi�ed to become /s*/ as in (13b) and (13d). However, when /s/ in /s ɑlɑmtɯli/ is not phrased together with the preceding word /mɑʨimɑk/ , /s/ is not tensi�ed, as in (13a) and (13c). For the same reason, /s/ in /moksoli/ is tensi�ed and thus pronounced [moks*oli]. Tese examples show that whether the segment is to be tensi�ed or not depends on where the sound occurs within the same prosodic structure. (We will discuss Post-obstruent ensi�cation in more detail in 8.2.1.) Words ound in the same phonological phrase orm a smooth, unbroken pitch curve. Figure 7.7 shows how a emale Seoul Korean speaker pronounces the sentence in (14). When there is a boundary between two phonological words as in (14b), there is a clear drop in pitch between the last syllable o the �rst word [nε] and the �rst syllable o the second word [ ʌ]. (See Figure 7.7a.) Yet, when there is no such boundary between the two words, there is no sudden drop in pitch between the two words, as in (14c). (See Figure 7.7b.) (14) a. ? (mijʌninɛ)ω (ʌmʌninɯn)ω (nɑ jʌnilɯl)ω (miw ʌhɑni)ω?/ ‘Does Mi.yeon’s mother hate Nayeon?’ b. () () ? [mijʌninɛ|ʌmʌninɯn| nɑ jʌnilɯl| miw ʌhɑni] c. () ? [mijʌninɛ ʌmʌninɯn| nɑ jʌnilɯl| miw ʌhɑni] Let’s consider the different pitch curves that can be ound in phonological phrases. In Seoul Korean, the two most common pitch patterns or our-syllable words are LHLH and HHLH (L = low and H = high). Whether the �rst syllable o a phonological phrase is produced with a high pitch or a low pitch depends on the syllable-initial phoneme. Phonological phrases in which the initial consonant is one o the ollowing: /p *, t*, k*, ʨ*, s*, ph, t h, k h, ʨh, h, s/ (i.e. tensed or aspirated stops and affricates, or any type o ricative) tend to start with a high pitch as in HHLH. Otherwise, they start with a low pitch as in LHLH. From now on, we will call those phonemes that
167
7.5
Phonological phrases
(a) 320
[nε]
) z H ( h c t i P
160
[]
0
Time (s)
30
(b) 320
[nε]
) z H ( h c t i P
160
Figure 7.7
[]
0
Time (s)
3.0
Two pitch curves of /(mijʌninɛ ) ( ʌmʌninɯn) (nɑjʌnilɯl) (miwʌhɑni)?/ spoken by a female speaker of Standard Korean
are responsible or high pitch in syllable-initial position high-tone group phonemes, and those that are responsible or low pitch in syllable-initial position low-tone group phonemes. Consider (15). (15)
|| |||? /(k*ɑtɑloun)ω| (ʌmʌninɯn)ω| (hɑnɯn)ω (ili)ω| (nʌmunʌmu)ω| (mɑntʰɑkonɯn)ω| (hɑsitʌnjɑ) ω?/ ‘Does your picky mother complain that she has lots o things to do?’
Figure 7.8 shows a pitch curve extracted rom the utterances o a emale Seoul Korean speaker. Tis utterance contains six phonological phrases. Consider the our circles in Figure 7.8, which show the �rst our phonological phrases. As expected, when the �rst phoneme is a high-tone phoneme, as in the �rst and third phonological phrases, HHLH pitch curves are observed.
168
Prosody
320
) z H ( h c t i P
160
Figure 7.8
0
Time (s)
4.0
A pitch curve of /(k*ɑtɑloun)| ( ʌmʌninɯn)| (hɑnɯn) (ili)| (nʌmunʌmu)| (mɑntʰɑkonɯn)| (hɑsitʌnjɑ ) ?/ spoken by a female speaker of Standard Korean
However, when the �rst phoneme is a low-tone phoneme, as in the second and ourth phonological phrases, LHLH pitch curves are observed. Based on this observation, we can conclude that the basic pitch curve or our-syllable words in Korean is (High/Low) HLH. 4 Whether an initial syllable is produced with a high or a low tone is determined by the syllable-initial phonemes. Figure 7.9 captures this. What, then, o words made up o ewer than our syllables, or more than our syllables? In either case, what determines the pitch pattern o a phonological phrase is the �rst and last tone o a phonological phrase, leaving the number o syllables irrelevant. Let’s consider cases where a word has three syllables or less. In such cases, the second or third tones o a phonological phrase will not be realised. Hence, HH, LH, HL, and LL are all possible pitch patterns. Figure 7.10 shows the typical pitch patterns o phonological phrases with less than our syllables. What about a phonological phrase that has more than our syllables? Tere is no difference in pitch patterns between our-syllable words and �ve-syllable words. Te difference between the two is simply which particular syllable tone is realised. In principle, the �rst syllable will have the �rst tone and the second syllable will have the second tone. Te third tone will then be realised rom the second syllable rom the end. All other syllables which are not assigned any particular tone will show a ‘between-pitch’, or interpolation. As the distance between the second pitch and the third pitch increases, the between-pitch or interpolation slope will be smooth as in Figure 7.11.
169
7.5
Phonological phrases
(a) L group
L
Figure 7.9
(b) H group
H
L
H
H
H
L
H
Typical pitch patterns of phonological phrases with four syllables in Standard Korean (a) L group
L
H
L
H
H
L
L
H
(b) H group
H
Figure 7.10
H
H
L
Typical pitch patterns of phonological phrases with fewer than four syllables in Standard Korean (a) L group
L
Figure 7.11
H
H
(b) H group
L
H
H
H
L
H
Typical pitch patterns of phonological phrases with more than four syllables in Standard Korean
In other words, those words with more than our syllables will show the same tonal pattern as our-syllable words. Te ewer syllables the word has, the stiffer the slope will be and vice versa. Te next question which comes to mind is how many syllables a phonological phrase in Korean has on average? Is it limitless? Figure 7.12 shows the results drawn rom a study o the spontaneous speech o 57 adult native Korean speakers, involving some 122,912 phonological phrases, as well as a study o one emale speaker’s reading o 3,000 sentences, which contains 33,719 phonological phrases.5 In the ormer study, up to 13 syllables were observed within a phonological phrase, and in the latter study, up to 11 syllables were observed. As we can see, the most requently observed were 3-syllable PPs, which constituted 32% o the spontaneous speech that was studied, and 25.3% o the emale speaker’s read speech. PPs o 2–5 syllables were also requent; in spontaneous speech, they constituted about 80.9%, and as or the speech
Prosody
170
35 Spontaneous speech
32.0
Read speech
30 24.1
25 ) % ( 20 y c n e u q 15 e r F
25.3 21.9 19.4 16.8
14.9 12.0 10.6
10
6.9 5.4
3.6
5 2.2
0
Figure 7.12
1
1.4
2.1
2
3
4
5 6 7 Number of syllables
0.1
0.3
0.7
0.2
8
9
0.1
10+
Number of syllables per phonological phrase in spontaneous and read speech
read aloud, they constituted about 85.5% o the total PPs. In both read speech and spontaneous speech, native Korean speakers tend to produce phonological phrases o 2–5 syllables. Interestingly, Figure 7.12 shows that the proportion o 1 or 2-syllable PPs in read speech (17.1%) was hal that o spontaneous speech (34.7%). Tis is because o the requent use o interjections and shortened orms in spontaneous speech. In addition, phonological phrases with more than 6 syllables were rare in both spontaneous speech (8.5%) and in read speech (12.3%). 7.6
Intonational phrases
Phonological phrases orm an intonational phrase. Te intonational phrase is where intonation is realised. Like phonological phrases, intonational phrases are responsible or various phonological processes. Consider the ollowing sentence (16): (16) ? /(mijʌk + kuk)ω (nu + -k ɑ)ω (mʌk- + -ni)ω?/ ‘Who is having seaweed soup?’ Te sentence in (16) has three phonological words and they can be phrased as in (17).
171
7.6
Intonational phrases
(17) a. ||? / (mijʌk + kuk)ω| (nu + -k ɑ)ω| (mʌk- + -ni)ω?/ [mi.jʌ.k*uŋ| nu.k ɑ| mʌŋ.ni] b. | ? /(mijʌk + kuk)ω∥ (nu + -k ɑ)ω| (mʌk- + -ni)ω?/ [mi.jʌ.k*uk ∥ nu.k ɑ| mʌŋ.ni] c. |? /(mijʌk + kuk)ω| (nu + -k ɑ)ω (mʌk- + -ni)ω?/ [mi.jʌ.k*uŋ| nu.k ɑ.mʌŋ.ni] d. ? /(mijʌk + kuk)ω∥ (nu + -k ɑ)ω (mʌk- + -ni)ω?/ [mi.jʌ.k*uk ∥ nu.k ɑ.mʌŋ.ni] According to the way the phonological word is incorporated into the larger phonological structure, the pitch pattern or the actual pronunciation o each segment can change as in (17a). When there is an intonational phrase boundary between the �rst and second word as in (17b, d), the �nal segment o the �rst phonological word /k/ is realised as [k], ollowing its basic pronunciation. However, when there is no intonational phrase boundary between the �rst and second phonological word as in (17a, c), the �nal segment o the �rst phonological word /k/ is nasalised and pronounced as [ŋ] because o the ollowing nasal sound /n/. Tis process is called Nasalisation. Te key thing here is that whether to nasalise /k/ into [ŋ] or not depends on the existence o an intonational phrase boundary. How, then, does one identiy an intonational phrase? What are the criteria or determining whether a phrase is intonational or not? Te lengthening o the �nal syllable ollowed by a pause or distinctive pitch pattern is called an intonational phrase boundary tone, and this can be used as an important cue in detecting an intonational phrase. As mentioned earlier, the intonational phrase is also the unit where intonation is realised. Intonation plays an important role in determining the grammatical and pragmatic meaning o a s entence. Consider (18): (18) ||. [mijʌninɯn|ʨʰinkutɯlɯl| miw ʌhɛ jo] (RISING /FALLING) ‘Miyeon hates her riends.’ Even i the sentence (18) has three phonological phrases and they are phrased together as in (18), the mood o the sentence can vary according to the intonation pattern, particularly in the last phonological phrase. In Figure 7.13 , we can see that the pitch curves o the two �gures are similar, except or the last
172
Prosody
330
(a)
(b)
) z H ( h c t i P
160 0
Figure 7.13
Time (s)
5.0
Pitch curves of two types of intonation patterns of [mij ʌninɯn| ʨʰinkutɯlɯl| miwʌhɛjo] ‘Miyoen hates her friends’
phonological phrases: both utterances have an LHLH pitch pattern or the �rst phonological phrase, HHLH or the second, and LHL or the third. At the �nal syllable, however, Figure 7.13a has a low tone – alling pitch; and Figure 7.13b has a high tone – rising pitch. Tis difference means the same sentence can either be made declarative (Figure 7.13a), or interrogative (Figure 7.13b). As discussed in Chapter 1, intonation plays an important role in realising mood (e.g. declarative, interrogative, imperative, propositive, etc.). Particularly in inormal speech, the most requently used sentence-�nal enders such as -ʌ(ɑ) –/ and -ʌ(ɑ)jo –/ 6 can be used to express all moods. Te speci�c mood is determined solely by the intonation o the last syllable in the �nal intonational phrase; or instance, in (19), - ʌ(ɑ) and -ʌ(ɑ)jo can be used as endings or declarative, interrogative, imperative, propositive and exclamatory sentences. (19) Intonation and Mood /(pɑp + -ɯl) (mʌk- + -ʌ + -jo)/ a. Declarative: low tone b. Interrogative: high tone c. Imperative: short low tone d. Propositive: low-high tone e. Exclamatory: high-low tone
‘(I) am having a meal.’ ‘Are (you) having a meal?’ ‘Have a meal!’ ‘Let’s have a meal.’ ‘(You) are having a meal!’
Finally, able 7.1 shows some statistical inormation on prosodic units in Korean based on data gathered rom the spontaneous speech o 57 adult Korean speakers. Te 35,439 utterances that were analysed contain a total o 55,927
173
7.6
Intonational phrases
Table 7.1 Statistical information relating to prosodic units
Utterance
Intonational phrase Phonological phrase Syllable Phoneme
1.58 3.47 11.39 24.22
Intonational phrase
Phonological phrase Syllable Phoneme Phonological phrase
2.20 7.22 15.35
Syllable Phoneme Syllable
3.28 6.98
Phoneme
2.13
intonational phrases, 12,919 phonological phrases, and 403,605 syllables. Tis means that, on average, one utterance contains 1.58 intonational phrases, 3.47 phonological phrases, 11.39 syllables, and 24.22 phonemes. One intonational phrase has an average o 2.20 phonological phrases, 7.22 syllables and 15.35 phonemes. One phonological phrase has 3.28 syllables and 6.98 phonemes. One syllable has an average o 2.13 phonemes.
DID YOU KNOW . . . ?
Why is it hard for Koreans to pronounce ‘www’?
You may notice something very interesting when you listen closely to Korean people read out website addresses. Despite there clearly being three ‘w’s, they may commonly only pronounce two. Why do Koreans pronounce ‘www.’ as [t*ʌpullju|t*ʌpulljuʨ*ʌm], instead o [t*ʌpullju|t*ʌpu llju|t*ʌpulljuʨ*ʌm]? O course, people who are used to the internet know that there are three ‘w’s in a web address, and may assume they heard three ‘w’s even though only two may have been pronounced. Te reason or this may be twoold. Te �rst is to do with production and rhythm. In a URL, the ‘www.’ part contains comparatively more syllables than other parts separated by dots. For example, www.korea.
174
Prosody
ac.kr or www.naver.com are each [t*ʌpullju|t*ʌpullju| t*ʌpulljuʨ*ʌm ‖ k ʰoliɑʨ*ʌm‖ ɛis*iʨ*ʌm‖ k ʰɛiɑl] (10, 4, 4, 3 syllables respectively), and [t*ʌpullju| t*ʌpullju| t*ʌpulljuʨ*ʌm‖ nɛipʌʨ*ʌm‖ khʌm] (10, 4, 1 syllables respectively), making the ‘www.’ segment unusually long. Pronouncing this unusually long segment in ull will disrupt natural speech rhythm. Reducing one ‘w’ still makes it the longest with seven syllables, but there is a considerable metrical difference between a seven-syllable and ten-syllable segment in the Korean language. Tis brings us to the second reason. As seen in able 7.1, Korean intonational phrases are comprised o an average o 2.20 phonological phrases and 7.22 syllables. Tereore, while reducing one ‘w’ will still leave this segment with the most syllables, it will be much easier to pronounce as it still conorms to Korean rhythmic patterns. Tereore it is thought that one ‘w’ is dropped in order to make the phrase better �t the pattern o Korean intonational phrases.
DID YOU KNOW . . . ?
Men, don’t speak!
Tere was an academic conerence to be held in Busan, so some students rom Seoul decided to take a trip. Tey all took taxis. One male student and two emale students rom Seoul ended up taking one taxi together. Afer stating their destination to the taxi driver, the three students began talking amongst themselves. Suddenly, the taxi driver interrupted them in a Busan dialect. “Young man, please stop speaking! You sound like a coward. Please just let the women speak!” People rom Gyeongsang province ofen say that the Seoul accent sounds cloying, effeminate and eeble. Tis is why people rom the Gyeongsang province, especially the men, consider the Seoul accent ill-suited and unbecoming or men. On the other hand, people rom Seoul say that the Gyeongsang dialect sounds blunt, brusque and angry. So people rom Seoul consider people rom the Gyeongsang province impolite and badmannered. Why might this be? Prosodically, the Gyeongsang dialect (which includes the Busan dialect) always has a low tone at the end o a sentence. Also, the last syllable in an utterance is never lengthened, making the end o a sentence abrupt;
175
7.6
Intonational phrases
even yes/no questions do not have a rising tone. In contrast to the Seoul dialect, the mood o a sentence is decided not by prosodic actors, but by morphological markings such as grammatical particles. Te act that the end o an utterance is always short, coupled with the quickly alling tone, makes speakers o Seoul or Standard Korean think that people rom Gyeongsang province sound angry and rude. In addition, in the Seoul dialect, abrupt speech endings are associated with commands, stubbornness and sternness. Lengthened speech endings and smooth pitch movement on the other hand represent kindness, ondness, and affection.
DID YOU KNOW . . . ?
Learning the prosody of a second language
Prosody is one o the �rst eatures acquired in one’s mother tongue, but one o the last learned in oreign language acquisition. Let’s think about some examples. Firstly, Japanese learners o Korean ofen realise the last syllable o an intonational phrase as too short. Tis is because there is no �nal lengthening in Japanese. As in the Gyeongsang dialect, short-ending at the end o an utterance may sound abrupt. Conversely, Korean speakers o Japanese tend to over-lengthen the �nal syllable o an utterance. What about English learners o Korean? Let’s think about the case o yes/no questions. In both Korean and English, the end o a question has a rising tone. However, how much o the question is spoken in a rising tone differs between the two languages. In Korean, only the last syllable o an utterance is raised. In English, however, it is ofen the case that the whole sequence o sounds afer the accented syllable o the last word is raised. I you are curious as to whether this is really true, ask a Korean riend to say the sentence below. ?
One o the most diffi cult parts o language acquisition or second language learners is acquiring the prosody o the target language. Tis is especially true or Korean and English, languages that differ signi�cantly in prosody. In Korean, there is no lexical stress, whereas there is a lexical stress to each word in English. Due to this difference, in Korean each syllable tends to be pronounced with similar length or weight.
Prosody
176
7.7
Summary
In this chapter, we have discussed the prosodic characteristics o Korean. Prosody includes the notions o pitch, loudness and length o a sound as well as pause. Tese are psychological rather than physical notions which are realised onto a sequence o segments. Different prosody can lead to a different interpretation o a word or a sentence. Although rising and alling tone in a Korean sentence make it either a question or a statement, it cannot be said that Korean is a tonal language like Chinese, since tone does not cause any changes in lexical meaning or individual words. Having a syllable structure with an onset, a nucleus and a coda is universal in all languages. Nevertheless, how actual syllable structure is ormed and realised differs rom language to language. Onset and coda consonants are optional, whereas a vowel as a syllable nucleus is obligatory in syllable structures. In Korean, we observed that there are eight possible syllable structures. Korean also has a prosodic structure where a single syllable, or a sequence o syllables, orms a phonological word, and a single phonological word, or a sequence o phonological words, orm a phonological phrase. Furthermore, a single phonological phrase, or a sequence o phonological phrases, may orm an intonational phrase. An utterance is ormed by a single intonational phrase, or a sequence o intonational phrases. EXERCISES 1
Explain why the same English word is adopted differently in Korean and Japanese, as in the case of the word milk, which was discussed earlier.
2
Discuss why Korean native speakers perceive the words below as having a different number of syllables from the number of syllables in English. news, boat, strike, silk, tractor, graph
3
Listen carefully to the following words in Korean and write down how many syllables they have.
4
Is Korean a tonal language? Answer the question with relevant examples.
5
Listen carefully to the following sentence and describe it using musical notes.
6
Listen carefully. How many phonological phrases or intonational phrases can
7
you hear in this sentence? Draw a diagram of the syllable structure of Korean and state which elements are obligatory and which are optional.
8
List examples of CV languages and CVC languages.
Exercises
177
9
Draw syllable structures for the following words. a. /sʌnsɛŋnim/ ‘teacher’ b. /pɑŋʌn/ ‘dialect’
10
c. /kɑmʌnisʌl/ ‘sweet talk’ What are the characteristics of the intonational phrase in Korean?
11
Listen carefully to the following sentences and state the mood of the sentence. a. declarative, interrogative, propositive, imperative, exclamatory b. declarative, interrogative, propositive, imperative, exclamatory c. declarative, interrogative, propositive, imperative, exclamatory d. declarative, interrogative, propositive, imperative, exclamatory
12
What are the most distinctive prosodic differences between English and Korean?
13
Are the following sequences of sounds possible syllable structures in Korean? If so, mark the syllable boundary for each word. If not, explain why? a. [tɑlk] b. [noin] c. [spolk] d. [pɑlɑm] e. [soŋɑʨi] f. [nʌks] g. [ʌlini] h. [kisk] i. [nos]
8
Phonological rules of Korean (I)
Te Korean alphabet, or Hangeul, is a phonemic alphabet. In other words, each letter corresponds to one phoneme. However, unlike the roman alphabet, Hangeul is not written letter by letter (e.g. or Hangeul), but in syllables. Tat is, two or three letters which orm one syllable are written together as one orthographic unit. In this book, we have displayed Hangeul with the corresponding IPA transcription alongside. However, knowing how each letter sounds is not enough when learning how to pronounce Korean texts. When the basic units o Hangeul orthography, which each correspond to one syllable, are put together to orm a morpheme or a unit bigger than a word, phonological rules will affect their pronunciation wherever they are applicable. For example, the words in (1) would not usually be pronounced syllable-by-syllable, unless perhaps they were being dictated to another person. For instance, the words in (1) show Nasalisation o the /k/ in the �rst syllable. (1)
/kuk + mul/ → [kuŋmul] ‘soup-liquid’ /sok + nɛ/ → [soŋnɛ] ‘intention’
Tis means that without knowing the Nasalisation rule in Korean, it is hard to pronounce the words in (1) properly. Native speakers apply phonological rules such as Nasalisation automatically, as such rules are innate in native speakers’ knowledge o their mother tongue. Sometimes, Korean speakers can even be heard to pronounce good morning as [kun moniŋ] rather than [gʊd mɔːrnɪŋ]. Tis occurs when Korean native speakers use the Nasalisation rule o Korean phonology in their pronunciation o English words. Native speakers acquire the phonological rules o their language without much effort, but learners o a second language need to study them one by one to be able to speak the language �uently. Tereore, anyone who wants to speak Korean �uently must be aware o the constraints placed on pronunciation and general phonological rules. 178
179
8.1
Phonological rules related to syllable structure constraints
In this book, phonological rules are divided into two types: (i) rules that can be applied without knowledge o the word’s morphological inormation; and (ii) rules that can be applied in consideration o the word’s morphological inormation. Te �rst type o rule is automatically applied to every sound sequence in Korean, but to apply the rules o the second type, one needs to know how a word is ormed. In this chapter, we will concentrate on the �rst type o phonological rule. Te second type is to be discussed in Chapter 9. Te �rst type o phonological rule can be urther classi�ed into (i) rules related to syllable structure constraints; and (ii) rules related to surace phonetic constraints. Rules relating to syllable structure constraints will be discussed in 8.1, and rules relating to surace phonetic constraints will be discussed in 8.2. DID YOU KNOW . . . ?
