A SEGMENT CONTACT ACCOUNT OF THE PATTERNING OF SONORANTS IN

CONSONANT CLUSTERS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree of Doctor of Philosophy in the

Graduate School of the Ohio State University

By

Misun Seo, B.A., M.A.

******

The Ohio State University

2003

Dissertation Committee: Associate Professor Elizabeth V. Hume, Adviser Professor Keith A. Johnson Professor Mary E. Beckman Approved by

___________________________

Adviser Department of Linguistics

ii

ABSTRACT

This dissertation explores sound alternations in a consonant cluster in which at least

one consonant is a sonorant (a son/C cluster, hereafter), that is, in sonorant plus sonorant,

obstruent plus sonorant, or sonorant plus obstruent sequences. In this study, I hypothesize

that phonological processes affecting son/C clusters are the result of segment contact rather

than the Syllable Contact Law as discussed in Vennemann (1988), Clements (1998), Rice &

Avery (1991), Baertsch & Davis (2000), among others. In explaining phonological

modifications in son/C clusters as resulting from segment contact, I argue that the

patterning of son/C clusters is closely related to speech perception. That is, as main factors

motivating the alternations in the cluster, I consider contrasts of weak perceptibility triggered

by phonetic similarity between two members of a cluster (Kawasaki 1982, Ohala 1992, 1993)

and the poor perceptibility of a segment in contexts with weaker phonetic cues. In addition,

I argue that a segment specific positional restriction can motivate phonological change in

son/C clusters. Based on the findings of a typological survey concerning alternations in

son/C clusters, I show that the segment contact account makes correct predictions regarding

the patterning of son/C clusters, while the syllable contact account does not.

This study is based on an extensive database of sound alternations found in son/C

clusters, so it provides a strong empirical base for cross-linguistic comparison and for the

study of phonotactic phenomena concerning sonorant consonant clusters. This study also

iii

lends robust empirical support to the segment contact approach to explaining phonological

processes.

iv

Dedicated to my parents

v

ACKNOWLEDGMENTS

My first thanks go to my advisor, Elizabeth Hume. Without her, my dissertation

would not have come about. Her advice both academic and personal has helped me to keep

focused over the years. I am so honored to be her student. I was also lucky to have great

committee members. Keith Johnson and Mary Beckman provided many insightful comments

and corrections particularly in the phonetics part of my dissertation.

I have also benefited from the support and friendship of Shelome Gooden, Tsan

Huang, Eun-Sook Kim, Sun-Hee Lee, Soyoung Kang, Kyuchul Yoon, Eun-Jong Kong, Jihyun

Park and Hyunsook Shin. I thank Jeff Mielke for generously sharing useful information about

sonorants in languages he researched. I have also been lucky to have friends who pray for me.

I thank Peggy Wong, Jimi Woo, Yuhjae Yoon and Jeonghyuk Choi for their incessant prayers.

Finally, thanks should go to my family in Korea. My parents have always been

supportive in everything I’ve done. I thank my sisters Mikyung, Kangjin, Sumin, Miyeon and

Bokyung and my brother Sangwoo, who are my best friends, for their encouragement. I thank

my one-year-old nephew Minhyun for his pictures that gave me a lot of joy throughout my

dissertation writing.

vi

CURRICULUM VITAE

February 16, 1971.................................................................. Born – Suncheon, Korea.

1993.......................................................................................... B.A. English language and literature,

Korea University, Seoul, Korea

1995.......................................................................................... M.A. English Linguistics,

Korea University, Seoul, Korea

1997 – present ........................................................................ Graduate Teaching and Research

Associate, The Ohio State University

PUBLICATIONS

Research Publication

to appear “Metathesis in Faroese and Lithuanian: From Speech Perception to Optimality Theory,” (with Elizabeth Hume). Nordic Journal of Linguistics.

2003 “A Perception-based Study of Sonorant Assimilation,” Japanese/Korean

Linguistics 11. 315-327.

2002 “The Realizations of the Korean Liquid in Onset Position,” Proceedings of the 1st International Conference on Speech Sciences (ICSS). 224-235.

2001a “A Comparative OT Account of Metathesis in Faroese and Lithuanian,” (first author with Elizabeth Hume). In E. Hume, N. Smith and J. van de Weijer (eds.), Surface Syllable Structure and Segment Sequencing. Leiden: Holland Institute of Generative Linguistics (HIL). 210-229.

2001b “A Perception-based Study of Sonorant Assimilation in Korean,” OSU

Working Papers in Linguistics 54. 43-69.

vii

2001c “A Cross-Linguistic Study of the Perception of Sonorant Sequences,” Journal of the Phonetic Society of Japan. Vol. 5. No. 3. 49-63.

1999 Cross-linguistic Study of Stop Place Perception,” (with Elizabeth Hume, Keith Johnson and Georgios Tserdanelis). Proceedings of the XIVth International Congress of Phonetic Sciences. 2069-2072.

FIELDS OF STUDY

Major Field: Linguistics

Minor Fields: Phonology, Phonetics

viii

TABLE OF CONTENTS

Page

Abstract .........................................................................................................................................................ii Acknowledgements .....................................................................................................................................v Vita ................................................................................................................................................................vi List of Tables...............................................................................................................................................xi List of Figures ...........................................................................................................................................xiii Chapters: 1. INTRODUCTION.......................................................................................................................1

1.1 Phonological processes involving son/C clusters........................................................2 1.2 Motivation............................................................................................................................3

1.2.1 Syllable contact account......................................................................................4 1.2.2 Segment contact account....................................................................................6

1.3 Theoretical framework – Optimality Theory ................................................................7 1.4 Organization of this study.................................................................................................7

2. MOTIVATION FOR PHONOLOGICAL PROCESSES IN SON/C SEQUENCES .............................................................................................................................................................9 2.1 Syllable contact account ................................................................................................. 10 2.1.1 Predictions of the Syllable Contact Law....................................................... 10

2.1.2 Problems............................................................................................................. 12 2.2 Segment contact account ............................................................................................... 20

2.2.1 Similarity ............................................................................................................. 24 2.2.1.1 Similarity ranking between homorganic and heterorganic

sequences .......................................................................................... 27

ix

2.2.1.2 Similarity ranking between sonorant/sonorant and obstruent/sonorant sequences...................................................... 30 2.2.1.3 Summary ........................................................................................... 33

2.2.2 Contextual factor .............................................................................................. 34 2.2.3 Predictions of the segment contact account................................................ 36

2.2 Summary............................................................................................................................ 38 3. CROSS-LINGUISTIC SURVEY OF PHONOLOGICAL PROCESSES INVOLVING SON/C SEQUENCES ................................................................................ 41

3.1 Major findings .................................................................................................................. 43 3.2 Homorganic vs. heterorganic sequences..................................................................... 46

3.2.1 Nasal/liquid sequences ...................................................................................... 47 3.2.1.1 Nasal plus liquid sequences ......................................................... 47 3.2.1.2 Liquid plus nasal sequences ......................................................... 51

3.2.2 Obstruent/liquid sequences .............................................................................. 53 3.2.1.1 Obstruent plus liquid sequences .................................................. 53 3.2.1.2 Liquid plus obstruent sequences.................................................. 56

3.2.3 Chi-squared Test ................................................................................................. 58 3.2.4 Summary ............................................................................................................... 59

3.3 Sonorant/sonorant vs. obstruent/sonorant sequences ........................................... 59 3.3.1 Nasal/liquid vs. obstruent/liquid sequences.................................................. 61

3.3.1.1 Nasal plus liquid vs. obstruent plus liquid sequences ............. 61 3.3.1.2 Liquid plus nasal vs. liquid plus obstruent sequences............. 63

3.3.2 Nasal/nasal vs. obstruent/nasal sequences.................................................... 66 3.3.3 Chi-squared Test ................................................................................................. 68

3.3.4 Summary ............................................................................................................... 69 3.4 Stop plus sonorant vs. sonorant plus stop sequences .............................................. 70

3.5 Homorganic vs. heterorganic non-liquid/nasal sequences...................................... 75 3.5.1 Nasal plus non-liquid sequences .................................................................... 76

3.5.2 Obstruent plus nasal sequences....................................................................... 78 3.5.3 Summary.............................................................................................................. 81

3.6 Kuman: articulatorily motivated change ..................................................................... 81 3.7 Summary ........................................................................................................................... 82

4. FORMAL ACCOUNT OF THE PATTENING OF SON/C CLUSTERS ............... 83 4.1 Motivation for Avoid C/X constraints....................................................................... 83

4.2 Avoid C/X constraints................................................................................................... 86 4.2.1 Homorganic vs. heterorganic C/liquid sequences ...................................... 87

4.2.2 Sonorant/sonorant vs. obstruent/sonorant sequences ............................. 93 4.2.3 Unreleased stop plus sonorant vs. sonorant/released stop

Sequences............................................................................................................ 96 4.2.4 Homorganic vs. heterorganic nasal plus non-liquid sequences ................ 99

4.3 Summary ......................................................................................................................... 101

x

5. CASE STUDY OF KOREAN............................................................................................. 103 5.1 Data.................................................................................................................................. 103

5.2 Formal account .............................................................................................................. 107 5.2.1 Motivation ........................................................................................................ 107

5.2.2 Analysis ............................................................................................................. 117 5.3 Summary........................................................................................................................... 133

6. Conclusion................................................................................................................................... 135 Appendix A: A typological survey of alternations in son/C clusters ........................................... 137 Bibliography............................................................................................................................................. 220

xi

LIST OF TABLES

Table Page

2.1 Acoustic properties of consonants....................................................................................... 25

3.1 Phonological patterning of homorganic vs. heterorganic nasal plus lateral or rhotic sequences in the surveyed languages ...............................................................................48 3.2 Phonological patterning of homorganic vs. heterorganic lateral or rhotic plus nasal sequences in the surveyed languages ...............................................................................51 3.3 Phonological patterning of homorganic vs. heterorganic obstruent plus lateral or rhotic sequences in the surveyed languages.....................................................................54 3.4 Phonological patterning of homorganic vs. heterorganic lateral or rhotic plus obstruent sequences in the surveyed languages .............................................................56 3.5 Contingency table of number of cases showing the same or different patterning of homorganic and heterorganic C/liquid sequences .........................................................58 3.6 Phonological patterning of nasal plus liquid vs. obstruent plus liquid sequences in the surveyed languages ......................................................................................................62 3.7 Phonological patterning of liquid plus nasal vs. liquid plus obstruent sequences in the surveyed languages ......................................................................................................64 3.8 Phonological patterning of nasal/nasal vs. obstruent/nasal sequences in the surveyed languages .............................................................................................................67 3.9 Contingency table of number of cases showing the same or different patterning of sonorant/sonorant and obstruent/sonorant sequences ............................................... 69 3.10 Phonological patterning of sotp plus liquid vs. liquid plus stop sequences in the surveyed languages ................................................................................................................... 70 3.11 Phonological patterning of stop plus nasal vs. nasal plus stop sequences in the surveyed languages .................................................................................................................... 71

-xii-

3.12 Phonological patterning of homorganic vs. heterorganic obstruent plus nasal sequences in the surveyed languages ......................................................................................76

3.13 Phonological patterning of homorganic vs. heterorganic nasal plus non-liquid Sequences in the surveyed languages .......................................................................................78

-xiii-

LIST OF FIGURES

Figure Page

2.1 Spectrograms of [anla], [amla] and [aNla] spoken by a male American English speaker..........................................................................................................................................28 2.2 Spectrograms of [anla] and [amla] spoken by a male Hindi speaker.................................29 2.3 Spectrograms of [anla], [atla] and [asla] spoken by an American English speaker..........................................................................................................................................31 2.4 Spectrograms of [a®na] and [a®ta] spoken by an American English speaker ...................32 2.5 Spectrograms of [amla], [amna] and [amta] spoken by an American English speaker..33

-1-

C H A P T E R 1

INTRODUCTION

This dissertation explores sound alternations in a consonant cluster in which at least

one consonant is a sonorant (a son/C cluster, hereafter), that is, in sonorant plus sonorant,

non-sonorant plus sonorant, or sonorant plus non-sonorant sequences. In this study, I

argue that phonological processes affecting son/C clusters result from segment contact

rather than from the Syllable Contact Law. In previous approaches, the Syllable Contact

Law has been considered as a factor motivating phonological change in son/C clusters

(Hooper 1976, Murray & Vennemann 1983, Vennemann 1988, Clements 1990, Rice &

Avery 1991, Davis & Shin 1999, among others). In explaining phonological modifications

in son/C clusters as resulting from segment contact, I argue that the patterning of son/C

clusters is closely related to speech perception. That is, as main factors motivating the

alternations in the cluster, I consider contrasts of weak perceptibility triggered by phonetic

similarity between two members of a cluster (Kawasaki 1982, Ohala 1992, 1993) and the

poor perceptibility of a segment in contexts with weaker phonetic cues. In addition, I argue

that a segment specific positional restriction can motivate phonological change in son/C

clusters. Based on the findings of a typological survey concerning alternations in son/C

clusters, I show that the segment contact account makes correct predictions regarding the

patterning of son/C clusters, while the syllable contact account does not.

-2-

Examples of phonological processes involving son/C clusters in different languages

are given in section 1.1. This study aims to explore what motivates phonological

modifications in son/C clusters, syllable contact or segment contact. In section 1.2, after

overviewing the syllable contact account, the segment contact account is proposed as an

alternative account. The theoretical framework that the present study employs is discussed

in section 1.3 and the organization of this dissertation is provided in section 1.4.

1.1. Phonological processes involving son/C clusters

In many different languages, son/C clusters are targeted by phonological processes.

For example, the /nl/ sequence is modified in a wide range of languages although types of

phonological processes applying to the sequence can be different depending on the

language. As illustrated in (1), by assimilation /nl/is realized as [ll] in Klamath, Ponapean,

Toba Batak, Moroccan Arabic, Leti, and Uyghur, while it is realized as [nn] in Tatar and

Yakut. In Kazakh, /nl/ surfaces as [nd] as the result of dissimilation.

(1) a. /nl/ [ll] Klamath: /honlina/ [hollina] ‘flies along the bank’ (Barker 1964, Rice & Avery

1991) Ponapean: /nan + lEN/ [nallEN] ‘heaven’ (Rehg & Sohl 1981, Rice & Avery 1991) Toba Batak: /lEan + lali/ [lEallali] ‘give a hen-harrier’ (Hayes 1986)

Moroccan Arabic: /ban + li/ [balli] ‘it seemed to me’ (Amakhmakh 1997) Leti: /na + losir/ [llosir] ‘3sg, to follow’ (van Engelenhoven 1995) Uyghur: /hayvan + lAr/ [hayvallar] ‘animals’ (Hahn 1991)

b. /nl/ [nn] Tatar: /khayvan + lÄr/ [khayvannar] ‘animals’ (Poppe 1963) Yakut: /oron + lAr/ [oronnor] ‘beds’ (Krueger 1962) c. /nl/ [nd] Kazakh: /aywan + lAr/ [aywandar] ‘animals’ (Bekturova & Bekturov 1996)

-3-

Likewise, obstruent plus sonorant sequences can be targeted by phonological

processes. For instance, /pl/ surfaces as [mn] in Korean, e.g. /p´p + li/ [p´mni]

‘pinciple of law’ (Davis & Shin 1999), and as [pt] in Yakut, e.g. /Iskaap + lAr/

[Iskaaptar] ‘cabinets’ (Krueger 1962). Sonorant plus obstruent sequences can also be

modified by different phonological processes. Thus, as illustrated in (2), an /lt/ sequence is

realized as [nt] in Ponapean, [ss] in Rendille, and [ll] in Finnish and Yakut.

(2) a. /lt/ [nt] Ponapean: /til + tilep/ [tintilep] ‘mend a roof (durative)’ (Rehg & Sohl 1981) b. /lt/ [ss] Rendille: /ˇe!l-t-a/ [ˇe!ssa] ‘you will give birth’ (Sim 1981) c. /lt/ [ll] Finnish: /kult + an/ [kullan] ‘gold, gen.’ (Sulkala & Karjalainen 1992) Yakut: /uol + tAAVAr/ [uollaaVar] ‘than a son’ (Kurueger 1962)

A goal of this dissertation is to provide a predictive account of the motivation underlying

the processes.

1.2. Motivation

There have been different approaches proposed employing different theoretical

frameworks to account for sound alternations involving son/C clusters. For example, Rice

& Avery (1991), Iverson & Sohn (1994), among others, make reference to features and

feature organization in a rule-based approach, and Davis & Shin (1999) employs an

Optimality Theoretic framework. However, each of these previous approaches is similar in

-4-

that they crucially rely on the notion of syllable contact in motivating the change. As will be

argued in chapter 2, the syllable contact account is problematic since it is unable to account

for observed patterns of son/C clusters in a unified way. Thus, as an alternative

phonological account, I propose that the processes are modified by segment contact rather

than syllable contact.

1.2.1. Syllable contact account

In the syllable contact account, the Syllable Contact Law has been considered a

factor motivating phonological processes in son/C clusters (Hooper 1976, Murray &

Vennemann 1983, Vennemann 1988, Rice & Avery 1991, Davis & Shin 1999, among

others). This law is based on the concept that speech sounds can be classified into different

categories according to their relative strength or sonority.1 Venneman (1988) provides a

Universal Consonantal Strength ordering as in (3), based on ‘degree of deviation from

unimpeded (voiced) air flow’.

1 There have been many proposals on the definition of sonority. For example, Ladefoged (1982) defines sonority as the loudness of a sound relative to that of other sounds with the same length, stress and pitch, and Clements (1990) defines it in terms of a set of major class features. Although there exist many competing sonority scales obtained based on different definitions of the sonority, it is generally agreed that liquids are more sonorous than nasals, which are more sonorous than obstruents.

-5-

(3) ↑ increasing Consonantal Strength

voiceless plosives voiced plosives voiceless fricatives voiced fricatives nasals lateral liquids (l-sound) central liquids (r-sound) high vowels mid vowels low vowels

By making use of the consonantal strength ordering above, Vennemann defines the Syllable

Contact Law as follows:

(4) Syllable Contact Law (Vennemann 1988: 40)

A syllable contact A.B is the more preferred, the less the consonantal strength of

the offset A and the greater the consonantal strength of the onset B.

The Syllable Contact Law can be rephrased as in (5) using the concept of sonority, which is

the reverse of the strength and more commonly used in current phonology.

(5) Syllable Contact Law (sonority version) (Davis & Shin 1999: 286)

A syllable contact A.B is the more preferred, the greater the sonority of the offset

A and the less the sonority of the onset B.

According to the Syllable Contact Law, among son/C clusters, heterosyllabic nasal

plus liquid sequences such as /nl/, /nr/, /ml/, /mr/, etc. are dispreferred sequences

-6-

because the sonority of the onset liquid is higher than the coda nasal in those sequences.

Heterosyllabic obstruent plus sonorant sequences such as /pn/, /pl/, etc. also violate the

Syllable Contact Law. Thus, it has been considered within the syllable contact account that

such types of son/C clusters undergo phonological modifications due to violations of the

Syllable Contact Law.

However, as will be argued in chapter 2, the syllable contact account is problematic

in that it cannot provide a unified account of phonological change in son/C clusters. In

addition, as will be discussed in chapter 3, a typological survey of alternations found in

son/C clusters shows that the syllable contact account makes wrong predictions regarding

the patterning of son/C clusters cross-linguistically.

1.2.2. Segment contact account

As an alternative, I propose that phonological processes involving son/C clusters

result from segment contact rather than from the Syllable Contact Law. In explaining

phonological modifications in son/C clusters as being motivated due to segment contact, I

I argue that low perceptual salience is a factor motivating phonological processes in son/C

clusters. Following Kawasaki (1982) and Ohala (1992, 1993), I hypothesize that a son/C

sequence will be perceptually less salient if two segments in the cluster are phonetically

similar to each other. That is, contrasts of weak perceptibility due to phonetic similarity

between two segments in a son/C cluster are argued to motivate phonological change in

the cluster. In addition, I argue that low perceptual salience can also be triggered by the

occurrence of a segment in contexts with less robust phonetic cues (Wright 1996, Côte!

2000, 2003, among others). Thus, if one of two segments in a son/C cluster occurs in

-7-

contexts with less robust perceptual cues, phonological modification may also apply to the

cluster due to the poor perceptibility of the segment. As will be discussed based on survey

findings in chapter 3, the segment contact account correctly predicts the patterning of

son/C clusters cross-linguistically.

In addition to perceptual factors, I argue that a segment specific positional

restriction can motivate phonological modifications in son/C clusters. For example, as will

be discussed in chapter 4, in Korean there is a restriction prohibiting /l/ in prevocalic

position and this positional restriction motivates phonological modifications in Korean

son/C clusters.

1.3. Theoretical Framework – Optimality Theory

The present study employs Optimality Theory (Prince and Smolensky 1993;

McCarthy and Prince 1993) as the main framework. In Optimality Theory, it is assumed

that the correct output is the one which best satisfies ranked and violable constraints. From

the typological survey, cross-linguistic general patterns of son/C clusters are found, which

can be argued to be triggered by perceptual factors. For a formal account of the patterns

within an Optimality Theoretic (OT) framework, I posit a universal ranking of perceptually

motivated constraints.

1.4. Organization of This Study

The organization of this dissertation is as follows. In Chapter 2, I discuss problems

of the Syllable Contact Law and its predictions regarding the patterning of son/C clusters.

-8-

Then, as an alternative account, I propose that phonological modifications in son/C

clusters are the result of segment contact. Under the hypothesis that low perceptual salience

triggered by phonetic similarity between two segments in a son/C cluster plays a key role in

motivating phonological change in the cluster, I establish a similarity ranking between two

members of son/C clusters based on phonetic properties of the segments involved. In

addition, it is hypothesized that low perceptual salience can also occur due to the poor

perceptibility of a segment in contexts with weak perceptual cues. Based on the established

similarity ranking between segments and a contextual factor influencing the identification of

a segment, predictions are made regarding the phonological patterning of son/C clusters.

In chapter 3, I draw on findings from a typological survey of alternations in son/C

clusters to show that the segment contact account correctly predicts the patterning of

son/C clusters, while the syllable contact account does not.

In chapter 4, constraints are posited to account for the general cross-linguistic

patterning of son/C clusters within an Optimality Theoretic (OT) framework.

In chapter 5, I provide an account of alternations involving son/C clusters in

Korean. I argue that these alternations are the result of segment contact and a segment

specific positional restriction, rather than resulting from syllable contact. Based on the

universal ranking of perceptually motivated constraints in chapter 3 and a segment specific

positional restriction, a formal account is given within an OT framework.

In Chapter 6, I provide a conclusion. Finally, Appendix A contains the patterns of

alternations involving son/C clusters found in each surveyed language.

-9-

C H A P T E R 2

MOTIVATION FOR PHONOLOGICAL PROCESSES INVOLVING SON/C

CLUSTERS: SEGMENT CONTACT VS. SYLLABLE CONTACT

In this chapter, I discuss problems with an account using the Syllable Contact Law

in explaining phonological processes involving a son/C cluster. Then, as an alternative

account, I propose that the change is the result of segment contact. With respect to sound

alternations observed in son/C clusters, I argue that low perceptual salience is a factor

motivating a phonological process in the cluster. I hypothesize that a son/C sequence will

be perceptually less salient if two segments in the cluster are phonetically similar or if one

of two segments in the cluster occurs in contexts with less robust phonetic cues.