Leg or moon? Liaison in Korean
Normally, phonemic letter systems like the English alphabet are written rom lef to right, horizontally, and one letter at a time. When the Korean word or ‘leg’, [t ɑli], is written in a like manner in Hangeul, it becomes . However, Hangeul is not written with each letter acting as an individual unit, as in . Instead, the sounds are divided into syllable units, and each syllable is written as one group. Te example o [tɑli] can be divided into two syllables – [tɑ] and [li] – and when it is written in syllable units, it becomes , not . In addition to writing in groups o syllables, Hangeul utilises a system o demarcating morphemes. Tis kind o orthography usually preserves each morpheme’s original orm rather than re�ecting the phonological changes it undergoes in a given context. Tus when the Korean word, ‘moon’, is combined with the subject particle - , it sounds like [tɑli], but is nonetheless written as (noun + subject particle), rather than . Although the o the subject particle, , [i], is written down in this case, it is an empty letter that has no phonetic value attached to it. Tus, the used here acts merely to point out the initial position o a syllable. Liaison reers to the phenomenon by which a word written as is realised phonetically as [t ɑ.li]. Although Liaison is not a phonemic rule, it is a necessary and central principle in reading Hangeul where the text must be read as a continuous linking o morphemes. In the case o + , the grouping o CVC + V in writing becomes [CV.CV] in pronunciation. As a result, ‘leg’ and + (‘moon’ + subject particle) are written differently, but pronounced the same.
180
Phonological rules of Korean (I)
Ten have these rules or writing existed rom the time when Hangeul was �rst invented? Hangeul was created in 1443 under the name, Hunminjeongeum , and rom that time it has always been written in groups o syllables, not in units o letters. Tis principle is outlined in ( ) Hunminjeongeum Haerye ‘Explanation and Examples o Hunminjeongeum ’, which was published in 1446. Although Hangeul is a phonemic system o writing, it was designed to be written top to bottom, and right to lef. Moreover, each letter was supposed to be arranged into units o syllables to �t in the space o a square. As can be seen in the picture, the reason why the committee behind Hunminjeongeum chose this method o writing was because o the powerul in�uence o Chinese characters. Tus the principle o grouping existed rom the very inception o Hangeul. However, when Hangeul was �rst created, the groupings were not based on morphemes but on pronunciation; in other words, syllables were marked not in units o morphemes, but in units o sound. In texts o this time, was still written as .
181
8.1
8.1
Phonological rules related to syllable structure constraints
Phonological rules related to syllable structure constraints As discussed in Chapter 7, syllables in Korean must have a vowel as their medial sound, and may optionally have consonants as their initial and/or �nal sounds. Tere are two important constraints in Korean syllable structures. Te �rst is that syllable-�nal consonants should be pronounced as unreleased sounds afer central closure. Te second is that only one consonant may appear in the initial and �nal positions o a syllable. Tese two constraints result in the ollowing phonological rules that we shall now discuss: Neutralisation (8.1.1) and Consonant-cluster Simpli�cation (8.1.2).
8.1.1
Neutralisation /(nɑʨʰ + -to) (tuk*ʌp- + -tɑ)/ → [nɑt*o | tuk*ʌpt*ɑ] in Korean is an idiomatic expression, meaning ‘thick-skinned’. However, when this phrase is heard, this can also reer to thickness o /nɑs/, meaning a sickle. In Korean, consonants in syllable-�nal position should never be released afer complete central closure. All ‘released’ sounds should be replaced with unreleased sounds pronounced in a similar place o articulation. Tis rule is called Neutralisation, because the lax/tense/aspirated contrast existing in Korean consonants is neutralised in syllable-�nal position. In addition, the ricative/affricate contrast is neutralised and syllable-�nal ricatives and affricates are all pronounced the same, as /t/. Tis is the cause o the ambiguity between /nɑʨʰ/ ‘ace’ and /nɑs/ ‘sickle’. Consider (2) and (3). Coda consonants in (2) and (3) are pronounced the same.
(2)
/k ɑk/ → [k ɑk] ‘angle’ /puʌk ʰ/ → [puʌk] ‘kitchen’ - /k*ɑk*-/ → [k*ɑk] ‘to carve’
(3)
/nɑs/ → [nɑt] ‘sickle’ - /nɑh-/ → [nɑt] ‘to bear’ /nɑʨ/ → [nɑt] ‘daytime’ /nɑʨʰ/ → [nɑt] ‘ace’
Te examples in (2) show the three-way distinction in stop sounds (e.g. /k, k ʰ, k */) being neutralised. Te examples in (3) show ricatives or affricates being neutralised into the alveolar sound /t/. Because all syllable-�nal consonants need to be unreleased and pronounced with complete central closure, ricatives and affricates, by de�nition released sounds cannot appear in syllable-�nal position. Tey need to be replaced by unreleased sounds. All ricatives
182
Phonological rules of Korean (I)
Table 8.1 Pronunciation of syllable-final consonants
Syllable-�nal Stop phonetic (plosive) Fricative Affricate Nasal Liquid realisation Example
Bilabial
/p/ /pʰ/ /m/
Alveolar
/t/ /tʰ/ /s/ /s*/ /n/ /ʨ/ /ʨʰ/
Alveolo-palatal
/l/ Velar
/k/ /k ʰ/ /k */ /ŋ/ /h/
Glottal a It
/p/
[pɑp] ‘rice’ [ɑp] ‘ront’
/m/
[pɑm] ‘chestnut’
/t/
- [tɑt] ‘to close’ [sot] ‘pot’ [nɑt] ‘sickle’ - [it] ‘to be’
/n/
[k ɑn] ‘liver’
/t/
[nɑt] ‘daytime’ [nɑt] ‘sickle’
/l/
[k ɑl] ‘knie’
/k/
[k ɑk] ‘angle’ [puʌk] ‘kitchen’ - [k *ɑk] ‘to carve’a
/ŋ/
[k ɑŋ] ‘river’
/t/
- [nɑt] ‘to bear’
is noticeable that there are no Korean words whose roots end with /p */, /t*/ or /ʨ*/.
(s, s*, h) and affricates (ʨ, ʨ*, ʨʰ) in syllable-�nal position are thereore replaced by the lax stop /t/, which is pronounced in the alveolar position, relatively close to where these ricatives and affricates are produced. Tis lax stop is then ultimately produced as an unreleased sound, [t ¬]. In sum, due to Neutralisation, /s/, /s*/, /h/, /ʨ/, /ʨʰ/ all become the lax stop /t/, and then as this lax stop is unreleased, the sound is realised as [t]. As a result o Neutralisation, the only distinctive sounds in syllable-�nal position are the ollowing seven sounds: /k, n, t, l, m, p, ŋ /.1 able 8.1 shows the patterns o Neutralisation. Syllable-�nal consonants however can be pronounced differently i grammatical morphemes starting with vowels are attached to each o the words above. Consider (4) and (5). (4) In syllable-�nal position a. noun /k ɑk/ → [k ɑk ̚] ‘angle’ /pɑp/ →[pɑp̚] ‘rice’
183
8.1
Phonological rules related to syllable structure constraints
b. verb - /pɑk/ → [pɑt̚] ‘to nail’ - /tɑt-/ → [tɑt̚] ‘to close’ (5) Followed by grammatical morphemes, starting with a vowel a. noun /k ɑk + -i/ → [k ɑki] ‘angle (subject particle)’ /pɑp + -i/ → [pɑpi] ‘rice (subject particle)’ b. verb /pɑk- + -ɑ/ → [pɑk ɑ] ‘to drive (connective)’ /tɑt- + -ɑ/ → [tɑtɑ] ‘to close (connective)’ For instance, in /k ɑk/ ‘angle’ in example (4), the last consonant should be pronounced as [k], where the body o the tongue presses against the velum. Likewise, the last sound o /tat-/ - ‘to close’ should be pronounced as [t ̚], with the end o the tongue touching the alveolar ridge. Te second /p/ in /pap/ should also be pronounced as [p ̚], with the upper and lower lips touching. However, in (5), syllable-�nal lax stops are all moved to the initial position o the next syllable. Hence, constraints upon syllable-�nal consonants are no longer applicable.
8.1.2
Consonant-cluster simplification Tere are no consonant clusters in Korean. Although orthography allows consonant clusters in the syllable-�nal position o a letter, as in /t ɑlk/ ‘chicken’ and /sɑlm/ ‘lie’, only one o the sounds in the consonant cluster will actually be pronounced. Consonant-clusters did exist in �feenth-century Korean in both syllable-initial and syllable-�nal position, but t hey are no longer ound in contemporary Korean.2 Consider the ollowing. (6)
/sɑlm/ → [sɑm] ‘lie’ /hɯlk/ → [hɯk] ‘mud’
In (6), /sɑlm/ is pronounced [sɑm], not [sɑl]. Likewise, /hɯlk/ is pronounced [hɯk], not [hɯl]. Te second sound in the cluster is however realised when it is ollowed by a grammatical particle starting with a vowel, as shown in (7). (7)
/salm + -i/ → [sɑlmi] ‘lie (subject particle)’ /hɯlk + -i/ → [hɯlki] ‘mud (subject particle)’
Phonological rules of Korean (I)
184
Table 8.2 Pronunciation of consonant clusters in syllable-final position
Followed by grammatical morphemes, starting with a vowel
Consonant cluster
Phonetic realisation In syllable-�nal position
/ks/
[k]
/moks/ → [mok] ‘portion’
/nʨ/
[n]
[moks*itɑ] /ɑnʨ- + -tɑ/ → [ɑnt*ɑ] ‘to sit’ /ɑnʨ- + -ɑ/ → [ɑnʨɑ]
/nh/
[n]
/ɑnh- + -nɛ/ → [ɑnnɛ]
/ɑnh- + -ɑ/ → [ɑnɑ]
[l]
‘negating verb’ /nʌlp- + -tɑ/ → [nʌlt*ɑ] ‘to be broad’
/nʌlp- + -ʌ/ → [nʌlpʌ]
[p]
/pɑlp- + -tɑ/ → [pɑpt*ɑ]
/pɑlp- + -ɑ/ → [pɑlpɑ]
/lp/
/ls/
[l]
/ltʰ/
[l]
/ps/
[p]
/lk/
[k] [l]
‘to step on’ /w ɛ- + kols/ → [w ɛkol] ‘single track’ /hɑltʰ- + -tɑ/ → [hɑlt*ɑ] ‘to lick’ /k ɑps/ → [k ɑp] ‘price’
/moks + -i- + -t ɑ/ →
/w ɛ- + kols + -i- + -t ɑ/
[w ɛkols*itɑ] /hɑltʰ- + -ɑ/ → [hɑltʰɑ] →
/k ɑps + -i/ → [k ɑps*i]
/hɯlk/→ [hɯk] ‘mud’
/hɯlk + -i/ → [hɯlki]
/mɑlk- + -k ɛ/ → [mɑlk *ɛ]
/mɑlk- + -ɑ/ → [mɑlk ɑ] /sɑlm + -i/ → [sɑlmi] /ɯlpʰ- + -ʌ/ → [ɯlpʰʌ]
/lm/
[m]
‘clearly’ /sɑlm/→[sɑm] ‘lie’
/lpʰ/
[p]
/ɯlph- + -tɑ/ → [ɯpt*ɑ]
‘to recite’ /lh/
[l]
/t*ulh- + -nɯn/ → [t*ullɯn]
/t*ulh- + -ʌ/ → [t*ulʌ]
‘to bore (adnominal)’
able 8.2 shows how consonant clusters are phonetically realised in speech. At this point one might ask whether there is any rule which guides the selection o the ‘one’ sound to be pronounced in consonant clusters. As we can see in able 8.2, it is not always the �rst sound in a sequence that is chosen to be pronounced. Tere are some rules, however, to which we shall now turn our attention. Consonant clusters can be divided into two types according to the phonetic characteristics o the consonants such as manner o articulation and place o articulation as in (8).
185
8.1
Phonological rules related to syllable structure constraints
(8) a. Sounds with [+sonorant] eatures will be chosen /ls/ → [l]: /w ɛ- + kols/ → [w ɛkol] ‘intently’ /ltʰ/ → [l]: /hɑltʰ- + -nɯn / → hɑlnɯn → [hɑllɯn] ‘to lick (adnominal)’ /nʨ/ → [n]: /ɑnʨ- + -tɑ/ → ɑnʨt*ɑ → [ɑnt*ɑ] ‘to sit (down)’ /lh/ → [l]: /t*ulh- + -nɯn/ → t*ulnɯn → [t*ullɯn] ‘to punch (adnominal)’ /nh/ → [n]: /mɑnh- + -nɛ/ → [mɑnnɛ] ‘to be plentiul’ b.
Sounds with [−coronal] eatures will be chosen /lm/ → [m]: /sɑlm- + -nɛ/ → [sɑmnɛ] ‘to boil’ /lk/ → [k]: /hɯlk/ → [hɯk] ‘mud’ /ks/ → [k]: /nʌks/ → [nʌk] ‘soul’ /ps/ → [p]: /k ɑps/ → [k ɑp] ‘price’
Let us consider the examples in (8) careully. First o all, we see that consonants with [+sonorant] or [ −coronal] eatures are almost always chosen. Tis is why in (8a), only sounds with [+sonorant] are chosen. However, in (8b), where the sounds are either both sonorants or both obstruents, non-coronals are chosen. However, this does not explain all cases. In act, the pronunciation o sequences such as /lk/ and /lp/ are very interesting since these are combinations o a sound with [+sonorant, +coronal] eatures (/l/) and a sound with [−sonorant, −coronal] eatures (/p, k/). We have shown that within a consonant cluster, a sound with [ −sonorant] or [+coronal] eatures is avoided. However, in a sequence o /lk/ and /lp/, each consonant has either a [ −sonorant] or [+coronal] eature. How, then, can the ‘one’ sound be chosen? It is not easy to prioritise between the two constraints that (i) [ −sonorant] should be avoided and (ii) [+coronal] should be avoided. (9) a. /lk/ → [l]: /mɑlk- + -k ɛ/ → mɑlkk*ɛ → [mɑlk*ɛ] ‘to be clean (conjunctive)’ /mɑlk- + -tɑ/ → mɑlkt*ɑ → [mɑkt*ɑ] ‘to be clean’ b. /lp/ → [l]: /nʌlp- + -tɑ/ → nʌlpt*ɑ → [nʌlt*ɑ] ‘to be wide’ /nʌlp- +- ko/ → nʌlpk*o → [nʌlk*o] ‘to be wide (connective)’ Further investigations are necessary in order to choose the ‘one sound’ in such cases. According to K.-W. Nam and J.-H. Oh (2009), in sequences o /lk/ and /lp/, the coronal consonant /l/ is ofen preerred over its counterparts when the ollowing verbal endings are /-ko/ – or /-ʨi/–. Tis shows a
186
Phonological rules of Korean (I)
Table 8.3 Dialectal differences in consonant cluster realisation +Sonorant >> −Sonorant
Coronal >> +Coronal −
Seoul /ilk- + -t ɑ/ /ilk- + -k ʌnɑ/ /ilk- + -nɛ/ /ilk- + -nɯn- + -tɑ/ Gyeongsang /ilk- + -t ɑ/ /ilk- + -k ʌnɑ/ /ilk- + -nɛ/ /ilk- + -nɯn- + -tɑ/ Jeolla /ilk- + -t ɑ/ /ilk- + -k ʌnɑ/ /ilk- + -nɛ/ /ilk- + -nɯn- + -tɑ/
Hierarchy is not clear
[ikt*ɑ] [ilk *ʌnɑ] iknɛ → [iŋnɛ] iknɯntɑ → [iŋnɯntɑ] [ilt*ɑ] [ilk *ʌnɑ] ilnɛ → [illɛ] ilnɯntɑ → [illɯntɑ] [ikt*ɑ] ikk *ʌnɑ → [ik *ʌnɑ] iknɛ → [iŋnɛ] iknɯntɑ → [iŋnɯntɑ]
growing tendency to choose a consonant with [+sonorant] eature in /lk/ and /lp/ clusters. Interestingly, some variation is observed between dialects due to the differing hierarchy between the constraints. For example, in the Gyeongsang dialect, the [+sonorant] constraint takes priority over the [ −coronal] constraint. On the other hand, in the Jeolla dialect, the [−coronal] constraint takes priority over the [+sonorant] constraint. Tis is why the pronunciations o /ilk- + -t ɑ/ , /ilk- + -k ʌnɑ/ , /ilk- + -nɛ/ , /ilk- + -nɯn- + -tɑ/ differ in those regions. Consider able 8.3, which shows dialectal differences in the pronunciation o consonant clusters.
DID YOU KNOW . . . ?
The simplification of consonant clusters is like playing musical chairs
Let’s compare the phenomenon o Consonant Cluster Simpli�cation with playing musical chairs. Musical chairs is a game where there is always one chair ewer than the number o people playing, and where the players have to go around the chairs in a circle until the music comes to an end. When
187
8.2
Phonological rules via surface phonetic constraints
the music stops, everyone must try to �nd a seat, and whoever does not get a seat is out o the game. Te game continues until there is only one chair lef, and whoever gets that seat is the winner o the game. Ten what characterises the people who are eliminated early on? Tey are probably less agile or weaker than the rest. Tose who have either o these traits would lose out to the person who does not. But i two people were playing musical chairs, and one person were slow and the other one weak, who would be the �rst to be eliminated? O course, it would depend on which actor contributes more to losing. I lack o agility mattered more, then the ‘slow’ person would lose to the ‘weak’ player; and i physical strength mattered more, the ‘weak’ person would lose to the ‘slow’ player. But what i such a clear hierarchy did not exist? Tis is indeed t he case in Seoul Korean. Since there is no clear hierarchy in Seoul Korean, sometimes the consonant with the [ −coronal] eature is chosen and sometimes the consonant with the [+sonorant] is chosen.
8.2
Phonological rules via surface phonetic constraints In this section, we will discuss the phonological rules related to phonetic constraints in Korean. Tese rules are applicable in any corresponding phonetic environment, regardless o morphological inormation o the sound sequence. Unlike some rules that we shall discuss in Chapter 9, rules relating to surace phonetic constraints are applied to pure-Korean, Sino-Korean and loan words. As we shall discuss, certain sound sequences are not available in Korean at all, whereas certain sounds and sound sequences are requently observed. Hence, even i Korean speakers do not clearly hear what is spoken, when they hear a certain sound sequence, even rom a distance, they will recognise it as Korean. Likewise, all native speakers will be able to do the same or their own languages. In this section, we will discuss phonological rules which are caused largely by surace phonetic constraints. In other words, the rules disc ussed in this section are applied automatically, regardless o morphological inormation in a given phonetic environment. Te rules to be discussed include: Post-obstruent ensi�cation (8.2.1) Obstruent Nasalisation (8.2.2) Liquid Nasalisation (8.2.3) Lateralisation (8.2.4) Non-coronalisation (Bilabialisation and Velarisation) (8.2.5) Aspiration (8.2.6)
188
Phonological rules of Korean (I)
Similar-place Obstruent Deletion (8.2.7) /j/ Deletion (8.2.8) /h/ Deletion (8.2.9) 8.2.1
Post-obstruent Tensification One o the popular winter dishes in Korean cuisine is /kuk + p ɑp/ ‘boiled rice served in soup’. You will hear Korean people pronouncing it with a tense [p*] as in [kukp*ɑp] rather than with a lax [p] as in [kukp ɑp]. Tis is due to Post-obstruent ensi�cation in Korean phonology. Post-obstruent ensi�cation reers to the process in which a lax obstruent becomes tensi�ed when it is preceded by stops, such as /p, t, k/ or instance. Post-obstruent ensi�cation is the result o a surace phonetic constraint which orbids a sequence o an obstruent and a lax stop. Post-obstruent ensi�cation is an obligatory rule. Consider (10). (10) a. /p, t, k/ + lax obstruent /pɑp + totuk/ → [pɑpt*otuk] ‘delicious side dish’ /tɑt- + -ko/ → [tɑtk*o] ‘to close (connective)’ /mʌk- + -tɑ/ → [mʌkt*ɑ] ‘to eat’ b. /pʰ, tʰ, k ʰ/ (→ replaced by /p, t, k/) + lax obstruent /jʌpʰ + ʨip/ → jʌpʨip → [jʌpʨ*ip] ‘next door’ /pɑtʰ + kolaŋ/ → pɑtkolɑŋ → [pɑtk*olɑŋ] ‘urrow’ /puʌk ʰ + paŋ/ → puʌkpɑŋ → [puʌkp*ɑŋ] ‘scullery’ c. /p*, t*, k*/ (→ replaced by /p, t, k/) + lax obstruent /k*ɑk*- + -tɑ/ → k*ɑktɑ → [k*ɑkt*ɑ] ‘to carve’ (Note that in Korean, no verbal stem ends with p* or t*.) d. /s, s*, ʨ, ʨʰ/ (→ replaced by /t/) + lax obstruent /os + pɑŋ/ → otpɑŋ → [otp*ɑŋ] ‘dress room’ /is*- + -ko/ → itko → [itk*o] ‘to be (connective)’ /iʨ- + -ko/ → itko → [itk*o] ‘to orget (connective)’ /k*oʨʰ + k ɑlu/ → k*otk ɑlu → [k*otk*ɑlu] ‘pollen’
DID YOU KNOW . . . ?
Finding the underlying form through playing word chains
‘Word chain’, also known as ‘grab on behind’ and ‘last and �rst’, is a word game in which one person says a word, and the next person has to say a new word starting with the last syllable o the previous word. I the next
189
8.2
Phonological rules via surface phonetic constraints
person cannot come up with a new word that starts with that particular syllable, he or she loses the game. I A and B were to play, or example, and had the chain A: [puŋ.ʌ] ‘carp’ → B: [ʌ.pu.pɑ] ‘piggybag’ → A: [pa.nɑ.nɑ] ‘banana’ → B: [nɑ.pi] ‘butter�y’ → A: [pi.nɯl] ‘scales’ going, A would win i B were unable to think o a new word starting with [nɯl]. Even when 4- to 6-year-olds play word chains, one can see that they have an awareness o base orms and the rules or ensi�cation. For example, i the word /os + kam/ [otk*ɑm] ‘cloth’ were given, the child might then answer, [k ɑmʨɑ] ‘potatoes’ recovering the base orm, /k ɑm/ , prior to ensi�cation, even though he would actually hear [k *ɑm] in [otk*ɑm] ‘cloth’. Tis shows that even young children understand the principle o ensi�cation and can recover the appropriate base orm beore �nding the new word that they want. 8.2.2
Obstruent Nasalisation /(pɑp) (mʌk- + -ʌs*- + -ni)/ → [pɑmmʌk ʌnni] is a common greeting in Korean, meaning, ‘Did you have a meal?’ However, this can also mean, ‘Did you eat chestnuts?’ Tis is because the /pɑp/ in /(pɑp) (mʌk- + -ʌs*- + -ni)/ is nasalised into [pɑm] beore /(mʌk- + -ʌs*- + -ni)/ , which starts with a sonorant sound. Nasalisation reers to the process by which an obstruent is nasalised when it is ollowed by a sonorant such as /m, n, l/.3 Nasalisation is the result o a surace phonetic constraint which orbids a sequence consisting o an obstruent and a sonorant. An obstruent sound ollowed by a sonorant is assimilated into a nasal that is articulated in the same place o articulation. Consider (11). Hence, /p/ is assimilated into /m/ in (11a), since both are bilabial sounds; /t/ is assimilated into /n/ in (11b), as both are alveolar sounds; and /k/ is assimilated into /ŋ/ in (11c), since both are velar sounds. Te examples in (11d) show cases where two phonological rules are applied: (i) Obstruent Nasalisation and (ii) Liquid Nasalisation.4 Firstly, syllable-�nal /p, k/ in / ɑp/ and /ɑk/ are nasalised as /m, ŋ/ in / ɑm/ and /ɑŋ/ and then the ollowing /l/ in /lj ʌk/ is nasalised into /n/ in /njʌk/.
(11) a. /p/ + /m, n/ → [m] + [m, n] /pɑp + mɑs/ → pɑpmɑt → [pammɑt] ‘taste o rice’ /kjʌp + nun/ → [kjʌmnun] ‘compound eyes’ b. /t/ + /n/ → [n] + [n] /tɑt- + -nɯn/ → [tɑnnɯn] ‘to close (adnominal)’
190
Phonological rules of Korean (I)
c. /k/ + /m, n/ → [ŋ] + [m, n] /k ɑk- + -mɑk/ → [k ɑŋmɑk] ‘cornea’ /mʌk- + -nun/ → [mʌŋnun] ‘to eat (adnominal)’ d. /p, k/ + /l/ → m, ŋ + l (Nasalisation o /p, k/) → [m, ŋ] + [n] (Nasalisation o /l/) /ɑp- + -ljʌk/ → ɑmljʌk → [ɑmnjʌk] ‘pressure’ /ɑk- + -ljʌk/ → ɑŋljʌk → [ɑŋnjʌk] ‘grip’ Consider the urther examples given in (12). Firstly, in (12a–c), aspirated stops are replaced by lax stops due to Neutralisation. Ten, they are urther nasalised. In (12d), tense stops are replaced by lax stops and then nasalised. In (12e), ricatives and affricates are neutralised into /t/ and then /t/ is urther nasalised into /n/. (12) Neutralisation and Nasalisation a. /pʰ/ → p → [m] /ɑpʰ + mɑtɑŋ/ → ɑpmɑtɑŋ → [ɑmmɑtɑŋ] ‘ront yard’ /ipʰ + nun/→ ipnun → [imnun] ‘lea bud’ b. /tʰ/ → t → [n] /kjʌtʰ + nun/ → kjʌtnun → [kjʌnnun] ‘a side glance’ /k ʌtʰ + mosɯp/ → k ʌtmosɯp → [k ʌnmosɯp] ‘appearance’ c. /k ʰ/ → k → [ŋ] /puʌk ʰ + mun/ → puʌkmun → [puʌŋmun] ‘a kitchen door’ d. /k*/ → k → [ŋ] /k*ɑk*- + -nɯn/ → k*ɑknɯn → [k*ɑŋnɯn] ‘to peel (adnominal)’ e. /s, s*, h, ʨ, ʨʰ/ → t → [n] /os + mɛpsi/ → otmɛpsi → [onmɛps*i] ‘style o clothes’ /is*- + -nɯn/ → itnɯn → [innɯn] ‘to be (adnominal)’ /noh- + -nɯn/ → notnɯn → [nonnɯn] ‘to lay (adnominal)’ /nɑʨ + mɑl/ → nɑtmɑl → [nɑnmɑl] ‘words spoken in the daytime’ /k*oʨʰ + mɑl/ → k*otmɑl → [k*onmɑl] ‘the language o �owers’ 8.2.3
Liquid Nasalisation When a sequence o a non-alveolar nasal and a liquid occurs in Korean, that is, two kinds o sonorant appear in a row, the liquid becomes a nasal. Te nasalisation o liquids is the result o a surace phonetic constraint in Korean,
191
8.2
Phonological rules via surface phonetic constraints
which orbids a sequence o a non-alveolar nasal plus a liquid /l/. Hence, [m–l] and [ŋ–l] sequences are replaced by [m–n] and [ŋ–n] sequences respectively, as shown in (13). (13) a. /m/ + /l/ → [m] + [n] /k ɯm- + -li/ → [k ɯmni] ‘interest rate’ /tɑm- + -lon/ → [tɑmnon] ‘discussion’ b. /ŋ/ + /l/ → [ŋ] + [n] /koŋ- + -li/ → [koŋni] ‘public interest’ /tɯŋ- + -lok/ → [tɯŋnok] ‘registration’ One might ask what happens in the case o the dental nasal /n/ plus liquid /l/. Just as in the case o non-dental nasals plus liquid /l/, a sequence consisting o /n/ and /l/ cannot be realised due to a surace phonetic constraint. However, as we shall see in Chapter 9, unlike the examples in (13), /n/ and /l/ can be assimilated into either a sequence o nasals, /n/ and /n/, or a sequence o liquids, /l/ and /l/. Since the phonetic realisation o the sequence o /n/ and /l/ depends on how the word is ormed rather than purely on phonetic environment, we will discuss this in Chapter 9.
DID YOU KNOW . . . ?
Who is [hɛmnit]?
Te nasalisation o liquid sounds is so ingrained in Korean that it is even applied to loan words, and affects the pronunciation o oreign words. Because o this principle, Hamlet cannot but be pronounced as [hɛmnit] in Korean. Korean speakers are restrained by the nasalisation principle even when they speak English, and cannot pronounce English words in the way they ought to be pronounced. Native Korean speakers who are learning English, unless they are particularly attuned to their own pronunciation, would pronounce Hamlet not as [hæmlɪt], but as [hεmnit]. Such changes in Korean occur even when there is a phonological phrase boundary, such as in /(kokoŋ) (lok ʰɛt)/ [kokoŋ| nok ʰɛt] ‘altitude rocket’. When Korean speakers are only saying the one word, /lok ʰɛt/, they would pronounce it as [lok ʰɛt], but when /lok ʰɛt/ is pronounced within the phrase o /(kokoŋ) (lok ʰɛt)/, regardless o whether a phonological phrase boundary exists or not, it is pronounced as [kokoŋnok ʰɛt] or [kokoŋ| nok ʰɛt], with [nok ʰɛt] in both cases.
192
8.2.4
Phonological rules of Korean (I)
Lateralisation Not only do liquids become nasalised as we have discussed in 8.2.3, but nasals also become lateralised when the two sounds co-occur, as shown in (14). Again, this rule is the consequence o a surace phonetic constraint in Korean, which orbids liquid–nasal sequences, /l/–/n/. (14) a. Sino-Korean words with /l/–/n/ sequence () /sil- + -nɛ/ → [sillɛ] ‘interior’ () /ʨil- + -njʌ/ → [ʨilljʌ] ‘a niece’ b. Pure-Korean words with /l/–/n/ sequence /kjʌul + nɑl/ → [kjʌullɑl] ‘winter days’ /kw ɑil + nɑmu/ → [kw ɑillɑmu] ‘ruit trees’
8.2.5
Non-coronalisation (Bilabialisation and Velarisation) In this section, we will discuss two types o phonological processes that turn coronal sounds into peripheral (= non-coronal) sounds. Simply speaking, sounds that are articulated in the deault (i.e. coronal) position are either assimilated into the ront position or into the back position, depending on the sounds that ollow them. Te �rst type o assimilation includes Bilabialisation and the second type includes Velarisation. Bilabialisation can be de�ned as an alveolar sound becoming a bilabial sound when ollowed by a bilabial sound. In a similar way, Velarisation is de�ned as an alveolar or bilabial sound becoming velar when ollowed by a velar sound. Examples are given in (15). We can see that the two options are equally applicable in most cases. Tis is because both processes are known to be optional rather than obligatory. Non-coronalisation is more requently observed in spontaneous, inormal speech, in particular, more in ast speech than in slow speech. (15) a. /n/ → [m] /sin- + -mun/ → [sinmun ∼ simmun] ‘newspaper’ /sin + pɑl/ → [sinpɑl ∼ simpɑl] ‘shoes’ b. /n / → [ŋ] /hɑn- + k ɑŋ/ → [hɑnk ɑŋ ∼ hɑŋk ɑŋ] ‘the Han river’ /k ɑn- + -kjʌk/ → [k ɑnkjʌk ∼ k ɑŋkjʌk] ‘interval’ c. /m/ → [ŋ] /k ɑmki/ → [k ɑmki ∼ k ɑŋki] ‘cold’ /imk ɯm + -nim/ → [imk ɯmnim ∼ iŋk ɯmnim] ‘king’
193
8.2
Phonological rules via surface phonetic constraints
DID YOU KNOW . . . ?
Why is there no change in [k ɑmnɑmu] ‘persimmon tree’?
In a chain o ‘an alveolar sound and a bilabial/velar sound’ (in other words, [−coronal] sounds,) such as /sin- + -mun/ ‘newspaper’, Hangeul /hɑn + k ɯl/, or in a chain o ‘a bilabial and a velar sound’ such as in /k ɑmki/ ‘cold’, Non-coronalisation occurs. However, in a chain o ‘a bilabial and an alveolar sound’, such as in /k ɑm + nɑmu/ ‘persimmon tree’ no change occurs. We can make two observations rom this. First, Bilabialisation is a process in which the ollowing sound in�uences the preceding sound. Tus, when a bilabial sound is placed in ront o an alveolar sound, such as in /k ɑm + nɑmu/ no change occurs. Secondly, patterns o in�uence, which are based on place o articulation, differ depending on the category o the consonant. Although the bilabial sound /m/ in�uences the alveolar /n/ in /sin- + -mun/ [sinmun ∼ simmun] ‘newspaper’, this is not the case i the consonants are reversed. So /k ɑm + nɑmu/ can never be pronounced as [k ɑnnɑmu]. In addition, since /kamki/ ‘cold’ is pronounced as [k ɑŋki] but /k ɑŋ + mul/ ‘river water’ is not pronounced as [k ɑmmul], we can see that although velar sounds can in�uence bilabials, the opposite does not hold true; bilabial sounds cannot in�uence velars. Tereore, even within a �xed place o articulation, there is a hierarchy o sounds that determines how sounds in�uence one another. Here, we have observed that the hierarchy, “velar > bilabial > alveolar”, exists among places o articulation.
8.2.6
Aspiration In Korean, when /h/ and a lax obstruent appear together in the same sequence, they are contracted. As a result, the lax obstruent becomes aspirated. Examples are shown in (16). Shown in (16a) are examples where /h/ precedes lax obstruents, and (16b) shows examples where /h/ ollows lax obstruents. (16) a. /h/ + /k, t, ʨ/ → [k ʰ, tʰ, ʨʰ] /noh- + -ko/ → [nok ʰo] ‘to put (connective)’ /noh- + -tɑ/ → [notʰɑ] ‘to put (declarative)’ /noh- + -ʨʰi /→ [noʨʰi] ‘to put (connective)’ b. /k, t, p/ + /h/ → [k ʰ, tʰ, pʰ] /tok- + -hɑk/ → [tok ʰɑk] ‘sel-study’
194
Phonological rules of Korean (I)
/mɑt- + hjʌŋ/ → [mɑtʰ jʌŋ] ‘the eldest brother’ /kop + -hɑ- + -ki/ → [kopʰɑki] ‘multiplication’
Te examples in (17) show the two steps o the phonological processes Neutralisation and Aspiration, occurring consecutively. (17) /ius + -ha- + -ta/ → iuthɑtɑ → [iutʰɑtɑ] ‘to be close to each other’ /k*oʨʰ + hjɑŋki/ → k*othjɑŋki → [k*otʰ jɑŋki] ‘the scent o a �ower’ 8.2.7
Similar-place Obstruent Deletion In Korean, when a sequence contains two consecutive consonants that are pronounced in similar positions, one o them is deleted or ease o pronunciation. In natural speech, the �rst consonant in the sequence o obstruents is deleted. Consider (18).5 (18) a. In a bilabial sequence o /p/ and /pʰ/, the �rst consonant, /p/ is deleted. /ʨip- + -pʰil/ → [ʨipʰil] ‘writing’ /pɑp + pʰul/ → [pɑpʰul] ‘rice paste’ b. In a velar sequence o /k/ and /k ʰ/, the �rst consonant, /k/ is deleted /sik + k ʰɑl/ → [sik ʰɑl] ‘kitchen knie’ /k ʰik + k ʰik/ → [k ʰi k ʰik] ‘giggle’ c. In an alveolar sequence as below, the �rst consonant is deleted. /nɑs + -to/ → nɑtto → nɑtt*o → [nɑt*o] ‘sickle-even’ /nɑʨ + -to/ → nɑtto → nɑtt*o → [nɑt*o] ‘daytime-even’ /nɑʨʰ+ -to/ → nɑtto → nɑtt*o → [nɑt*o] ‘ace-even’ /os + ʨɑŋ/ → otʨɑŋ → otʨ*ɑŋ → [oʨ*ɑŋ] ‘closet’ /nɑʨ + -ʨʰʌlʌm/ → nɑtʨʰʌlʌm → [nɑʨʰʌlʌm] ‘like daytime’ /os + sol/ → otsol → ots*ol → [os*ol] ‘clothes brush’
FURTHER STUDY
Pronouncing and : dictionary transcription vs. real pronunciation
How are /ʨʌʨ + so/ ‘cow’ and /os + sol/ ‘clothes brush’ pronounced?
195
8.2
Phonological rules via surface phonetic constraints
Te Korean dictionary stipulates that the pronunciation o is [ʨʌts*o] and the pronunciation o is [ots*ol]. However, these are unlikely to be how the words are actually pronounced, because obstructive sounds in similar articulation positions are not eliminated. When Korean speakers say these words naturally, they pronounce them as [ ʨʌ.s*o] and [o.s*ol] respectively by eliminating the stop sounds. It is hard to conceive o [ʨʌt.s*o] or [ot.s*ol] as natural pronunciations o the words. Te �gure shows a spectrogram o a standard Korean speaker’s pronunciation o /os + sol/ . As this shows, the ricative region (marked by the bold square in the picture) can only be observed between vowels in the way that /os + sol/ is actually pronounced. I it were pronounced as [ots *ol], a stop region (marked by a white space on the spectrogram) ought to be detected prior to the ricative region, but this does not happen in natural speech. Essentially, the actual pronunciation o /os + sol/ is [os*ol]. Spectrogram 8000
Hz
6000 4000 2000
Waveform
o
s*
o
l
0
1.0 Time (s)
8.2.8
/j/ Deletion In Korean, [ʨ, ʨ*, ʨʰ] cannot appear with a [j] sound. Tereore, no words in Korean have any o the ollowing sequences: [ ʨ jʌ, ʨ* jʌ, ʨʰ jʌ] or [ʨ jo, ʨ* jo, ʨʰ jo], etc. Hence, when verbs whose stems have /ʨ, ʨ*, ʨʰ/ and an /i/ vowel such as /ʨi- + -tɑ/ ‘to lose’, /ʨ*i- + -tɑ/ ‘to steam’ or /ʨʰi- + -tɑ/ ‘to hit’ are in�ected with a verbal ending starting with – /-ʌ/, they are pronounced with an /ʌ/ vowel instead o an /j ʌ/ vowel as in (19), as [j] is deleted. (19) /ʨi- +-ʌsʌ/ → ʨ jʌsʌ → [ʨʌsʌ] ‘to lose (connective)’ /ʨ*i- + -ʌsʌ/ → ʨ* jʌsʌ → [ʨ*ʌsʌ] ‘to steam (connective)’ /ʨʰi- + -ʌsʌ/ → ʨʰjʌsʌ → [ʨʰʌsʌ] ‘to hit (connective)’
196
Phonological rules of Korean (I)
However, [j] is not deleted in all cases, as shown in (20), where the preceding consonants are not /ʨ, ʨ*, ʨʰ/. (20) /kuki- + -ʌsʌ/ → [kukjʌsʌ] ‘to wrinkle (connective)’ /k*umi- + -ʌsʌ/ → [k*umjʌsʌ] ‘to decorate (connective)’
8.2.9
/h/ Deletion In natural speech in Korean, an /h/ appearing between voiced sounds is avoided. Tis is more requently observed in spontaneous, inormal speech, in particular, more in ast speech than in slow speech. Examples are given in (21). (21) a. /h/ between vowels /w ɛ- + hɑlmʌni/ → [w ɛhɑlmʌni ∼ w ɛɑlmʌni] ‘grandmother’ /tɛ- + -hɑk/ → [tɛhɑk ∼ tɛɑk] ‘university’ b. /h/ between a vowel and a sonorant consonant /jʌŋ- + -hjɑŋ/ → [jʌŋhjɑŋ ∼ jʌŋjɑŋ] ‘in�uence’ /mun- + -hw ɑ/ → [munhw ɑ ∼ munw ɑ] ‘culture’ /sil- + -hjʌn/ → [silhjʌn ∼ siljʌn] ‘realisation’
FURTHER STUDY
Writing perfect Korean
Writing in Korean is not easy because the way people pronounce words and the way they write them down are not the same. Dictation training is one o the key eatures o Korean language education at school. However, even i one is thoroughly trained in dictation, it is still not easy to write Korean without mistakes. ake the example o /(kokmul + -ɯl) (pɑʨʰi- + -tɑ)/ [koŋmulɯl | pɑʨʰitɑ] ‘to offer grain’. Tis sounds exactly the same as /(koŋmul + -ɯl/) (pɑʨʰi- + -tɑ)/ [koŋmulɯl | pɑʨʰitɑ] ‘to offer an official gif’. Tere are many examples like this. For instance, /ɑk*i- + -tɑ/ [ɑk*itɑ] ‘to save’ sounds exactly the same as /ɑk-+ -ki + -tɑ/ [ɑk*itɑ] ‘(it is a) musical instrument’. Likewise, /k ɯm- + -li/ [k ɯmni] ‘interest’ sounds exactly the same as /k ɯm + ni/ [k ɯmni] ‘gold teeth’ due to Nasalisation. Te only way to
197
8.3
Summary
determine the correct orthography rom two identical sound sequences is to determine the correct meaning or the given context.
8.3
Summary In this chapter, we have seen phonological rules that are particularly moti vated by syllable structure constraints and surace phonetic constraints. Rules affected by syllable structure constraints include Neutralisation and Consonant-cluster Simpli�cation. Rules affected by surace phonetic constraints include (i) Post-obstruent ensi�cation; (ii) Obstruent Nasalisation; (iii) Liquid Nasalisation (iv) Lateralisation; (v) Non-coronalisation (Bilabialisation and Velarisation); (vi) Aspiration; (vii) Similar-place Obstruent Deletion; (viii) /j/ Deletion; and (ix) /h/ Deletion. able 8.4 shows a summary o the phonological rules discussed in this chapter.