In section 2.1, I discuss predictions of the syllable contact account regarding the

patterning of son/C clusters and problems associated with this approach. In section 2.2, I

provide a similarity ranking between two members of son/C clusters based on phonetic

properties of the segments involved. In addition, I discuss the poor perceptibility of a

segment in contexts with weak perceptual cues as another factor motivating phonological

processes in a son/C cluster. Finally, I make predictions regarding the phonological

patterning of son/C clusters based on the established similarity ranking between segments

and a contextual factor influencing the identification of a segment.

-10-

2.1 Syllable contact account

As shown in chapter 1, phonological modifications of son/C clusters have often

been explained by relying on the notion of syllable contact. In section 2.1.1, I discuss

predictions of the syllable contact account regarding the patterning of son/C clusters,

which are tested against a typological survey in chapter 3. In section 2.1.2, I argue that the

syllable contact account is problematic since it cannot provide a unified account of the

patterning of son/C clusters. In addition, I argue that syllable contact is not crucial in

accounting for the patterns of son/C clusters in languages such as Yakut, Kazakh and

Korean although the patterns found in the languages have been argued to be the result of

syllable contact (Davis & Shin 1999, Baertsch & Davis 2000 and Baertsch 2002).

2.1.1 Predictions of the Syllable Contact Law

According to the Syllable Contact Law, a coda is supposed to be more sonorous

than the following onset (Murray & Vennemann 1983, Vennemann 1988, Clements 1990).

Based on the Syllable Contact Law, two predictions can be made regarding the patterning

of son/C clusters. First, son/C clusters violating the Syllable Contact Law are less preferred

regardless of whether they are homorganic or heterorganic. For example, all nasal plus

lateral sequences are less preferred regardless of the place of articulation of the nasal since

the sequences violate the Syllable Contact Law. Thus, it is predicted that within a language

homorganic and heterorganic nasal plus lateral sequences (e.g. /ml/, /Nl/ vs. /nl/) will

show no difference in being targets of phonological modifications unless a segment specific

phonotactic constraint is in play.

-11-

A second set of predictions emerges from the proposal that violations of the

Syllable Contact Law are not equally serious (Murray & Vennemann 1983, Clements 1990)

“[…] the optimality of two adjacent, heterosyllabic segments increases in proportion to the extent that the first outranks the second in sonority. In this view, a sequence such as am.la, for example, constitutes a lesser violation than a sequence such as at.ya.” (Clements 1990: 319)

Thus, in a heterosyllabic A.B sequence a violation is serious in proportion to the difference

between the sonority of A and that of B. Based on this, we predict that if a sequence with a

more serious violation is modified in response to the Syllable Contact Law, we would also

expect a sequence with a less serious violation to be modified as well. For example, if a

nasal plus lateral sequence is modified, we would expect that a stop plus lateral sequence

also gets modified since the latter violates the Syllable Contact Law more seriously than the

former. We therefore would not expect a language where a sequence with a lesser violation

of the Syllable Contact Law gets modified while a sequence violating the law more seriously

remains unchanged.

These predictions are tested against a typological survey of 31 different languages in

chapter 3. According to the survey, the predictions made within the syllable contact

account are not supported. In C/liquid sequences, homorganic sequences are more likely

targets of phonological modifications than heterorganic sequences, while the opposite

pattern is found in nasal plus nasal or obstruent sequences. In addition, the survey

illustrates that within a language a sonorant plus sonorant sequence with a less serious

violation of the Syllable Contact Law is more likely to undergo phonological processes than

an obstruent plus sonorant sequence with a more serious violation of the law.

-12-

2.1.2 Problems

In this section, I argue that the syllable contact account has two problems. First, the

syllable contact account cannot explain phonological processes involving a son/C cluster in

word-initial position where a syllable boundary does not occur between two members of

the cluster. Second, within the syllable contact account, a unified account cannot be given

to the same types of phonological modifications found, for example, in liquid plus nasal

and nasal plus liquid sequences. Since intervocalic liquid plus nasal sequences do not violate

the Syllable Contact Law, phonological modifications in the sequences would have to be

motivated by a factor other than syllable contact.

The syllable contact account is problematic, first of all, since it cannot explain why a

son/C cluster is targeted by a phonological process when a syllable boundary cannot be

assumed between two segments in the cluster, for example, in word-initial position. Recall

that the syllable contact account crucially refers to a syllable boundary in explaining

phonological change of son/C clusters by assuming that the change is motivated to avoid

cases where the onset B has higher sonority than the coda A in a heterosyllabic A.B

sequence.

As illustrated below, in Leti the /nl/ sequence is realized as [ll] whether it occurs

word-initially or intervocalically. Note that vowel deletion occurs in the given examples.

(1) Leti (van Engelenhoven 1995, Hume et al. 1997) /na + losir/ [llosir] ‘3sg, to follow’ /a:na + leti/ [a:lleti] ‘Alety (clan)’ (‘child’ + ‘Leti’)

In line with the view that the change of a son/C cluster is motivated due to the Syllable

Contact Law, the word-initial geminate [ll] in [llosir] might be considered to be bisyllabic, as

represented below:

(2) σ σ | | µ µ | | l o

-13-

However, as discussed in Hume et al. (1997), the representation in (2) has a problem. In

Leti, there is the minimal word requirement that lexical words must be minimally bimoraic

(ibid.). With the representation of the geminate in (2), the minimal word requirement is

satisfied in words consisting of an initial geminate and a vowel, for example [p.pe]. Thus, it

is expected that there will exist words with such a structure in Leti. However, none are

attested. On the other hand, if word-initial geminates are assumed to be part of the onset of

a single syllable, words containing initial geminates such as [ppu.na] ‘nest’ are bisyllabic,

conforming to the dominant tendency that lexical words are made up of two syllables in

Leti. Thus, Hume et al. (1997) propose that geminates in Leti have phonological structure

with a root node multiply-linked to two timing slots, as in (3).

(3) Geminates and long vowels in Leti (Hume et al. 1997) GV VGV σ σ σ | | | µ µ µ | | | X X X X X X X | | | l o [llo] a l e [alle]

-14-

Considering the minimal word requirement, it cannot be assumed that an initial

geminate is bisyllabic in Leti. Therefore, the modification of the word-initial tautosyllabic

/nl/ sequence in Leti cannot be motivated by the Syllable Contact Law since a syllable

boundary cannot be referred to. On the other hand, phonological change of the intervocalic

/nl/ sequence in Leti could be accounted for by relying on the avoidance of rising sonority

over a syllable boundary. Thus, even though both heterosyllabic and tautosyllabic /nl/

sequences show the same pattern in Leti, the syllable contact account cannot provide a

unified account of the modification.

Phonological modifications of both word-initial and intervocalic son/C clusters are

also attested in Moroccan Arabic, as illustrated below:

(4) Moroccan Arabic (Heath 1987) /nl/ [nl] or [ll] /n-l?´b/ [nl?´b] or [ll?´b] ‘I play’ /ban-l-i/ [banli] or [balli] ‘I thought’ /nr ¢/ [nr ¢] or [r ¢r ¢] /n-r ¢mi/ [nr ¢mi] or [r ¢r ¢mi] ‘I throw’ /fin + r ¢ak/ [f ¢er ¢r ¢a¢k ¢] ‘where are you’ /lr/ [lr] or [rr] /l-rusiy-a/ [lrusiya] or [rrusiya] ‘to Russia’ /l-ri©/ [lri©] or [rri©] ‘the wind’ /ln/ [ln] or [nn] /l-na/ [lna] or [nna] ‘to us’ /dyal+ na/ [dyalna] or [dyanna] ‘ours’

In Moroccan Arabic, son/C clusters such as /nl/, /nr¢/ and /lr/ can optionally undergo

assimilation in word-initial or intervocalic position, that is, regardless of whether or not the

clusters occur across a syllable boundary. Since a syllable boundary cannot be referred to in

word-initial position, it needs to be assumed within the syllable contact account that the

modification of a word-initial son/C cluster is motivated by a factor other than syllable

-15-

contact. In addition, intervocalic and word-initial /ln/ sequences optionally undergo

assimilation in Moroccan Arabic. Intervocalic /ln/ sequence is fine with respect to the

Syllable Contact Law since sonority falls over a syllable boundary. Thus, the alternation

cannot be motivated by syllable contact. Therefore, the syllable contact account is

problematic in that a unified account cannot be given to the modifications of son/C

clusters in Moroccan.

The syllable contact account is also problematic since it cannot provide a unified

account of the same types of phonological changes found in nasal plus liquid and liquid

plus nasal sequences in languages such as Boraana Oromo, Korean, Leti and Toba Batak.

(5) Modifications of nasal/liquid sequences a. Boraana Oromo (Stroomer 1995) /nl/ [ll] /hin + lool + a/ [hilloola] ‘I, he will fight’ /ln/ [ll] /kofl + na/ [kofalla] ‘we smile’ /nr/ [rr] /hin + riik’ + a/ [hirriik’a] ‘I, he will grind’ /rn/ [rr] /har + na/ [harra] ‘we sweep’ b. Korean (Davis & Shin 1999) /nl/ [ll] /non-li/ [nolli] ‘logic’ /ln/ [ll] /s´l + nal/ [s´llal] ‘New Year’s Day’ c. Leti (van Engelenhoven 1995, Hume et al., 1997) /nl/ [ll] /a:na + leti/ [a:lleti] ‘Alety (clan)’ (‘child’ + ‘Leti’) /ln/ [ll] /vulan/ [vulla] ‘moon’ d. Toba Batak (Nababan 1981) /nl/ [ll] /láOn + láOn/ [làOlláOn] ‘eventually’ /ln/ [ll] /bál + na/ [bálla] ‘his ball’

As illustrated in (5a), in Boraana Oromo /nl/ and /nr/ sequences surface as [ll] and [rr],

respectively. Within the syllable contact account, these are expected changes since the

sequences violate the Syllable Contact Law. However, the same changes apply to the /ln/

-16-

and /rn/ sequences in the language although the sequences are fine with respect to the

Syllable Contact Law. Thus, it has to be assumed within the syllable contact account that

the modification of nasal plus liquid sequences and that of liquid plus nasal sequences result

from different factors although the same types of phonological changes occur in both types

of sequence. Boraana Oromo is not the only language having the same types of alternations

in nasal/liquid sequences. As shown in (5b)-(5d), the same type of phonological change

applies to both /nl/ and /ln/ sequences in Korean, Leti and Toba Batak. In these

languages, /nl/ and /ln/ are realized as [ll], and within the syllable contact account the

modification of /ln/ would have to be motivated by a factor other than syllable contact

although both /ln/ and /nl/ undergo the same type of phonological change.

Finally, I argue against the syllable contact account by showing that syllable contact

is not crucial in accounting for phonological modifications of son/C clusters in languages

such as Yakut, Kazakh and Korean although the patterns found in the languages have been

argued to be the result of syllable contact (Davis & Shin 1999, Baertsch & Davis 2000 and

Baertsch 2002). In Yakut, /l/ does not occur in post-consonantal onset position except

when it is preceded by another /l/, as illustrated below:

(6) Yakut (Krueger 1962) a. /ll/ [ll] /küöl + lAr/ [küöller ] ‘lakes’ b. /nl/ [nn] /oron + lAr/ [oronnor] ‘beds’ /ml/ [mn] /olom + lAr/ [olomnor] ‘fords’ /Nl/ [Nn ] /XatIN + lAr/ [XatINnar] ‘birches’ c. /yl/ [yd ] /ubay + lAr/ [ubaydar] ‘elder brothers’ /sl/ [st] /muos + lAr/ [muostar] ‘horns’ /Xl/ [Xt ] /tuoX + lAr/ [tuoXtar] ‘what things’ /pl/ [pt ] /Iskaap + lAr/ [Iskaaptar] ‘cabinets’ /kl/ [kt] /balIk + lAr/ [balIktar] ‘fishes’ /tl/ [tt] /at + lAr/ [attar] ‘horses’

-17-

d. /rl/ [rd] /kötör + lAr/ [kötördör] ‘birds’ (stem-final /r/ + /l/-initial suffix) /rl/ [ll] ahaa- ‘to eat’ ahIIr ‘he eats’ ahIIllar ‘they eat’ (/r/-final suffix + /l/-initial suffix)

A lateral surfaces as [n] after a nasal and as an alveolar stop after a glide or an obstruent.

When stem-final /r/ is followed by /l/-initial suffix, the /rl/ sequence is realized as [rd].

On the other hand, when /r/-final suffix is followed by an /l/-initial suffix, the /rl/

sequence is modified to [ll].

Baertsch & Davis (2000) motivate the change in consonant plus /l/ sequences by

assuming that syllable contact is respected in Yakut. Thus, to account for the patterning of

nasal or obstruent plus /l/ clusters in Yakut within an Optimality Theoretic (OT) framework,

they posit the following constraint.

(7) SyllCon: Avoid rising sonority over a syllable boundary.

However, the constraint SyllCon cannot be assumed to motivate all phonological changes

involving son/C clusters in Yakut. The modifications of /yl/ and /rl/ cannot be due to

SyllCon since the sequences do not have rising sonority over a syllable boundary. In

addition, phonological change is attested in Yakut when /l/ occurs before a homorganic

consonant. Thus, although the /ld/ and /lt/ sequences are fine with respect to SyllCon,

both sequences are realized as [ll], e.g. /il + dI + lar/ [illIlar] ‘they took’, /küöl + tA/

[küölle] ‘a lake’, while there is no such change when /l/ occurs before a heterorganic

consonant, e.g. /dZilga/ [dZilga] ‘in a year’ (Krueger 1962). It is assumed within Baertsch

-18-

& Davis (2000)’s analysis that the change of /yl/, /ld/ and /lt/ is motivated by the highly

ranked constraint ICC [lat].

(8) ICC [lat]: A sequence of consonants must be identical in laterality. (Pulleyblank 1997)

However, employing ICC [lat] makes the constraint SyllCon redundant in accounting for

the patterning of son/C clusters in Yakut. That is, with the highly ranked constraint ICC

[lat], there is no reason to posit the highly ranked constraint SyllCon. Phonological change

of C/lateral sequences in Yakut can be motivated by ICC [lat] alone.

Baertsch & Davis (2000) posit another type of syllable contact constraint in (9) to

account for the modification of stem-final /r/ plus /l/ sequences in Yakut.

(9) SCS (Syllable Contact Slope): Have falling sonority over a syllable boundary.

As illustrated in the tableau below, within Baertsch & Davis’ analysis the constraint SCS is

crucially required to prevent a stem-final /r/ plus /l/ sequence from surfacing as *[rr].

(10) /kötör + lAr/ [kötördör] ‘birds’ (Baertsch & Davis 2000) /kötör+lAr/ SyllCon ICC[nas] ICC[lat] FaithStem FaithOnset SCS a. kötörlör *! b. kötörnör *! * c. kötöllör *! * d. kötörrör * *! e. kötördör *

According to Baertsch & Davis, candidate (a) is ruled out since it violates ICC [lat]. When

/l/ in the /rl/ sequence is modified to [n] to satisfy ICC [lat], the change triggers a critical

-19-

violation of ICC [nas] which requires that a sequence of consonants should be identical in

nasality (Pulleyblank 1997). Candidate (c) is excluded by the constraint FaithStem which

prohibits a segment of a stem from being modified. Both candidates (d) and (e) violate

FaithOnset which penalizes the modification of an onset consonant. Thus, within Baertsch

& Davis’ analysis, the constraint SCS is crucially required to rule out candidate (d) with [rr].

However, given that [rr] is not attested in Yakut, it is unnecessary to posit SCS to prevent

the /rl/ sequence from being modified to [rr]. Without SCS, the case can be explained in

terms of structure preservation (Kiparsky 1985). That is, since [rr] is not attested in Yakut,

the modification of /rl/ to [rr] is avoided. As discussed thus far, unlike Baertsch & Davis’

argument that syllable contact is active and motivates phonological change of son/C

clusters in Yakut, syllable contact does not play a crucial role in explaining the patterning of

son/C clusters in their account. As an alternative, I argue that phonological modifications

of son/C clusters in Yakut are motivated by two factors: (a) the avoidance of /l/ in post-

consonantal position and (b) the avoidance of perceptually similar segments in a sequence.

In chapter 4, I provide further discussion regarding this proposal.

As discussed thus far, the syllable contact account cannot provide a unified account

of the same types of phonological changes found in son/C clusters. In addition, syllable

contact is not relevant to phonological patterning of son/C clusters in Yakut despite the

use of syllable contact in previous accounts. Korean and Kazakh have also been assumed to

be languages where modifications of son/C clusters are motivated by syllable contact

(Baertsch & Davis 2000, Davis & Shin 1999). However, I argue that as in Yakut, the

modifications occur in the languages due to segment contact as well as a positional

constraint on /l/. In chapter 4, analyses of the patterning of son/C clusters in Korean are

-20-

given. In addition, I provide detailed discussion regarding the positional constraints on /l/

in Yakut and Kazakh.

2.2 Segment contact account

As an alternative account, I propose that phonological processes found in son/C

clusters result from segment contact rather than from the Syllable Contact Law. I consider

contrasts of weak perceptibility triggered by phonetic similarity between two members of a

cluster as a key factor motivating the alternations in the cluster. In addition, I argue that

phonological change of a son/C cluster can occur due to the poor perceptibility of a

segment in contexts with weaker phonetic cues.

According to Kawasaki (1982) and Ohala (1992, 1993), a given speech signal is

detected better when there is a large modulation in the signal. This is because a sharp

change in the signal increases the salience of cues in the portion of the signal where the

modulation takes place. On the other hand, if a small modulation occurs between two

sounds in a sequence due to the acoustical and auditory similarity of two sounds, they

would be subject to confusion and modification. As noted in Hume & Johnson (2001),

when repair strategies apply to a sound sequence with a small modulation, that is, with

weakly perceptible contrast, they can result in enhancing or sacrificing contrast. Among

repair strategies, dissimilation, epenthesis and metathesis tend to enhance contrast. For

example, dissimilation in Greek can be seen as strengthening contrast among consonants.

As illustrated in (11), dissimilation can occur optionally in Greek consonant clusters

composed of two stops or two fricatives.

-21-

(11) Dissimilation in Greek (Newton 1972, Tserdanelis 2001) a. [pt] ~ [ft] [epta] ~ [efta] ‘seven’ b. [fT] ~ [ft] [fTinos] ~ [ftinos] ‘cheap’ (masc. Nom.)

Weak contrast between members of a consonant cluster consisting of two stops or two

fricatives is enhanced in Greek by increasing the modulation between two adjacent

segments, that is, by giving two adjacent segments different manners of articulation.

On the other hand, a sound sequence with a small modulation can be optimized by

sacrificing contrast with repair strategies such as deletion and assimilation. Deletion in

consonant clusters is an example of avoiding similar consonants in a sequence by sacrificing

contrast. According to Côte! (2003), for example, in Hungarian a stop consonant deletes

optionally after a homorganic nasal consonant, while it is not deleted after a liquid.

(12) Stop deletion in Hungarian: [lOmbdO] ~ [lOmdO] ‘lambda’ [rPndgEn] ~ [rPndgEn] ‘X-ray’ [domptEtP:] ~ [domtEtP:] ‘hilltop’ No deletion: [bOzOltkP:] *[ bOzOlkP:] ‘basalt stone’ [sErptP:] *[ sErtP:] ‘from (a) Serb’

Côte! (ibid.) argues that members of a homorganic nasal plus stop sequence are more similar

to each other than those of a liquid plus stop sequence since members of the former

contrast in [sonorant] only while those of the latter contrast in [sonorant] and

[approximant]. Thus, deletion can occur optionally in a homorganic nasal plus stop

sequence to avoid similar consonants in a sequence.

-22-

With respect to sound alternations observed in son/C clusters, following Kawasaki

(1982) and Ohala (1992, 1993), I hypothesize that contrasts of weak perceptibility triggered

by phonetic similarity between two segments in a son/C cluster play a crucial role. In

addition, I argue that phonological change of a son/C cluster can occur due to the poor

perceptibility of a consonant, which is triggered by the occurrence of the consonant in

contexts where the cues for its identification are weak. The perceptibility of a segment in

contexts with relatively weak perceptual cues has been considered as a factor motivating

various phonological processes such as neutralization (Steriade 1997), metathesis (Hume

1998, 2001), consonant deletion (Mielke 2001), dissimilation (Ohala 1992), etc.

According to Hume (1998, 2001), by metathesis a segment shifts to a context with

more robust phonetic cues. In Faroese, for instance, the sequence /sk/ metathesizes when

it is followed by a stop consonant, e.g. /baisk + t/ [baikst] *[baiskt] ‘bitter, neut.sg.’

(Jacobsen & Matras 1961, Lockwood 1995, Rischel 1972, Hume 1998, 2001). Hume (1998,

2001) indicates that the unmetathesized form *[baiskt] is problematic since the stop [k] is

positioned between two consonants. According to Wright (1996), the perceptual cues for

place and manner of articulation for a stop consonant come mostly from contextual cues

such as release burst and vowel formant transitions. Thus, if a stop consonant occurs

between two consonants, the perceptibility of the stop would be poor since vowel formant

transitions are absent in the context. In Faroese, according to Hume (1998, 2001), the

perceptibility of the stop consonant is enhanced by shifting it from interconsonantal

position to post-vocalic position. Since the fricative consonant has strong internal cues to

place and manner of articulation, it fares better in interconsonantal position (Wright 1996).

-23-

The fricative consonant with relatively strong internal cues, however, can have poor

perceptibility and be targeted by a phonological modification when it occurs in contexts

with weak perceptual cues. Turkish /h/ deletion illustrated in (13) is such an example.

(13) /h/ deletion in Turkish (Mielke 2001) a. after an aspirated stop [ethem] ~ [etem] proper name [Süphe] ~ [Süpe] ‘suspicion’ cf. no /h/ deletion before a stop [kahpe] *[ka:pe] ‘harlot’ [sahte] *[sa:te] ‘counterfeit’ b. adjacent to a fricative [safha] ~ [safa] ‘sleep’ [tahsil] ~ [ta:sil] ‘education’

Mielke (2003) presents experimental evidence to show that in Turkish /h/ is less

perceptible after an aspirated stop and adjacent to a fricative, and thus it optionally deletes.

Regarding sound alternations attested in son/C clusters, I argue that they can be

triggered by the poor perceptibility of a consonant with weak contextual cues. For example,

I hypothesize that phonological modifications of stop plus sonorant sequences occur to

avoid the weak perceptibility of a stop.

In section 2.2.1, I establish a similarity ranking between homorganic and

heterorganic consonants and that between sonorant-sonorant and obstruent-sonorant by

considering the acoustic properties of segments. In section 2.2.2, sound alternations in

son/C clusters are discussed by considering the perceptibility of a segment in different

contexts. Based on the established similarity ranking between segments and the

-24-

perceptibility of a segment in different contexts, predictions concerning the patterning of

son/C clusters are outlined in section 2.2.3.

2.2.1 Similarity

In accounting for phonological modifications using similarity between segments, a

degree of similarity between segments has been computed by considering contrast among

segments (Frisch, Broe & Pierrehumbert 1995, Frisch 1996), a subset-superset relation

between two sets of contrasting features and contextual cues (Côte! 2003) and phonetic

factors such as the relative difference between segments on a shared perceptual dimension

(Steriade 2003). However, to my knowledge, no previous research provides a scale of

similarity between sonorants which can be employed to explain sound alternations found in

son/C clusters. Thus, in the present study, I establish a scale of similarity between segments

occurring in a son/C cluster by considering acoustic properties of the segments. I assume

that two segments sharing more acoustic properties are phonetically more similar to each

other than those sharing fewer acoustic properties. However, even if two segments have

the same acoustic cues, the two segments can be perceptually dissimilar to each other due

to the different quality of the cues. For example, two vowels having the same acoustic

properties would be perceptually separable if one is longer in duration. Thus, in addition to

the number of shared acoustic properties, the quality of the properties needs to be

considered in establishing a similarity scale.

Acoustic properties of different types of consonants can be summarized as follows:

-25-

Internal cues Contextual cue

Segments formants anti-

formants silence

release burst

frication vowel transition

a. Laterals yes yes no no yes [l] coronal [K] velar

b. Nasals yes yes no no no [n] coronal [m] labial [N] velar

c. Rhotics no no no [®] yes no coronal [|] no yes coronal [r] yes no coronal

d. Stops no no yes yes/no no [t] coronal [p] labial [k]

e. Fricatives no no no no yes [s] coronal [f] labial [x] velar [h] glottal

Table 2.1 Acoustic properties of consonants

First of all, the presence or absence of formants is considered as one internal cue

contributing to phonetic similarity between segments. By this, laterals, nasals, the

approximant [®] and the trilled [r] are distinguished from obstruents and the flap. However,

the segments with formant structure are different from one another regarding the quality of

formant structure. That is, laterals and the approximant [®] have clear formants throughout

their durations (Wright 1996), while the trilled [r] has disconnected formant structure by

having pulses made by vibrations (Lindau 1985).

Anti-formants are important in distinguishing laterals and nasals from other

sonorants. Laterals and nasals have anti-formants which result from the oral side cavity in

-26-

the production of nasals and the supra-lingual side cavity in laterals (Johnson ibid.).

According to Espy-Wilson et al. (2000), anti-resonance is introduced in the spectrum of the

approximant [®] by the sublingual space. However, unlike the lateral [l] and nasals, anti-

resonance occurs in high frequency regions, that is, between 5-6 kHz in the case of a non-

retroflex approximant [®] and between 3-4 kHz in the case of a retroflex [ï]. Thus, anti-

resonance in high frequency regions may contribute to the distinction of [®] and [ï] which

generally have low F3.

Stops are characterized by silence and the possibility of release burst (Wright 1996).

In the case of release burst, it can be present or absent in preconsonantal or word-final

position depending on the language. Unlike other consonants, fricatives have frication

noise as their internal cue.

As contextual cues of consonants, vowel transitions which include place cues are

considered in hypothesizing a similarity scale between two segments in a son/C cluster. In

addition, a spectral difference between two consonants in a sequence will be considered.

The purpose of establishing a similarity scale between two segments in a son/C

cluster in this study is to compare predictions regarding the patterning of son/C clusters in

the segment contact account with the ones in the syllable contact account. Recall that in

section 2.1.1, predictions regarding the patterning of homorganic vs. heterorganic son/C

clusters and that of sonorant/sonorant vs. obstruent/sonorant sequences were made within

the syllable contact account. Thus, instead of establishing a similarity scale for all

consonants, only the similarity rankings between homorganic and heterorganic consonants

and those between sonorant-sonorant and obstruent-sonorant will be discussed in the

present study.

-27-

2.2.1.1 Similarity ranking between homorganic and heterorganic consonants

With respect to homorganic and heterorganic consonants, I hypothesize that

homorganic consonants will be phonetically more similar to each other than heterorganic

consonants when the same types of consonants are involved. This is based on the

observation that compared with heterorganic consonants, homorganic consonants have an

additional shared acoustic property, e.g. vowel formant transitions for the same place of

articulation, assuming that they are adjacent to a vowel. Thus, for example, the alveolar [l]

and [n] will be phonetically more similar to each other than [l] and a non-coronal nasal, and

the similarity between the alveolar [l] and [t] will be greater than that between [l] and a non-

coronal stop.

If [l] and [n] are phonetically more similar to each other than [l] and a non-coronal

nasal, it is expected that a smaller modulation will occur between [n] and [l] in a sequence,

making it potentially difficult to tell them apart. On the other hand, a modulation between

the two sounds in the m/l or N/l sequence will be larger than that between [n] and [l]. Thus,

it is hypothesized that it will be easier to distinguish the nasal and lateral in an m/l or N/l

sequence than in an n/l sequence. The spectrograms below, of [anla], [amla], and [aNla] in

figure 2.1, provide supporting evidence for this prediction. These were produced by a

native American English speaking phonetician.

[anla]

[amla]

[aNla]

Figure 2.1 Spectrograms of [anla], [amla] and [aNla] spoken by a male American English speaker

In the spectrogram of [anla], it is hard to tell the boundary between [n] and [l] while a

clearer transition from [m] or [N] to [l] is detected in the spectrograms of [amla] and [aNla],

as marked with an arrow.

rated in figure

.2 and they also show that a transition from [m] to [l] is clearer than the one from [n] to [l].

Spectrograms of [anla] and [amla] spoken by a Hindi speaker are illust

2

-28-

[anla]

[amla]

Figure 2.2 Spectrograms of [anla] and [amla] spoken by a male Hindi speaker

Another piece of evidence supporting the idea that [n] and [l] are phonetically more

similar to each other than a non-coronal nasal and [l] comes from perception experiment

results in Seo (2001a, 2001b, 2003). The experiment was conducted with native speakers of

Korean, Moroccan Arabic and Swedish to investigate the influence of speech perception on

the direction of sonorant assimilation in nasal plus lateral sequences. In the experiment,

paired stimuli such as [anla]/[alla] and [anla]/[anna] recorded by a male Hindi speaker were

presented and listeners were asked to determine whether members of a pair of stimuli they

heard were the same or different. The goal was to compare the perceptual distances

between the pairs [anla]/[alla] and [anla]/[anna], for example. To compare perceptual

istances between pairs, reaction times were measured. The hypothesis was that

perceptually confusable pairs would slow down the reaction time. The results showed that

d

-29-

-30-

the sti

2.2.1.2 Similarity ranking between sonorant/sonorant and obstruent/sonorant sequences

With regard to sonorant/sonorant and obstruent/sonorant seqeunces, I

hypothesize that a sonorant consonant is phonetically more similar to another sonorant

consonant than to an obstruent. This is based on the observation that sonorants have

formants and/or anti-formants in common, while obstruents do not have such internal

cues. In addition, as shown in Table 2.1, obstruents have internal cues such as silence,

release burst or frication noise which will make them distinct from sonorants. For instance,

[l] and [n] are expected to be phonetically more similar to each other than [l] is to [t] or [s].

Spectrograms of [alna], [alta] and [asla] spoken by an American phonetician in figure 2.3

show that the transition from [l] to [t] or [s] can be more easily detected than that from [l]

to [n].

mulus [anla]/[alla] (717.4 ms) had a longer reaction time than the stimulus

[amla]/[alla] (680.1 ms) or [aNla]/[alla] (684.0 ms). Since the second member of the three

stimuli pairs was consistently [alla], different reaction times would have been triggered by

differences in the first part of each stimulus pair. Thus, it can be said that listeners showed

a longer reaction time in detecting the transition in [anla] than the one in [amla] or [aNla].

That is, the transition in [anla] was more confusable to the listeners than the one in [amla]

or [aNla]. Given our discussion above, this is because the transition from [n] to [l] is less

distinct than the one from [m] or [N] to [l].

[anla]

-31- -31-

[atla]

Figure 2.3 Spectrograms of [anla], [atla] and [asla] spoken by an American English speaker

As with the lateral [l], the approximant [®] will be phonetically more similar to

another sonorant than to an obstruent. Spectrograms of [a®na] and [a®ta] by an American

phonetician are illustrated in figure 2.4 and they show that a transition from [®] to [t] is

learer than one from [®] to [n].

c

[a®na]

[a®ta]

Figure 2.4 Spectrograms of [a®na] and [a®ta] spoken by an American English speaker

In addition, the degree of similarity between a nasal and a sonorant will be greater

than that between a nasal and an obstruent since the former have formants and/or anti-

formants in common while the latter do not. Spectrograms of [amla], [amna] and [amta]

meric pho etican figu 2.5 silence, release burst and

e ansition from [m] to [t] more noticeable,

compared with the one from [m] to [l] or [n].

spoken by an A an n in re show that

aspiration noise of the stop [t] makes th tr

-32-

[amla]

[amna]

-33-

[amta]

c ogram a], [am a] and en by an American English k r

Based on the acoustic properties of segments, the following two similarity rankings

Figure 2.5 Spe tr s of [aml n [amta] s spea

pok e

2.2.1.3 Summary

are proposed:

-34-

more similar less similar

b. sonorant/sonorant sequences, e.g. [nl] sonorant/obstruent sequences,

First, members of a homorganic son/C cluster (e.g. /nl/) are phonetically more similar

to each other than those of a heterorganic son/C cluster (e.g. /ml/, /Nl/). Second,

members of sonorant plus sonorant sequences (e.g. /nl/, /ml/, etc.) are phonetically more

similar to each other than are those of sonorant/obstruent sequences (e.g. /tl/, /sl/, /pl/).

2.2.2

ions in son/C clusters can be

motivated by the poor perceptibility of one of the segments. When a segment occurs in

contexts where most of its perceptual cues are absent or reduced, the perceptibility of the

segment will be weak and phonological changes may occur. For example, Côte! (2000, 2003)

argues that due to the poor perceptibility of stops, stops are deleted in consonant clusters in

n, Hungarian and Que!bec French. The perceptibility of a

can be or poor depending on contexts where it occurs.

ternal cues and thus its identification

rticulation for stop consonants come mostly from contextual cues such as release burst

(14) Generalizations regarding the similarity between segments in a cluster

a. homorganic sequences, e.g. [nl] heterorganic sequences, e.g. [ml]

e.g. [tl]

Contextual factor

In addition to phonetic similarity, sound alternat

languages such as English, Catala

stop consonant relatively strong

This is because stop consonants have relatively weak in

mainly depends on contextual cues. That is, the perceptual cues for place and manner of

a

and vowel formant transitions, as indicated below based on Wright (1996):

-35-

Type Cue Distribution ma

racteristics for a particular place of articulation are enhanced when

vocalic position as a

cond member of a consonant cluster, both release burst and vowel formant transitions

position for the

e ther nd, a stop is frequently unreleased in

preconsonantal position. Thus, when a stop occurs in the context of V__CV, that is, as a

(15) Perceptual cues to stop consonants

nner: silence internal release burst contextual: consonant release

transition duration contextual: VC, CV transitions place: F2 transition contextual: VC, CV transitions burst spectrum contextual: consonant release

The identification of place of articulation is less accurate in unreleased stop consonants

than in released consonants (Blumstein & Stevens 1979, Stevens & Blumstein 1978, Halle

et al. 1957, Male!cot 1956, Wang 1959). In addition, as pointed out by Blumstein & Stevens

(1979), the spectral cha

both vowel transitions and release bursts are present.

In the context of VC__V, that is, when a stop occurs in pre

se

are present for a stop consonant. Thus, prevocalic position is a favorable

perceptibility of a stop. On th o ha

first member of a consonant cluster and the burst is potentially absent, the perceptibility of

a stop may be weak.

Unlike stops, fricatives have relatively strong internal cues to both place and

manner of articulation, as illustrated in (16) (Wright 1996).

-36-

Type Cue Distribution

noise duration internal

frication amplitude internal

2.2.3 Predictions of the segment contact account

As discussed above, it is argued in this study that phonetic similarity between

segments in a son/C cluster motivates phonological change in the cluster. According to

Kawasaki (1982) and Ohala (1992, 1993), if two sounds in a sequence are acoustically and

auditorily similar, the degree of distinctiveness of the two sounds would be diminished and

thus they would be subject to modification. Based on this, I hypothesize that a son/C

cluster whose members are phonetically more similar to each other is more likely to

undergo a phonological process, compared with one whose members are phonetically less

similar to each other.

By considering the acoustic properties of segments, the following generalizations

were made regarding the similarity between segments in section 2.2.1.

(17) Generalizations regarding the similarity between segments in a cluster

a. homorganic sequences, e.g. [nl] heterorganic sequences, e.g. [ml]

e.g. [tl]

(16) Perceptual cues to fricatives

manner: frication noise internal

place frication spectrum internal

F2 transition contextual, VC, CV transitions

Thus, unlike a stop consonant, a fricative will have relatively good perceptibility in the

context of V__CV, that is, as a first member of a son/C cluster.

more similar less similar

b. sonorant/sonorant sequences, e.g. [nl] sonorant/obstruent sequences,

-37-

rs. First, homorganic son/C clusters will be more

likely to

sonorant/sonorant sequences will be more likely targets of phonological

process

ct /ml/, whose members are more similar to each other, to

lso be modified. As discussed in section 2.1, the opposite prediction is made within the

llable contact account. Recall that a violation of the Syllable Contact Law is assumed to

e serious in proportion to the difference between the sonority of A and that of B in the

1990). We would

likely to change

he Syllable Contact

By considering the contextual factors discussed in section 2.2.2, predictions can also

/VplV/, /VknV/, etc.) will be more likely targets of phonological processes than

Based on the generalizations above, two predictions can be made concerning the

phonological patterning of son/C cluste

be modified than heterorganic son/C clusters. This suggests that if in a given

language a heterorganic sequence is modified due to phonetic similarity, we would also

expect a homorganic sequence to be modified. For example, if m/l or N/l sequences were

targeted by a phonological process in a language, n/l sequences would also be targeted in

the language. This prediction does not hold for the syllable contact account where all

nasal/lateral sequences are considered equal in terms of violating the Syllable Contact Law.

Second,

es than sonorant/obstruent sequences within a language. For example, if /pl/,

whose members are less similar to each other, were targeted by a phonological process in a

language, then we would expe

a

sy

b

heterosyllabic A.B sequence (Murrary & Vennemann 1983, Clements

therefore expect that an obstruent plus sonorant sequence would be more

than a sonorant plus sonorant sequence because the former violates t

Law more seriously than the latter.

be made regarding the patterning of son/C clusters. For example, when a stop consonant is

ot released in preconsonantal position, intervocalic stop plus sonorant sequences (e.g. n

-38-

oblematic in that it cannot provide a

intervocalic sonorant plus stop sequences (e.g. /VlpV/, /VnkV/, etc.) due to the low

perceptibility of a stop in preconsonantal position.

2.3 Summary

In this chapter, it is argued that phonological processes occur in son/C clusters as

the result of segment contact. That is, phonological change can be motivated by an

occurrence of acoustically and auditorily similar segments in a sequence. Phonological

alternations in son/C clusters can also arise due to the poor perceptibility of a segment

which is triggered by the occurrence of the segment in a position where perceptual cues to

its identification are absent or reduced.

As discussed, the syllable contact account is pr

unified account of the patterning of son/C clusters. Phonological modifications found in

word-initial and word-final position should be motivated by a factor other than syllable

contact. In addition, it has to be assumed within the syllable contact account that the

modification of nasal plus liquid sequences and that of liquid plus nasal sequences result

from different factors although the same types of phonological changes occur in both types

of sequences.

As we saw, the segment contact account and the syllable contact account make

different predictions regarding the patterning of intervocalic son/C clusters within a

language. They are as follows:

-39-

Segment contact account Syllable contact account homorganic ≈ heterorganic

(e.g. nl > ml, Nl) (e.g. nl, ml, Nl) b. son

(e.g. ml > pl) (e.g. pl > ml)

unrel. stop + son > rel. stop + son (where A > B means that the A sequence is more likely to undergo a phonological process than the B sequence, and A ≈ B means that there is no difference between the A sequence and the B sequence in being a target of a phonological process)

As stated in (18a), the segment contact account predicts that a homorganic son/C cluster is

more likely to be targeted by a phonological process than a heterorganic son/C cluster.

Conversely, the syllable contact account predicts that both will be considered the same in

either violating or not violating the Syllable Contact Law.

he two accounts also predict exactly the opposite patterns regarding sonorant plus

sonorant vs. obstruent plus sonorant sequences. Within the segment contact account, it is

expected that sonorant plus sonorant sequences will be more likely to undergo

phonological processes since members of the sequences are more similar to each other. On

the other hand, the syllable contact account predicts that heterosyllabic obstruent plus

sonorant sequences will be more likely to undergo phonological processes since their

violation of the Syllable Contact Law is more serious.

As indicated in (18c), within the segment contact account, it is predicted that stop

plus sonorant sequences will be more likely targeted by phonological processes when a stop

consonant is unreleased before another consonant since the perceptibility of unreleased

stop consonants can be assumed to be relatively poor in this context. If a stop consonant is

released preconsonantally in a language and no segment specific phonotactic constraint is

(18) Predictions regarding the patterning of intervocalic son/C clusters

a. homorganic > heterorganic

+ son > obst + son obst + son > son + son

c. unrel. stop + son > son + rel. stop stop + son > son + stop

T

-40-

involved, stop plus sonorant sequences can remain the same since released stop consonants

in preconsonantal position have relatively good perceptibility. According to the syllable

contact account, heterosyllabic stop plus sonorant sequences violate the Syllable Contact

Law by having a rising sonority over a syllable boundary while heterosyllabic sonorant plus

stop sequences with a falling sonority are good with respect to the law. Thus, as in the

segment con plus

sonorant sequences will be less stabl However, the

syllable contact account differs from the segment contact account in that it predicts stop

plus so

tact account, it is expected that within the syllable contact account stop

e than sonorant plus stop sequences.

norant sequences to be vulnerable to phonological processes even when stop

consonants are released preconsonantally.

In chapter 3, predictions of the segment contact and the syllable contact account

are tested against the results of a typological survey in 31 different languages.

-41-

S

C H A P T E R 3

CROSS-LINGUISTIC SURVEY OF PHONOLOGICAL PROCESSES

INVOLVING SON/C SEQUENCE

Based on findings from a typological survey of alternations involving son/C

clusters, I show in this chapter that the segment contact account makes correct predictions

regarding the patterning of son/C clusters, while the syllable contact account does not.

Recall that within the segment contact account, perceptual similarity of two segments

involved in a sequence and the poor perceptibility of a segment in contexts with weak

phonetic cues are considered as factors motivating phonological processes in the clusters.

As discussed in chapter 2, the segment contact and the syllable contact accounts

make different predictions concerning the patterning of intervocalic son/C clusters within a

language. They are as follows:

-42-

(1) Pred Segment contact account Syllable contact account

(e.g. nl > ml, Nl) (e.g. nl, ml, Nl)

no modification in ‘son + obst .’

unrel. stop + son > rel. stop + son

(where A > B means that the A sequence is more likely to undergo a phonological process

31 languages, alphabetically ordered as follows: Choctaw,

chee, DEg, Finnish, Fur, Kazakh, Kikamba, Kimatuumbi, Klamath, Koasati, Korean,

Leti, Mara, Moroccan Arabic, Mutsun, Oromo, Ponapean, Rendille, Selayarese, Sidamo,

Batak, Turkish, Udi, Uyghur, Yakut, Yindjibarndi, and Zoque.

The m

honological patterns were also considered so that generalizations regarding the patterning of

s of the same family showing the same

f the usters Thus, Kazakh, Tatar, Turkish,

differ ca usters, all of them were

ges of the same

family illustrate the same phonological patterning of son/C clusters, only one of them was

chosen.

ictions regarding the patterning of son/C clusters

a. homorganic > heterorganic homorganic ≈ heterorganic

b. son/son > obst./son (e.g. ml > pl) obst. + son > son + son (e.g. pl > ml)

c. unrel. stop + son > son + rel. stop stop + son > son + stop

than the B sequence, and A ≈ B means that there is no difference between the A sequence and the B sequence in being a target of a phonological process)

I show that the segment contact account correctly predicts the patterning of son/C clusters

through the typological survey of

Chuk

Somali, Swedish, Tatar, Toba

ain reason why these languages were chosen is that the patterns attested in the

languages are relatively well-studied. (For details of the alternation patterns found in each

surveyed language, see the appendix.) In choosing languages from the same family,

p

son/C clusters were not biased due to language

phonological patterns o cl . for example, since

Uyghur and Yakut show ent phonologi l patterns of son/C cl

chosen although they are all Turkic languages. On the other hand, if langua

-43-

es. In section 3.6, I provide an account of the

pattern

3.1 Major findings

The survey findings show that the segment contact account makes correct

predictions regarding which son/C cluster is more or less likely to be targeted by a

phonological process. The following general patterns were found involving son/C clusters.

(2) The patterning of intervocalic son/C clusters mor a phonological process a phonological process

b. sonorant/sonorant obstruent/sonorant

sonorant + rel. stop

In section 3.1, major findings from the typological survey are given. In section 3.2

through section 3.5, I provide a detailed discussion of the patterning of son/C clusters in

the languages surveyed. That is, I show that the segment contact account makes correct

predictions regarding the patterning of different types of son/C clusters; that is,

homorganic vs. heterorganic C/liquid, sonorant/sonorant vs. obstruent/sonorant, stop

plus sonorant vs. sonorant plus stop sequences and homorganic vs. heterorganic nasal plus

non-liquid (i.e. nasal or obstruent) sequenc

ing of lateral plus consonant sequences in Kuman, which conflicts with the general

pattern that homorganic sequences are more likely to undergo phonological modifications

than heterorganic sequences.

e likely to undergo less likely to undergo

a. homorganic C/liquid heterorganic C/liquid

c. unrel. stop + sonorant rel. stop + sonorant

d. heterorganic nasal + non-liquid homorganic nasal + non-liquid

-44-

at

both ho

uences, the typological survey shows that stop

plus so

e phonological patterning of stop plus sonorant sequences in Choctaw and Toba

Batak supports the prediction within the segment contact account that unreleased stop plus

In C/liquid sequences, as predicted by the segment contact account, homorganic sequences

are more likely to undergo a phonological process than heterorganic sequences. Recall that

this is predicted by the segment contact account since perceptual similarity between two

segments in a sequence is considered as a factor motivating phonological modifications in a

son/C cluster. On the other hand, the syllable contact account incorrectly predicts th

morganic and heterorganic sequences will be considered the same in either violating

or not violating the Syllable Contact Law.

As shown in (2b), it is also observed that sonorant/sonorant sequences are more

vulnerable to phonological modification than obstruent/sonorant sequences. This is the

predicted pattern within the segment contact account since members of a

sonorant/sonorant sequence are more similar to each other than members of an

obstruent/sonorant sequence. On the other hand, the syllable contact account incorrectly

predicts that an obstruent plus sonorant sequence will be more likely to be a target of a

phonological process than a sonorant plus sonorant sequence since the former violates the

Syllable Contact Law more seriously than the latter. In addition, it is expected within the

syllable contact account that sonorant plus obstruent sequences will not be modified since

they do not violate the Syllable Contact Law.

With respect to stop/sonorant seq

norant sequences are more likely to undergo phonological modifications than

sonorant plus stop sequences when the sonorant is a liquid. This is the pattern predicted

within both the segment contact and the syllable contact accounts. When the sonorant

consonant involved is a nasal, this asymmetrical pattern was not found.

Th

-45-

ed than released stop plus sonorant

sequenc

Wright (1996), I argue that place assimilation

involvin

ed to avoid the articulatory difficulty of coordinating the production

n, the typological survey shows that predictions regarding the

ile

sion concerning

e patterning of specific types of son/C clusters is given in the remaining sections of this

chapter.

sonorant sequences are more likely to be modifi

es due to the weak perceptibility of the unreleased stop consonant. However, the

syllable contact account incorrectly predicts that stop plus sonorant sequences are

vulnerable to phonological processes regardless of whether or not stop consonants are

released preconsonantally since both type of sequence violate the Syllable Contact Law.