Table 8.4 Phonological rules of Korean (I)
Rule
Core example: underlying form
Core example: pronunciation
Neutralisation
/puʌk ʰ/
Consonant-cluster simpli�cation Post-obstruent tensi�cation Obstruent nasalisation
/sɑlm/
[puʌk] ‘kitchen’ [sɑm] ‘lie’
/mʌk- + -tɑ/ /jʌpʰ + ʨip/ → jʌpʨip
[mʌkt*ɑ] ‘to eat’ [jʌpʨ*ip] ‘next door’
/pɑp + mɑs/ → papmɑt /kjʌp + nun/
Liquid nasalisation Lateralisation
/k ɯm- + -li/
[pɑmmɑt] ‘taste o rice’ [kjʌmnun] ‘compound eyes’ [kɯmni] ‘interest rate’
/kjʌul + nɑl/
[kjʌullɑl] ‘winter days’
Non-coronalisation (bilabialisation and velarisation)
/sin- + -mun/ /hɑn- + k ɑŋ/ /k ɑmki/
Aspiration
/noh- + -tɑ/ /tok- + -hɑk/
[sinmun∼simmun] ‘newspaper’ [hɑnk ɑŋ∼hɑŋk ɑŋ] ‘the Han river’ [k ɑmki∼k ɑŋki] ‘cold’ [notʰɑ] ‘to put’ [tok ʰɑk] ‘sel-study’
Similar-place obstruent deletion
/ʨip- + -pʰil/ /os + ʨɑŋ/ →
[ʨipʰil] ‘writing’ [oʨ*ɑŋ] ‘closet’
/j/ deletion
otʨɑŋ → otʨ*ɑŋ /ʨi- +-ʌsʌ/ → ʨ jʌsʌ
[ʨʌsʌ] ‘to carry (connective)’
/tɛ- + -hɑk/ /jʌŋ- + -hjɑŋ/
[tɛhɑk ∼tɛɑk] ‘university’ [jʌŋhjɑŋ ∼ jʌŋjɑŋ] ‘in�uence’
/h/ deletion
Phonological rules of Korean (I)
198
EXERCISES 1
2
3 4
5
6
7
8
9
10
11 12
Explain why /pɑt-/ - ‘to accept’ and /p ɑtʰ/ ‘field’ are pronounced the same. Explain why the /tʰ/ in /pɑtʰ/ and /pɑtʰ + -ɛ/ ‘field (prepositional particle)’ or /pɑtʰ + -ɯl/ ‘field (object particle)’ are pronounced differently. Explain why a consonant cluster cannot occur in syllable-final position. Is it natural to pronounce /os/ ‘clothes’ as [os] and /oʨʰ/ ‘sap of the lacquer tree’ as [oʨʰ] in Korean? If not, explain why. Please pronounce the following and explain the processes of phonetic realisation for these words. a. /ɑnʨ- + -ko/ ‘to sit down (connective)’ b. /mɑnh- + -so/ ‘to be plentiful (declarative)’ c. /sɑlm- + -tɑ/ ‘to boil (declarative)’ d. /nʌlp- + -ko/ ‘to be wide (connective)’ e. /ɯlpʰ- + -nɛ/ ‘to recite (exclamative)’ Explain why /s/ in /sɛk/ ‘colour’ is tensified in (a) but not in (b). a. /nok + sɛk/ ‘green’ b. /norɑn + sɛk / ‘yellow’ Illustrate the phonological processes by which /pu ʌkʰ + -to/ ‘kitchen (additional meaning particle)’ is pronounced as [pu ʌkt*o], using relevant phonological rules. Illustrate the phonological processes by which /kuk + mul/ ‘soup-liquid’ is pronounced as [kuŋmul] using relevant phonological rules. Illustrate the phonological processes by which /ko ŋ- + -li/ ‘public interest’ is pronounced as [koŋni] using relevant phonological rules. Illustrate the phonological processes by which /kj ʌul + nɑl/ ‘winter days’ is pronounced as [kjʌullɑl] using relevant phonological rules. Why is /ʨip- + -pʰil/ ‘writing’ pronounced as [ʨipʰil] instead of [ ʨippʰil]? Why is /h/ in the following words very often not pronounced? a. /mun- + -hwɑ/ ‘culture’ b. /tɛ- + -hɑk/ ‘university’ c. /wɛ- + hɑlɑpʌʨi/ ‘maternal grandfather’ d. /ju- + -hjʌŋ/ ‘type’
9
9.1
Phonological rules of Korean Korean (II) (II )
Background In Chapter 8, 8, we discussed phonological rules related mainly to surace phonetic constraints and syllable structure constraints. Te phonological rules discussed in Chapter 8 are applied to a sequence o sounds regardless o morphological inormation. For instance, a /k/–/m/ sequence in Korean automatically automatically changes into a /ŋ/–/m/ sequence regardless o the morphological structure or origin o a word. Tis is the case or all the ollowing types o word. In this chapter, we will use + or morpheme boundary, # or word boundary and - to indicate bound morpheme. (1)
a. Morphological Morphological structure Compounding: /kuk # mul/ → [kuŋmul] ‘soup liquid’ Derivation: /ʨʰʌs- # nɑl/→ ʨʰʌtnɑl →[ʨʰʌnnɑl] ‘�rst day’ b. Word origin orig in Pure Korean words: /kuk # mul/ → [kuŋmul] ‘soup liquid’ Sino-Korean words: () /kok- + -mul/→ [koŋmul] [koŋmu l] ‘grain’
However, not all phonological rules can be applied in this way. Indeed, knowing how and by what types o morpheme a word is ormed is crucial in understanding Korean phonology. phonology. For instance, ensi�cation is affected by the underlying morphological inormation o the word. Te nature o the morpheme boundary in a given type o a word also in�uences the application o ensi�cation. Hence, although orthographically the pairs o words in able 9.1 each look the same, only in the ormer is there ensi�cation, not in the latter. Tis shows that the way in which a word is ormed determines the pronunciation o that word. In this chapter, as above, we will discuss the phonological rules or which morphological inormation plays a crucial part in their application. Beore going urther, we will brie�y re-cap grammatical terms, 199
200
Phonological rules of Korean (II)
Table 9.1 Tense vs. lax contrast contrast for words with the same orthography orthography
Words
Second syllable onset consonant being tensi�ed
Morphologic Morphological al boundary: /ʨɑm # ʨɑli/ [ʨɑmʨ*ɑli] ‘bed’
Within a compound: /pɑl # pjʌŋ/ [pɑlp* jʌŋ] ‘sore eet’
Pure Korean verbal in�ection: /sin- + -ko/ [sink *o] ‘to put on (connective)’
Second syllable onset consonant not being tensi�ed (remains as lax consonant)
No morphological morphological boundary: boundary : /ʨɑmʨɑli/ [ʨɑmʨɑli] ‘dragon�y’ Within a simple word: () /pɑl- + -pjʌŋ/ [pɑlpjʌŋ] ‘to become sick’ Sino-Korean stem: () /sin- + -ko/ [sinko] ‘declare’
since understanding how and rom which elements a word is ormed is crucial in understanding its phonological realisation. Te examples in (2) and (3) show how a morpheme, the smallest meaning-bearing unit, can be urther classi�ed. Morphemes can be divided into either ree or bound morphemes and into lexical or grammatical morphemes. Below we provide examples o each type. (2)
(3)
a. Free morpheme (i.e. can be used independently) /nun/ ‘eye’ /sɑlɑm/ ‘person’ ‘person’ /ʨɑmʨɑli/ ‘dragon�y’ b. Bound morpheme (i.e. can’t can’t be used independently) indepen dently) Pure Korean pre�x: - /ʨʰʌs/ ‘�rst’ as in /ʨʰʌs- # sɑlɑŋ/ ‘�rst love’ Pure Korean suffi x: - /-k*ili/ ‘together’ as in /ʨʰinku # -k*ili/ ‘together with riends’ Sino-Korean pre�x: ()- /ʨʰo-/ ‘extremely’ as in /ʨʰo- # hohw ɑ/ ‘Extremely luxurious’ Sino-Korean Sino -Korean suffi x: - () /-sik/ ‘style’ as in /jʌŋkuk # -sik/ ‘English style’ In�ection In�ec tional al suffi xes: -- /-si-/ ‘subject honori�cation’, -- /-ʌs*-/ ‘past tense’, - /-tɑ/ ‘declarative’ a. Lexical morpheme (i.e. has lexical meaning) /nun/ ‘eye’ /sɑlɑm/ ‘person’ /ʨɑmʨɑli/ ‘dragon�y’ - /hɑ-/ ‘to do’ - /mʌk-/ ‘to eat’
9.2
201
Rules where two lexical lexical morphemes morphemes meet
b. Grammatical morpheme morpheme (i.e. has only grammatical grammatical unction) (i) particles: - /-i/ ‘subject particle’ par ticle’, - /-ɯl/, ‘object particle’, /-to/ ‘too, also’ (ii) derivational derivationa l suffixes: - /-i/ /-i / ‘causati ‘caus ative ve suffi s uffi x’, - /-hi/ ‘passiv ‘pas sivee suffi x’ (iii) in�ectional in�ect ional suffi xes: --/-si-/ -/-si- / ‘subject honori�cation honori�cati on’’, - /-ʌs*-/ ‘past tense’, - /-tɑ/ ‘declarative’ In the ollowing, in 9.2, we will discuss disc uss the phonological rules applied where one lexical morpheme meets another and in 9.3, we will discuss rules r ules that are applied applied where a lexical morpheme meets a grammatical morpheme.
9.2
Rules where two lexical morphemes meet In this section, we will discuss dis cuss phonological rules that are applied applied at the boundary between two lexical morphemes. Lexical morphemes coincide with ree morphemes in most cases but not all the time. As noted above, Sino-Korean vocabulary constitu c onstitutes tes a major part o Korean vocabulary. Te phonological phonological rules to be discussed here are mainly observed between Sino-Korean lexical morphemes. In particular, particular, we will ocus on the t he Lateralisation and Nasalisation o Sino-Korean words (9.2.1), the ensi�cation o Sino-Korean words (9.2.2), /t/ Insertion (9.2.3) and /n/ Insertion (9.2.4).
9.2.1
Lateralisation and Nasalisation of Sino-Korean words
Why is the /nl/ sequence sometimes pronounced [ll] and at other times [nn]? When /n/ and /l/ appear in a row, /n/ is assimilated into /l/ when a morpheme boundary occurs between the two sounds.1 Tis rule is not, however, applicable when a word boundary occurs between the two sounds. In such cases, assimilaassim ilation occurs in the opposite direction, with /l/ becoming /n/. Consider (4). (4)
a. /n/ becomes /l/ when there is a morpheme boundary between the sounds () /kw ʌn- + -ljʌk/ → [kw ʌlljʌk] ‘power’ () /non- + -li/ → [nolli] ‘logic’ b. /l/ becomes /n/ when there is a word boundary between the sounds () /(pʰɑn- + -tɑn) # -ljʌk/ → [pʰɑntɑnnjʌk] ‘decision’ () /(ɯm- + -un) # -lon/ → [ɯmunnon] ‘phonology’
able 9.2 shows examples o Lateralisation and Nasalisation.
202
Phonological rules of Korean (II)
Table 9.2 Phonetic realisation of /nl/: /nl/: Lateralisation Lateralisation vs. Nasalisation
Words
/nl/ → [ll] (i.e. Lateralisation)
/nl/ → [nn] (i.e. Nasalisation)
() /(ɯm- + -un) #
N/A
[ɯmunnon]
N/A
[koŋk *w ʌnnjʌk]
N/A
[toŋwonnjʌŋ]
[kw ɑŋhɑllu]
N/A
[nɑllo]
N/A
-lon/ ‘phonology’ () /(koŋ- + -kw ʌn) #
-ljʌk/ ‘governmental authority’ ()
/(toŋ- + -won) # -lj ʌŋ/ ‘mobilisation order’ () /kw ɑŋ- + -hɑn- +
-lu/ ‘a tower in Namweon’ () /nɑn- + -lo/ ‘a heater’
DID YOU KNOW . . . ?
Knowing word structure is important in pronouncing a word!
Sometimes whether the /nl/ sequence should be lateralised [ll] or nasalised [nn] is ambiguous. Moreover, morphological inormation plays an important role in deciding the appropriate phonetic realisation, as shown in the example below. below. In (a), i one means a particular noodle eaten in the Silla period in Korean history (668–935 AD), then the word structure is /(sin- + -lɑ) # mjʌn/ and /sin- + -lɑ/ orms one word together. As a result, Lateralisation will be applicable and the pronunciation o the word will be [sillɑmjʌn]. On the other hand, i one means a particular instant noodle (ramen) brand, ‘Sin-ramen’ (), then the word structure is /sin- # (lɑmjʌn)/ and /sin/ and /lɑ/ do not orm one word together. As a result, Nasalisation will be applicable and the pronunciation o the word will be [sinnɑmjʌn]. a. (# ) /(sin- + -lɑ) # mjʌn/ → [sillɑmjʌn] (Lateralisation applicable, applicable, Nasalisation Nasalisation not applicable) b. (# ) /sin- # lɑmjʌn/ → [sinnɑmjʌn] (Lateralisation not applicable, Nasalisation applicable)
203
9.2.2
9.2
Rules where two lexical lexical morphemes morphemes meet
Tensification of Sino-Korean words
Why is /kjʌl + tɑn/ ( ) pronounced [kjʌlt*ɑn]? A sequence consisting o an /l/ and a lax consonant with [+coronal] eatures such as /t, s, ʨ/ may not appear in a Sino-Korean word. Instead, those lax consonants are pronounced as /t*, s*, ʨ*/. However, when /l/ is ollowed by [−coronal] lax consonants in a Sino-Korean word, this rule is not applicable. Consider (5). (5)
ensi�cation ensi�cati on o Sino-Korean words a. /l/ ollowed by coronal lax consonant: ensi�cation ensi�cat ion applicable () /kjʌl- + -tɑn/ [kjʌlt*ɑn] ‘decision’ () /kjʌl- + -sʌn/ [kjʌls*ʌn] ‘�nal (competitio ( competition)’ n)’ () /pɑl- + -ʨʌn/ [pɑlʨ*ʌn] ‘improvement’ b. /l/ ollowed by non-coronal lax consonant /k/ or /p/: ensi�cation ensi�cat ion non-applicable () /kjʌl- + -kw ɑ/ [kjʌlkw ɑ] ‘result’ () /pɑl- + -kjʌn/ [pɑlkjʌn] ‘�nding’ () /kjʌl- + -pɑk/ [kjʌlpɑk] ‘strap’
In (5a), /t, s, ʨ/ become /t*, s*, ʨ*/ according to the rule. However, in (5b), the rule is not applied. In addition, this rule is applied to Sino-Korean words only.
9.2.3
/t/ Insertion
Why is the same /pi/ pronounced differently in /pom # pi/ and /nun # pi/ ? In Chapter 8, 8, we saw that lax consonants are tensi�ed when ollowed by another obstruent consonant due to the surace phonetic constraint which orbids a sequence o obstruents. Tis rule (Post-obstruent ensi�cation) is automatically applied regardless o how the word is ormed. However, However, i we turn our attention attent ion to the pronunciation o /pom /p om # pi/ [pomp *i] ‘spring rain’ rain’ and /nun # pi/ [nunpi] ‘snow and rain’, the ensi�cation rule appears to be less predictable except in the cases o obstruent–obstruent sequences. Te same /pi/ is tensi�ed in /pom # pi/ ‘spring rain’ but not in /nun # pi/ ‘snow and rain’. Why is this the case? Is there any way o making a prediction as to whether ensi�cation should be b e applied? Some argue that ensi�cation ensi�cation via /t/ Insertion Inserti on
204
Phonological rules of Korean (II)
can be predicted, in that it occurs when a subordinate relation holds between sub-compounds instead o a coordinate relation. In /pom # pi/ ‘spring rain’, /pom/ ‘spring’ modi�es /pi/ ‘rain’. Hence, a subordinate relation holds in the compound. As the result o this, /pi/ ‘rain’ is tensi�ed. On the other hand, in /nun # pi/ ‘snow and rain’, a coordinate, parallel relation exists between /nun/ ‘snow’ and /pi/ ‘rain’. Hence, /pi/ ‘rain’ is not tensi�ed. (6)
a. /t/ Insertion occurs in subordinated compound words. /pom # pi/ → pom # t # pi (/t/ Insertion) → pom t p* i (Post-obstruent ensi�cation) pom t p* i (Consonant-cluster Simpli�cation) → [pomp*i] ‘spring rain’ /sɑŋ # tɑli/ → sɑŋ # t # tɑli (/t/ Insertion) → sɑŋ t t*ɑli (Post-obstruent ensi�cation) → sɑŋ t*ɑli (Consonant-cluster Simpli�cation) → [sɑŋt*ɑli] ‘table legs’ /sɑn # sɛ/ → sɑn # t # s ɛ (/t/ Insertion) → sɑn t s*ɛ (Post-obstruent ensi�cation) → sɑns*ɛ (Consonant-cluster Simpli�cation) → [sɑns*ɛ] ‘mountain bird’ b. /t/ Insertion does not occur in coordinated compound words. /nun # pi/ → [nunpi] ‘snow and rain’ /pom # k ɑɯl/ → [pomk ɑɯl] ‘spring and all’ /son # pɑl/ → [sonpɑl] ‘hands and eet’
In this book, we argue that the pronunciation o /pom # pi/ becoming [pomp*i] cannot be simply explained by ensi�cation, but must be explained as ensi�cation via /t/ Insertion. Tis is due to examples such as /wi # os/ [witot] ‘top clothes’ or /ɑlɛ # mɑɯl/ [ɑlɛnmɑɯl] ‘next village’, where ensi�cation itsel cannot be applicable, since the word-initial sound or the second sub-compound starts with a vowel or a nasal consonant instead o a lax consonant. In order to allow or examples such as /wi # os/ ‘top clothes’ becoming [witot] and /ɑlɛ # mɑɯl/ ‘next village’ becoming [ ɑlɛnmɑɯl], we argue that /pom # pi/ becoming [pomp*i] is not merely the result o ensi�cation, but o ensi�cation via /t/ Insertion. (7) shows the phonological processes resulting in the surace orm o these two words. (7)
Phonological processes or /wi # os/ and /ɑlɛ # mɑɯl/ /wi # os/ → wi # t # os → witot (Neutralisation) → [witot] ‘top cloth’ /ɑlɛ # mɑɯl/ → ɑlɛ # t # mɑɯl (/t/ Insertion) → ɑlɛnmɑɯl (Obstruent Nasalisation) → [ɑlɛnmɑɯl] ‘next village’
However, counter-examples can easily be ound. For instance, /koki/ is pronounced differently in /mul # koki/ ‘�sh’ and /pul # koki/ ‘marinated bee ’: it is tensed in [mulk*oki] but not in [pulkoki]. More examples
205
9.2
Rules where two lexical morphemes meet
Table 9.3 Tensification via /t/ Insertion variation among subordinated compounds
Compounds where the preceding word is the same
/t/ Insertion applicable
/t/ Insertion not applicable
/pul # piʨʰ/ → pul # t
/pul # pjʌtʰ/ →
# piʨʰ
* → [pulp it]
‘light’
/tol # tʌŋʌli/ →
Compounds where the ollowing word is the same
[pulpjʌt] ‘burning sun’ /tol # kituŋ/ →
tol # t# tʌŋʌli → [tolt*ʌŋʌli] ‘rock’
[tolkituŋ] ‘stone pillar’
/s*ɑl # k ɑk ɛ/ → s*ɑl
/s*ɑl # pɑp/ →
# t # k ɑk ɛ → [s*ɑlk *ɑk ɛ] ‘rice shop’
[s*ɑlpɑp] ‘boiled rice’
/k *um # kjʌl/ → k *um
/k *um # ʨɑli/ →
# t # kjʌl → [k *umk * jʌl] ‘a dreamy state’
[k *umʨɑli] ‘a dream’
/k ʰoŋ # k ɑlu/ →
/k ʰoŋ # kilɯm/ →
k ʰoŋ # t # k ɑlu → [k ʰoŋ k *ɑlu] ‘bean �our’
[k ʰoŋkilɯm] ‘soybean oil’
/k ɑŋ # pɑtɑk/
/kol # pɑtɑk/ →
k ɑŋ # t # pɑtɑk → [k ɑŋp*ɑtɑk] ‘river bed’
[kolpɑtɑk] ‘the lowest part o the valley’
/k ɯmɯm # tɑl/
/pɑn # tɑl/ →
k ɯmɯm # t # tɑl → [k ɯmɯmt*ɑl] ‘dark moon’
[pɑntɑl] ‘hal moon’
/mul # sɛ/ → mul # t #
/pʰɑlɑŋ # sɛ/ →
→
→
sɛ
* → [muls ɛ]
‘water bird’
[pʰɑlɑŋ sɛ] ‘bluebird’
/kjʌul # ʨɑm/
/tupʌl # ʨɑm/ →
kjʌul # t # ʨɑm → [kjʌulʨ*ɑm] ‘hibernation’
[tupʌlʨɑm] ‘the second sleep that one alls into afer the �rst sleep’
/mɑl # ʨɑŋnɑn/ →
/pul # ʨɑŋnɑn/
→
mɑl # t # ʨɑŋnɑn → [mɑl ʨ*ɑŋnɑn] ‘pun’
[pulʨɑŋnɑn] ‘play with �re’ →
like this one can be ound in able 9.3. able 9.3 shows cases where the ensi�cation rule resulting rom /t/ Insertion is applied as well as the cases where the rule is not applicable.