Unlike in C/liquid sequences, in nasal plus non-liquid sequences the opposite

pattern is found in the languages surveyed: heterorganic sequences are more likely targets of

phonological change than homorganic sequences. When a heterorganic nasal plus non-

liquid sequence is modified, the nasal assimilates to a following consonant in place in most

of the languages surveyed. Following

g heterorganic nasal plus non-liquid sequences is motivated by a nasal’s weak

internal cues to place in preconsonantal position. In addition, following Kohler (1990),

place assimilation in the sequences can also be motivated by articulatory simplication.

According to Kohler, a reduction in the number of organs is favored in speech production

so that energy expenditure is minimized. Winters (2003) provides strong empirical evidence

that the vulnerability of heterorganic nasal plus stop sequences to a phonological

modification is motivat

of two or more gestures at the same time.

In conclusio

patterning of son/C clusters made within the segment contact account are supported, wh

those made within the syllable contact account are not. A detailed discus

th

-46-

3.2

ological process in son/C clusters. Thus, the segment contact

account predicts that a homorganic son/C cluster is more likely to be targeted by a

phonological process than a heterorganic son/C cluster within a language. However, the

syllable contact account predicts that both will be considered the same in either violating or

not violating the Syllable Contact Law.

According to the typological survey, as predicted by the segment contact account,

in C/liquid sequences homorganic sequences are more likely to be targeted by a

phonological process than heterorganic sequences within a language. The general pattern in

(3) is found in C/liquid sequences from the typological survey. As will be shown in section

3.2.3, the Chi-squared test shows that the generalization based on the phonological

ng C/liquid sequences in the surveyed languages is

statistic

Homorganic vs. heterorganic sequences

As discussed in chapter 2, the segment contact and the syllable contact accounts

make different predictions with respect to the patterning of homorganic and heterorganic

son/C clusters in a language. Within the segment contact account, weak perceptibility of

segments due to phonetic similarity between the two segments in a sequence is considered

as a factor motivating a phon

patterni of homorganic vs. heterorganic

ally significant.

(3) Generalization I

In a given language, phonological change involving heterorganic C/liquid sequences

implies that of homorganic C/liquid sequences.

-47-

that of

obstruent/liquid sequences. Depending on whether the liquid involved is a lateral or a

ustrate diff s in homorganic and heterorganic

C/liquid sequences. Thus, I present the patterning of C/lateral sequences and that of C/

rhotic sequences separately.

liq

As predic ent co sequences,

homorganic sequ re likely targets of phonological

rocesses than heterorganic sequences (i.e. /lm/, /rm/, etc.) within a language. This

f homorganic sequences with that of

heterorganic sequences in nasal plus liquid and liquid plus nasal sequences.

1 languages, in nasal plus liquid sequences,

homorg

The generalization above regarding the target of a phonological process in C/liquid

sequences is made by looking into the patterning of nasal/liquid sequences and

rhotic, most surveyed languages ill erent pattern

3.2.1 Nasal/ uid sequences

ted by the segm ntact account, in nasal/liquid

ences (i.e. /ln/, /rn/, etc.) are mo

p

generalization is made by comparing the patterning o

3.2.1.1 Nasal plus liquid sequences

According to the typological survey of 3

anic sequences are more vulnerable to phonological modification than heterorganic

sequences regardless of whether a liquid involved is a lateral or a rhotic. This generalization

can be made based on the observation that, among the surveyed languages, there is no

language with a modification in heterorganic nasal plus liquid sequences while homorganic

nasal plus liquid sequences remain unchanged.

-48-

arized as follows:

The phonological patterning of homorganic vs. heterorganic nasal plus lateral or

rhotic sequences found from the surveyed languages can be summ

Target of phonological processes Languages Neither homorganic nor DEg, Somali, Swedish heterorganic Only homorganic Klamath, Moroccan Arabic, Turkish, Leti,

Mara, Uyghur Only heterorganic none

nasal + lateral

heterorganic Korean, Ponapean, Tatar, Toba Batak, Yakut, Both homorganic and Choctaw, Fur, Kazakh, Kimatuumbi, Koasati,

Yindjibarndi Neither homorganic nor Leti heterorganic Only homorganic Moroccan Arabic and Toba Batak Only heterorganic none

nasal + rhotic

Both homorganic and heterorganic

Fur, Mara

Table 3.1 Phonological patterning of homorganic vs. heterorganic nasal plus lateral or

rohibiting a liquid in prevocalic or post-

onsonantal position. Thus, by the restriction, both homorganic and heterorganic nasal plus

, Mutsun, Oromo, Rendille, Selayarese, Sidamo, Udi

teral

us lateral

se on uen in the

xample, in Kikamba the only underlying nasal/liquid sequence found is

/Nl/ surfacing as [nd]. Thus, the phonological patterning of homorganic and heterorganic

rhotic sequences in the surveyed languages

As will be discussed in chapter 5, in languages such as Korean, Tatar, Yakut, Yindjiardni

and Mara, there is a positional restriction p

c

liquid sequences are modified in the languages.

In Finnish, Kikamba, Kuman

and Zoque, it is impossible to determine whether or not homorganic nasal plus la

sequences are more likely targets of phonological change than heterorganic nasal pl

sequences. This is becau ly homorganic or heterorganic seq ces are attested

languages. For e

-49-

ces cannot be contrasted. In Mutsun, only the heterorganic /ml/

which surfaces unmodified, is attested (Okrand 1977).

With respect to nasal plus rhotic sequences, only homorganic or heterorganic

tested rese, Somali, Swedish

and Zoque, making it impossible to compare phonological patterning of homorganic

sequences with that of heterorganic sequences in the languages. Among 31 languages

surveyed, nasal plus rhotic sequences are not attested in 18 languages since a rhotic is not

included in the sound inventory of the language (e.g. Choctaw, Kikamba, Klamath, Koasati,

and Korean) or a rhotic does not occur in onset position in native words of a language (e.g.

Turkish, Tatar, Uyghur, Yakut and Yindjibarndi). In Finnish, Kimatuumbi, Kuman, Oromo,

Sidamo and Udi, nasal plus rhotic sequences are not attested although a rhotic is included

in a sound inventory.

Moroccan Arabic and Uyghur are representative cases illustrating the pattern that

homorganic nasal plus liquid sequences are more likely to undergo a phonological process

than heterorganic nasal plus liquid sequences.

(4) Representative cases with a modification only in homorganic nasal plus liquid sequences

(Nasals: /m, n/, Liquid:/l, r, ¢r ¢/)

nasal plus lateral sequen

sequence,

sequences are at in DEg, Mutsun, Ponapean, Rendille, Selaya

a. Moroccan Arabic (Amakhmakh 1997, Heath 1987)

Change ? ? ?

/ban + li/ [balli] ‘it seemed to me’

/fin + r ¢ak/ [f ¢er ¢r ¢a¢k ¢] ‘where are you’

/nl/ [nl] or [ll] /n-l ´b/ [n l ´b] or [l l ´b] ‘I play’

/nr ¢/ [nr ¢] or [r ¢r ¢] /n-r ¢mi/ [nr ¢mi] or [r ¢r ¢mi] ‘I throw’

/nr/ [nr] or [rr] /m´n + r´Zl +u/ [m´rr´Zlu] ‘from his leg’

No Change

(foot, etc.)’ /ml/ [ml] /z¢t¢m + l + ha/ [z¢t¢mlha] ‘he stepped on her

/mr ¢/ [mr ¢] /m-r ¢aZ?a/ [mr ¢aZ?a] ‘review’

-50-

(Nasals: /m, n, N/, Lateral: /l/) b. Uyghur (Hahn 1991)

Change /nl/ [ll] /hayvan + lAr/ [hayvallar] ‘animals’

No Change

/ml/ [ml] /khanIm + lAr/ [khanImlar] ‘ladies’

[ [ SinyaNliq ] ‘(person) of Xinjiang’

anguages with a modification only in

omorganic sequences or in both homorganic and heteororganic sequences: however, there

ion only i nic sequences. Thus, the survey

findings illustrate that in a given language, a phonological process involving heterorganic

nasal plus liquid sequences implies that of

pattern is predicte ntact a which takes into account the phonetic

similarity between two segments in a sequence; the vulnerability of a homorganic nasal plus

liquid sequence to l process can

phonetic similarit segme

observation that plus li es are more likely to undergo

phonological proc not predicted within

red

/Nl/ Nl] /SinyaN+liG/

As shown in (4a), in Moroccan Arabic homorganic nasal plus liquid sequences optionally

undergo assimilation, while heterorganic nasal plus liquid sequences surface unmodified. In

Uyghur, the homorganic /nl/ sequence surfaces as [ll] as the result of assimilation, while the

heterorganic /ml/ and /Nl/ sequences do not undergo any phonological change. In Uyghur,

/r/ does not occur in onset position in native words of a language (Hahn 1991).

As can be seen in Table 3.1 above, there are l

h

is no language with a modificat n heterorga

homorganic nasal plus liquid sequences. This

d by the segment co ccount

a phonologica be understood as being a consequence of the

y between the two nts involved in the sequence. However, the

equenchomorganic nasal quid s

esses than heterorganic nasal plus liquid sequences is

the syllable contact account. This is because all nasal plus lateral sequences are conside

equal in terms of violating or not violating the Syllable Contact Law. Thus, some additional

-51-

3.2.1.2 Liquid plus nasal sequences

The typological survey also shows that the segment contact account makes a correct

ncerning the patterning of liquid plus nasal sequences. As predicted,

ces are more likely targets of phonological processes than heterorganic

in eral or a rhotic.

here is no surveyed language where a heterorganic liquid plus nasal sequence undergoes a

rg ces unmodified,

s can be seen from the survey findings in Table 3.2:

logical Languages

formal mechanism would be needed to account for the patterning of homorganic vs.

heterorganic nasal plus liquid sequences within the syllable contact account.

prediction co

homorganic sequen

sequences within a language regardless of whether a liquid volved is a lat

T

phonological process while a homo anic liquid plus nasal sequence surfa

a

Target of phono processes Neither homorganic nor Klamath, Swedish heterorganic Only homorganic Choctaw, Finnish, Koasati, Korean, Leti,

Mara, Moroccan Arabic, Rendille, Sidamo, Somali, Toba Batak, Turkish, Udi

Only heterorganic none

lateral + nasal

Both homorganic and none heterorganic Neither homorganic nor Fur, Kazakh, Leti, Moroccan Arabic,

Yakut, Zoque heterorganic Swedish, Tatar, Toba Batak, Turkish, Uyghur,

Only homorganic DEg, Finnish, Mara, Rendille, Sidamo, Somali, Udi

Only heterorganic none

rhotic + nasal

Both homorganic and Ponapean heterorganic

nasal sequences in the surveyed languages Table 3.2 Phonological patterning of homorganic vs. heterorganic lateral or rhotic plus

-52-

) Representative cases with a modification only in a homorganic lateral plus nasal sequence

c 1993) (Nasals: /m, n/) C

The patterns found in Somali and Choctaw are given in (5) as representative cases

showing the pattern that homorganic liquid plus nasal sequences are more likely to be

targeted by phonological processes than heterorganic liquid plus nasal sequences within a

language.

(5

a. Somali (Zaborski 1986, Saeed 1993, Zor

hange

/ln/ [ll] /dilnay/ [dillay] ‘(we) killed’

/rn/ [rr] /furniin/ [furriin] ‘divorce’

No change

/lm/ [lm] /maalmo/ [maalmo] ‘days’ /rm/ [rm] /irmaanaa/ [irmaanaa] ‘in milk’

. Choctaw (Nicklas 1974)

(Nasals: /m, n/) Change

b

anna] ‘to be weaved’

/l-noti/ [nnoti] ‘tooth’ No change

/ln/ [nn] /ta-l-na/ [t

ences undergo

ssimilation whether the liquid involved is a lateral or a rhotic. In Choctaw, the homorganic

ln/ sequence surfaces as [nn] as the result of assimilation while the heterorganic /lm/

quence remains unchanged. A rhotic is not attested in Choctaw.

The segment contact account thus correctly predicts that a homorganic liquid plus

asal sequence is more likely to undergo a phonological modification than a heterorganic

quid plus nasal sequence within a language. The segment contact account correctly

/lm/ [lm] /a-l-mo/ [almo] ‘to be gathered’

In Somali, as illustrated in (5a), homorganic nasal plus liquid sequ

a

/

se

n

li

-53-

yllable contact account does not. The syllable contact

account does not. This is because all liquid plus th

respect to the Syllable Contact Law by having sonority over a syllable boundary.

Thus, some additional formal mechanism is requ

un

3.2.2 Obstruent/li

The segment contact a rrectly predicts the patterning of obstruent/liquid

of phonological processes than heterorganic

d sequences

c sequences are

ore likely targets of phonological processes than heterorganic sequences.

predicts this pattern, while the s

nasal sequences are considered good wi

falling

ired to account for the pattern within the

syllable contact acco t.

quid sequences

ccount also co

sequences. The survey findings show that homorganic obstruent/liquid sequences (i.e. /lt/,

/rt/, /tl/, /tr/, etc.) are more likely targets

obstruent/liquid sequences sequences (i.e. /lp/, /rk/, /pl/, /kr/, etc.) within a language.

3.2.2.1 Obstruent plus liqui

As shown in Table 3.3, in obstruent plus liquid sequences there is no language with

a modification in heterorganic sequences while homorganic sequences do not undergo any

phonological process. Thus, in obstruent plus liquid sequences, homorgani

m

-54-

Languages Target of phonological processes Neither homorganic nor heterorganic

Klamath, Moroccan Arabic, Mutsun, Somali, Tatar, Turkish, Uyghur, Zoque

O rgannly homo ic Udi and Leti Only heterorganic none

obstruent + lateral

Both homorganic and h

Choctaw, Kazakh, Koasati, Korean, Mara, an, Yakut, eterorganic Ponape Toba Batak,

Yindjibarndi N omorgheterorganic

, Let oqueither h anic nor Kuman i, Mutsun, Z e

Only homorganic none Only heterorganic none

obstruent + rhotic

Both homorganic and Ponapean, Rendille, Toba Batak heterorganic

rhotic sequences in the surveyed languages

a. Udi (Schulze 2001)

h/)

Table 3.3 Phonological patterning of homorganic vs. heterorganic obstruent plus lateral or

In the other surveyed languages not shown in the table above, only homorganic or

heterorganic obstruent plus liquid sequences occurred in each language. Thus, it was

impossible to determine the phonological patterning of homorganic vs. heterorganic

obstruent plus liquid sequences.

Among the surveyed languages, Udi and Leti are representative cases illustrating the

pattern that homorganic obstruent plus lateral sequences are more likely targets of

phonological processes than heterorganic obstruent plus lateral sequences.

(6) Cases with modification only in homorganic obstruent plus lateral sequences

(Obstruents: /p, p’, t, t’, k, k’, q, q’, b, d, g, ts, ts’, dÛ, tÇ, tÇ’, tS, tS’, dZ, f, v, s, z, Ç, Û, S, X, Â,

Change /dl/ [dd] / / [ oddu ] ‘woods’ No change

Xodlu X Â

/pl/ [pl] /pläkän/ [pläkän] ‘stairs, ladder’ /ql/ [ql] /qoqla/ [qoqla] ‘egg’ /Sl/ [ Sl] /aSl-aXun/ [aSl-aXun] ‘work-abl.’

-55-

(Obstruents: /p, t 1, d, k, v, s/) b. Leti (van Engelenhoven 1995)

Change

/kuda + let 1i/ [kullet 1i] /dl/ [ll] /dudal/ [dulla] ‘horn’

‘Letinese horse’ No change

/vl/ [vl] /ai + vlakar/ [aivlakar] ‘wood + crossed’ /pl/ [pl] /kapla/ [kapla] ‘ship’

Okla] ‘limepot’ /sl/ [sl] /sEsla/ [sEsla] kind of gecko

As illustrated in (6a), in Udi the homorganic /dl/ sequence undergoes assimilation and

surfaces as [ll]. On the other hand, heterorganic obstruent plus lateral sequences do not

undergo any phonological process. In Udi, /r/ does not occur in word-initial or post-

n (

a coronal stop /d/, the sequence is targeted b process. Thus, the /dl/

sequence is realized a f assim n the case of the /t1l/ sequence,

lateral assimilation is not attested. Instead, voicing assimilation occurs, e.g. /pO!t1la/

[pO!t1l ≤a] ‘bottle’ (Van Engelenhoven 1995). On the

coronal consonant, lateral or voicing assimilation is not found in the sequence. The

homorganic /sl/ sequ fication either. To my

nowledge, there are no other heterorganic obstruent plus lateral sequences in the

ange is found whether

the preceding obstruent is homorganic or heterorganic to /r/, e.g. /motru/ [motru]

‘motorboat’, /m-dupra/ [mdupra] ‘you flee’, /sikri/ [sikri] ‘Sikir’ (van Engelenhoven

1995).

/kl/ [kl] /sOkla/ [s

consonantal positio Schulze 2001). In Leti, as shown in (6b), when a lateral is preceded by

y a phonological

s [ll] as the result o ilation. I

other hand, when a lateral follows a non-

ence does not undergo any phonological modi

k

languages. When /r/ is preceded by an obstruent, no phonological ch

-56-

ences, the general pattern is that

ences heterorganic

logical s ges

The typological survey thus shows that homorganic obstruent plus liquid sequences

are more vulnerable to phonological change than heterorganic obstruent plus liquid

sequences, as predicted by the segment contact account.

3.2.2.2 Liquid plus obstruent sequences

In the case of liquid plus obstruent sequ

homorganic sequ are more likely to undergo phonological processes than

sequences. This is illustrated in Table 3.4.

Target of phono processe LanguaN orgheterorganic

, Mutsun, Rendille, Sidamo, Tatar, Toba Batak, Turkish, Zoque

either hom anic nor Kazakh, Klamath, Korean, Mara

Only homorganic Koasati, Rendille and Somali Only heterorganic Kuman

lateral + obstruent

Both homorganic and Choctaw, Finnish, Ponapean, Yakut heterorganic Neither homorganic nor DEg, Kazakh, Mara, Mutsun, Rendille,

Yakut, Zoque heterorganic Sidamo, Somali, Tatar, Toba Batak, Udi,

Only homorganic none Only heterorganic none

rhotic + obstruent

heterorganic Both homorganic and Ponapean, Finnish

rhotic plus obstruent sequences in the surveyed languages Table 3.4 Phonological patterning of homorganic vs. heterorganic lateral plus obstruent or

In some surveyed languages, only homorganic or heterorganic liquid plus obstreunt

sequences are attested. Thus, the patterning of homorganic sequences and that of

heterorganic sequences cannot be compared.

-57-

e showing the pattern that homorganic lateral plus obstruent

sequenc

Change

As a representative cas

es are more vulnerable to phonological modifications than heterorganic lateral plus

obstruent sequences, patterns found in Rendille are given below:

(7) Rendille (Sim 1981, Zaborski 1986): a modification only in homorganic lateral plus obstruent sequences (Obstruents: /b, t, d, ˇ, tÇ, d◊, k, g, f, s, x, ©, h/)

/golol-s-aˇ/ [golossaˇ] ‘eat [ss] /d¢e!l-t-a/ [d¢e!ssa] ‘you will give birth’

/maxabal-ti/ [maxabassi] ‘the woman who…’ No change

/ls/ [ss] /tolol-s-aˇ/ [tolossaˇ] ‘stand up’

/lt/

/lx/ [lx] /bo!lxe/ [bo!lxe] ‘feel pain, I/ he’ /l©/ [l©] /jil©-o!/

/lk/ [lk] /kulku!ld¢-/ corat

In Rendille, the homorganic /lt/ and /ls/ sequences do not occur within a morpheme, and

hen the sequences occur across a morpheme boundary, they undergo phonological

t sequences

surface unmodified.

Among 31 languages surveyed, Kuman is the only language illustrating the pattern

that homorganic liquid plus obstruent sequences are more likely to undergo phonological

processes than heterorganic liquid plus obstruent sequences. In Kuman, the heterorganic

/K b/ sequence is targeted by a phonological process while the homorganic /Kk/ sequence

is not. An account of this pattern is provided in section 3.7.

[jil©o] ‘charcoals’[kulku!ld¢-] ‘de e self’

w

change (Sim 1981). On the other hand, the heterorganic lateral plus obstruen

m

-58-

nces and that of obstruent/liquid

quences in the surveyed languages, homorganic C/liquid sequences are more likely to

undergo phonological processes than heterorganic C/liquid sequences within a language. A

Chi-squared test was run on the results in Table 3.1 through Table 3.4. The generalization

based on the phonological patterning of homorganic vs. heterorganic C/liquid sequences in

the surveyed languages is shown to be statistically significant. The results of the test are as

follows:

Heterorganic sequences

3.2.3 Chi-squared Test

Based on the patterning of nasal/liquid seque

se

no modification modification

same as heterorganic sequences

I 52 (60.51)

II 32 (23.49)

Homorganic sequences

different from heterorganic sequences

III 33 (24.49)

IV 1 (9.51)

Total deviance = 14.85 Table 3.5 Contingency table of number of cases showing the same or different patterning of homorganic and heterorganic C/liquid sequences

In Table 3.5, numbers that are not in parentheses represent observed frequencies and the

ones in parentheses are expected frequencies. Cells I and II include the number of cases

showing the same patterns in both homorganic and heterorganic C/liquid sequences. On

the seq

the other hand, cell III and IV contain the number of cases illustrating different patterns in

uences. That is, cell III represents the case where a homorganic C/liquid sequence

undergoes phonological modification while a heterorganic C/liquid sequence remain

unmodified, and cell IV the case with the reversed pattern. The test shows that the

-59-

3.2.4 Summary

According to the typological survey, within a language homorganic sequences are

more likely to be targeted by a phonological process than heterorganic sequences in

C/liquid sequences whether the consonant involved is a nasal or an obstruent. This is the

predicted pattern within the segment contact account.

However, the syllable contact account predicts that both homorganic and

heterorganic C/liquid sequences are the same in either violating or not violating the Syllable

Contact Law. In addition, phonological processes involving liquid plus nasal or obstruent

sequences cannot be motivated by the Syllable Contact Law since the sequences do not

violate the law by having falling sonority over a syllable boundary. Therefore, to account for

the asymmetrical patterning of homorganic vs. heterorganic C/liquid sequences and

phonological modifications of liquid plus nasal or obstruent sequences, some additional

formal mechanism is needed. On the other hand, within the segment contact account they

s and

at of obstruent/sonorant sequences. The segment contact account predicts that

patterning of heterorganic C/liquid sequences and that of homorganic C/liquid sequences

within a language is significantly different (p < 0.001).

can be explained in terms of phonetic similarity between two segments in a sequence.

3.3 Sonorant/sonorant vs. obstruent/sonorant sequences

As discussed in chapter 2, the segment contact and the syllable contact accounts

make different predictions regarding the patterning of sonorant/sonorant sequence

th

-60-

ed by phonological processes than

sonoran

icted by the segment contact account,

norant/sonorant sequences are more likely to be modified by a phonological process

. As will be shown in

section

) Generalization II

In a given language, phonological modification involving obstruent/sonorant sequences

consonant is involved.

he generalization above is made by comparing the patterning of nasal/liquid sequences

ith that of obstruent/liquid sequences given in section 3.3.1, and the patterning of

asal/nasal sequences with that of obstruent/nasal sequences given in section 3.3.2.

sonorant/sonorant sequences are more likely to be target

t/obstruent sequences within a language since members of the former are

phonetically more similar to each other than those of the latter. On the other hand, the

syllable contact account predicts that obstruent plus sonorant sequences will be more likely

to undergo phonological processes than sonorant/sonorant sequences since the former

violate the Syllable Contact Law more severely. Since sonorant plus obstruent sequences do

not violate the Syllable Contact Law, this account predicts that the sequences will not be

subject to phonological modification due to syllable contact.