206
Phonological rules of Korean (II)
Table 9.4 Pronunciation of days of the week showing /n/ Insertion in Gyeongsang dialect
9.2.4
Words
Seoul
/w ʌl # joil/ ‘Monday’
[wʌ ljoil]
wʌ l # n # joil → [w ʌlljoil] (urther Lateralisation)
/mok # joil/ ‘Tursday’
[mokjoil]
mok # n # joil → [moŋnjoil] (urther Nasalisation)
/k ɯm # joil/ ‘Friday’
[kɯ mjoil]
kɯ m # n # joil → [k ɯmnjoil]
Gyeongsang
/n/ Insertion /n/ Insertion can happen when the preceding word ends with a consonant and the ollowing word starts with /i/ or /j/.Tereore, even though two words may be ormed in a similar way, i their phonological environments differ, /n/ Insertion may occur in one word but not in the other. (8) shows examples o /n/ Insertion in compounding and derivation. (8)
a. /n/ Insertion in compounding /som # ipul/ → som # n # ipul → [somnipul] ‘a cotton-wool duvet.’ /tɯl # il/ → tɯl # n # il → tɯlnil → [tɯllil] ‘arm work’ (urther Lateralisation) /puk # julʌp/ → [pukjulʌp] ∼ [puŋnjulʌp] ‘Northern Europe’ /sɛk # jʌnpʰil/ → [sɛkjʌnpʰil] ∼ [sɛŋnjʌnpʰil] ‘coloured pencil’ b. /n/ Insertion in derivation /nɯʨ- # jʌlɯm/ → nɯʨ-n # jʌlɯm → nɯtnjʌlɯm → [nɯnnjʌlɯm] ‘late summer’ /hɑn- # jʌlɯm/ → hɑn- n # jʌlɯm → [hɑnnjʌlɯm] ‘midsummer’ /kjʌp- # ipul/→ kjʌp- n # ipul → kjʌpnipul →[kjʌmnipul] ‘layered blanket’ /mɑk- # il/ → mɑk- n # il → mɑknil → [mɑŋnil] ‘physical labour’ (urther Nasalisation)
Te /n/ Insertion rule varies rom dialect to dialect. In North Korean dialects, /n/ Insertion is not as requently observed. For instance, in North Korean, /po- + -l # il/ ‘something to do’ is pronounced as [polil], instead o [pollil] as in the South. On the other hand, /n/ Insertion is widespread in the Gyeongsang dialect. So, or instance, in the examples in able 9.4, reerring to days o the
9.3
207
Rules between lexical morphemes and grammatical morphemes
week, Seoul Korean speakers will pronounce each word without /n/ Insertion, but Gyeongsang dialect speakers will pronounce them with /n/ inserted.
9.3
Rules between lexical morphemes and grammatical morphemes Some phonological phenomena are observed between lexical and grammatical morphemes. Tese include Palatalisation (9.3.1); Verbal Suffix ensi�cation (9.3.2); Glide Formation in Verbal In�ection (9.3.3); /h/ Deletion in Verbal In�ection (9.3.4); /ɯ/ Deletion in Verbal In�ection (9.3.5); / ɑ/ or /ʌ/ Deletion in Verbal In�ection (9.3.6).
9.3.1
Palatalisation
Why is /kɑtʰ + i/ never pronounced as [kɑtʰi] but as [kɑʨʰi]? When a lexical morpheme ending with /t, t ʰ/ is ollowed by a grammatical morpheme beginning with /i/, /t, t ʰ/ is palatalised into / ʨ, ʨʰ/. (9)
a. Noun + subject particle /patʰ + -i/ → [pɑʨʰi] ‘�eld (subject particle)’ b. Noun + copula /patʰ + -i- + -tɑ/ → [pɑʨʰitɑ] ‘�eld (copula)’ c. Noun + conjunctive /patʰ + -ilɑŋ/ → [pɑʨʰilɑŋ] ‘�eld (conjunctive)’ d. Adverbial suffi x /k ɑtʰ- + -i/ → [k ɑʨʰi] ‘together’ /kut- + -i/ → [kuʨi] ‘willingly’ e. Nominalising suffi x /mitɑt- + -i/ → [mitɑʨi] ‘sliding door’ . Causative suffi x /putʰ- + -i- + -tɑ/ → [puʨʰitɑ] ‘to stick’
However, can be pronounced not only as [pɑʨʰilɑŋ] but also as [pɑnnilɑŋ]. Consider (10). In (10a), /patʰ + -ilɑŋ/ is pronounced [pɑʨʰilɑŋ]. Tat is, /t/ is palatalised. But in (10b), /pɑtʰ # ilɑŋ/ is pronounced [pɑnnilɑŋ], with /n/ Insertion. (10) a. Palatalisation /patʰ + -ilɑŋ/ → [pɑʨʰirang] ‘�eld (conjunctive particle)’ ( /-ilɑŋ/ is a grammatical morpheme meaning ‘and’)
208
Phonological rules of Korean (II)
b. Neutralisation and /n/ Insertion /pɑtʰ # ilɑŋ/ → pɑtʰ # n # il ɑŋ (/n/ Insertion) → pɑtnilɑŋ (Neutralisation) → [pɑnnilɑŋ] (Obstruent Nasalisation) ( /ilɑŋ/ is a lexical morpheme meaning ‘urrow’) Why is this the case? Tis is because Palatalisation occurs between a lexical morpheme and a grammatical morpheme. In (10a), /-il ɑŋ/ is used as a conjunctive particle, that is, a grammatical morpheme. Tereore, Palatalisation is applied and the pronunciation becomes [p ɑʨʰilɑŋ]. On the other hand, in (10b), /ilɑŋ/ is used as a lexical morpheme. Hence, Palatalisation is not applicable and the pronunciation becomes [p ɑnnilɑŋ] as a result o applying /n/ Insertion and subsequent Neutralisation and Obstruent Nasalisation.
9.3.2
Verbal suffix Tensification: after a nasal-ending stem When a lax consonant, such as /k, t, s, ʨ/, is preceded by a verbal stem ending with a nasal /n, m/, the lax consonant is tensed and pronounced as /k *, t *, s *, ʨ*/. Tis ensi�cation rule is applicable only in verbal in�ections between the stem o a verb and verbal suffix. Consider the ollowing. (11) Verbal stem + suffi x a. - /ɑn-/ ‘to hug’ /ɑn- + -ko/ → [ɑnk*o] ‘to hug (connective)’ /ɑn- + -tɑ/ → [ɑnt*ɑ] ‘to hug (declarative)’ /ɑn- + -so/ → [ɑns*o] ‘to hug (declarative)’ /ɑn- + -ʨɑ/ → [ɑnʨ*ɑ] ‘to hug (propositive)’ b. - /k ɑm-/ ‘to coil’ /k ɑm- + -ko/ → [k ɑmk*o] ‘to coil (connective)’ /k ɑm- + -tɑ/ → [k ɑmt*ɑ] ‘to coil (declarative)’ /k ɑm- + -so/ → [k ɑms*o] ‘to coil (declarative)’ /k ɑm- + -ʨɑ/ → [k ɑmʨ*ɑ] ‘to coil (propositive)’ c. - /sɑlm/ ‘to boil’ /sɑlm- + -ko/ → sɑlmk*o → [sɑmk*o] ‘to boil (connective)’ /sɑlm- + -tɑ/ → sɑlmt*ɑ → [sɑmt*ɑ] ‘to boil (declarative)’ /sɑlm- + -so/ → sɑlms*o → [sɑms*o] ‘to boil (declarative)’ /sɑlm- + -ʨɑ/ → sɑlmʨ*ɑ → [sɑmʨ*ɑ] ‘to boil (propositive)’
209
9.3
Rules between lexical morphemes and grammatical morphemes
FURTHER STUDY
In addition to the ensi�cation discussed above, another type o ensi�cation occurs between a grammatical morpheme and a lexical morpheme. When a noun starting with a lax consonant ollows the adnominal ending /-ɯl/ or /-l/ –/-, the lax consonant is tensi�ed. When the ollowing lexical morpheme is a dependent morpheme (noun), ensi�cation is obligatory. However, when the ollowing lexical morpheme is an independent morpheme, ensi�cation occurs only when the two words orm one phonological phrase. a. Adnominal ending plus dependent nouns (orming one phonological word) /k ɑ- + -l # -kos/ → [k ɑlk*ot] ‘the place to go’ /nol- + -l # -tɛ/ → [nolt*ɛ] ‘the place to play’ b. Adnominal ending plus independent nouns (orming one phonological phrase) /k ɑ- +- l # sɑlɑm/ → [k ɑls*ɑlɑm] ‘someone to go’ /nol- + -l # ʨɑŋso/ → [nolʨ*ɑŋso] ‘the place to play’ c. Adnominal ending plus independent nouns (not orming one phonological phrase) /k ɑ- + pʌli- + -l # s ɑlɑm + -ɛk ɛ/ → [k ɑpʌlil | sɑlɑmɛk ɛ] ‘to someone to go’ /mʌk- + -ko #nol- + -l # ʨɑŋso + -ɛsʌ/ → [mʌk*onol | ʨɑŋsoɛsʌ] ‘at the place to eat and play’
9.3.3
Glide Formation in verbal inflection When a verbal stem ending with a consonant and /i/ sequence is in�ected with a suffi x beginning with /ʌ/, /i/ ofen becomes a glide. Tis is because vowel– vowel sequences are avoided in Korean. /j/ Glide Formation is more requently observed when the verbal stem in question has more than two syllables. (12) a. Words with one syllable verbal stem: /i/ Glide Formation is optional. /ki- + -ʌsʌ/ ∼ [kiʌsʌ] ∼ [kjʌsʌ] ‘to crawl (connective)’ b. Words with more than two syllable verbal stem: /j/ Glide Formation is more requently observed.
210
Phonological rules of Korean (II)
/sɛki- + -ʌsʌ/ → [sɛkjʌsʌ] ‘to carve (connective)’ /k ɑlik ʰi- + -ʌsʌ/ → [k ɑlik ʰ jʌsʌ] ‘to point (connective)’
In addition, when the stem o a verb ends with a vowel, /w/ Glide Formation occurs. Consider (13). (13) a. Verbal stem ending with a vowel: /w/ Glide Formation is obligatory. /o- + -ɑsʌ/ → [w ɑsʌ] ‘to come (connective)’ /pɛu- + -ʌsʌ/ → [pɛw ʌsʌ] ‘to learn (connective)’ b. Verbal stem ending with a CV sequence: /w/ Glide Formation is optional. /k ɑtu- + -ʌsʌ/ → [k ɑtuʌsʌ] ∼ [k ɑtw ʌsʌ] ‘to imprison (connective)’ /milu- + -ʌsʌ/ → [miluʌsʌ] ∼ [milw ʌsʌ] ‘to imprison (connective)’ 9.3.4
/h/ Deletion in verbal inflection When /h/ occurs at the end o a verbal stem, it is deleted automatically i ollowed by verbal endings starting with a vowel. Te examples in (14) show how /h/ is dropped both when used alone, as in (14a), and also when it orms part o a cluster and is ollowed by a verbal ending which begins with a vowel, as in (14b). (14) a. /ʨoh- + -ɑsʌ/ → [ʨoɑsʌ] ‘to like (connective)’ b. /silh- + -ʌsʌ/ → [silʌsʌ] ‘to hate (connective)’
9.3.5
/ɯ/ Deletion in verbal inflection As noted above, vowel–vowel sequences are avoided in Korean. Because o this, the /ɯ/ vowel is deleted when it appears at the end o a verbal stem and is ollowed by a verbal ending which begins with a vowel. Tis rule is only applied between a verbal stem and a verbal ending as in (15a), and not between a noun and a particle as in (15b). Consider the ollowing. (15) a. /k*ɯ- + -ʌsʌ/ → [k*ʌsʌ] ‘to turn off (connective)’ b. /k ʰɑtɯ + -ɛ/ → [k ʰɑtɯɛ] ‘card (prepositional)’ Verbs which display /ɯ/ Deletion are listed in (16). (16) - /k*ɯ-/ ‘to turn off’ - /t*ɯ-/ ‘to �oat’ - /s*ɯ-/ ‘to write’
9.4
211
Rules relating to the word-initial liquid or nasal in Sino-Korean words
- /k ʰɯ-/ ‘to be big’ - /tʰɯ-/ ‘to sprout’ 9.3.6
/ɑ/ or /ʌ/ Deletion in verbal inflection When a verbal stem ending with / ɑ/ or /ʌ/ is ollowed by a verbal ending starting with the same vowel / ɑ/ or /ʌ/, one o them is deleted. (17) a. /ɑ/ Deletion: when the verbal stem ends with / ɑ/, this rule is obligatory. /k ɑ- + -ɑsʌ/ → [k ɑsʌ] ‘to go (connective)’ b. /ʌ/ Deletion: when the verbal stem ends with / ʌ/, this rule is obligatory. /sʌ- + -ʌsʌ/ → [sʌsʌ] ‘to stand (connective)’ c. /ʌ/ Deletion: when the verbal stem ends with / ɛ/, this rule is optional. /pɛ- + -ʌsʌ / → [pɛʌsʌ] ∼ [pɛsʌ] ‘to cut (connective)’ /k ɛ- + -ʌsʌ / → [k ɛʌsʌ] ∼ [k ɛsʌ] ‘to old up (connective)’ /tw ɛ- + -ʌsʌ/ → [tw ɛʌsʌ] → [tw ɛsʌ] ‘to become (connective)’
9.4
Rules relating to the word-initial liquid or nasal in Sino-Korean words wo speci�c phonological rules are applicable to word-initial liquid or nasal /n/ in Sino-Korean words. Tat is, the word-initial liquid sound /l/ is changed into nasal /n/. Word-initial nasal /n/ is deleted beore /i/ or /j/. Tese two processes occur sequentially.
9.4.1
/l/ into /n/ in Sino-Korean words In Pure Korean /l/ cannot appear in word-initial position. However, Sino-Korean words can start with /l/. Sino-Korean words that start with /l/ are nasalised into, or substituted by /n/, which is pronounced in the same place o articulation. Consider able 9.5.
9.4.2
/n / Deletion in Sino-Korean words Another rule stipulates that /n/ cannot appear in word-initial position when the ollowing sounds start with /i/ or /j/. In such cases, /n/ is deleted automatically. Consider the words in able 9.6.
212
Phonological rules of Korean (II)
Table 9.5 Pronunciation of word-initial and medial /l/
Underlying form
Word-initial /l/ → /n/
Non-word-initial /l/
() /lɑk / ‘joy’
() [nɑkw ʌn]
() [k ʰw ɛlɑk ]
‘paradise’ () [nɛil] ‘tomorrow’ () [notoŋ] ‘labour’
‘delight’ () [milɛ] ‘uture’
() /lɛ/ ‘to come’ () /lo/ ‘troublesome’
() [pʰilo] ‘atigue’
Table 9.6 Pronunciation of word-initial and medial /n/
Underlying form
Word-initial /n/ deletion Non-word-initial /n/
() /njʌ/
() [jʌʨɑ]
‘woman’ () /njʌn/ ‘year’ () /nik/ ‘to hide’
‘woman’ () [jʌnsɛ] ‘age’ () [mɛnjʌn] ‘annually’ () [iŋmjʌŋ] () [ɯnnik] ‘being ‘anonymous’ hidden’
() [mɑnjʌ] ‘witch’
Table 9.7 Three types of application of the rules governing word-initial /l/ and/n/
(i) l → n
(ii) n → ø
() →
() → /njʌ- + () → -ʨɑ]/ → [jʌʨɑ] ‘woman’ → /li- + -ju/ →
/lɑk- + -w ʌn/ → [nɑkw ʌn] ‘paradise’ () → /lɛ- +
() → /njʌn-
-il/ → [nɛil] ‘tomorrow’ + -sɛ/ → [jʌnsɛ]‘age’ () → /lo- + -toŋ/ → [notoŋ] () → /nik- + -mjʌŋ/ → [ikmjʌŋ] ‘labour’ ‘unanimous’
(iii) l → n → ø
niju → [iju] ‘reason’ () → → /ljʌ- + -k ɛk/ → njʌk ɛk → [jʌk ɛk] ‘passenger’ () → → /ljɑŋ- + -sim/ → njɑŋsim → [jɑŋsim] ‘conscience’
Word-initial /l/ is �rst nasalised into /n/, which is then deleted, as in the case o /li- + -ju/ () ‘reason’. Consider (18). (18) a. First step: application o /l/ becoming nasalised into /n/ () /li- + -ju/ ‘reason’ → niju b. Second step: /n/ Deletion niju → [iju] More examples along with three types o application o the rules governing the word-initial /l/ or /n/ are provided in able 9.7.
213
9.4
Rules relating to the word-initial liquid or nasal in Sino-Korean words
Table 9.8 Sino-Korean vocabularies with word-initial /l/ or /n/: South vs. North
l → n South
No change in the North
n → ø South
No change in the North
l → n → ø South
No change in the North
()
()
()
()
()
()
/lɑk- + -w ʌn/ → [nɑkw ʌn] () /lɛ- + -il/ → [nɛil] () /lo- + -toŋ/ → [notoŋ]
/lɑk- + -w ʌn/ → [lɑkw ʌn] () /lɛ- + -il/ → [lɛil] () /lo- + -toŋ/ → [lotoŋ]
/njʌ- + -ʨɑ]/ → [jʌʨɑ] () /njʌn- + -sɛ/ → [jʌnsɛ] () /nik- + -mjʌŋ/ → [ikmjʌŋ]
/njʌ- + -ʨɑ]/ → [njʌʨɑ] () /njʌn- + -sɛ/ → [njʌnsɛ] () /nik- + -mjʌŋ/ → [nikmjʌŋ]
/li- + -ju/ → niju → [iju] () /ljʌ- + -k ɛk/ → njʌk ɛk → [jʌk ɛk] () /ljɑŋ- + -sim/ → njɑŋsim → [jɑŋsim]
/li- + -ju/ → [liju] () /ljʌ- + -k ɛk/ → [ljʌk ɛk] () /ljɑŋ- + -sim/ → [jɑŋsim]
Examples in able 9.8 show differences in Sino-Korean vocabulary in North and South Korea due to different application o the rules or word-initial /l/ and /n/.
FURTHER STUDY
Word-initial /l/ and /n/ North vs. South
Afer their division, or a while North and South Korea both used the same orthographic system, which was called Standard Korean Orthography in 1933, until the North published their own Cultured Korean Orthography in 1966, which was revised in 1987 and 2000. In the South, a partially revised edition o the Standard Korean Orthography was published in 1988. As a result, the North and South now have many disparities in the number o basic letters (South 24, North 40), the ordering o the letters in the dictionary and the rules or pronunciation and writing. Among these differences, a particularly prominent one is the way that each writes and pronounces the Sino-Korean word, . In the South, it is pronounced as [nɑkw ʌn] and written as , whilst in the North, it is pronounced as [lɑkw ʌn] and written as . At the moment, the North and the South are compiling a Uni�ed Korean Language Dictionary (Inormation about this dictionary can be ound at www.gyeoremal.or.kr:8080/navigator?act=index ). In order to compile a uni�ed dictionary, it is necessary to have a uni�ed set o orthographical
214
Phonological rules of Korean (II)
Table 9.9 Phonological rules of Korean (II)
Rule
Lateralisation Nasalisation ensi�cation /t/ Insertion /n/ Insertion Palatalisation
Morphological information
Core example: Core example: Lexical origin underlying form pronunciation
Morpheme boundary between /l/ and /n/
Sino-Korean
Word boundary between /l/ and /n/
Sino-Korean
Within a lexical and unctional morpheme Compound word
Sino-Korean
/ɯ/ Deletion in verbal in�ection
Verbal stem + ending
/ɑ, ʌ/ Deletion in verbal in�ection
Verbal stem + ending
/(kjʌl- +
[kjʌlt*ɑnnjʌk]
-tɑn) # -ljʌk/ /kjʌl- + -tɑn/ /pom # pi/
[pomp*i]
/som #
[somnipul]
[kjʌlt*ɑn]
ipul/ /k ɑtʰ- + -i/
[kɑʨʰ i]
/ɑn- + -ko/ [ɑnk *o]
Verbal suffi x Verbal stem + ending ensi�cation Glide Formation in Verbal stem + ending verbal in�ection
Verbal stem + ending
[kw ʌlljʌk]
-ljʌk/
Derivative or compound word Between lexical morpheme and grammatical morpheme
/h/ Deletion in verbal in�ection
/kw ʌn- +
+
[nɑmkjʌsʌ]
/nɑmki- + -ʌsʌ/ + /pɛu- + -ʌsʌ/
[pɛw ʌsʌ]
+ /ʨoh-
[ʨoɑsʌ]
+ -ɑsʌ/ + /k *ɯ-
[k *ʌsʌ]
+ -ʌsʌ/ + /k ɑ- + -ɑsʌ/ +
/l/ into /n/ in SinoKorean
Word-initial position
Sino-Korean
/n/ deletion in Sino-Korean
Word-initial position
Sino-Korean
/l/ → /n/ → ø in Sino-Korean
Word-initial position
Sino-Korean
/sʌ + ʌsʌ/ /lɑk- + -w ʌn/ /njʌ+ -ʨɑ/ /li- + -ju/
[k ɑsʌ] [sʌsʌ] [nɑkw ʌn] [jʌʨɑ] [iju]
Exercises
215
rules as well; one o the greatest challenges in the uni�cation o sp elling necessary or this dictionary is the question o how to write Sino-Korean words such as . Because neither country is willing to change their position, it seems more likely that both spellings – ‘ ’ and ‘’ – will be accepted, rather than one being chosen over the other.
9.5
Summary In this chapter, we have seen phonological rules that require morphological inormation. We have classi�ed these rules based on the morphological characteristics o combined morphemes as ollows: (i) lexical morpheme and lexical morpheme; and (ii) lexical and grammatical morpheme. In able 9.9 we present a summary o each rule with its core example.
EXERCISES 1
2
3
4
5
State whether the letter-initial consonants in underlined letters should undergo Tensification or not and explain why. a. ( ) /kɑl- + -ʨɯŋ/ ‘thirst’ b. ( ) /tɑl- + -sʌŋ/ ‘achievement’ c. ( ) /pɑl- + -tɑl/ ‘development’ d. ( ) /(ko- + -kɑl) # -twɛ- + -tɑ/ ‘to be exhausted’ e. ( ) /pɑl- + -kjʌn/ ‘discover’ f. ( ) /( ʨʰim- + -mol) + -to/ ‘even sinking’ Pronounce the following words, write down the Standard Korean Pronunciation for each and explain why it is pronounced like this. a. ( ) /(sɛŋ- + -sɑn) + -ljɑŋ/ ‘output’ b. ( ) /pun- + -ljɑŋ/ ‘amount’ c. ( ) /wʌn- + -lim/ ‘garden’ d. ( ) /( ʨɑ- + -jʌn) + -lim/ ‘natural forest’ Illustrate the phonological processes by which /pom # pi/ ‘spring rain’ is pronounced [pomp*i] using relevant phonological rules. Illustrate the phonological processes by which /h ɑn- # jʌlɯm/ is pronounced [hɑn njʌlɯm] using relevant phonological rules. Pronounce the following words and explain the relevant phonological rules. a. /kɑtʰ- + -i/ ‘together with’ b. /putʰ- + -i- + -tɑ/ ‘to stick’ c. /kut- + -i/ ‘willingly’
Phonological rules of Korean (II)
216
6
7
Discuss the dialectal differences between North and South regarding the pronunciation of word-initial liquids. Please write down the pronunciation of each inflected verbal form given below and state which phonological rule should be applied. If more than one phonological rule should be applied, list them in order. Note: - /kɯs-/ ‘to draw’ and - /nɑs-/ ‘to recover’ in Korean are /s/ ( ) irregular verbs. That is, in these verbs, /s/ is deleted in verbal inflection when a vowel-initial ending follows the stem.
Stem
Ending
Pronunciation Phonological rule
- /nɑh-/
[
]
‘to bear’
- /-ɑsʌ/ ‘connective’
- /nɑh-/
- /-ko/ ‘connective’ [
]
[
]
‘to grow’
- /-ɑsʌ/ ‘connective’
- /nɑ-/
- /-ko/ ‘connective’ [
]
[
]
‘to recover’
- /-ɑsʌ/ ‘connective’
- /nɑ-/
- /-ko/ ‘connective’ [
]
[
]
‘to be big’
- /-ɑsʌ/ ‘connective’
- /k ʰɯ-/
- /-ko/ ‘connective’ [
]
[
]
‘to draw’
- /-ɑsʌ/ ‘connective’
- /k ɯs-/
- /-ko/ ‘connective’ [
]
‘to bear’ - /nɑ-/
‘to grow’ - /nɑ-/
‘to recover’ - /k ʰɯ-/
‘to be big’ - /k ɯ-/
‘to draw’
10
Loanword phonology
Loanwords are words which, although oreign in origin, have become like native words and are requently used. In Korean, Sino-Korean words undoubtedly orm the major part o Korean vocabulary. In the history o the Korean language, Sino-Korean words started to be included rom as early as the second to third century and since then they have ormed a major part o Korean vocabulary. According to the Standard Korean Language Dictionary , around 57 per cent o Korean vocabulary consists o Sino-Korean loanwords. During the twentieth century, loanwords, particularly those o English origin, rapidly increased in number in line with the Westernisation o the country. In this chapter, thereore, we will ocus on phonological aspects o English loanwords. Since the Korean alphabet is a phonetic alphabet, on the whole, it is suitable or realising a variety o sounds in English. Nevertheless, in naturalising English loanwords into Korean, there are some unavoidable limitations. First, the sound inventories o the two languages are different. In addition, the phonologies o the two languages, including syllable structure, are not the same. For instance, �lm is spelled and pronounced with a [pʰ] in [pʰill ɯm]. Tis happens simply because [] doesn’t exist in Korean, so the sound [pʰ] which is pronounced in a similar position to [], is substituted or it. Inevitably, once loanwords are nativised, they ollow the phonology o the target language, unless they are already used so requently that their pronunciation is set in stone. For instance, the French word renaissance is pronounced as ‘re-Nay-sonce’ by British English speakers but also as ‘Ren-a-sonce’ just as it is pronounced in France. In the ollowing, we will discuss origin and orthography o loanwords (10.1) and how different sound inventories (10.2, 10.3) and different syllable structures (10.4) in�uence the Korean pronunciation o English loanwords. In 10.5 217
218
Loanword phonology
there are examples o how phonological rules in Korean are applied to English loanwords and we provide a summary in 10.6. 10.1
Origin and orthography of loanwords Tere are different ways or oreign words to be nativised. In the ollowing, we will discuss brie�y how loanwords are adopted, spelled and pronounced. Ofen, loanwords are adopted when there are no native words to express a particular meaning. Tese days, Korean has many I-related loanwords – most o which do not have any corresponding Korean words. On the other hand, when loanwords have corresponding words in the native language, their adoption could cause urther meaning division. For instance, these days in Korean the word / ʨʰik ʰin/ ‘chicken’ does not mean chicken in general, but reers only to ried or seasoned chicken that is ofen delivered to one’s home. Hence, when cooking chicken at home, Koreans will use the native Korean word /tɑlk + tʰwikim/ ‘ried chicken’ instead o the loanword /ʨʰik ʰin/ ‘chicken’. Moreover, i you are ordering ried chicken rom a restaurant, you will hear the word instead o ‘chicken’. Consider the examples below. Loanwords which do not have Korean counterparts are given in (1a) and those which do have Korean counterparts, but with a difference in meaning between the Korean words and English loanwords, are given in (1b). Generally speaking, the examples included here are sufficiently nativised that they may be ound in the Standard Korean Language Dictionary . (1) a. No equivalent words in Korean [k ʰʌmpʰ jutʰʌ] ‘computer’ [k ʰʌpʰi] ‘coffee’ [lɑtio] ‘radio’ [mɑus*ɯ] ‘mouse’ [intʰʌnɛt] ‘internet’ b. Semantic difference between Korean words and English loanwords [tɑktʰwikim] ‘ried chicken dish’ vs. [ʨʰīk ʰin] ‘home delivered ried chicken’ [kjʌŋki] ‘sports game’ vs. [k*ɛim] ‘computer game’ [sʌn] ‘line’ vs. [lɑin] ‘personal connections’ At the beginning o the twentieth century, English words were adopted via Japan, but gradually, more and more English words have been adopted directly, rather than via Japan. Examples are given in (2) o some English loanwords that
219
10.2
Consonant correspondence between English and Korean
were adopted via Japan, along with versions o the same words adopted directly rom English into Standard Korean. (2)
English loan words vs. English loan words via Japan. [pɛtʰʌli] ‘battery’ vs. [p*ɑt*ɛli] [s*ɛllʌtɯ] ‘salad’ vs. [sɑlɑtɑ] [mʌpʰɯllʌ] ‘muffl er’ vs. [mɑhulɑ] [pʰɯlɑipʰɛn] ‘rypan’ vs. [hulɑipʰɛn]
Te way in which loanwords are adopted, including Sino-Korean words, is one o the major differences between North and South Korean dialects. As shown in Chapter 1, the North ofen chose to translate both Sino-Korean and English words into Pure Korean. Moreover, many loanwords used in the North are o Russian origin due to the in�uence o the Soviet Union afer the Korean War. Consider the ollowing. (3)
South Korean loanwords vs. North Korean loanwords. [k ʰɛipɯl] ‘cable’ vs. [k*ɑpɛl] [k ʰɛmpʰɛin] ‘campaign’ vs. [k*ɑmp*ɑniɑ] [tʰɯlʌk] ‘truck’ vs. [t*ɯlɑkt*olɯ]
In this chapter, we will discuss in particular how directly adopted English loanwords are spelled and pronounced. Loanwords in Korean were initially spelled according to how they sounded in their original language using only the twenty-our basic letters o Hangeul. Nevertheless, some discrepancies can be seen between the spelling and pronunciation o those loanwords. One o the major problems lies in the prohibition o the tense consonants /k *, t *, p *, s*, ʨ*/. Interestingly, although the original English words do not contain tense consonants, when they are nativised into Korean, they tend to be tensi�ed. However, Korean orthography prohibits the use o these consonants in the written orms o the loanwords, as the letters representing the tense sounds are not included among the basic letters. Hence, the English word banana is spelled as , but is pronounced more like [p *ɑnɑnɑ] than [pɑnɑnɑ]. For those cases, we will provide additional explanation as we go along.
10.2
Consonant correspondence between English and Korean As shown in able 10.1, Korean has a limited range o ricative sounds when compared to English. On the other hand, Korean stop consonants possess three subdivisions, that is, the lax–tense–aspirated contrast between stop consonants (i.e. k, k*, k ʰ – t, t*, tʰ – p, p*, pʰ) and affricates (i.e. ʨ, ʨ*, ʨʰ). One thing to note
Loanword phonology
220
Table 10.1 Consonant correspondence charts of English and Korean
English consonants Korean consonants
Stop
Fricative
Affricate Nasal Approximant Lateral approximant
Examples
Voiceless p t k
pʰ tʰ k ʰ
paint: [ pʰɛintʰɯ]
Voiced
p/p* t/t* k/k*
ballad: [ pɑlɑtɯ]; banana: [p*ɑnɑnɑ]
Voiceless f θ
pʰ s/s*/tʰ
s ʃ
s/s* s/s*+j
oul: [ pʰɑul] thrill: [sɯlil]; health: [hɛls*ɯ]; marathon: [mɑlɑtʰon] sports: [sɯpʰoʨʰɯ]; mask: [mɑs*ɯkʰɯ] shower: [sjɑwʌ] hall: [hol]
bdg
h Voiced v ð z ʒ Voiceless ʧ Voiced ʤ mnŋ r l
Glide
jw
h
virus: [ pɑilʌs*ɯ] rhythm: [litɯm] zipper: [ʨipʰʌ] massage: [mɑs*ɑʨi]
p t ʨ ʨ ʨʰ ʨ mnŋ l l jw
cheese: [ʨʰiʨɯ] jelly: [ʨɛlli] mouse: [mɑus*ɯ] radio: [lɑtio] league: [likɯ]
yacht: [ jotʰɯ] quiz: [kʰwiʨɯ]
is that, as discussed in Chapter 5, glides such as /w/ and /j/ are considered consonants in English, but vowels in Korean. How each o the English consonants in loanwords is realised in Korean is demonstrated in able 10.1 . 10.2.1
Correspondence between /p, t, k/ in English and /pʰ, tʰ, kʰ/ in Korean Te voiceless stops /p, t, k/ in English correspond to /p ʰ, tʰ, k ʰ/ in Korean, regardless o the position in which they occur. Consider the ollowing. (4) a.
Word-initial paint: [pʰɛintʰɯ] tennis: [tʰɛnis*ɯ] key: [k ʰi]
221
10.2
Consonant correspondence between English and Korean
b.
Elsewhere rope: [lopʰɯ] note: [notʰɯ] knock: [nok ʰɯ]
Correspondence between /b/, /d/, /g/ in English and /p, p*/, /t, t*/, /k, k*/ in Korean 10.2.2
Te voiced stops /b, d, g/ in English correspond to the lax stops /p, t, k/ in Korean. But, in certain circumstances, they can be tensi�ed and pronounced as [p *, t*, k*]. However, as mentioned above, Korean orthography does not allow the spelling o loanwords to re�ect their ensi�cation. In other words, gown is spelled instead o , even though it is pronounced [k*ɑun]. (5) /b, d, g/ correspond to lax stop /p, t, k/ a. Word-initial ballad: [pɑllɑtɯ] date: [tɛitʰɯ] guide: [k ɑitɯ] b. Elsewhere ribbon: [lipon] model: [motɛl] league: [lik ɯ] (6)
Word-initial /b, d, g/ in English can be tensed particularly in spontaneous speech banana: [p*anana panana] dance: [t*ɛns*ɯ ∼ tɛns*ɯ] gown: [k*ɑun ∼ k ɑun]
Correspondence between /f/, /s/, /θ/, / ʃ / in English and /pʰ/, /s, s*/, /s, s*/, /s + j, s* + j/ in Korean 10.2.3
Te voiceless ricatives //, /s/, /θ/, / ʃ / in English correspond to /pʰ/, /s, s*/, /s, s*/, /s + j, s* + j/ respectively in Korean. Let’s �rst look at //. Consider (7). Regardless o its position, // corresponds to /pʰ/ in Korean. (7) // corresponds to /pʰ/ oul: [pʰɑul] oundation: [pʰɑuntɛisjʌn] coffee: [k ʌpʰi] muffler: [mʌpʰɯllʌ]
222
Loanword phonology
Te case o /s/ is a little complicated. English /s/ corresponds to [s] or [s *]. Consider the ollowing. (8) Pronunciation o /s/ a. /s/ in word-initial cluster in English: /s/ in English corresponds to /s/ in Korean1 sports: [sɯpʰoʨʰɯ] style: [sɯtʰɑil] screen: [sɯk ʰɯlin] b. /s/ afer a vowel in English: /s/ in English corresponds to /s*/ in Korean Christmas: [k ʰɯlis*ɯmɑs*ɯ] boss: [pos*ɯ] mask: [mɑs*ɯk ʰɯ] c. /s/ beore a vowel in English: /s/ in English corresponds to /s/ or /s*/ in Korean cider: [sɑitɑ] syrup: [silʌp] Cinderella: [sintɛlellɑ] soprano: [s*opʰɯlɑno] soa: [s*opʰɑ] solo: [s*ollo] English /θ/ usually corresponds to /s/ or /s */ in Korean. However, in certain circumstances, /θ/ is spelled with and is pronounced /tʰ/. (9) thrill: [sɯlil] health: [hɛls*ɯ] marathon: [mɑlatʰon] Te sound / ʃ /, usually spelled in English as ‘sh’ or ‘ti’, is spelled and pronounced /s, s*/ plus a diphthong such as /j ɑ/, /jo/, /jʌ/ or /ju/ in Korean. (10) shower: [sjɑw ʌ] shock: [s* jok ʰɯ] lotion: [losjʌn]
Correspondence between /v/, /z/, / ʒ/ and /ð/ in English and /p/, /ʨ/, /ʨ/, /t/ in Korean 10.2.4
Te voiced ricatives / v/, /z/, / ʒ/, and /ð/ in English correspond to / p/, /ʨ/, /ʨ/, /t/ in Korean respectively. None o these consonants exists in Korean and
223
10.2
Consonant correspondence between English and Korean
thereore they are replaced by approximate consonants pronounced in a similar position. (11) /v/ corresponds to /p/ virus: [pɑilʌs*ɯ] oven: [opɯn] (12) /z/ and / ʒ/ correspond to /ʨ/ zero: [ʨɛlo] zipper: [tɕipʰʌ] beige: [pɛiʨi] massage: [mɑs*ɑʨi] (13) /ð/ corresponds to /t/ in Korean. rhythm: [litɯm] 10.2.5
Correspondence between /t ʃ / and /d ʒ/ in English and /ʨʰ/ and /ʨ/ in Korean Te affricates /t ʃ / and /d ʒ/ in English correspond to / ʨʰ/ and /ʨ/ in Korean. (14) a. b.
10.2.6
/t ʃ / corresponds to /ʨʰ/ cheese: [ʨʰiʨɯ] chicken: [ʨʰik ʰin] chocolate: / ʨʰok ʰollis/ → [ʨʰok ʰollit] /dʒ/ corresponds to /ʨ/ juice: [ʨus*ɯ] jogging: [ʨokiŋ] jelly: [ʨɛlli]
Correspondence between /l/ and /r/ in English and /l/ in Korean Te English liquid sounds /l/ and /r/ correspond to /l/ in Korean. But, when /l/ occurs between vowels in word-medial position, ofen /l/ is doubled to /ll/ in Korean. (15) a.
/l/ corresponds to /l/ league: [lik ɯ] bell: [pɛl] liter: [litʌ]
b. /l/ → /ll/ jelly: [ʨɛlli] club: [k ʰɯllʌp] cola: [k ʰollɑ]
Loanword phonology
224
c. 10.2.7
/r/ → /l/ radio: [lɑtio] Rome: [lomɑ] orange: [olɛnʨi]
Correspondence between /j/ and /w/ in English and /j/ and /w/ in Korean Glides in English, such as /j/ and /w/ correspond to /j/ and /w/ in Korean. Glides are regarded as consonants in English, but as vowels in Korean (see discussion in Chapter 5). (16) a. b.
10.3
/j/ corresponds to /j/ yacht: [jotʰɯ] yoghurt: [jokulɯtʰɯ] /w/ corresponds to /w/ twist: [tʰɯwis*ɯtʰɯ] quiz: [k ʰwiʨɯ]
Vowel correspondence between English and Korean able 10.2 shows vowel correspondence between Korean and English with respect to requently used English loanwords. Te pronunciation or English words is based on RP. Te ollowing characteristics can be drawn rom able 10.2. Te English ront-high vowels /ɪ/ and /i/ corresponds to /i/ in Korean. Te distinction between tensed vowels and lax vowels disappears in Korean. (ii) Te distinction in English between /e/ and /æ/ is not present in Korean (though it may be maintained in Korean orthography). Instead the two are merged into /ɛ/ in Korean pronunciation. (iii) Te distinction in English between // and / ɒ/ is not present in Korean; the two invariably correspond to / / in Korean. (iv) Te distinction between /u/ and /ʊ/ is not present. Both correspond to /u/ in Korean. (v) Te distinction between /ʌ/ and /ɜ/ is not present. Both correspond to /ʌ/ in Korean. (i)
In addition, some diphthongs in English correspond to two syllables in Korean, because no Korean words contain off-glides such as /ej/. Examples can be seen in (17).
10.4
225
Syllable structure and loanword phonology
Table 10.2 Vowel correspondence examples between English and Korean
English vowels
ɪ e ӕ ɒ ʌ ʊ i: u: ɑ: ɔ: ɜ: eɪ aɪ ɔɪ əʊ aʊ ɪə eə ʊə
English words
hit [hɪt] dress [dres] manner [mӕnə] documentary [dɒkjumentəri] muffler [mʌə] cushion [kʊʃn] league [li:g] boomerang [bu:mərӕŋ] card [kɑ:d] ork [fɔ:k ] burner [ bɜ:nə] date [deɪt] guide [gaɪd] boiler [ bɔɪlə] boat [ bəʊt] out [aʊt] earphone [ɪəfəʊn] hardware [hɑ:dweə] tour [tʊə]
Corresponding Korean vowel
i ɛ ɛ ɑ ʌ u i u ɑ o ʌ ɛi ɑi oi o ɑu iʌ ɛʌ uʌ
Loanword examples [hitʰɯ] [tɯlɛs*ɯ] [mɛnʌ]
[tɑkʰjumɛntʰʌli] [mʌpʰɯllʌ] * ] [kʰus jʌn
[likɯ] [ pumɛlɑŋ] [kʰɑtɯ] [ pʰokʰɯ] [ pʌnʌ] [tɛitʰɯ] [kɑitɯ] [ poillʌ] [ potʰɯ] [ɑut] [iʌpʰon]
[hɑtɯwɛʌ] [tʰuʌ]
(17) cake: [k ʰɛik ʰɯ] date: [tɛitʰɯ] 10.4
Syllable structure and loanword phonology Syllable structure has been previously discussed in 7.3, where it was noted that a syllable is an abstract and psychological unit embedded in a native speaker’s knowledge o their mother tongue. Te general structural skeleton o a syllable is language-universal. Hence, a syllable nucleus orms the essential part o a syllable, an onset precedes the syllable nucleus, and a coda ollows it. However, rom language to language, the way in which syllables are structured is different. Even the same sound sequence will represent different syllable structures when heard by people who speak different languages. For instance, a word
226
Loanword phonology
like bus would be regarded as a one-syllable word by a native English speaker, but native Korean speakers would regard it as having two syllables. Tereore, understanding the syllable structure o the target language is important in understanding loanword pronunciation. Te syllable structure o Korean is shown in Figure 7.3. All syllable nuclei in Korean have a vowel. A glide can precede that vowel, but this is optional. In addition, the ollowing constraints are ound in the syllable structure o Korean. (18) Syllable structure restrictions a. Syllable-initial position (= onset) Neither /ŋ/ nor any other consonant cluster may occur. b. Syllable-medial position (= nucleus) Off-glides may not occur. Tereore, (C) V (G)(C) cannot appear as a syllable structure, and a sequence such as / j/ cannot exist in Korean. c. Syllable-�nal position (= coda) (i) Consonant clusters may not appear. Only one consonant may be realised in that position. (ii) Consonants should not be released afer central closure. Tereore, ricatives or affricates cannot appear in syllable-�nal position. As a result, only seven o the stop sounds which exist in Korean, namely /k, n, t, l, m, p, ŋ/ may occur, since they remain distinct even afer central closure. Tese constraints in�uence the pronunciation o English loanwords. First, since consonant clusters in syllable-initial position are not allowed in Korean phonology, a / ɯ/ vowel is inserted into the consonant cluster. Hence, all the ollowing sound sequences in English are regarded as a sequence with two syllables in Korean. (19) a. b. c.