The typological survey shows that, as pred

so

than obstruent/sonorant sequences within a language, as stated in (8)

3.3.3, a Chi-squared test shows that differences in the phonological patterning of

sonorant/sonorant and obstruent/sonorant sequences in the surveyed languages is

statistically significant.

(8

implies that of sonorant/sonorant sequences when the same type of a sonorant

T

w

n

-61-

nt/liqu es

predicted by the segment cont uences are more likely

to be subject to ses t

This pattern is found by comparing the patterning of nasal plus liquid sequences with that

nt plus liquid sequences, and erning of liquid plus nasal sequences with

that of liquid plus obstruent sequences. I present the patterning of nasal or

bstruent/lateral sequen

Nasal plus liquid sequences are more likely targets of phonological processes than

obstruent plus liquid sequences within a language. There is no surveyed language where

obstruent plus liquid sequences are modified by phonological processes while nasal plus

3.3.1 Nasal/liquid vs. obstrue id sequenc

As act account, nasal/liquid seq

phonological proces han obstruent/liquid sequences within a language.

of obstrue the patt

o ces and that of nasal or obstruent/ rhotic sequences separately

since most surveyed languages illustrate different patterns in the sequences depending on

whether the liquid involved is a lateral or a rhotic.

3.3.1.1 Nasal plus liquid vs. obstruent plus liquid sequences

liquid sequences are not, as can be seen from Table 3.6:

-62-

Target of phonological processes Languages Both Kazakh, Koasati, Korean, Leti, Mara, Toba

Batak, Yakut, Yindjibarndi Neither DEg, Somali, only nasal + lateral Klamath, Moroccan Arabic, Tatar, Turkish,

Uyghur

nasal + lateral

obst + lateral vs.

only obst + lateral none Both Moroccan Arabic, Ponapean, Rendille, Toba

Batak Neither Leti, Somali only nasal + rhotic DEg

nasal + rhotic vs. obst + rhotic

only obst + rhotic none Table 3.6 Phonological patterning of nasal plus liquid vs. obstrue the surveyed languages

nt plus liquid sequences in

Surveyed languages not shown in the table above are indeterminate since only nasal plus

Obstruents: /p, b, t, d, k, g, /, c, tS, dZ, f, v, s, z, S, Z, x, h/) ences

liquid or obstruent plus liquid sequences are attested.

Patterns found in Tatar are provided below as representative of cases with a

modification only in nasal plus lateral sequences.

(9) Tatar (Poppe 1968) (Nasals: /m, n, N/, nasal + lateral sequ /nl/ [nn] /khayvan + lÄr/ [khayvannar] ‘animals’ /ml/ [mn] /khanIm + lÄr/ [khanImnar] ‘ladies’ obstruent + lateral sequences /kl/ [kl] /balIk + lÄr/ [balIklar] ‘fish’ /tl/ [tl] /at + lÄr/ [atlar] ‘horses’

In Tatar, /nl/ and /ml/ surface as [nn] and [mn], respectively. On the other hand, /kl/ and

/tl/ do not undergo any phonological modification. To my knowledge, there are no other

nasal or obstruent plus lateral sequences attested in the language.

-63-

ows t us lateral sequences are more likely

phonological processes than o ral sequences within a language.

Within the segment contact account, this icted pattern since members of nasal plus

lateral sequences are more phonetically similar to each other than those of obstruent plus

e , th orrectly predicts the

r truent plus la

phonological processes than a nasal pl se the former

iolates the Sylla re se than the latter.

.3.1.2 Liquid plus nasal vs. liquid plus obstruent sequences

hile liquid plus nasal sequences are not, as can be

en from Table 3.7:

The typological survey sh hat nasal pl

targets of bstruent plus late

is a pred

lateral sequenc s. On the other hand e syllable contact account inc

opposite patte n: an obs teral sequence will be more likely to undergo

us lateral sequence. This is becau

v ble Contact Law mo verely

3

It is shown from the survey that, as predicted by the segment contact account,

liquid plus nasal sequences are more vulnerable to phonological modification than liquid

plus obstruent sequences within a language. There is no surveyed language where liquid

plus obstruent sequences are modified w

se

-64-

Languages Target of phonological processes B Choctaw, Finnish, Koasati, Kumanoth , Ponapean,

li Rendille, SomaNeither DEg, Kazakh, Kla ath, Swedish, Tatam r, Uyghur,

Zoque o teral + na Korean, Leti, Ma , Moroccannly la sal ra Arabic, Sidamo,

Toba Batak, Turkish, Udi

lateral + nasal vs. lateral + obst

only lateral + obs none t B Finnish, Pona Swedisoth pean, h Neither Kazakh, Klama eti, Moroccan Arath, L bic, Somali,

jibarndi, Zoque Tatar, Toba Batak, Uyghur, Yakut, Yind

only rhotic + nasal DEg, Mara, Rendille, Sidamo, Udi

rhotic + nasal vs. rhotic + obst

only rhotic + obst none

nasal vs. liquid plus obstruent sequences in the surveyed languages

s nasal sequences

go ph l chan

ara, some lateral plus nasal sequences are modified. However, all lateral plus obstruent

, =, ≠ N/, Ob b, d , Ô,

Table 3.7 Phonological patterning of liquid plus

Mara is a representative case illustrating the pattern that lateral plu

are more likely to under onologica ge than lateral plus obstruent sequences. In

M

sequences attested surface unmodified.

(10) Mara (Heath 1981) (Nasals: /m, n , struents: / , d1, ˇ g/) lateral + nasal sequences

Change: /ln/ [l] /bal + niwi-Ôa≠i/ [baliwiÔa≠i] ‘we wrote it’

/n/ [] /ba + niwi-ini/ [baiwiini] ‘we sneaked along’

No change:

/lm/ [lm] /balma=a/ [balma=a] ‘hat’

/l≠/ [l≠] /≠il-≠iluna/ [≠il≠iluna] ‘trevally fish’ /wu-≠in/ [wu≠in] ‘genital cover’

/lN/ [lN] /ˇalNunÔi/ [ˇalNunÔi] ‘frilled lizard’ /N/ [N] /buNu/ [buNu] ‘ashes’

/l=/ [l=] /wul + =i|i|i/ [wul=i|i|I] ‘small ones’

/m/ [m] /mbimu|/ [mbimu|] ‘grub sp.’

/≠/ [≠]

-65-

late ral + obstruent sequences

/lˇ/ [lˇ] /ˇil + ˇil/ [ˇilˇil] ‘feathered stick’

/b/ [b] /ubab / [ubab] ‘juvenile euro’

/b/ [b] /biba|/ [biba|] ‘neck’

/lg/ [lg] /balgi≠/ [balgi≠] ‘salty’ /lÔ/ [lÔ] /gulÔi/ [gulÔi] ‘grinding stone’ /Ô/ [Ô] /waÔa/ [waÔa] ‘dugong and turtles’

Patterns found in Rendille are given below as representative of cases with a

modification only in rhotic plus nasal sequences.

/ld/ [ld] /guldu|u/ [guldu|u] ‘bonefish’

/ˇ/ [ˇ] /ˇima + ˇima/ [ˇimaˇima] ‘(cloud) to appear’

/lb/ [lb] /balba/ [balba] ‘river’

/ g/ [ g] /ba gan/ [ba gan] ‘mangrove sp.’

(11) Rendille (Sim 1981, Zaborski 1986) (Nasals: /m, n, ≠, N/, Obstruents: /b, t, d, ˇ, k, g, tÇ, dÛ, f, s, x, ©, h/) rhotic + nasal sequences /rn/ [rn] or [rr] /fu !rne/ [fu!rre] or [fu !rne] ‘we opened’ /rm/

[rm] /karamu/ [karmo] ‘calabash (pl.)’

rhotic + obstruent sequences /rt/ [rt] /isle!i !rten/ [isle!i !rten] ‘they went together’ /rb/ [rb] /garbo !/ [garbo !] ‘shou /rˇ/ [rˇ] /fi !rˇe/ [fi !rˇe] ‘escap

lder blades’ e, I/he’

/rg /rf/

/wa!rase/ [wo!rse] ‘I mixed’ /ka!rtÇa/ [ka!rtÇa] ‘cook! (pl.)’

In Rendille, only the homorganic /rn/ sequence can optionally be a target of phonological

change. Rhotic plus obstruent sequences never undergo phonological change. To my

knowledge, rhotic plus nasal or obstruent sequences which are not given in (20) do not

occur in the language.

/ [rg] /ˇa!rge/ [ˇa!rge] ‘be full, I/he’ [rf] /dÛi !rfe/ [dÛi !rfe] ‘plait, I/he’

/rs/ [rs] /rtÇ/ [rtÇ]

-66-

gs disc n generalize that in a given language, a

ical process involving liquid plu uent sequences implies that of liquid plus

obstruent seque ed by

lateral plus nasal sequence are phonetically more similar to each other than those of a lateral

t s ce, the former is

th owever, with ount, it is predicted

incorrectly that ill ap us nasal or lateral plus obstruent

gical

sal or obstruent sequences need to be motivated by a factor

ther than the Syllable Contact Law.

uage.

Among

Based on the survey findin ussed, we ca

phonolog s obstr

nces. This is predict the segment contact account. Since members of a

plus obstruen equen expected to be more vulnerable to phonological

t accchange than e latter. H in the syllable contac

no modification w ply to lateral pl

sequences since the sequences respect the Syllable Contact Law. Thus, phonolo

modifications of liquid plus na

o

3.3.2 Nasal/nasal vs. obstruent/nasal sequences

The typological survey illustrates that nasal/nasal sequences are more likely to

undergo phonological modification than obstruent/nasal sequences within a lang

31 languages surveyed, the opposite pattern is found only in Leti.

The patterning of nasal/nasal vs. obstruent/nasal sequences in the surveyed

languages is as follows:

-67-

cesses Languages Target of phonological proBoth Chukchee, Korean, Ponapean, Sidamo, Toba

Batak Neither , Oromo, Somali DEgonly nasal + nasal Choctaw, Koasati, Mara, Selayarese, Tatar,

Uyghur

nasal + nasal vs. obst. + nasal

only obst + nasal none Both Choctaw, Ch ee, Kimatuumbi, ukch Koasati,

Tatar, Korean, Ponapean, Selayarese, Sidamo,Toba Batak, Uyghur, Yakut

Neither DEg, Kuman, Somali only nasal + nasal Kazakh, Mara

nasal + nasal vs. nasal + obst.

only obst + nasal Leti Table 3.8 Phonological patterning of nasal/nasal vs. obstruent/nasal sequences in the

assimilation. In other

rveyed languages, nasal/nasal or obstruent/nasal sequences are not attested. Those

languages are considered as indeterminate cases and not shown in the tableau above.

The patterns found in Uyghur and Koasati are given as representative cases

illustrating the pattern that nasal/nasal sequences are more likely to undergo phonological

change than obstruent/nasal sequences:

(12) Representative cases with a modification only in nasal/nasal sequences

(Nasals: /m, n, N/, Obstruents: /p, b, t, d, k, g, q, G, /, tS, dZ, f, s, z, S, Z, x, h/) sal sequences

surveyed languages

In some surveyed languages, heterorganic nasal/nasal sequences are not attested while

heterorganic nasal/obstruent sequences are modified by place

su

a. Uyghur (Hahn 1991)

nasal + na /nm/ [mm] /kälgänmän/ [kälgämmän] ‘I have come’

nasal + obstruent sequenes

/tm/ [tm] /kätma/ [kätma] ‘go, negative

/sm/ [sm] /ismim/ [ismim] ‘my name’ (loanword from Farsi)

-68-

. Koasati (Kimball 1991) (Nasals: /m, n/, Obstruents: /p, b, c, k, /, F,

n

bÒ, s/)

nasal + nasal seque ces /am-na:sinc /mn/ [nn] a !:ka/ [anna:sinca!:k ‘my book’

obstruent + nasal sequences /Fn/ [Fn] /ho !Fnan/ omething’ /sn/ [sn] /aka!snon/ [aka!snon] ‘to be hungry (pl.)’

ng’

/kn/ [kn] /acokna!/ [acokna!] ‘pick [tool]’ /sm/ [sm] /mismi !hlin/ [mismi !hlin] ‘to flutter the eyelids’

/hm/ [hm] /okso !hmon/ [okso!hmon] ‘to be slushy’

at something’

3.3.3 Chi-squared Test

Based on the patterning of nasal/liquid vs. obstruent/liquid sequences and that of

nasal/nasal vs. obstruent/nasal sequences in the surveyed languages, I conclude that within

a language sonorant/sonorant sequences are more likely to undergo phonological processes

ant sequences. A Chi-squared test was performed on the combined

finding

[ho !Fnan] ‘to smell s /hn/ [hn] /Fotohna!:cin/ [Fotohna!:cin] ‘to be good-smelli /tn/ [tn] /matna:Òi !:ka/ [matna:Òi !:ka] ‘radio’

/tm/ [tm] /matmici!ksin/ [matmici!ksin] ‘to raise the head and look

/km/ [km] /i !kmot/ [i !kmot] ‘They did not gather it’

In Koasati, the /mn/ sequence surfaces as [nn] as the result of place assimilation. However,

no phonological process applies to obstruent plus nasal sequences.

than obstruent/sonor

s in Table 3.6 through Table 3.8. The test illustrates that the result is statistically

significant.

-69-

son/son sequences no modification modification

same as son/son I 28 (20.45) II 39 (46.55) sequences

obst/son sequences

different from III 1 (8.55) IV 27 (19.45) obst/son sequences

Total deviance = 13.60 Table 3.9 Contingency table of number of cases showing the same or different

erved frequencies and the

ones in parentheses are expected frequencies.

orant sequences within a language (p <

mma

According to the survey, sonorant/sonorant sequences are more likely to be

ph ical process than ent/sonorant sequences within a language.

ed by the segment contac t

incorrectly pred nt pl (e.g. /tl/) is more likely to

patterning of sonorant/sonorant and obstruent/sonorant sequences

In Table 3.9, numbers that are not in parentheses represent obs

Cell I and II include cases showing the same

patterns in both sonorant/sonorant and obstruent/sonorant sequences. On the other hand,

cases illustrating different patterns in the sequences are in cell III and IV. That is, cell III

represents the case where a sonorant/sonorant sequence undergoes phonological

modification while a obstruent/sonorant remain unmodified, and cell IV the case with the

reversed pattern. The test shows that the patterning of sonorant/sonorant sequences is

significantly different from that of obstruent/son

0.001).

3.3.4 Su ry

targeted by a onolog obstru

This is predict t account. However, the syllable contact accoun

icts that an obstrue us sonorant sequence

undergo a phonological process than a sonorant plus sonorant sequence (e.g. /nl/) since

-70-

ent sequences, a factor

ther than the Syllable Contact Law needs to be assumed. This is because the sequences do

Law ave falling sonority over a syllable

With respect to stop/sonorant sequences, both the segment contact and the syllable

contact

Languages

the former violates the Syllable Contact Law more severely than the latter. In addition, to

account for phonological processes involving sonorant plus obstru

o

not violate the Syllable Contact since they h

boundary.

3.4 Stop plus sonorant vs. sonorant plus stop sequences

accounts predict that, all else being equal, stop plus sonorant sequences will be

more vulnerable to phonological modification than sonorant plus stop sequences. The

typological survey shows that this prediction is supported when the sonorant consonant

involved is a liquid. Among the surveyed languages, there is no language modifying only

liquid plus stop sequences, while there are languages with a modification only in stop plus

liquid sequences, as can be seen from Table 3.10.

Target of phonological processesBoth Choctaw, Ponapean, Yakut Neither DEg, Finnish, Leti, Mutsun, Somali, Tatar,

Turkish, Udi, Uyghur only stop + lateral none

stop + lateral vs. lateral + stop

only lateral + stop Kazakh, Koas orean, Mara, Tobati, K a Batak Both none Neither DEg, Finnish, Kazakh, Leti, Mutsun, Somali,

Turkish onl top + rho none y s tic

stop + rhotic vs. rhotic + stop

only rhotic + stop Rendille, Toba Batak Table 3.10 Phonological patterning of stop plus liquid vs. liquid plus stop sequences in the surveyed languages

-71-

Languages

However, in the surveyed languages, no differences in patterning are found in stop

plus nasal vs. nasal plus stop sequences, as illustrated in Table 3.11.

Target of phonological processesBoth Finnish, Korean, Leti, Mara, Mutsun, Ponapean,

Sidamo, Toba Batak Neither DEg, Rendille, Selayarese, Somali only stop + nasal Choctaw, Koasati, Swedish, Turkish, Uyghur

stop + nasal vs.

nasal + stop

only nasal + stop Chukchee, Oromo, Udi Table 3

ergo phonological modification, while

liquid plus released stop sequences do not, as illustrated below:

.11 Phonological patterning of stop plus nasal vs. nasal plus stop sequences in the surveyed languages

In addition, within the segment contact account, it is predicted that, in a given

language, unreleased stop plus sonorant sequences will be more likely targets of

phonological processes than sonorant plus released stop sequences since the perceptibility

of unreleased stop consonants are relatively weak in this context. For example, in Toba

Batak unreleased stop plus liquid sequences und

(13) Toba Batak (Nababan 1981) stop + liquid /tl/ [kl] /huhu!t + la!o/ [huhu!kla!o] ‘go at the same time’ /tr/ [kr] /ju !t + roha !na/ [ju!kroha !na] ‘he is exasperated’ /pl/ [kl] /rap + la!o/ [rakla!o] ‘go together’ /pr/ [kr] /rap + ro !/ [rakro] ‘come together’ liquid + stop /lt/ [lt] /palta!k/ [palta!k] ‘clearly visible’ /rp/ [rp] /ga!rpu/ [ga!rpu] ‘fork’ /lp/ [lp] /talpe/ [talpe] ‘on the brink’ /rt/ [rt] /a!rta/ [a!rta] ‘treasure’

-72-

ith a vowel. […] A released final /p/ sounds foreign to a Batak ear […]. Like the /p/, /t/

final position is as a rule not released. […] The /k/ in final position, in constrast with the other two voiceless stops /p/ and /t/, is always released.” (Nababan 1981:

Thus, l ch fo words are followed by a liquid can

be und ein of unreleas the

case of liquid plus equences, the stop has relatively good cues to its

identification by occurring in prevocalic position. Thus, the sequences surface unmodified. As

will be discussed, word-final /k/ does not undergo any phonological modification in Toba

Batak when it is followed by a liquid-initial word.

The segment contact and the syllable contact accounts also make different

predictions concerning the patterning of stop plus sonorant sequences when a stop

consonant is released in preconsonantal position. The segment contact account predicts

that if a stop consonant is released preconsonantally in a language and no segment specific

phonotactic constraint is involved, stop plus sonorant sequences can remain unchanged

since released stop consonants in preconsonantal position have relatively good

perceptibility. On the other hand, the syllable contact account predicts stop plus sonorant

sequences to be vulnerable to phonological processes even when stop consonants are

released preconsonantally. Phonological patterning of released stop plus sonorant vs.

In Toba Batak, word-final /t/ and /p/ are realized as [k] when a liquid-initial word follows.

On the other hand, no phonological change applies to liquid plus /t/ or /p/ sequences.

Nababan (1981) notes that in Toba Batak /k/ is different from /t/ and /p/ in that it is

always released in final position.

“When final the /p/ is never released except when the following word begins w

in

2-3)

phonologica ange und when /t/ or /p/-final

erstood as b g motivated by the weak perceptibility ed /t/ or /p/. In

intervocalic /t/ or /p/ s

-73-

is released

efore a nasal or a liquid, and the release takes the form of a very short vowel of the same

t/ are

not released before a nasal or a liquid. As illustrated in (14), /p/ or / sonorant and

/k/ plus sonorant sequences show different phonological patterning in the language.

(14) ation sto aw (N

unreleased stop plus sonorant sequences in Choctaw and Toba Batak supports the

prediction within the segment contact account. In Choctaw, the velar stop /k/

b

quality as the vowel preceding the /k/ (Nicklas 1974). On the other hand, /p/ and /

t/ plus

Sound altern s in p plus sonorant sequences in Choct icklas 1974) /tl/ [ll] /akatli/ [akalli] ‘to patch’ /pl/ [bl] /tapli/ [tabli] ‘to cut in two’ /kl/ [kl] /haklo/ [hakalo] ‘to hear’ /kn/ [kn] /pakna/ [pakana] ‘top’ /km/ [km] /hokmi/ [hokomi] ‘to burn up’ /p/ + nasal not attested /t/ + nasal not attested

Unre (i.e. / r /t rface modif re not attested,

hile released stop (i.e. /k/) plus sonorant sequences surface unmodified. I argue that this

atterning is related to the released or unreleased realization of a stop consonant. As

nd in Toba Batak. As illustrated in (15), word-

leased stop p/ o /) plus sonorant sequences su ied or a

w

p

discussed in chapter 2, when a stop consonant is unreleased in preconsonantal position, the

perceptibility of the stop is relatively weak (Wright 1996). On the other hand, when a stop

consonant is released preconsonantally, the perceptibility of the stop is relatively good. Thus,

released stop plus sonorant sequences are more stable and remain unchanged.

Recall from above that a similar pattern involving unreleased stop plus sonorant vs.

released stop plus sonorant sequences is fou

-74-

as [k] when another word beginning with a nasal or a liquid follows.

l or a liquid.

5) Sound alternations in stop plus sonorant sequences in Toba Batak (Nababan 1981)

final /t/ and /p/ surface

On the other hand, word-final /k/ surfaces unchanged before a nasa

(1/tm/ [km] /huhu!t + ma!rsak/ [huhu!k ma!rsak] ‘sad at the same

time’ /tn/ [kn] /huhu!t + no!lnol/ [huhu!k no!llol] ‘stare at the same

time’ /tN/ [kN] /huhu!t + No!t/ [huhu!k No!t] ‘rise at the same

time’ /tl/ [kl] /huhu!t + la!o/ [huhu!kla!o] ‘go at the same time’/tr/ [kr] /ju!t + roha !na/ [ju!kroha !na] ‘he is exasperated’

/pm/ [km] /raŸp + mo!dom/ [raŸkmo!dom] ‘sleep together’ /pn/ [kn] /a!lap + na/ [a!lakna] ‘get the one’ /pN/ [kN] /rap + No!t/ [rakNo!t] ‘get up together’ /pl/ [kl] /rap + la!o/ [rakla!o] ‘go together’ /pr/ [kr] /rap + ro !/ [rakro] ‘come together’

/km/ [km] /didok + ma + maddok + au/

[didokmamaddokau] ‘(he) said to me…’

/kn/ [kn] /halak + na + modom + i/ [halaknamodomi]

‘the sleeping man’

/kl/ [kl] /lak + lak/ [laklak] ‘rind’ /kr/ [kr] /rEk + rEk/ [rEkrEk] ‘shelf’

Similar to the situation in Toba Batak, in Choctaw, it can be argued that phonological

change of /t/ or /p/ plus sonorant sequences are motivated by the weak perceptibility of

o modification is found in /k/ plus sonorant sequences since /k/

is re

, the following

eneral patterns are found from the survey.

unreleased /t/ or /p/. N

leased and can be assumed to have relatively good cues to its identification.

With respect to released or unreleased stop/sonorant sequences

g

-75-

(16) Gea. In a given language, phonological modification involving liquid plus released stop

b. In a given language, phonological modification involving released stop plus sonorant

.5 Homorganic vs. heterorganic non-liquid/nasal sequences

ing of homorganic and heterorganic nasal/non-

liquid s

/,

tc.) are more likely targets of phonological modification than homorganic sequences (i.e.

following general pattern is found in nasal plus non-liquid

sequences from the typological survey.