Words with an ‘sk-’ sequence skate: [sɯk ʰɛitʰɯ] Words with an ‘st-’ sequence star: [sɯtʰɑ] Words with an ‘sp-’ sequence sports: [sɯpʰoʨʰɯ]
Off-glides may not occur in syllable-medial position. Tereore, off-glides are regarded as two syllables. Examples include: (20) mike: [mɑikʰɯ] nylon: [nɑillon]
227
10.4
Syllable structure and loanword phonology
(a) First step: find a nucleus and build the bone structure of a syllable sɯ
th ɯ
l ɑ
i
k h ɯ
N
N
N
N
N
σ
σ
σ
σ
σ
(b) Second step: link the onset by the onset-first principle sɯ
th ɯ
l ɑ
i
k h ɯ
ON
ON
ON
N
ON
σ
Figure 10.1
σ
σ
σ
σ
Syllabification of strike (O = onset, N = nucleus)
Consonant clusters may not appear in syllable-�nal position, just as in syllable-initial position. Moreover, in syllable-�nal position, consonants should not be released afer central closure. Tereore, ricatives and affricatives cannot occur in syllable-�nal position, since they are released sounds. English loanwords which end with released consonant have an additional / ɯ/ vowel added in order to make it onset. Tis is one o the major causes o discrepancy in the number o syllables between English and Korean. Example words are shown in (21). Numbers in parentheses are the number o syllables. (21) bus (1) vs. [p*ʌs*ɯ] (2) sports (1) vs. [sɯpʰoʨʰɯ] (3) strike (1) vs. [sɯtʰɯlɑikʰɯ] (5) Tis is why an English word like strike, which is a one-syllable word in English, is regarded as a �ve-syllable word in Korean. In Figure 10.1 we show how strike is syllabi�ed by native Korean speakers. Regarding the pronunciation o strike, some readers may have heard native speakers o Korean pronouncing it as [sɯ.tʰɯ.l.ik] (i.e. a our-syllable word) instead o [sɯ.tʰɯ.l.i.k ʰɯ] (i.e. a �ve-syllable word). For instance, in a baseball game, you may have heard [s ɯtʰɯlɑik] instead o [sɯtʰɯrɑik ʰɯ]. Why is strike pronounced differently? Tis is because the word strike has an
Loanword phonology
228
alternative way o syllabi�cation. Tat is, the �nal consonant (i.e. /k/ in the case o strike) can be placed as the coda o the preceding syllable instead o creating an extra syllable by inserting a new / ɯ/ vowel. (22) a. strike: [sɯ.tʰɯ.l.i.k ʰɯ] ∼ sɯ.tʰɯ.l.ik] cake: [k ɛ.i.k ʰɯ] ∼ [k ɛ.ik] b. graph: [kɯlɛpʰɯ] vs. [k ɯlɛp] (never pronounced as this) mark: [mɑkʰɯ] vs. [mɑk] (never pronounced as this) In addition, o all the stop sounds present in Korean, only seven, /k, n, t, l, m, p, ŋ/, can appear in syllable-�nal position, since they remain distinct even afer central closure. Te same rule is applied to loanwords. We will discuss this urther in the next section.
10.5
Application of Korean phonological rules Once loanwords are nativised, they are subject to the same phonological rules as native words. In this section, we will discuss some examples to show how phonological rules that have been discussed in Chapters 8 and 9 can be also applied to loanwords.
10.5.1
Why is out pronounced as [ɑut] not [ɑus]? Tis is due to Neutralisation (see 8.1.2) in syllable-�nal position. Te evidence or this can be seen by adding a series o case particles to the word, as shown in (23). Te underlying orm or is /ɑus/, but in syllable-�nal position, it is realised as [ɑut]. (23) a. /ɑus +-i-+ -tɑ/ → [ɑusitɑ] ‘to oul off (in baseball)’ b. /ɑus/ → [ɑut] ‘out’
10.5.2
Why is rugby pronounced as [lʌkp*i]? Tis is because two lax obstruent consonants cannot occur in a row in Korean. I the second obstruent is a lax consonant, it is tensi�ed (see 8.2.1). Tis is why the /p/ in /lʌkpi/ ‘rugby’ is tensi�ed. (24) /lʌkpi/ → [lʌkp*i] ‘rugby’
229
10.5.3
10.5
Application of Korean phonological rules
Why is Good morning pronounced as [kun moni ŋ]? Korean speakers who do not pay close attention are not able to pronounce Good morning as English speakers do. Most Korean speakers will say this morning greeting as [kun moniŋ] rather than [ gʊd mɔːrnɪŋ]. Tis is because Korean speakers automatically change obstruent–nasal sequences like ‘d–m’ to a nasal–nasal sequence like ‘n–m’, ollowing Nasalisation rules in Korean phonology. As discussed in 8.2.2, an obstruent–sonorant sequence cannot appear in a row in Korean, unlike in English. Similar examples include: (25) /puk + mɑkʰɯ/ → [ puŋmɑkʰɯ] ‘book mark’ /ʌp + lotɯ/ → ʌmlotɯ → [ʌmnotɯ] ‘upload’
10.5.4
Why is Hamlet pronounced as [hɛmnit]? As discussed in 8.2.3, a sequence o a non-dental nasal and a liquid sound is assimilated into a nasal–nasal sequence. In the case o Hamlet , /l/ becomes a nasal /n/, showing Liquid Nasalisation. Similar examples include: (26) /ɑŋli/→[ɑŋni] ‘Henry’ (French name)
10.5.5
Why is Finland pronounced as [p ʰillɑntɯ]? In the case o / pʰinlɑntɯ/, the preceding nasal /n/ is assimilated into the liquid /l/, giving [ pʰillɑntɯ]. Similar examples include: (27) /hɛnli/ → [hɛlli] ‘Henry’ (English name) Sometimes, a sequence o a nasal and a liquid can be assimilated in both ways, i.e. the nasal can be assimilated into the liquid or the liquid can be assimilated into the nasal. For instance, online can be pronounced either as [ollɑin] or [onnɑin]. In addition, the ollowing phonological rules are also applied to English loanwords: (28) Bilabialisation and Velarisation in�ation: [inpʰɯllɛisjʌn] ∼ [impʰɯllɛisjʌn] concrete: [kʰonkʰɯlitʰɯ] ∼ [kʰoŋkʰɯlitʰɯ] camcorder: [kʰɛmkʰotʌ] ∼ [kʰɛŋkʰotʌ] (29) Aspiration black hole: / pɯllɛk +hol/→[ pɯllɛkʰol]
Loanword phonology
230
(30) Similar-position Obstruent Deletion background: / pɛk +kɯlɑuntɯ / → pɛkk *ɯlɑuntɯ → [ pɛk *ɯlɑuntɯ] hot dog: /hɑstokɯ/ → hɑttokɯ → hɑtt*okɯ → [hɑt*okɯ] (31) /h/ Deletion Bohemian: / bohɛmiɑn/ → [ bohɛmiɑn] ∼ [ boɛmiɑn] manhole: /mɛnhol / → [mɛnhol ] ∼ [mɛnol] 10.6
Summary In this chapter, we have ocused on English loanwords and discussed how English loanwords have been adapted in Korean in terms o spelling and pronunciation. First, we examined both consonant and vowel correspondence between Korean and English. In addition, we have shown that, once nativised, English loanwords ollow the rules o Korean phonology, which are discussed in Chapters 8 and 9. Some English loanwords were adopted via Japan, particularly in the early twentieth century. However, since then, most loanwords have been imported directly. In principle, the original pronunciation o loanwords is respected, although nativisation o their pronunciation, in accordance with Korean phonology, is inevitable. In particular, discrepancies between the original English word and its nativised loanword in Korean are caused by dierences in sound inventory as well as syllable structure between the two languages. For instance, Korean has ewer ricatives and liquids, hence, it is inevitable that English words containing ricatives and liquids will be pronounced differently in Korean. In terms o syllable structure, unlike in English, syllable�nal consonants are not released in Korean afer central closure. In addition, consonant clusters are not allowed in syllable-initial position in Korean. Tese rules cause extra / ɯ/ vowels to be inserted into English loanwords.
EXERCISES 1
State how many syllables the following English words have when used as loanwords in Korean and whether there is any difference in the number of syllables. If so, explain why. desk, sports, milk, film, mouse
2
Compare the consonant inventories of English and Korean.
3
Compare the syllable structures of Korean and English with examples.
Exercises
231
4
Why is Finland pronounced [pʰillɑntɯ] in Korean?
5
Why is download pronounced [t ɑunnotɯ] in Korean?
6
Why is concrete pronounced [k ʰoŋkʰɯlitʰɯ] in Korean?
7
Why is manhole pronounced [m ɛnol] in Korean?
8
Why is outside pronounced [ ɑus*ɑitɯ] in Korean?
Notes
2
Production of sounds
1 Although Korean stops and affricates have three phonation types, ricatives have only two. For more inormation see section 4.1.
3
Basic concepts of phonology
1 O course, the view that a phoneme is a unit o sound which cannot be divided urther has itsel been modi�ed with the the advent o eature theory. Feature theory assumes that phonemes can themselves be urther divided into smaller units called phonological eatures (or distinctive eatures) which can be deended as linguistically signi�cant properties o sounds. However, even afer the advent o eature theory, the phoneme continues to be used as a linguistic unit. In eature theory, the phoneme is de�ned as a bundle o phonological eatures. For urther discussion o phonological eatures, see 3.6. 2 O course, when pronouncing the /p/ in nip, the lips can be opened afer closure. However, we will leave this case out o the current discussion, and return to it in 3.3 when we discuss ree variation. 3 Because writing in itsel is abstract, these written phonetic orms are ar more likely to be allophones than to be physical phonetic sounds. 4 Tat is why Korean people who eat are not actually eating a ‘lice burger’ but rather a ‘rice burger’. 5 Prosodic units will be discussed in detail in Chapter 7. In this chapter, only a ew basic terms relating to syllables will be brie�y introduced. 6 A consonant that can orm the nucleus o a syllable is called a “syllabic consonant”. In English, whether or not a consonant is syllabic depends on the sonority o its surrounding segments. 7 For a more detailed look at the difference between the syllabic systems o English and Korean, see Chapter 7. 8 Here, an arrow → indicates what is changing to what, a orward slash / signi�es the end o the change, and underlining denotes the place where the change occurs. For instance, k → ŋ/ _ m would mean that ‘k’ changes to ‘ŋ’ when it precedes ‘m’. 9 In Korean an obstruent that precedes a nasal becomes a nasal itsel, and this is called nasalisation. For a more detailed discussion o this phenomenon, see Chapter 8. 232
233
Notes to pages 56–63
4
Consonants
1 “Liquids” is the name given to the letters o the alphabet notated as ‘r’ and ‘l’, which encompass sounds produced through various different manners o articulation, including approximants, taps and trills. In spite o the act that these sounds are articulated in many different ways, they are either grouped together as “liquids” in many languages, or the category “liquids”, under which these sounds are grouped, is used or convenience’s sake in phonology, because they are realised as allophones o one phoneme. As will soon be seen, in Korean, the phoneme /l/ has two allophones; it may be realised either as a lateral approximant or a tap according to the phonological environment in which it is ound. 2 As will be seen in more detail in 8.1.1, this is due to a restriction in the syllable structure o Korean, namely that syllable-�nal coda sounds must not be released. Because consonants in word-�nal position correspond to codas in the syllable structure, stop sounds in word-�nal position must be realised as unreleased stop sounds, in accordance with this principle. Because o this, distinctions in phonation type made on the basis o whether or not a sound is released become irrelevant. 3 All the spectrograms in this chapter show samples rom the speech o the same speaker. 4 It is impossible to determine accurately the starting point o closure in a word-initial stop sound by analysing a sound signal. Tis is because the period o closure created by stoppage in the oral cavity and the period o silence prior to vocalisation cannot be distinguished using sound signal alone. However, since it is necessary to show the point o closure or convenience o explanation, the point at which closure occurs has been set arbitrarily. Tereore, (a), (b) and (c) on the diagram bear no relation to the true starting points o closure. 5 Te time difference between the release o a stop sound and the point at which vibration begins in the vocal cords in order to produce the ollowing vowel is known as Voice Onset ime, or VO or short. Te length o VO is usually proportional to breathiness. Te more breathiness a stop sound has, the longer it takes until the vocal olds begin to vibrate or the ollowing vowel, because the vocal olds are separated urther during closure. Breathiness is caused by turbulent air escaping through the glottis, which is narrowed as the vocal olds come together; hence it appears on the spectrogram as an irregular, blurry mark, and on the waveorm as an aperiodic wave. In the example in Figure 4.1 , the VO or the lax, tense and aspirated sounds are 47 ms, 13 ms, and 76 ms respectively. 6 In the case o tense sounds, vibrations in the vocal olds begin almost at the same time as the release o the closure, because these sounds are articulated with the two vocal olds almost touching during closure (Kagaya 1974). Tereore, there is almost no aspiration. Despite the closeness o the vocal olds, they do not vibrate during the closure itsel, because they are tensed. 7 Except in special cases, all phonetic notation here uses broad transcription. Tis will also be used in explaining phonological processes. 8 Te closure period between vowel sounds is clearly visible on the spectrogram as a white space where sound energy suddenly drops between the vowels, and on the waveorm as the amplitude suddenly drops and nearly approaches zero. 9 In the example in Figure 4.2, the closure duration or the lax, tense and aspirated stops are 80 ms, 148 ms and 134 ms respectively.
234
Notes to pages 64–73
10 In the examples in Figure 4.2, the VOs or the tense and aspirated sounds are 8 ms and 45 ms respectively. In word-medial position, VO tends to be shorter, and this phenomenon is universal to all languages. Since lax s ounds are realised as voiced sounds during the closure period, it is pointless to measure their VO. 11 In Figure 4.2, the lengths o the vowels preceding the lax, tense and aspirated stops are 132 ms, 80 ms, and 72 ms respectively. 12 Because Korean syllable structure does not allow consonant clusters in the onset or coda positions, a maximum o two consonants may appear in between vowels. Lax sounds may only appear between vowels or between a sonorant consonant and a vowel, and a lax sound afer an obstruent changes into a tense sound as it undergoes Post-obstruent ensi�cation, an essential phonological process in Korean. See 8.2.1 or more inormation. 13 In Korean, because stop consonants in word-�nal position are realised as unreleased stops, the length o their closure cannot be measured by s ound signal alone. Nor can the length o the closure and the period o silence that ollows it be distinguished using sound signal alone. Tereore, the end point o the closure period, indicated by (b) on Figure 4.4, has been determined arbitrarily, as in the case o the start points o word-initial stops. 14 / k ɑk ɑk/ is a Sino-Korean noun meaning ‘street corner ( )’. Although not a commonly used word, it is useul here or demonstrating the differences between lax stops in different positions within a word. 15 As explained above, aspirated consonants have all the properties o glottal ricatives. Because /h/ is a glottal ricative, it could o course be said to be an aspirated sound. For a more detailed discussion o phonological properties please reer to 4.4. 16 In Korean, ricatives are only articulated at alveolar and glottal places, in contrast to English, which has ricatives articulated in various places. English has nine voiced and unvoiced ricatives, /, v, θ, ð, s, z, ʃ , ʒ, h/, with �ve different places o articulation; labiodental, dental, alveolar, palato-alveolar and g lottal. Tereore, English ricatives are always a struggle or native Korean speakers learning English. For more inormation on this, please see the section entitled “Why are ricative sounds so diffi cult or Korean speakers?” on p. 75. 17 In order to show the differences in energy distribution, a noise requency band has been drawn at around 3,700 Hz. As can be seen on the diagram, it is characteristic o alveolar ricatives that riction noise begins to be observed in the high requency band above 3,700 Hz and that the highest energy levels are observed above 6,500 Hz. Te energy distribution o similar riction noise is linked to the particular eatures o their place o articulation. 18 Te reason why glottal ricatives display similar energy distribution to that o the ormant structure o the vowel which ollows them is linked to the act that glottal ricatives do not have a unique place o articulation within the vocal tract. As word-initial glottal ricatives already assume the orm o the vowel that ollows them at the start o articulation, strong energy can be seen in the ormant requency band o the ollowing vowel. For this reason, glottal ricatives have many allophones depending on the ollowing vowel, as will be described later. 19 Te period o aspiration can be identi�ed on the spectrogram as the irregular noise energy observed in the ormant area o the ollowing vowel. 20 Looking closely at the spectrogram, it can be seen that lax alveolar ricatives in intervocalic position are not articulated with stronger aspiration. Te basis or
235
Notes to pages 73–86
21 22 23
24
25
26
27
28
29
30
this is that, unlike in word-initial position, little difference can be seen between the onset and offset o the ricative, and an F1 (�rst ormant) transition can be seen at the onset o the ollowing vowel. In the presence o aspiration, one would not expect to observe an F1 transition, and this eature is known as an F1 cutback. Tis can easily be observed when compared to a vowel ollowing a word-initial lax alveolar ricative, as seen in Figure 4.6. In Figure 4.9 the riction durations or the lax and tense sounds were 106 ms and 177 ms respectively. In Figure 4.9, the duration o the vowels preceding the lax and tense s ounds were 94 ms and 60 ms respectively. Previous studies on the realisation o Korean ricatives in various phonetic environments have observed that, like other obstruents, lax ricatives between voiced sounds are also realised as voiced sounds (K.-H. Lee 2001). In act, the results o the experiments in Kagaya (1974) need to be interpreted differently. Looking careully at the results o this research, we see a tendency or glottal width, which differs greatly between lax and tense types o all obstruents in word-initial position, to be almost the same. A similar tendency can be seen in alveolar ricatives. I this is taken to mean that Korean ricatives should be subclassi�ed into aspirated and tense variants, then large differences in glottal width between /s/ and /s */ in word-medial position should be observed, as they are in other obstruents. Since the starting point o the closure period o affricates in word-initial position cannot be determined using sound signal alone, the starting point o closure has been set arbitrarily as or the stop sounds above. Tereore the starting points o (a), (b) and (c) on the diagram bear no relation to the true starting points o closure. In Figure 4.11 the lengths o the closure periods or lax, tense and aspirated sounds were 47 ms, 130 ms, and 106 ms respectively. Te respective lengths o riction, including aspiration, were 45 ms, 47 ms, and 89 ms, while the respective lengths o the preceding vowels were 127 ms, 87 ms and 81 ms. In word-initial position such weak nasality seems to be related to the strong voicelessness exhibited by word-initial stop sounds in Korean. All Korean stop sounds are voiceless, and as this voicelessness is most pronounced in word-initial position, Korean speakers are unlikely to �nd it diffi cult to distinguish nasals rom other stop sounds, even when their nasality is not emphasised in word-initial position. According to Yoshinaga (2002), based on the results o measurements o air�ow volume through the nasal cavity at the time o articulation o three syllables in Korean and Japanese, namely /a/, /ama/ and /ana/, Korean nasals do not have much effect on the vowels that precede and ollow them, whereas Japanese nasals do. When the coda consonant is anything other than /n/, /l/ turns into /n/. When the coda consonant is /n/, depending on the morphological environment, either /l/ will turn to /n/ as with other consonants, or the preceding consonant will turn into /l/. For a more detailed discussion see chapter 8.2.3. Generally speaking i a word boundary cannot be placed between the /n/ and the /l/ they are realised as /ll/, and i the word boundary can be placed there then they will be realised as /nn/. In the two examples in (8b), none o the morphemes {n ɑn} and {lo} or {kw ɑn} and {ljo} can exist as words in their own right. Tereore the
236
Notes to pages 86–119
31
32
33 34
35
5
word boundary cannot be placed between them, and as such /n/–/l/ will be realised as /ll/. On the other hand, in the examples in (8c), the suffixes {-lon} and {-ljʌk} have been added to the words ‘pikw ɑn’ and ‘hɯpin’, made up o {pi} and {kw ɑn} and {hɯp} and {in} respectively, to create new words. Tereore, a word boundary can be placed between /n/ and /l/, and they will be realised as /nn/. Te realisation o /n/–/l/ as /ll/ as in the examples in (8b) is related to the phonological process known as Lateralisation, and realisation as /nn/ is related to another phonological process, known as Nasalisation o laterals. For more details see Chapter 9. Te most noticeable phonetic difference between taps and nasals is whether or not the nasal passage remains open. However, because stops, like taps, are usually articulated with the nasal passage blocked, one has to understand the phonetic eatures that distinguish not only nasals and taps but also nasals and stops, and stops and taps, to explain why taps turn nasals into taps or why taps turn into nasals. I laterals are set as the underlying orms o liquids, then Korean alveolar nasals can be distinguished rom liquids depending on whether or not they are lateral. Tereore, it is possible to explain why alveolar nasals b ecome liquids when they earn the lateral eature, and why liquids become alveolar nasals when they lose the lateral eature, and so it is also possible to explain the reasons behind this phonological phenomenon. For more details see Chapter 9. Coronal consonants reer to consonants which have [+coronal] eatures. See the next section or the [coronal] eature. In this book the de�nition o the eature [tense] is different rom that given in SPE. In SPE, tense and lax vowels are distinguished using the eature [tense], and Korean tense sounds are distinguished using the eature [glottal constriction], which is classi�ed as one o the sonorant cavity eatures. Tis eature can also be termed [constricted glottis], as in Halle and Stevens (1971). Te eature de�ned as [aspirated] in this book is not present in SPE. SPE uses the eature [heightened subglottal pressure], which is classed as one o the sound source eatures or aspirated sounds. Tis eature may also be termed [spread glottis], as is ound in Halle and Stevens (1971).