(17) Generalizatio

id

neralization III

sequences implies that of unreleased stop plus liquid sequences.

sequences implies that of released stop plus sonorant sequences

Within the segment contact account, the different patterning of released stop plus sonorant

vs. unreleased stop plus sonorant sequences in Choctaw and Toba Batak is predicted.

However, the patterning is unexpected within the syllable contact account.

3

In this section, I discuss the pattern

equences, that is, nasal/nasal or obstruent sequences. As in C/liquid sequences, in

obstruent plus nasal sequences, homorganic sequences are more likely to be subject to

phonological processes than heterorganic sequences. However, nasal plus non-liquid

sequences illustrate the opposite pattern: heterorganic sequences (i.e. /nm/, /mn/, /np

e

/nt/, /mp/, etc.). That is, the

n IV

In a given language, a phonological process involving homorganic nasal plus non-liqu

sequences implies that of heterorganic nasal plus non-liquid sequences.

-76-

t, as in C/liquid sequences,

omorganic sequences will be more likely to undergo phonological change than

s non-liquid sequences. I argue that the patterning of

liquid consonant sequences given in (17) is

and

The typological survey shows that in obstruent plus nasal sequences, homorganic

sequences are more likely to be subject to phonological modification than heterorganic

sequences, as in C/liquid sequences. There is no surveyed language with a modification

only in heterorganic obstruent plus nasal sequences, as can be seen from the Table 3.12.

Target of phonological change Languages

When we consider perceptual similarity between two segments in a sequence as a

factor motivating phonological change in son/C clusters, the opposite pattern is expected

within the segment contact account. That is, it is expected tha

h

heterorganic sequences in nasal plu

homorganic and heterorganic nasal plus non-

motivated by a nasal’s weak internal cues to place in postconsonantal position

articulatory simplification.

3.5.1 Obstruent plus nasal sequences

Neither homorganic nor Klamath, Koasati, Moroccan Arabic, Mutsun, heterorganic Somali, Tatar, Yindjibarndi, Zoque Only homorganic Leti, Ponapean Only heterorganic None

obst + nasal

Both homorganic and Chuchee, Korean, Rendille, Sidamo, Toba heterorganic Batak

Table 3.12 Phonological patterning of homorganic vs. heterorganic obstruent plus nasal sequences in the surveyed languages

-77-

ses than heterorganic sequences.

) , n

Homorganic obstruent + nasal sequences

Among the surveyed languages, Ponapean is representative of cases showing the

pattern that in obstruent plus nasal sequences, homorganic sequences are more likely

targets of phonological proces

(18) Ponapean (Rehg & Sohl 1981 (Nasals: /m , N, mw/, Obstruents: /p, pw, t, s, k, c#/)

/tn/ [tV netenet] ‘smell’ /pm/ [m + kelik /pmw/ [mwm ] /kehp + m ot/ ehm m ot] ‘variety of yam’ /pwmw/ [mw t of breath’ /kN/ [NN Nget/ [e saiN Net] ‘He’s not yet out of

breath’ Heterorganic obstruent + nasal sequences

n] /net + net/ [m] /kehp + meir

w w/ [kehmmeirkelik] ‘variety of yam’ [k w w

mw] /mwopw + mwopw/ ] /e saik

[mwomwmwop] ‘ou

/km/ [k an/

In Ponapean, whe t is followed

r the obstruent changes into a nasal. On the other hand, when a heterorganic nasal follows

These patterns are predicted by the segment contact account when perceptual

nt, it is incorrectly predicted that all

are homorganic or heterorganic.

id sequences

n-liquid (that is, nasal or obstruent) sequences, a

honological process involving homorganic sequences in a language implies that of

m] /ak + manam [akmanaman] no gloss given

n an obstruen by a homorganic nasal, a vowel epenthesizes

o

an obstruent, no modification occurs.

similarity between two segments involved in a sequence is considered as a factor motivating

phonological change. Within the syllable contact accou

heterorganic obstruent plus nasal sequences will undergo phonological change regardless of

whether the sequences

3.5.2 Nasal plus non-liqu

As stated in (17), for nasal plus no

p

-78-

n is made based on the observation that there is

no surveyed language where only homorganic nasal plus non-liquid sequences are modified,

as can be seen in Table 3.12:

Target of phonological processes Languages

heterorganic sequences. This generalizatio

Neither homorganic nor heterorganic

Kuman, Yindjibarndi, Zoque

Only homorganic none Only heterorganic Fur, Koasati, Oromo, Rendille, Sidamo,

Yakut

nasal + nasal

Both homorganic and heterorganic

Chukchee, Kazakh, Mara

Neither homorganic nor anic

Klamath, Somali heterorgOnly homorganic none Only heterorganic Chukchee, Kazakh, Klamath, Moroccan

Arabic, Mutsun, Oromo, Rendille, Sidamo, Swedish, Uyghur

nasal+obstruent

innish, Fur, Kikamba, Kimatuumbi, Koasati, Korean, Ponapean, Tatar, Toba Batak, Yakut,

Both homorganic and Fheterorganic

Zoque Table 3.12 Phonological patterning of homorganic vs. heter ganic nasaor l plus non-liquid sequences in the surveyed languages

Surveyed languages not seen in the table above are indeterminate because only homorganic

or heterorganic nasal plus non-liquid sequences occur.

Koasati is a representative case showing modification in only heterorganic nasal

plus nasal sequences.

(19) Koasati (Kimball 1991)

Homorganic nasal + nasal sequences (Nasals: /m, n/)

/mm/ [mm] /am-ma!:ma/ [amma!:ma] ‘my mother’

-79-

Heterorganic nasal + nasal sequences /mn/ [nn] /am-na:sinca !:ka/ [anna:sinca!:k ‘my book’

In Koasati, the heterorganic /mn/ sequence is realized as [nn] while the homorganic /mm/

sequence surfaces unchanged. To my knowledge, other nasal plus nasal sequences are not

attested in the language.

The patterns in Uyghur illustrate modification in only heterorganic nasal plus

obstruent sequences:

(20) Uyghur (Hahn 1991)

Homorganic nasal + obstruent sequences

(Nasals: /m, n, N/, Obstruents: /p, b, t, d, k, g, q, G, /, tS, dZ, f, s, z, S, Z, x, h/)

/nd/ [nd] /män + dä/ [mändä] ‘at my place’

Heterorganic nasal + obstruent sequences

/np/ [mp] /qalGan + pul/ [qalGmpul] ‘(the) remaining money’

/nk/ [Nk] /bügünki/ [bügüNki] ‘today’s’

In Uyghur, the homorganic /nd/ surfaces unmodified. On the other hand, when /n/ is

followed by a heterorganic obstruent, place assimilation applies to the sequence.

Unlike in C/liquid sequences, the typological survey results illustrate that

heterorganic

/nb/ [mb] /on + bir/ [ombir] ‘eleven’

/ng/ [Ng] /körüngän/ [körüNgän] ‘having appeared’

sequences are more likely to be targets of phonological processes than

omorganic sequences in nasal plus non-liquid sequences. In most of the languages

surveyed, the nasal in a nasal plus non-liquid sequence undergoes place assimilation. Thus

far, perceptual similarity of two segments has been considered as the main factor in

explaining phonological change found in son/C clusters. If only perceptual similarity is in

h

-80-

ce of

rticulation in postconsonantal position. As he points out, nasals have fairly weak internal

cues to place in this position and thus a loss of place contrasts in coda nasals is observed in

different languages. In addition, following Kohler (1990), I assume that place assimilation

involving nasal plus non-liquid sequences can be motivated by articulatory simplification.

According to Kohler, a reduction in the number of articulators is favored in speech

production so that energy expenditure is minimized. Winters (2003) provides strong

empirical evidence that the vulnerability of heterorganic nasal plus stop sequences to

phonological change is motivated to avoid the articulatory difficulty of coordinating the

production of two or more gestures at the same time. Based on production experiments in

peakers produced VCCV type stimuli, he found that

or ce

enc rs also found that speakers

nces by producing them as homorganic

equences are more likely targets

of a phonological process than heterorganic sequences in obstruent plus nasal sequences as

play, however, we would expect homorganic nasal plus nasal sequences to be more likely to

undergo phonological modification than heterorganic nasal plus nasal sequences within a

language.

Following Wright (1996), I argue that place assimilation involving heterorganic

nasal plus non-liquid sequences is motivated by the weak perceptibility of a nasal’s pla

a

which Dutch and English s

heterorganic nasal plus stop sequences were more difficult f speakers to produ

accurately than homorganic nasal plus stop sequ es. Winte

simplified heterorganic nasal plus stop seque

sequences.

3.5.3 Summary

According to the typological survey, homorganic s

-81-

in C/liq

3.6 Kuman: articulatorily motivated change

As mentioned in section 3.2, Kuman illustrates a pattern conflicting with the

generalization that homorganic sequences are more likely targets of phonological processes

than heterorganic sequence in liquid plus consonant sequences. In Kuman, heterorganic

sequences are modified while the homorganic /Kk/ sequence

surface

(Lynch 1983, Trefry 1969)

/Kmb/ [mb] /praKmbuka/ [prambuka] ‘hear, 3 sg. future’ /kaK-na/ [katna] ‘leg, 1 sg.’

/Kn/ [r] (In 3 nouns) /agiK-na/ [agira] ‘brother, 1 sg.’ /Km/ [m] /praKmnga/ [prmnga] ‘hear, 1 pl. future’

Alternations found in liquid plus consonant sequences given above cannot be

argued to be motivated by syllable contact since the sequences do not violate the Syllable

Contact Law. In addition, the modification of /Kn/ into [tn], shown above, illustrates that

the Syllable Contact Law is not respected in the language.

uid sequences. I assume that this pattern is motivated by phonetic similarity of two

segments in a sequence. On the other hand, unlike in C/liquid clusters, heterorganic

sequences are more likely to undergo phonological modification than homorganic

sequences in nasal plus nasal or obstruent sequences. I argue that this pattern is motivated

by nasals’ weak internal cues to place in coda position and articulatory simplification.

lateral plus consonant

s unmodified, as illustrated below:

(21) Sound alternations in lateral plus obstruent sequences in Kuman

/Kk/ [Kk] /yunaKka/ [yunaKka] ‘I will bring’

/Kn/ [tn]

-82-

Based on the observation that the velar lateral /K/ only occurs before the velar /k/

in Kuman, I argue that the alternations above are motivated to avoid the articulatory

complexity of coordinating two or more gestures at the same time similar to the

modifications seen in nasal plus nasal or obstruent sequences in section 3.6.2. However, the

articulatory factor is not alone in motivating phonological modification of son/C clusters in

Kuman: the homorganic /rn/ sequence is modified to [tn] i he

articulatory factor were responsible for the change, we would expect the homorganic /rn/

regarding the patterning of the clusters

patterns in son/C clusters.

n the language. If only t

sequence to remain unchanged. In this case, it can be said that the change occurs due to

perceptual similarity of two segments involved in a sequence.

3.7 Summary

Findings from the typological survey show that the patterning of son/C clusters is

correctly predicted and explained by the segment contact account. On the other hand, the

syllable contact account makes incorrect predictions

and, as a result, some additional formal mechanisms are needed to account for the observed

-83-

ed in Hume (1998).

C H A P T E R 4

FORMAL ACCOUNT OF THE PATTERNING OF SON/C CLUSTERS

In this study, a formal account of the patterning of son/C clusters will be provided

within

ns found in son/C clusters. I use Avoid C/X constraints, as

propos

an Optimality Theoretic (OT) framework, which assumes that a given surface form

is derived from an input by means of a universal set of ranked constraints (Prince &

Smolensky 1993; McCarthy & Prince 1993, 1995). To account for the general cross-

linguistic patterning of son/C clusters from the typological survey, I posit Avoid C/X

constraints. In this chapter, I first discuss the motivation for these constraints and then

posit specific universally ranked Avoid C/X constraints to account for the general

patterning of son/C clusters in different languages.

4.1 Motivation for Avoid C/X constraints

As discussed in chapter 3, it is assumed in the present study that the cross-linguistic

patterning of son/C clusters is motivated by speech perception. Following Steriade (1995a,

b, 1997), Jun (1995) and Hume (1998), I posit perceptually based constraints to formally

account for the general patter

-84-

perceptually weak.

ted for by

mal account of phonological change occurring in

tautosy

(1) AVOID C/X Avoid positioning a consonant (C) in a context (X) in which it is

For example, to account for the general pattern that a sonorant/sonorant sequence tends to

be avoided, I posit the constraint *son/____ son (or *son/son___). In this case, the

context X is defined as adjacent to another sonorant. However, since X can be any context

in which the perceptibility of a segment is weak, X need not to be a segment; it could also

be the end of a word or phrase, for example.

Avoid C/X constraints can be said to be more general than the Syllable Contact

Law in that they cover not only phonological modifications involving son/C clusters which

can be accounted for by syllable contact, but also those which cannot be accoun

the law. For example, as discussed in chapter 2, phonological processes involving

tautosyllabic son/C clusters and liquid plus nasal or obstruent sequences are problematic

cases for the Syllable Contact Law; they do not violate syllable contact since there is no

syllable boundary between the two members of the cluster. Further, in the case of

heterosyllabic liquid plus nasal or obstruent sequences, there would be falling sonority over

a syllable boundary and thus would not violate syllable contact. Therefore, the Syllable

Contact Law cannot provide a for

llabic son/C clusters and liquid plus nasal or obstruent sequences. However, since

Avoid C/X constraints do not crucially refer to a syllable boundary or a sonority contour,

phonological modifications occurring in such types of son/C clusters can be analyzed by

employing these constraints.

Some Avoid C/X constraints look much alike OCP constraints, such as

*son/___son as noted above. However, Avoid C/X constraints have wider coverage than

the Obligatory Contour Principle (OCP) which is defined as in (2):

-85-

At the melodic level, adjacent identical elements are prohibited.

According to Pierrehumbert (1993), the OCP operates in accordance with general

principles in the perception of similarity. As she points out, the strength of an OCP effect

on place in Arabic increases with perceived similarity between two target segments. For

example, according to Pierrehumbert (ibid.), in Arabic the labials /f/ and /m/ can be said

to be more similar to each other than the coronals /s/ and /n/. This is because the two

labials differ only in the two features [sonorant] and [continuant], while the two coronals

differ in more features. Thus, /f/ and /m/, which are more similar to each other than /s/

and /n/, are subject to the OCP for place in Arabic, while /s/ and /n/ are not.

Avoid C/X constraints are similar to the OCP in that they can prohibit the

occurrence of phonetically similar segments in a sequence. The avoidance of a homorganic

consonant sequence, for example, can be formulated in the form of Avoid C/X constraints

n be considered as formal

he v ew tha a between two sounds

ac ustical tw sounds is avoided

ven though the members of the cluster do not share any common

atures. As discussed in chapters 2 and 3, the weak perceptibility can lead to phonological

change in a son/C cluster. Remember that an unreleased stop before a sonorant consonant

in Choctaw and Toba Batak underwent phonological processes, while no phonological

process occurred when a released stop was before a sonorant consonant in the languages.

(2) Obligatory Contour Princple (OCP) (McCarthy 1986: 208)

or the OCP. Thus, both Avoid C/X constraints and the OCP ca

mechanisms representing t i t a sm ll modulation occurring

in a sequence due to the o and auditory similarity of o

(Kawasaki 1982, Ohala 1992, 1993).

However, Avoid C/X constraints are different from the OCP in that they can be

employed to account for phonological modifications involving a son/C cluster which are

perceptually poor e

fe

-86-

al plus non-liquid sequences

tend to

CP. In fact, OCP

onstraints are subsumed by Avoid C/X.

revea d the following general patterns in son/C clusters.

less likely to undergo ess a phonological process

liquid heterorganic C/liquid

on-liquid homorganic nasal + non-liquid

nts and

general cross-linguistic patterning of son/C clusters within

In addition, the typological survey shows that heterorganic nas

be avoided in intervocalic position due to the poor perceptibility of nasals in coda

position. Each of these cases can be represented in simple terms by means of Avoid C/X

constraints. For example, to account for the pattern that an unreleased stop plus sonorant

sequence tends to be avoided, we can posit the constraint *stopunrel/___son which prohibits

the occurrence of an unreleased stop before a sonorant consonant. The OCP, on the other

hand, has nothing to say regarding these cases since the segments do not crucially share any

common features. Thus, Avoid C/X has wider coverage than the O

c

4.2 Avoid C/X constraints

The typological survey le

(3) The patterning of son/C cluste more likely to undergo

rs a phonological proc. homorganic C/a

b. sonorant/sonorant obstruent/sonorant + rel. stop c. unrel. stop + sonorant sonorant

t rel. stop + sonorant d. unrel. stop + sonoran heterorganic nasal + ne.

In sections 4.2.1 through 4.2.5, I posit universally ranked Avoid C/X constrai

provide a formal account of the

an OT framework.

-87-

uences are more

likely t

morganic

quences with that of heterorganic sequences in nasal/liquid sequences and

es. To account for the pattern that homorganic nasal/liquid

se al p he nasal/liquid

sequences, the following ranked Avoid C/X constraints can be posited.

) a. *Hom nas/liq

| |

| | consonant.

positioning a liquid after a heterorganic nasal

[α pl] [β pl]

sal

The constraint *Hom nas/liq prohibits a

4.2.1 Homorganic vs. heterorganic C/liquid sequences

The typological survey illustrates that homorganic C/liquid seq

argets of phonological processes than heterorganic C/liquid sequences within a

language. This generalization was made by comparing the patterning of ho

se

obstruent/liquid sequenc

quences are more likely to undergo a phonologic rocess than terorganic

(4

*liquid/nasal _____ : Avoid positioning a liquid after a homorganic nasal consonant.

[α pl] [α pl]

*liquid/_____ nasal: Avoid positioning a liquid before a homorganic nasal

[α pl] [α pl] b. *Het nas/liq *liquid/nasal _____ : Avoid | | consonant.

iquid before a heterorganic na *liquid/_____ nasal: Avoid positioning a l | | consonant. [α pl] [β pl] c. *Hom nas/liq >> *Het nas/liq

liquid from occurring before or after a

homorganic nasal consonant. In addition, the constraint *Het nas/liq penalizes the

occurrence of a liquid before or after a heterorganic nasal consonant. The universal ranking

-88-

ile the heterorganic /ml/ and /Nl/ sequences do not.

Change

in (4c) formalizes the observation that homorganic nasal/liquid sequences are more likely

targets of phonological change than heterorganic nasal/liquid sequences.

For example, as illustrated below, in Uyghur the homorganic /nl/ sequence

undergoes assimilation, wh

(5) Uyghur (Hahn 1991)

No change

/nl/ [ll] /hayvan + lar/ [hayvallar] ‘animals’

/ml/ [ml] /khanIm + lar/ [khanImlar] ‘ladies’

rical patterning of homorganic vs. heterorganic nasal plus lateral sequences in

ore important to preserve a lateral than a

asal. Thus, assuming that the features [lateral] and [nasal] are privative, the following two

al]: Every feat e [lateral] of an input segment has an identical correspondent in the output.

an input segment has an identical correspondent th

l] >> Max [n al]

According to Lombardi (1995), the Max[feature] family of constraints plays a crucial

le in explaining the different patterns of alternation in response to coda restrictions on

[voice] and place of articulation. Following Lombardi (ibid.), I make use of Max [feature]

/Nl/ [Nl] /SinjaN+liq/ [SinjaNliq] ‘(person) of Xinjiang’

The asymmet

Uyghur can be explained by the ranked constraints in (4c). However, to select the optimal

output [ll] from the input /nl/, additional constraints are required. The observation that

/nl/ surfaces as [ll] in Uyghur suggests that it is m

n

ranked constraints can be posited.

(6) a. Max [later ur b. Max [nasal]: Every feature [nasal] of in e output. c. Max [latera as

ro

-89-

(7) Iden

yghur it is more important to keep a lateral than

nasal when the /nl/ sequence is modified, the constraints Max [lateral] and Max [nasal]

Through the interaction of *Hom nas/liq, *Het nas/liq, Max [lateral] and Max

rorganic nasal plus lateral

quences in Uyghur can be analyzed as follows:

hayvan + lar/ [hayvallar] ‘animals’ iq

constraints as opposed to Ident [feature]. This is because the constraints Ident [lateral] and

Ident [nasal] given in (7) cannot account for why assimilating /n/ to /l/ is a more optimal

modification than assimilating /l/ to /n/ since both repairs trigger violations of Ident

[lateral] and Ident [nasal].

t [lateral]: Corresponding segments are identical with respect to the feature [lateral].

Ident [nasal]: Corresponding segments are identical with respect to the feature [nasal].

Thus, to account for the pattern that for U

a

are posited.

[nasal], the asymmetrical patterning of homorganic vs. hete

se

(8) i. / /hayvan + lar/ *Hom nas/liq Max [lateral] Max [nasal] *Het nas/l a. hayvanlar *! b. hayvannar *! c. hayvallar * ii. /khanIm + lar/ [khanImlar] ‘ladies’

/khanIm + lar/ *Hom nas/liq Max [lateral] Max [nasal] *Het nas/liq a. khanImlar *

b. khanImnar *! c. khanIllar *!

-90-

ith [ll] is selected as optimal.

n the other hand, as illustrated in tableau in (8ii), a heterorganic nasal plus lateral sequence

adjacent to an obstruent, as opposed

ore likely to

gical than ences

consonant.

r after a heterorganic obstruent

consonant.

*liquid/obst _____ : Avoid positioning a liquid after a heterorganic obstruent | | consonant.

____ obst: Avoid sitioning a liquid efore a heterorg nic obstruent | conso nt.

[α pl] [β pl]

q

As shown in tableau in (8i), candidate (a) with /nl/ is ruled out by the highly ranked

constraint *Hom nas/liq. Due to Max [lateral] ranked above Max [nasal], modifying /nl/ to

[nn], as in candidate (b), is non-optimal. Thus, candidate (c) w

O

does not violate any highly ranked constraint and thus surfaces unmodified.

Consider now sequences in which the liquid is

to a sonorant. To account for the pattern that homorganic sequences are m

undergo phonolo processes heterorganic sequ in such sequences, I posit the

following ranked Avoid C/X constraints.

(9) a. *Hom obst/liq: Avoid positioning a liquid before or after a homorganic obstruent *liquid/obst _____ : Avoid positioning a liquid after a homorganic obstruent | | consonant. [α pl] [α pl] *liquid/_____ obst: Avoid positioning a liquid before a homorganic obstruent | | consonant. [α pl] [α pl]

b. *Het nas/liq: Avoid positioning a liquid before o

[α pl] [β pl] *liquid/_ po b a | na c. *Hom obst/liq >> *Het obst/li

-91-

obstruent sequences t, d, ˇ, tÇ, d◊, k, g, f, s, x, ©, h/)

Change

With the ranked constraints in (9c), we can account for why homorganic lateral plus

obstruent sequences undergo phonological change in Rendille, while heterorganic lateral plus

obstruent sequences do not, as the examples below illustrate.