Vowels
1 English data in Figure 5.7 are from Wells (1962). In Wells (1962), 25 native speakers of RP English were asked to read 11 words twice, each word containing a monophthong inserted between /h/ and /d/. Te value of F1 and F2 in Figure 5.7 is the average for 50 measured values. 2 For the de�nition of the neutral position, see (11) in Chapter 4. 6
Frequency trends of Korean sounds
1 Words that are not included in Standard Korean Language Dictionary are not considered to be part o Standard Korean vocabulary.
237
Notes to pages 152–222
7
Prosody
1 Te change in pitch realised at a sentential level is called intonation. 2 A “phonological phrase” is exactly the same as an “accentual phrase” according to S.-A. Jun (1993, 2000). However, the term “accentual phrase” could lead to misunderstanding, as it suggests that Korean is a type o pitch accent language. Tis book thereore does not use the term “accentual phrase” but has replaced it with “phonological phrase”. 3 Counting syllables is relatively easy in Korean because one complete character in Korean corresponds to one syllable. For instance, ‘angle’ has one syllable and ‘school’ has two syllables. 4 = tone. It is interesting that no other dialects except Seoul Korean have a rising tone at the end o the phonological phrase. 5 Te data on spontaneous speech were drawn rom the SLILC that we discussed in Chapter 6, and the data o speech read aloud was extracted rom Korea elecom’s speech synthesis database. 6 Te difference between -ʌ/ɑ and -ʌ/ɑ + jo is that the latter is an ending or listener-honori�cation, unlike the - ʌ/ɑ ending.
8
Phonological rules of Korean (I)
1 See Chapter 4 also or discussions on phonetic characteristics o word-�nal stops. 2 In �feenth-century Korean, a maximum o three consonants could appear in syllable-initial position as in [pst ɑ j] ‘time’ and two consonants in syllable-�nal position as in [olmko] ‘to move (connective)’. 3 In Korean, there are our sonorant sounds, /m, n, ŋ, l/, but /ŋ/ cannot occur in syllable-initial (= onset) position. 4 See 8.2.3 or the discussion on Liquid Nasalisation. 5 H.-S. Park (2007) ound that there is no phonetic or articulatory difference between the V–C–V sequences o /p ɑ.k ʰɑ/ and /pɑ.k*ɑ/ and the corresponding V–C–C–V sequences /pɑk.k ʰɑ/ and /pɑk.k*ɑ/. Both /pɑ.k ʰɑ/ and /pɑk.k ʰɑ/ as well as /pɑ.k*ɑ/ and /pɑk.k*ɑ/ are pronounced the same as [p ɑ.k ʰɑ] and [pɑ.k*ɑ] respectively.
9
Phonological rules of Korean (II)
1 Tis phonological rule is only applicable to Sino-Korean or English loanwords, as the sequence described above does not appear in any Pure Korean word.
10
Loanword phonology
1 As we shall discuss again shortly, consonant clusters starting with ‘sp-’, ‘st-’ or ‘sk-’ in English are pronounced in Korean with the / ɯ/ vowel inserted between.
Bibliography
Works in Korean
Bae, J.-C. (1996). An Introduction to Korean Phonology (Gugeoeumunrongaeseol). Seoul: Singumunhwasa. (2003). Te Sounds of Korean (Hangugeoui Bareum). Seoul: Sangyeongmunhwasa. (2008). Systematize the Teory of Korean Phonology (Gugeo Eumunnoui Chegyehwa). Seoul: Hangungmunhwasa. Cha, J.-E., Jung, M.-S., and Shin, J.-Y. (2003). “A phonetic and phonological study on /h/-realization between sonorant sounds (Gongmyeongeum saiui / /ui silhyeone daehan eumseong, eumunnonjeok gochal)”, Eoneo, 28(4): 765–84. Choi, H.-W. (2002). A Survey on the Pronunciation of Standard Korean (Pyojuneo Baleum Siltae Josa). Seoul: Gungnipgueowon. Choi, M.-O. (1989). “Te history of study on umlaut of Korean (Gueo umlautui yeongusajeok gochal)”, Jusigyeonghakbo, 3: 7–39. (2004). Korean Phonology (Gugeoemunnon). Seoul: aehaksa. Heo, U. (1965). Korean Phonology (new edn.) (Gueoeumunhak (gaekosinpan) ). Seoul: Jeongeumsa. Jang, H.-J. and Shin, J.-Y. (2006). “An acoustic study on the generational difference of the monophthongs in the Daegu dialect (Daegu bangeon danmoeumui sedae gan chaie daehan eumhyang eumseonghakjeok yeongu)”, Malsori, 57: 15–31. Jang, H.-J. and Shin, J.-Y. (2007). “Gender difference of standard language orientation throughout the monophthongs by Daegu dialect speakers in their twenties (Daegu bangeon 20dae hwajaui danmoeum silhyeon yangsange natanan pyojuneo jihyangseongui seongbyeoljeok chai)”, Hangugeohak, 36: 289–314. Kang, B.-M. and Kim, H.-K. (2009). Frequency of Korean Words (Hangukeo sayong bindo). Seoul: Hansinmunhwasa. Kang, O.-M. (2003). Korean Phonology (Hangukeo Eumunnon). Seoul: aehaksa. Kim, M.-R. (1992). Korean Phonology (Gugeoeumunnon). Seoul: Hansinmunhwasa. Kim, R.-J. (2005). “Morphological representation of one-syllable Sino-Korean base (Ilumjeol hanjaeo eogiui hyeongtaeronjeok jaehaeseok)” Eomunronjip, 52: 97–120. Kim, Y.-S. (1981). A Study on Korean Sounds (2nd edn.) (Urimal Soriui Yeongu (gochinpan)). Seoul: Gwahakssa. Ko, Y.-K. and Koo, B.-K. (2008). Grammar of Korean (Urimal Munbeopnon). Seoul: Jipmundang. Lee, H.-Y. (1996). Korean Phonetics (Gueoeumseonghak). Seoul: aehaksa. 238
239
Bibliography
Lee, K.-H. (2001). “A Study on Korean Fricatives (Gugeo Machaleum Yeongu)”. PhD dissertation: Korea University. Lee, K.-M. (1998). An Introduction to the History of Korean (new edn.) (Gueosagaeseol (sinjeongpan)). Seoul: aehaksa. Lee, K.-M., Kim, C.-W., and Lee, S.-O. (2000). Korean Phonology (new edn.) (Gueoeumunnon (Jeungbopan)). Seoul: Hagyeonsa. Lee, S.-N. (1960). “A study on the accent of contemporary Korean (Hyeondae Seoulmalui accentui gochal)”, in Anthology of Korean Linguistics (Isungnyeonggugeohakseonjip (eumunpyeon III)). Seoul: Mineumsa, pp. 11–69. Moon, S.-J. (1997). “An acoustical study of Korean ‘s’ (Gueo ‘s’ eumgae taehan eumhyanghakjeok yeongu)”, Malsori, 33–4: 11–22. Munhwagwankwangbu (2004) Regulation of Korean (23rd edn.) (Gugeo Eomun Gujeongjip (23pan) ). Seoul: Daehankyogwaseojusikoesa. Nam, K.-W. and Oh, J.-H. (2009). “Te analysis of the reasons and aspects of pronunciation of ‘/’ (‘’, ‘’ ui bareum silhyeon yangsanggwa wonin bunseok)”. Hangugeohak, 42: 123–53. Park, B.-C. (1989). A History of the Korean Language (Gugeobaldalsa). Seoul: Seyeongsa. Park, H.-S. (2007). “emporal structure of Korean bisyllabic words, with special reference to the types of obstruents in intervocalic positions. (Hangugeo 2eumjeol daneoui sigan gujo -moeum gan jangaeeum nyuhyeonge ttareun chaireul jungsimeuro-)”. Eoneohak, 49: 349–83. Shi, C.-K. (1993). Principles of Korean Word Formation (Gueoui Daneohyeongseong Weolli). Seoul: Hangungmunhwasa. Shin, J.-Y. (1999a). “A diachronic study on the Korean diphthong / (Ijungmoeum /ui tongsijeok yeongu)”, Minjongmunhwa, 32: 473–97. (1999b). “Prosodic units and tensi�cation in Korean (Hangueoui unyul danwiwa gyeongeumhwa hyeonsang)”, Hangugeohak, 10: 27–45. (2000a). Understanding Speech Sounds: For the Basis of Phonetics and Phonology (Malsoriui ihae). Seoul: Hangungmunhwasa. (2000b). “Phonological implications of the realization of Korean plain obstruents: focusing on plain stops (Gueo Pyeongeumui eumseongjeok silhyeone daehan eumunnonjeok haeseok)”, Hangeul , 250: 5–41. (2008). “Phoneme and syllable frequencies of Korean based on the analysis of sp ontaneous speech data (Seongin jayu balhwa jaryo bunseogeul batangeuro han hangueoui eumso mit eumjeol gwallyeon bindo) ”, Eoneocheonggakjangaeyeongu, 13(2): 193–215. (2009). “Frequency related information and syllable structure constraints on Sino-Korean (Hanguk hanjaeumui bimdo gwallyeon Jeongbo mit eumjeol gujo jeyak)”, Masoriwaeumseonggwahak, 2: 129–40. (2010). “Phoneme and syllable frequencies based on the analysis of the entries of a Korean dictionary (Hangueo sajeon pojeeo bareumui eumso mit eumjeol bindo)”, Eoneocheonggakjangaeyeongu, 15(1): 94–106. Shin, J.-Y. and Cha, J.-E. (2003). Te Sound Pattern of Korean (Urimal Soriui Chejye). Seoul: Hangungmunhwasa. Shin, J.-Y., Kim, M.-J., and Kim, K.-H. (2000) “Te phonetic realization of the accentual phrase and resyllabi�cation in Korean (Hangueo gangseguui eumseonggeok silhyeon yangsanggwa jaeeumjeolhwa)”, Eoneo, 25(3): 383–403. Yoshinaga, Y. (2002). “A phonetic comparative study on nasals in Korean and Japanese: Trough the acquisition of nasals in syllable-�nal positions by Japanese speakers
240
Bibliography
of Korean (Hangueowa Ilbneoui Jaeum ap bieume gwanhan eumseonghakjeok bigyo yeongu)”. MA thesis: Seoul National University. Works in English
Abberton, E. (1972). “Some laryngographic data for Korean stops”, Journal of the International Phonetic Association, 2(2): 67–78. Abramson, A. S. and Lisker, L. (1972). “Voice timing in Korean Stops”, Proceedings of the Seventh International Congress of Phonetic Science, 439–46. Ashby, M. and Maidment, J. (2005). Introducing Phonetic Science. Cambridge University Press. Borden, G. J., Harris, K. S., and Raphael, L. J. (1994). Speech Science Primer (3rd edn.). Baltimore: Williams & Wilkins. Catford, J. C. (1977). Fundamental Problems in Phonetics. Bloomington: Indiana Uni versity Press. (1988). A Practical Introduction to Phonetics. Oxford University Press. Cho, M. H. and Shin, S. L. (1997). “A Comparison between [h] in English and Korean ”, Korean Journal of Linguistics, 24(4): 653–73. Cho, Y.-M. Y. (1990). “Syntax and phrasing in Korean”, in Te Phonology–Syntax Connection, ed. S. Inkelas and D. Zec. University of Chicago Press, pp. 47–62. Chomsky, N. and Halle, M. (1968). Te Sound Pattern of English. New York: Harper and Row. Clark, J. and Yallop, C. (1990). An Introduction to Phonetics and Phonology . Oxford: Blackwell. Connell, B. and Arvaniti, A. (1995). Phonology and Phonetic Evidence (Papers in Laboratory Phonology IV). Cambridge University Press. Crystal, D. (1995). Te Cambridge Encyclopedia of the English Language. Cambridge University Press. Cutler, A. and Ladd, D. R. (1983). Prosody: Models and Measurements. Berlin: Springer. Dart, S. N. (1987). “An aerodynamic study of Korean stop consonants: measurements and modeling”, Journal of the Acoustical Society of America, 81(1): 138–47. Denes, P. B. and Pinson, E. N. (1993). Te Speech Chain: Te Physics and Biology of Spoken Language (2nd edn.). New York: W. H. Freeman and Company. Fox, A. (2000). Prosodic Features and Prosodic Structure: Te Phonology of Suprasegmentals. New York: Oxford University Press. Fry, D. B. (1947). “Te frequency of occurrence of speech sounds in Southern English”, Archives Néerlandaises de Phonétique Expérimentale, 20. Halle, M. and Stevens, K. (1971). “A note on laryngeal features”, Quarterly Progress Report MI Research Laboratory of Electronics, 101: 198–213. Han, M. S. and Weitzman, R. S. (1970). “Acoustic features of Korean /P, , K/, /p, t, k/ and /ph, th, kh/”, Phonetica, 22: 112–28. Hardcastle, W. J. (1973). “Some observations on the tense–lax distinction in initial stops in Korean”, Journal of Phonetics, 1(3): 263–72. Hirose, H., Lee, C. Y., and Ushijima, . (1974). “Laryngeal control in Korean stop production”, Journal of Phonetics, 2(2): 145–52. Hong, K., Niimi, S., and Hirose, H. (1991). “Laryngeal adjustments for the Korean stops, affricates and fricatives: an electromyographic study”, Annual Bulletin Research Institute of Logopedics and Phoniatrics, 25: 17–31.
241
Bibliography
International Phonetic Association (1999). Handbook of the International Phonetic Association: A Guide to the Use of the International Phonetic Alphabet . Cambridge University Press. Iverson, G. K. (1983). “Korean s”, Journal of Phonetics, 11: 191–200. Jakobson, R., Fant, G., and Halle, M. (1952). Preliminaries to Speech Analysis. Cambridge, MA: MI Press. Jun, S.-A. (1993). “Te Phonetics and Phonology of Korean Prosody”. PhD dissertation: Ohio State University. (1995). “Asymmetrical effects on the laryngeal gesture in Korean”, in Connell and Arvaniti (eds.), pp. 235–53. (1998). “Te accentual phrases in the Korean prosodic hierarchy”, Phonology , 15(2): 189–226. (2000). “K-oBI (Korean oBI) labelling conventions: Version 3”, Speech Sciences, 7: 143–69. [Version 3.1 is published in UCLA Working Papers in Phonetics, 99: 149–73. www.linguistics.ucla.edu/people/jun/sun-ah.htm] Kagaya, R. (1974). “A �berscopic and acoustic study of the Korean stops, affricates, and fricatives”, Journal of Phonetics, 2: 161–80. Kenstowicz, M. (1994). Phonology in Generative Grammar . Oxford: Blackwell. Kim, C.-W. (1965). “On the autonomy of the tensity feature in stop classi�cation ”, Word , 21: 339–59. (1970). “A theory of aspiration”, Phonetica, 21: 107–16. Kim-Renaud, Y.-K. (1974). “Korean Consonantal Phonology”. PhD dissertation: University of Hawaii. Ladefoged, P. (2001). Vowels and Consonants. Oxford: Blackwell. (2006). A Course in Phonetics (5th edn.). Boston, MA: Tomson Wadsworth. Ladefoged, P. and Maddieson, I. (1996). Te Sounds of the World’s Languages. Oxford: Blackwell. Lass, R. (1984). Phonology: An Introduction to Basic Concepts. Cambridge University Press. Laver, J. (1994). Principles of Phonetics. Cambridge University Press. Lisker, L. and Abramson, A. S. (1964). “A cross-language study of voicing in initial stops: acoustical measurements”, Word , 20: 384–422. Odden, D. (2005). Introducing Phonology . Cambridge University Press. Pullman, G. K. and Ladusaw, W. A. (1986). Phonetic Symbol Guide. University of Chicago Press. Roca, I. and Johnson, W. (1999). A Course in Phonology . Oxford: Blackwell. Selkirk, E. (1984). Phonology and Syntax: Te Relation between Sound and Structure. Cambridge, MA: MI Press. Silva, D. J. (1992). “Te Phonetics and Phonology of Stop Lenition in Korean”. PhD dissertation: Cornell University. Spencer, A. (1996). Phonology . Oxford: Blackwell. Vovin, A. (2008). Korea-Japonica: A Re-evaluation of a Common Genetic Origin. Honolulu: University of Hawaii Press. Wells, J. C. (1962). A Study of the Formants of the Pure Vowels of British English. MA thesis: University of London. Yoon, Y. B. and Derwing, B. L. (2001). “A language without a rhyme: syllable structure experiments in Korean”, Te Canadian Journal of Linguistics, 46(3–4): 187–237.
Index
/ɑ /, /ʌ/ Deletion in verbal in�ection, 211 /h/ Deletion, 196 /h/ Deletion in verbal in�ection, 210 /j/ Deletion, 195 /j/ Glide Formation, 209 /l/ into /n/ in Sino-Korean Words, 211 /ɯ/ Deletion in verbal in�ection, 210 /n / Deletion, 211 /n/ Insertion, 206 /t/ Insertion, 203 /w/ Glide Formation, 210 adnominal ending, 209 affricate, 31 in English, 223 in Korean, 77 allophone, 43 alveolar ridge, 18 anterior, 90 approximant, 30 articulation, 26 aspirated sounds, 33, 92 aspiration, 229 Aspiration, 193
242
d elayed release, 88
ejective, 21 English loanwords via Japan, 219 ortis, 33 ree morpheme, 200 ree variation, 45 ricative, 30 in English, 75, 221, 222 in Korean, 68 glide in English, 224 in Korean, 96 glottalic egressive, 21 glottalic ingressive, 21 glottis, 17 grammatical morpheme, 201 Gyeongsang dialect, 206 hard palate, 18 high, 116
back, 116 Bilabialisation, 192 bound morpheme, 200
implosive, 21 Initiation, 20 initiator o speech sound, 20 intervocalic position, 48 intonational phrase, 170
click, 21 coda, 47 complementary distribution, 45 consonant, 27 in Korean, 56 consonant cluster, 227 consonantal, 52 Consonant-cluster Simpli�cation, 183 continuant, 87 coordinated compound word, 204 coronal, 90
larynx, 16 lateral, 88 Lateralisation, 192 Lateralisation and Nasalisation o Sino-Korean Words, 201 lenis, 33 lexical morpheme, 200 liquids, 233 in English, 223 in Korean, 83 Liquid Nasalisation, 190, 229
243
Index
loanword, 217 low, 116 major class eature, 51 manner o articulation, 28 minimal pair, 44 nasal cavity, 18 Nasalisation, 229 nasals, 29 in Korean, 81 neutral position, 90 Neutralisation, 181 North Korean loanword, 219 nucleus, 47 obstruent, 26 Obstruent Nasalisation, 189 onset, 47 onset-�rst principle, 161 oral cavity, 18 Palatalisation, 207 palate, 18 pharyngeal cavity, 18 phonation, 23 phone, 41 phoneme, 41 phonetic similarity, 45 phonological eatures, 51 phonological phrase, 165, 209 phonological rule, 228 phonological word, 164 place o articulation, 27 Post-obstruent Tensi�cation, 188 prosodic eature, 151 prosodic structure, 155 prosodic unit, 155 prosody, 151 pulmonic egressive, 21 rounded vowels, 37, 116 similar-place Obstruent Deletion, 194 similar-position Obstruent Deletion, 230 Sino-Korean words, 217 sof palate, 18 sonorant, 26, 53 in Korean, 80 Standard Korean orthography, 213
stops, 29 in English, 220, 221 in Korean, 57 subordinated compound word, 204 supra-segmental eature, 151 surace orm, 48 surace representation, 48 syllabic, 52 syllabi�cation, 163 syllable, 156, 225 syllable structure, 158, 226 syllable type, 160 tap, 29 tense, 91 Tensi�cation o Sino-Korean Words, 203 Tensi�cation via /t/ Insertion, 203 tongue, 17 backness, 36 blade, 18 body, 18 height, 35 root, 18 tip, 18 trill, 29 unaspirated sounds, 33 underlying orm, 48 underlying representation, 48 Uni�ed Korean Language Dictionary, 213 unreleased, 65 unrounded vowels, 37 uvula, 18 velaric ingressive, 21 Velarisation, 192 velum, 18 Verbal Suffix Tensi�cation, 208 vocal olds, 16 vocal organs, 16 vocal tract, 26 Voice Onset Time (VOT), 63, 233 voiced sounds, 23, 33 voiceless sounds, 23 VOT, 63, 233 vowel, 35 in Korean, 95 word-�nal position, 48 word-initial position, 48 word-medial position, 48