(10) Rendille (Sim 1981, Zaborski 1986): a modification only in homorganic lateral plus

(Obstruents: /b,

/ls/ /golol-s-a / [golossa ] ‘eat

[ss] /tolol-s-aˇ/ [tolossaˇ] ‘stand up’ ˇ ˇ

/lt/ [ss] /ˇe!l-t-a/ [ˇe!ssa] ‘you will give birth’ /maxabal-ti/ [maxabassi] ‘the woman who…’ No change /lx/ [lx] /bo!lxe/ [bo!lxe] ‘feel pain, I/ he’ /l©/ [l©] /jil©-o!/ [jil©o] ‘charcoals’ /lk/ [lk] /kulku!ld¢-/ [kulku!ld¢-] ‘decorate self’

In Rendille, the homorganic /lt/ and /ls/ sequences do not occur within a morpheme, and

s occur across a morpheme boundary, they undergo phonological

es

ranked constraints in (9c), the realizations of /ls/ and /lx/ as

1) l-s-aˇ/ [tolossaˇ] ‘stand up’

*Hom obst/liq Max [lateral] *Het obst/liq

when the sequence

change (Sim 1981). On the other hand, the heterorganic lateral plus obstruent sequenc

surface unmodified. With the

[ss] and [lx], respectively, can be analyzed as follows:

(1i. /tolo /tolol-s-aˇ/ a. tololsaˇ *! b. tolossa * ˇ ii. /bo!lxe/ [bo!lxe] ‘feel pain, I/ he’ /bo!lxe/ *Hom obst/liq Max [lateral] *Het obst/liq a. bo!lxe * b. bo!xxe *!

-92-

te (a) with the homorganic /ls/ cannot be selected as

ptimal since it violates the highly ranked constraint *Hom obst/liq. Thus, candidate (b) is

ly ranked constraint *Het obst/liq, while candidate (b)

iolates the faithfulness constraint Max [lateral] (or Ident [lateral]). Thus, candidate (a) is

selected as optimal.

As we have seen, the typological survey shows that within a language homorganic

sequences are more likely to be targeted by a phonological process than heterorganic

sequences in C/liquid sequences. This is the case regardless of whether the consonant

involved is a nasal or an obstruent. The constraints in (4) and (9) can thus be considered as

specific instances of the more general constraints in (12). This then unifies the observation

that C may be a nasal or an obstruent.

(12) a. *Hom C/liq

C _____ : Avoid positioning a liquid after a homorganic consonant.

*liquid / ___ C : Avoid positioning a liquid before a heterorganic consonant. | |

[

[α pl] [β pl]

As shown in tableau (11i), candida

o

selected as optimal. On the other hand, as illustrated in tableau (11ii), candidate (a) with the

heterorganic /lx/ violates the low

v

*liquid/_____C : Avoid positioning a liquid before a homorganic consonant. | | [α pl] [α pl] *liquid/ | | [α pl] [α pl]

b. *Het C/liq

α pl] [β pl]

*liquid / C ___ : Avoid positioning a liquid after a heterorganic consonant. | |

-93-

he constraint ranking in (12c) holds when C is identical in both homorganic and

es.

bstruent/sonorant sequences

tion, 13) t servation

nd n e honological change

nerally,

e

e type

f a sonorant consonant is involved.

*son/_____son >> *son/___obst

n (13c) account for assimilation occurring in nasal

lus nasal sequences in Koasati. Recall that obstruent plus nasal sequences surface

c. *Hom C/liq >> *Het C/liq

T

heterorganic C/liquid sequenc

4.2.2 Sonorant/sonorant vs. o

In this sec I posit the ranked constraints in ( o account for the ob

that nasal/liquid a asal/nasal sequ nces are more likely to undergo p

than obstruent/liquid and obstruent/nasal sequences, all else being equal. More ge

we can state that within a language sonorant/sonorant sequences are more likely to b

targeted by a phonological process than obstruent/sonorant sequences when the sam

o

(13) a. *son/son *son/_____son: Avoid positioning a sonorant consonant before another sonorant consonant. *son/ son ____: Avoid positioning a sonorant consonant after another sonorant consonant b. *son/obst *son / ___ obst: Avoid positioning a sonorant consonant before an obstruent. *son / obst ___: Avoid positioning a sonorant consonant after an obstruent.

c. *son/son >> *son/obst

*son/son ___ >> *son/obst ___

To illustrate, the ranked constraints i

p

-94-

rved.

ll 1991) (Nasals: /m, n/, Obstruents: /p, b, c, k, /, F, Ò, s/)

nasal + nasal sequences

unmodified. In Koasati, to my knowledge, nasal plus nasal sequences other than /mn/ and

a geminate nasal are not obse

(14) Koasati (Kimba

/mn/ [nn] /am-na:sinca !:ka/ [anna:sinca!:k] ‘my book’

l se obstruent + nasa quences /Fn/ [Fn] /ho !Fnan/ [ho !Fnan] ‘to smell something’

/aka!snon/ [aka!snon] ‘to be hungry (pl.)’ /Fotohna!:cin/ [ otohna!:cin] ‘to be good-smelling’

/tn/ [tn] /matna:Òi !:ka/ [matna:Òi !:ka] ‘radio’ n] /acokna!/ [acokna!] ‘pick [tool]’

/mismi n/ [mismi !hlin] ‘to flutter the eyelids’ hm] /okso !hmon/ [okso!hmon] ‘to be slushy’

/matmici!ksin/ [matmici!ksin] ‘to raise the head and look at something’

/km/ [km] /i !kmot/ [i !kmot] ‘They did not gather it’

constraints and their

ranking

b. Pres-Lab: Preserve a labial. c. Pres-Dor: Preserve a dorsal.

d. Pres-Cor, Pres-Dor >> Pres-Lab

/sn/ [sn] /hn/ [hn] F /kn/ [k /sm/ [sm] !hli /hm/ [ /tm/ [tm]

To account for the observation that /mn/ is realized as [nn], not as *[mm], we can assume

that for Koasati it is more important to preserve a coronal than to preserve a labial. This is

also manifested in the modification of /mt/ and /ml/ into [nt] and [mm], respectively, e.g.

/am-towa!/ [antowa!] ‘my onion’, /am-layki!/ [al-layki!] ‘my dung’ (ibid.). Thus, to

account for the modification of /mn/ to [nn], I posit the following

.

(15) a. Pres-Cor: Preserve a coronal.

-95-

oasati, a dorsal never undergoes place assimilation, implying that Pres-Dor is ranked

e Pres-La

e constraints in (13) and (15), the modif can be

ollows:

!:k] ‘my book’ /am-na:sinca!:ka/ *son/son ___ Pres-Cor Pres-

Dor *son/obst___ Pres-Lab

In K

abov b as in (15d).

With th ication of /mn/ to [nn]

analyzed as f

(16) /am-na:sinca!:ka/ [anna:sinca

a. amna:sinca!:k *! b. am ma:sinca!:k

[+son]

*!

c. an na:sinca!:k

[+son] [+son]

*! * | |

!:k

[+son]

* d. an na:sinca

In the tableau above, candidate (a) is ruled out by violating *son/son___. Candidate (b)

does not violate *son/son___ since [mm] is a geminate sharing the feature [+sonorant].

However, candidate (b) cannot be selected as optimal since it violates Pres-Cor. Candidate

) with a non-geminate [nn] is excluded by violating *son/son___, and thus candidate (d)

ot violate the

r example, [i!kmot] is chosen as the optimal

n (17i):

(c

violating the lowly ranked constraint Pres-Lab is selected as optimal.

Obstruent plus nasal sequences surface unmodified since they do n

highly ranked constraint *son/son___. Thus, fo

output of the input / i!kmot/, as shown in tableau i

-96-

i. /i !kmot/ [i !kmot] ‘They did not gather it’

/i !kmot/ [son] Dor

(17)

*son/son___ Ident- Pres-Cor Pres- *son/obst___ Pres-Lab

a. i !kmot * b. i !kkot *! * c. i!mmot *! * * ii. Ident-[son]: Corresponding segments are identical with respect to the feature [±sonorant].

As can be seen from he tableau abov te (a), which is most faithful to the inpt e, candida ut,

th timal. On

ent-[son] by

l into an obstruent. Candidate (c) is ruled out by Ident-[son] and Pres-Dor.

leas uences

To account for the general pattern that an unreleased stop plus sonorant sequence

re a sonorant.

sitioning a released stop after a sonorant.

c. *stopunrel/___ son >> *stoprel/son ____

l patterning of unreleased stop plus

liquid vs. liquid plus released stop sequences in Toba Batak can be explained. As discussed

does not violate any of the highly ranked constraints, and us is selected as op

the other hand, candidate (b) cannot be selected as optimal since it violates Id

changing the nasa

4.2.3 Unreleased stop plus sonorant vs. sonorant/re ed stop seq

is more likely to be subject to phonological modification than a sonorant/released stop

sequence in a given language, I posit the ranked constraints in (18).

(18) a. *stop /___ son: Avoid positioning an unreleased stop befounrel

b. *stoprel/son ____: Avoid po

d. *stopunrel / ___ son >> *stoprel / ___ son

With the constraints in (18c), the phonologica

-97-

in secti

an 1981). When /t/ or /p/-final words are followed by a liquid, the

nreleased stop /t/ or /p/ is modified to [k], as illustrated in (19). On the other hand, no

released by appearing in prevocalic

intervocalic liq plus /t/ or /p/ equences surfac unmodified.

ons 3.4 and 3.5, in Toba Batak /t/ and /p/ are not released in word-final position

while /k/ is (Nabab

u

phonological modification is found when /t/ or /p/ is

position. Thus, uid s e

(19) Toba Batak (Nababan 1981) stop + liquid /tl/ [kl] /huhu!t + la!o/ [huhu!kla!o] ‘go at the same time’

/pl/ [kl] /rap + la!o/ [rak /tr/ [kr] /ju !t + roha !na/ [ju!kroha !na] ‘he is exasperated’

la!o] ‘go together’ /pr/ [kr] /rap + ro !/ [rakro] ‘come together’ liquid + stop /lt/ [lt] /palta!k/ [palta!k] ‘clearly visible’ /rp/ [rp] /ga!rpu/ [ga!rpu] ‘fork’

[talpe] ‘on the brink’ [rt]

To account for the asymmetrical patterning of unreleased stop plus sonorant vs. sonorant

plus released stop sequences in Toba Batak, I posit the following additional constraints.

(20) a. Ident-Ons: Corresponding segments in the onset are identical.

c. Ident-Ons >> Ident-Coda

The constraint ranking in (20c) is based on the view that some positions such as prevocalic

position facilitate perception of contrasts more than other positions such as postvocalic

position (Steriade 1993, Beckman 1997, among others).

/lp/ [lp] /talpe/ /rt/ /a!rta/ [a!rta] ‘treasure’

b. Ident-Coda: Corresponding segments in the coda are identical.

-98-

With the constraints in (18c) and (20), the modification of /tl/ into [kl] can be

s:

s / Id Ident-Coda

analyzed as follow

(21) /huhu!t + la!o/ [huhu!kla!o] ‘go at the same time’ /huhu!t + la!o/ * topunrel ___ son ent-Ons *stoprel/son ____ a. huhu!t}la!o *!

b. huhu!kla!o * *

In the tableau above, candidate (a) is ruled out since it has the unreleased stop [t}] plus

sonorant sequence. Candidate (b) violates the lowly ranked constraints *stoprel/son ____

and Ident-Coda, and thus is selected as optimal.

The realization of /lt/ as [lt] can be analyzed as follows:

/palta!k/ *stopunrel/___ son Ident-Ons *stoprel/son ____ Ident-Coda (22) /palta!k/ [palta!k] ‘clearly visible’

a. palta!k * b. palka!k *! *

To rule out candidate (b), Ident-Ons needs to be ranked above *stop/son ____. Otherwise,

candidate (a) is incorrectly selected as optimal.

constraints in (18d), the patterning of released stop plus sonorant vs. With the

unreleased stop plus sonorant sequences in Toba Batak can be analyzed as follows:

-99-

(23i. / [a!lakna] ‘‘get the one’

) a!lap + na/

/a!lap + na/ *stopunrel/ ___ son *stoprel/____ son a. a!lap}na *! b. a!lakna * ii. /halak + na + modom + i/ [halaknamodomi] ‘the sleeping man’ /halak + na …/ *stopunrel/ ___ son *stoprel/____ son

a. halakna… * b. halap}na… *!

In Toba Batak, the occurrence of an unreleased stop before a sonorant is ruled out by

strategy, the unreleased stop is modified into [k] which

released word-finally in the language. On the other hand, /k/ plus sonorant sequences

organic vs. heterorganic nasal plus non-liquid sequences

4) a. *Het NC

p u n c ited.

ic nasal plus obstruent sequence is

l]

*stopunrel/___ son. Thus, as a repair

is

remain unmodified since the sequences do not violate *stopunrel/___ son.

4.2.4 Hom

The typological survey shows that within a language heterorganic sequences are

more apt to undergo a phonological process than homorganic sequences in nasal plus nasal

or obstruent sequences. To account for the pattern, I posit the following ranked constraints.

(2 *N N (*Het NN): A heterorganic nasal l s asal sequen e is prohib | | [α pl] [β pl] *N O (*Het NO): A heterorgan | | prohibited. [ pl] [β pα

-100-

*N N (*Hom NN): A homorganic nasal plus nasal sequence is prohibited.

| | [α pl] [α pl]

O ): n b nc | proh ited. l] [α pl]

C >> *Ho NC

*Het NN >> *Hom NN

N

he constraints *Het NN and *Het NO can be considered as specific instances of the more

Hom NC.

(Nasals: /m, n, N/, Obstruents: /p, b, t, d, k, g, q, G, /, tS, dZ, f, s, z, S, Z, x, h/)

b. *Hom NC

*N (*Hom NO A homorga ic nasal plus o struent seque e is | ib [α p c. *Het N m

*Het NO >> *Hom O

T

general constraint *Het NC. In addition, *Hom NN and *Hom NC can be collapsed as

*

With the ranked constraints in (24), the asymmetrical pattern of Uyghur

homorganic and heterorganic nasal plus obstruent sequences can be analyzed. In Uyghur, as

illustrated in (25), the homorganic /nd/ surfaces unmodified. On the other hand, when /n/

is followed by a heterorganic obstruent, place assimilation applies to the sequence.

(25) Uyghur (Hahn 1991)

Homorganic nasal + obstruent sequences /nd/ [nd] /män + dä/ [mändä] ‘at my place’

Heterorganic nasal + obstruent sequences

/on + bir/ [ombir] ‘eleven’

/np/ [mp] /qalGan + pul/ [qalGmpul] ‘(the) remaining money’ /ng /nk/ [Nk] /bügünki/ [bügüNki] ‘today’s’

/nb/ [mb]

/ [Ng] /körüngän/ [körüNgän] ‘having appeared’

-101-

rn can be analyzed as follows:

NO

By using the constraints in (24), the asymmetrical patte

(26) i. /on + bir/ [ombir] ‘eleven’

H t Ident-place /on + bir/ * e Ident-Ons Ident-[son] *Hom NO a. onbir *!

b. ombir * * c. ondir *! d. onnir *! * e. obbir *!

/män + dä/ *Het NO Ident-Ons Ident-[son] *Hom NO Ident-place ii. /män + dä/ [mändä] ‘at my place’

a. mändä * * b. männä *! c. mäddä *!

pondenet segments should be identical with respect to place features. (Davis & Shin 1999)

As illustrated in the tableau in (26i), candidate (a) with heterorganic /nb/ cannot be

selected as optimal because it violates the highly ranked constraint *Het NO. Assimilating

the onset /b/ to the coda /n/ in place or manner, as in candidates (b) and (c), cannot be an

optimal repair stratey due to the highly ranked *Ident-Ons. Candidate (c) is ruled out by

Ident-[son], and thus candidate (b) having violations of lowly ranked constraints is selected

as the optimal output. On the other hand, as shown in the tableau in (26ii), a homorganic

nasal plus obstruent sequence does not violate *Het NO. Thus, it surfaces unmodified.

4.3 Summary

In the formal account of the patterns found in son/C clusters, I proposed the

following universal ranking of perceptually motivated Avoid C/X constraints.

iii. Ident-place: Corres

-102-

(27) A summary of proposed Avoid C/X constraints more likely to undergo less likely to undergo Constraints a phonological process a phonological process a. homorganic C/liquid heterorganic C/liquid *Hom C/liq >> *Het C/liq b. sonorant/sonorant obstruent/sonorant *son/son >> *son/obst c. unrel. stop + sonorant sonorant + rel. stop *stopunrel/ ___ son >> *stop/son ____ d. unrel. stop + sonorant rel. s unrel/___ son >> *stoprel/ ___ son e. heterorganic nas. + homorganic nas. + *Het NC >> *Hom NC nas. or obst. nas. or obst.

With the proposed constraints and their universal rankings as a basis, I provide an analysis

of phon

top + sonorant *stop

ological modifications of son/C clusters in Korean in the following chapter. These

alternations are of particular interest since they have previously been used as motivation for

the Syllable Contact Law (Rice & Avery 1991, Iverson & Sohn 1994, and Davis & Shin

1999, among others).

-103-

C H A P T E R 5

contact and a segment specific positional

constraints in chapter 4 and a

striction prohibiting a lateral in prevocalic position.

r nasal is adjacent

When a ce zed as the nasal

a pattern asal or sto

/ml/ [mn] /sam + lyu/ [samnyu] ‘third rate’ /Nl/ [Nn] /y´N + lak/ [y´Nnak] ‘downfall’

CASE STUDY OF KOREAN

In this chapter, I provide an account of alternations involving son/C clusters in

Korean, which I argue are the result of segment

restriction, rather than resulting from syllable contact (see, e.g. Davis & Shin 1999). A

formal account is given within an Optimality Theoretic (OT) framework based on the

universal ranking of perceptually motivated Avoid C/X

re

5.1 Data

In Korean, phonological processes are observed when a lateral o

to a consonant. lateral is pre ded by a consonant, the lateral is reali

[n] or the consonant preceding the lateral becomes a lateral, as shown in (1).

(1) Phonologic l ing of n p + lateral sequences (Davis & Shin 1999) a. nasal plus lateral sequences /nl/ [ll] /non + li/ [nolli] ‘logic’

-104-

l’’ w’

. Thus, both /ml/ and /pl/ are

ttested except for the /ln/ sequence which is realized as [ll]. Since in Korean

N/ does not occur in onset position, /lN/ is not attested.

uen Shin 1999)

b. stop plus lateral sequences /tl/ [ll] /tikÈt + liÈl/ [tikÈlliÈl] ‘the letters ‘t’ and ‘ /pl/ [mn] /p´p + li/ [p´mni] ‘pinciple of la /kl/ [Nn] /ky´k + li/ [ky´Nni] ‘separation’

As illustrated above, when a lateral is preceded by a coronal consonant /n/ or /t/, the

coronal consonant becomes a lateral. Thus, both /nl/ and /tl/ surface as [ll]. On the other

hand, when a lateral is preceded by a non-coronal consonant, the lateral surfaces as [n]

accompanying the modification of the obstruent into a nasal

realized as [mn], and both /Nl/ and /kl/ as [Nn].

However, when a lateral occurs in pre-consonantal position, no phonological

modification is a

/

(2) Phonological patterning of lateral plus nasal or stop seq ces (Davis & a. lateral plus nasal sequences Change /ln/ [ll] /pul + nÈN/ [pullÈN] ‘incapability’ /s´l + nal/ [s´llal] ‘New Year’s Day’

No change

/tÓÈl + ni/ [tÓÈlli] ‘denture’

/lm/ [lm] /kal + maN/ [kalmaN] ‘desire’

/tÇal + mot/ [tÇalmot] ‘fault’

b. In lateral plus stop sequences2

No c

/pal + my´N/ [palmy´N] ‘invention’

hange

/mal + tÈN/ [maltÈN] ‘horse back’

/lt/ [lt] /kalte/ [kalte] ‘reed’

2 /kalte/ and /mal + tÈN/ surface as [kalt’e] and [malt’ÈN], respectively. If a voiceless stop is not tensified

el, it is voiced. Thus, for example, /kalpi/ and /kalku/ are realized as [kalbi] and algu], respectively. Depending on the word, a voiceless stop is tensified or voiced between a sonorant

consonant and a vowel, e.g. /mul + koki/ [mulk’ogi] ‘fish’ vs. /pul + koki/ [pulgogi] ‘bulgogi’. In accounting for the patterning of Korean son/C clusters in this study, such changes will be ignored.

between a lateral and a vow[k

-105-

bs’ [mulpola] ‘a spray of water’

and (2 eed to account for why

l p an occ antal position

except when followed by /n/. In addition, we need to provide an account of the modification

of an underlying C/l sequence into a particular surface form.

Phonological change is also attested in Korean when a nasal is adjacent to an

obstruent. When an obstruent is followed by a nasal, the obstruent becomes nasal, as

shown below:

(3) Phonological patterning of obstruent plus nasal sequences

/tn/ [nn] /mat + nui/ [mannui] ‘first sister’

/pn/ [mn] /sip + ny´n/ [simny´n] ‘ten years’ ppetite’ (<‘ rice’ + ‘taste’)

/kn/ [Nn] /m´k + nÈn/ [m´NnÈn] ‘eat, adj.’ /km/ [Nm] /kuk + mul/ [kuNmul] ‘soup-broth’

ce

Unlike in obstruent plus nasal sequences, in a nasal plus obstruent sequence the

/lp/ [lp] /kalpi/ [kalpi] ‘ri /mul + pola/ /lk/ [lk] /kal + ku/ [kalku] ‘desire’ /mal + kalki/ [malkalki] ‘a horse’s mane’

With respect to the C/lateral sequences illustrated in (1) ), we n

/l/ is not allowed in post-consonanta osition, while it c ur in pre-conson

(Davis & Shin 1999, Iverson & Lee 1995)

/tm/ [mm] /mat + my´nÈli/ [mammy´nÈ|i] ‘first daughter-in-law’

/pm/ [mm] /pap + mas/ [pammat] ‘a

In the /tm/ sequence, in addition to the change of an obstruent into a nasal, pla

assimilation occurs. Thus, /tm/ surfaces as [mm].

obstruent is not realized as a nasal. Rather, the nasal may be targeted by place assimilation,

as the examples below illustrate:

-106-

(Davis & Shin 1999, Iverson & Lee 1995)

/np/ [mb] /sin + pal/ [simbal] ‘shoes(-foot)’ Ng] /han + kaN/ [haNgaN] ‘Han river’

b. /mt/ [md] /tÇam + tÈlta/ [tÇamdÈlda] ‘fall asleep’ /mp/ [mb] /kim + pap/ [kimbap] ‘maki’ (< ‘laver’ + ‘rice’)

/mk/ [Ng] /kam-ki/ [kaNgi] ‘cold’ c. /Nt/ [Nd] /toN + tÈN/ [toNdÈN] ‘equality’

pi/ [tÇ´Nbi] ‘maintenance’ /Nk/ [Ng] /tÇaN + kun/ [tÇaNgun] ‘general’

A coronal nasal assimilates to a following labial or velar. In addition, a voiceless segment is

voiced between voiced segments in Korean. Thus, /VnpV/ and /VnkV/ surface as

[VmbV] and [VNgV], respectively. When a labial nasal is followed by a velar, the labial

assimilates to the velar. Thus, /VmkV/ surfaces as [VNgV]. However, as illustrated in (4c),

no place assimilation is observed when a velar nasal is followed by a consonant. Place

assimilation is also observed in nasal/nasal sequences. As the result of assimilation of a

coronal nasal to a following labial, the /nm/ sequence is realized as [mm], while other

nasal/nasal sequences surface unchanged, as illustrated below:

(5) Phonological patterning of nasal/nasal sequences

/mn/ [mn] /pam + namu/ [pamnamu] ‘chestnut tree’

With respect to the patterning of obstruent/nasal sequences, a phonological account must

predict why obstruent plus nasal sequences are modified into nasal plus nasal sequences

(4) Phonological patterning of nasal plus obstruent sequences3

a. /nt/ [nd] /kun + te/ [kunde] ‘army’

/nk/ [

/Np/ [Nb] /tÇ´N +

/nm/ [mm] /kun + mantu/ [kummantu] ‘fried dumpling’

/Nn/ [Nn] /tÇaN + nal/ [tÇaNnal] ‘market day’ /Nm/ [Nm] /tÇaN + ma/ [tÇaNma] ‘rain season’

ignored in this study in accounting for the patterning of Korean son/C clusters. 3 Voicing of a voiceless segment between voiced consonants, which is a low-level allophonic change, will be

-107-

5.2 Formal account

We now turn to a formal analysis of the data. In presenting an analysis of the

alternations of son/C clusters in Korean, factors motivating the alternations will be

discussed first and then an analysis of different types of son/C clusters in Korean will be

given within an OT framework.

5.2.1 Motivation

Previous accounts have employed the Syllable Contact Law to motivate

phonological modifications of nasal plus lateral sequences and those of obstruent plus

sonorant sequences in Korean (Davis & Shin 1999, Iverson & Sohn 1994, Rice & Avery

1991). That is, it has been assumed that a heterosyllabic nasal plus lateral sequence is

unacceptable since a lateral is more sonorous than a nasal and thus the sequence violates

the Syllable Contact Law. Likewise, obstruent plus sonorant sequences have been assumed

to be modified into sonorant plus sonorant sequences due to the Syllable Contact Law.

One problem with this account, however, is that in Korean the intervocalic /ln/ sequence

is modified into [ll], just as /nl/ is, although it is perfectly good with respect to the Syllable

while such a change does not occur in nasal plus obstruent sequences. In addition, the

patterns of place assimilation found in nasal plus nasal or obstruent sequences need to be

explained.

Contact Law. Thus, while syllable contact motivates the change of /nl/ [ll], a factor

other than syllable contact is required to motivate the change of the /ln/ sequence. In

Davis & Shin (1999), it is assumed that the /ln/ sequence is modified into [ll] to avoid

-108-

of syllable contact and

e avoidance of two similar consonants in a sequence.

n of syllable contact is also

ions regarding the patterning of son/C

lusters cross-linguistically. First, the syllable contact account cannot account for the

syllable contact account makes incorrect predictions regarding the patterning of

sonorant/sonorant sequences and that of obstruent/sonorant sequences. Within the

syllable contact account, it is predicted that, in a given language, an obstruent plus sonorant

sequence is more likely to be modified than a sonorant plus sonorant sequence since the

former violates the Syllable Contact Law more severely than the latter. However, according

to the typological survey, the opposite pattern is found: sonorant/sonorant sequences are

more likely to be targeted by a phonological process than obstruent/sonorant sequences

within a language. Finally, the syllable contact account cannot explain why an unreleased

stop plus sonorant sequence is more likely to undergo phonological change than a released

stop plus sonorant sequence as in Choctaw and Toba Batak.

Given this, I argue that the observed patterns in Korean are accounted for by

segment contact constraints and a segment specific positional constraint. I begin by

discussing the latter. I suggest that the change of nasal or obstruent plus lateral sequences in

having two similar consonants in a sequence. According to them, /l/ and /n/ are similar to

each other since both are sonorant coronal consonants. Thus, Davis & Shin explain the

alternations of son/C clusters in Korean by relying on the notion

th

As discussed in chapter 3, an account using the notio

problematic in that it makes incorrect predict

c

observation that within a language homorganic C/liquid sequences are more likely to be

targeted by a phonological process than heterorganic C/liquid sequences since both

sequences are considered the same with respect to the Syllable Contact Law. Second, the

Korean is motivated, first of all, by a restriction prohibiting a lateral in prevocalic position.

-109-

a. /lon + tÇeN/ [non

As discussed in section 5.1, in Korean a lateral does not occur in post-consonantal position.

In addition, a lateral is avoided in word-initial or intervocalic position as we can see from

the alternations of the Sino-Korean morpheme /lon/ ‘to discuss’.

(6) Alternations of Sino-Korean morpheme /lon/ ‘to discuss’ tÇeN] ‘argument’

b. /i + lon/ [iRon] ‘theory’

in (6a s position and it

p ow l surfaces in Korean

As t e e ample ) ho s, /l/ a ar as a asal [n] in word-initial h x s w ppe n

surfaces as a flap intervocalically as the exam le in (6b) s s. A laterah

only after another lateral as in /ky´l + lon/ [ky´llon] ‘conclusion’. Such realization

lateral in

n exc a

Independent evidence for the lateral restriction comes from the fact that Korean is

[a{alar] ‘fathers’ (Krueger

1962). The lateral can also occur in word-final position, e.g. /suol/ [suol] ‘road’, /dZilga/

patterns of /l/ indicate that in Korean there is a restriction disallowing a

prevocalic pos io ept when it art of gemina eit is p t .

not the only language with such a restriction. A similar restriction is observed in other

languages such as Turkish, Yakut, Kazakh, Kuman and Yindjibarndi. These five languages

are similar to Korean in that a lateral has a positional restriction although they differ from

one another in the scope of the restriction. In Turkish, for example, words of native

vocabulary do not begin with a lateral while a lateral can occur in other positions (Kornfilt

1997, Lewis 1975). Unlike Korean, Turkish also has /r/ and the same positional restriction

is found with /r/. Thus, the generalization is that in Turkish a liquid is prohibited in word-

initial prevocalic position. On the other hand, in Yakut /l/ can occur word-initially and

intervocalically, e.g. /laaX/ [laaX] ‘lacquer’, /a{a + lAr/

-110-

[dZil

en /l/ occurs in post-consonantal position as the result

f morpheme concatenation, the lateral is modified, as illustrated below:

of /l/ in post-consonantal onset position in Yakut (Krueger 1962)4

/

/Nl/ [Nn] /XatIN + lAr/ [XatINnar] ‘birches’ ll] /barar + lar/ [barallar] ‘they go’

hers’

r/ ngs’ /pl/ [pt] /Iskaap + lAr/ [Iskaaptar] ‘cabinets’ /kl/ [kt] /balIk + lAr/ [balIktar] ‘fishes’

/tl/ [tt] /at + lAr/ [attar] ‘horses’

/y/. When /l/ occurs after /r/, the /rl/ sequence can surface as either [ll] or [rd]

depending on the word. Thus, in Yakut there is a restriction prohibiting a lateral in post-

consonantal position. A similar positional restriction is attested with /r/ in Yakut; there is

no native Yakut word with /r/ in word-initial or post-consonantal position, although /r/

can occur in intervocalic or word-final position, e.g. /oron/ [oron] ‘land, country’,

/dZaXtar/ [dZaXtar] ‘woman’ (Krueger 1962). Other Yakut segments in post-consonantal

position surface unmodified, e.g. /oyun + tAn/ [oyuntan] ‘shaman, ablative’,

/udarnikk n/ [udarnikk n] ‘you are a shockworker’ (ibid.).

ga] ‘in a year’ (ibid.), and in pre-consonantal position, e.g. /küöl + KA/ [küölge]

‘to the lake’. However, /l/ is prohibited in post-consonantal position in Yakut unless it

occurs as a part of a geminate lateral, e.g. /barar + lar/ [barallar] ‘they go’, /küöl + lAr/

[küöller] ‘lakes’ (ibid.). Thus, wh

o

(7) Realization /nl/ [nn] /oron + lAr [oronnor] ‘beds’ /ml/ [mn] /olom + lAr/ [olomnor] ‘fords’ /rl/ [ /rl/ [rd] /kötör + lAr/ [kötördör] ‘birds’ /yl/ [yd] /ubay + lAr/ [ubaydar] ‘elder brot /sl/ [st] /muos + lAr/ [muostar] ‘horns’ /Xl/ [Xt] /tuoX + lA [tuoXtar] ‘what thi

In Yakut, /l/ is realized as [n] after a nasal, as [t] after an obstruent, and as [d] after a glide

I I

4 /A/ is the archiphoneme of [a], [e], [o], and [ö].

-111-

(Bekturova & Bekturov 1996, Kara 2002)

In Kazakh, as in Turkish, a lateral does not occur in word-initial position. In

addition, a lateral does not occur in post-consonantal position when the preceding

consonant is non-approximant. When it does occur in this context, it is realized as an

alveolar stop, as illustrated below:

(8) Realizations of /l/ in post-consonantal onset position in Kazakh

No Change

/rl/ [rl] /qiyar + lA /yl/ [yl] /ay + lAr/ [aylar] ‘moons, months’

r/ [qiyarlar] ‘cucumbers’ Change

/nl/ [nd] /aywan + lAr/ [aywandar] ‘animals’

is /köl +

[köldär] ‘lakes’ (ibid.), even though /l/ is an approximant. This can be explained by

.). Therefore, it can be said that Kazakh has a restriction prohibiting

/l/ in onset position unless the preceding consonant is an approximant.

/ml/ [md] /khanIm + lAr/ [khanImdar] ‘ladies’ /ql/ [qt] /balIq + lAr/ [balIqtar] ‘fish’ /tl/ [tt] /at + lAr/ [attar] ‘horses’

When /l/ is preceded by another /l/, the /ll/ se n modified to [ld], as inque ce

lAr/

the observation that a sonorant geminate is not allowed in Kazakh. Thus, a nasal geminate

is also modified, as in /minnet/ [mindet] ‘duty’, /kymmat/ [kymbat] ‘expensive’

(ibid.). The phonological changes illustrated above are not attested in other Kazakh

segments. So, for example, /m/ surfaces unmodified regardless of whether it is preceded

by an approximant or a non-approximant, as in /körmespin/ [körmespin] ‘I may not

see’, /kyzmet/ [kyzmet] ‘service’ (ibid.). In Kazakh, a lateral can occur intervocalically

and in pre-consonantal position, e.g. /alma + lAr/ [almalar] ‘apples’, /al + mA/

[alma] ‘don’t take’ (ibid

-112-

/nl/ [nd] /maygan + la/ [mayganda] ‘my gum tree’ (locative)

/µl/ /ˇl/

In Kuman, there are two different laterals: an alveolar lateral /l/ and a velar lateral

/K/. The two laterals do not occur in prevocalic position while they do occur in pre-

consonantal position, e.g. /olto/ [olto] ‘long’, /yunaKka/ [yunaKka] ‘I will bring’, and

in intervocalic position, e.g. /yalo/ [yalo] ‘plant, plural’, /yaKo/ [yaKo] ‘plant, singular’

(Lynch 1983, Trefry 1969). Yindjibarndi illustrates the same type of positional restriction of

a lateral; as in Kuman, neither the retroflex nor alveolar lateral occurs word-initially or in

prevocalic position.5 Thus, when /l/ occurs before a vowel, phonological processes apply,

as the following examples show:

(9) Realizations of /l/ in post-consonantal onset position in Yindjibarndi (Wordick 1982)

/¯l/ [¯Î] /yu|u¯ + la/ [yu|un¯Îa] ‘hair’ (locative) [µÔ] /yuÎa|iµ + lu/ [yuÎa|iµÔu] (proper name, instrumental) [ˇ] /ku¯ˇ + la/ [ku¯T

T \ T \

As illustrated above, when /l/ occurs in prevocalic position, it becomes a stop at the same

articulation as the preceding consonant. There is no geminate in Yindjibarndi, and

thus a single consonant surfaces in the last three examples. The rhotics /r/ and /|/ in this

language have a similar positional restriction: there is no word having a rhotic in prevocalic

position. Other segments do not have such a positional restriction. Thus, for example, /m/

can occur in onset position without any restriction, e.g. /maÊa|/ [maÊa|] ‘red ochre’,

/kara + = + mara/ [kara=mara] ‘thick scrub or woods’, /Nalinma/ [Nalinma] ‘to ask

to come along’ (ibid.).

ata] ‘daughter’ (locative) /Ôl/ [Ô] /kaNkaÔ + la/ [kaNkaÔa] ‘slack’ (locative) /|l/ [|] /ma a + la/ [ma a a] ‘red ochre’ (locative)

place of

5 In Kuman, the alveolar lateral /l/ is found only in the word /loopu/ [loopu] ‘Friday’ (Lynch 1983).

-113-

n differ

epending on the language. By including the prohibition against /l/ in prevocalic position in

f nasal or obstruent plus lateral sequences, we can provide a unified

count of the change of /l/ in post-consonantal, word-initial and intervocalic onset position

in Korean. That is, the modification of /l/ in these positions can be understood as a strategy

to avoid a lateral in prevocalic position. Note that in Davis & Shin (1999), the non-

occurrence of a lateral after a nasal or an obstruent is explained by employing the notion of

syllable contact. Since the Syllable Contact Law is not relevant when /l/ occurs in word-initial

and intervocalic position, it must be assumed within their account that a factor other than the

Syllable Contact Law motivates the change of a word-initial and intervocalic /l/.

Based on the argument that in Korean there is a restriction disallowing a lateral in

prevocalic position, the following constraint can be posited for the analysis of the

alternations of consonant plus lateral sequences.

eV-l): A lateral in prevocalic position is prohibited. [lateral]

I oe late *PreV-l since it has a

di e from is gle eral, a geminate

lateral has a single feature [lateral] multiply linked to two prosodic positions, as shown in

As discussed above, a positional restriction of /l/ (or more generally liquids) is

attested in languages other than Korean although the scope of the restriction ca

d

motivating the change o

ac

6

(10) *CV (*Pr |

assume that an intervocalic geminate lateral d s not vio

fferent structur the one given in (10). That , unlike a sin ton lat

6 According to the production experiment results in Seo (2002), in Sino-Korean words, /l/ in prevocalic position is mainly realized as a nasal, suggesting that the constraint prohibiting a lateral in prevocalic is still rigid. On the other hand, in loanwords it is mainly realized as a lateral, implying that the prohibition against a lateral in prevocalic position is not active in loanword pronunciation.

-114-

V [lateral]

In addition to the positional restriction of /l/, I assume that segment contact

influenced by perceptual factors motivates phonological alternations involving son/C

clusters, as discussed in chapter 2. That is, I assume that /nl/ and /ln/ sequences are

e perce ually similar to each other. This receives additional

pport from languages such as Uyghur where only the /nl/ sequence is targeted by a

phonological process while other consonant plus lateral sequences remain unchanged, as

follows:

(12) Uyghur (Hahn 1991)

(11). This is consistent with the predictions of the Linking Constraint (Hayes 1986) which

states that association lines in structural descriptions are interpreted as exhaustive.

(11) V C C

modified since /n/ and /l/ ar pt

su

Change

/nl/ [ll] /hayvan + lar/ [hayvallar] ‘animals’

No change

/Nl/ [Nl] /SinjaN+liq/ [SinjaNliq] ‘(person) of Xinjiang’

/zl/ [zl] /qIz + lar/ [qIzlar] ‘the girls’

/tl/ [tl] /at + lar/ [atlar] ‘horses’

/ml/ [ml] /khanIm + lar/ [khanImlar] ‘ladies’

/sl/ [sl] /äslik/ [äslik] ‘a ~ the memory’

/ql/ [ql] /beliq + lar/ [beliqlär] ‘fish’

/pl/ [pl] /äswa:b+lar/ [äswaplar] ‘tools’ /kl / [kl] /tSötSäg+lik/ [tSötSäklik] ‘person of Chughuchaq’

-115-

To account for the change of /nl/ and /ln/ in Korean, I employ the two ranked

guistic

attern that homorganic C/liquid sequences are more likely to undergo phonological

hange than heterorganic C/liquid sequences in a given language:

(13) a. *Hom C/liq: Avoid positioning a liquid before or after a homorganic consonant.

b. *Het C/liq: Avoid positioning a liquid before or after a heterorganic consonant.

c. *Hom C/liq >> *Het C/liq

I consider the perceptibility of a stop as a factor motivating a phonological

modification of intervocalic clusters of stop plus sonorant in Korean. When a stop occurs

in word-final or pre-consonantal position, it is not released in Korean. Given that the

perceptual cues for place and manner of articulation for stop consonants come mostly from

contextual cues such as release burst and vowel formant transitions (Blumstein & Stevens

1979, Halle et al. 1957, Male!cot 1956, Stevens & Blumstein 1978, Wang 1959, Wright

1996), the perceptibility of unreleased stop consonants without a release burst will be

latively weak, compared with that of released stop consonants with a release burst. The

ifferent phonological patterning of a released stop and an unreleased stop before a

sonorant consonant in Choctaw and Toba Batak, discussed in chapter 3, illustrates that the

As discussed in chapter 3, motivating the change of /nl/ and /ln/ in terms of similarity

also makes it possible to provide a unified account of the change of both heterosyllabic and

tautosyllabic /nl/ and /ln/ sequences as is found in languages such as Moroccan Arabic

and Leti.

constraints in (13). Recall from section 4.2.1 that these reflect the general cross-lin

p

c

re

d

-116-

e of a release burst can be a factor motivating phonological change.

Thus, I

rel

unrel

According to the ranked constraints in (14c), it is worse for an unreleased stop consonant

in nasal plus nasal or obstruent sequences in

the coda nasal and to simplify articulation, as

discussed in chapter 3. Recall from the typolog

onstraints were posited to account for this

general pattern:

(15) a. *Hom NC: A homorganic sequence of a nasal and a consonant is prohibited. b. *Het NC: A heterorganic sequence of a nasal and a consonant is prohibited. c. *Hom NC >> *Het NC

presence vs. absenc

would suggest that the change of stop plus nasal sequences in Korean is motivated

due to the low perceptibility of an unreleased stop consonant. I make use of the following

two ranked constraints, as proposed in chapter 4:

(14) a. *stopunrel / ___ son : Avoid positioning an unreleased stop before a sonorant consonant. b. *stop / son ____ : Avoid positioning a released stop before a sonorant consonant. c. *stop / ___ son >> *stop/son ____

to occur before a sonorant consonant than the case that a stop consonant occurs after a

sonorant consonant.

I assume that place assimilation found

Korean serves to enhance the perceptibility of

ical survey that heterorganic sequences are

more vulnerable to modification than homorganic sequences in nasal plus nasal or

obstruent sequences. The following ranked c

-117-

is section, I p ide an analysis o the alternations of son/C clusters in Korean

e interaction of constraints wit an OT framework. Let us first examine the

intervocali [ll]. As discussed above, I assume that the

is motivated to avoid having two similar consonants in a sequence. The

bservation that /ln/ surfaces as [ll] rather than as [nn] in Korean implies that it is more

portant to keep a lateral than a nasal when the /ln/ sequence is modified. Thus,

5.2.2 Analysis

In th rov f

in terms of th hin

change of the c /ln/ sequence into

change of /ln/

o

im

assuming that the features [lateral] and [nasal] are privative, the following two ranked

constraints can be posited.

(16) a. Max [lateral]: The feature [lateral] from an input segment is realized in the output. b. Max [nasal]: The feature [nasal] from an input segment is realized in the output. c. Max [lateral] >> Max [nasal]

Through the interaction of *Hom C/liq, *Het C/liq, Max [lateral] and Max [nasal],

we can select [s´llal] as the optimal output from the input /s´lnal/, as shown in tableau

(17):

-118-

/s´l + nal/ *Hom C/liq Max [lateral] Max [nasal] *Het C/liq (17) /s´l + nal/ [s´llal] ‘New Year’s day’

a. s´lnal *! b. s´ltal *! * c. s´nnal *! d. s´tnal *! e. s´l lal | | [cor] [cor]

*! *

f. s´l lal [cor]

*

In the tableau above, candidate (a), which is most faithful to the input, cannot be chosen as

the optimal output since it violates the constraint *Hom C/liq. Candidate (b) is also ruled

out by violating the constraint *Hom C/liq. Although candidates (c) and (d) do not violate

*Hom C/liq, they cannot be selected as optimal since they violate the constraint Max

[lateral]. Candidate (e) with a non-geminate [ll] is ruled out by *Hom C/liq. Candidate (f)

with a geminate lateral does not violate *Hom C/liq and it violates the lowly ranked

constraint Max [nasal]. Thus, it is chosen as the optimal output. In selecting candidate (f) as

optimal, it is important that Max [nasal] is ranked below the constraints *Hom C/liq and

Max [lateral]. Recall that *Het C/liq was ranked below *Hom C/liq to account for the

general cross-linguistic pattern that homorganic C/liq sequences are more likely to undergo

phonological change than heterorganic C/liq sequences in a given language. At this point,

no evidence is found that there is any ranking relation between Max [nasal] and *Het C/liq

and between Max [lateral] and *Het C/liq. Thus, we can establish the following constraint

ierarchy: h

-119-

[nasal]

C/liq *Het C/liq

We now turn to the analysis of the pattern found when a non-coronal nasal is

adjacent to a lateral. Recall that in Korean, /l/ surfaces as [n] when it is preceded by /m/ or

/N/, e.g. /sam + lyu/ [samnyu] ‘third rate’, /y´N + lak/ [y´Nnak] ‘downfall’. It is

assumed in this study that the change is motivated due to the prohibition against a lateral in

prevocalic position, which is formulated as the constraint *PreV-l. In Korean, /l/ never

occurs in prevocalic position, suggesting that *PreV-l is undominated. Given that Max

[lateral] is ranked above Max [nasal], when assimilation is used as a repair strategy, it is

expected that, as in the /ln/ sequence, a nasal will assimilate to a lateral, changing /ml/ and

/Nl/ into [ll]. However, in the /ml/ and /Nl/ sequences a lateral assimilates to a nasal. Thus,

/ml/ and /Nl/ are realized as [mn] and [Nn], respectively. I argue that /ml/ and /Nl/ are not

modified to [ll] since the change would involve both place and manner of articulation. In

the case of /ln/, on the other hand, both /n/ and /l/ are coronals, and thus changing /ln/

place. Based on this argument,

llowing Davis & Shin (1999), I employ the constraint Ident-place for the analysis of /ml/

Ident-place, correspondenet segments

ould be identical with respect to place features. As we can see from the tableau in (19),

(18) a. *Hom C/liq, Max [lateral] >> Max

b. *Hom >>

into a geminate lateral does not involve a change in

fo

and /Nl/ to [mn] and [Nn], respectively. According to

sh

Max [lateral] is ranked below *PreV-l and Ident-place:

-120-

/ -l - [later x [

(19) /kam + li/ [kamni] ‘supervision’ /kam + li *PreV Ident place Max al] Ma nasal] a. kamli *! b. kalli *! * c. kamni *

If Max [lateral] were to dominate *PreV-l or Ident-pl, candidate (a) or (b) would incorrectly

be selected as optimal. At this point, no evidence is found that there is any crucial ranking

relation between *Hom C/liq and a set of the constraints *PreV-l, Ident-place and Max

[lateral]. Thus, we can posit the following constraint ranking:

te ai -l

anking: PreV-l, Id t-place > Max [la ral]

er strate to avoid aving a la l in prevocalic position would be to modify

o *[kam|i]. To rule this out, we need to note that a flap occurs only

R

& Hahn (2000) posit the undominated constraint VflapV.

(21) VflapV: A flap can only appear intervocalically.

By VflapV, the candidate *[kam|i] can be correctly ruled out. In the tableau below, there is

eral] and *Het C/liq.

(20) a. Undomina d constr nt: *PreV

b. Crucial r * en > te

Anoth gy h tera

/kamli/ int

intervocalically in Korean, as in /i-lon/ [i- on] ‘theory’. Based on this observation, Kang

no crucial ranking between Max [lat

-121-

/kam +li/ VflapV *PreV-l Ident- Max *Hom Max *Het (22) /kam + li/ [kamni] ‘supervision’

place [lateral] C/liq [nasal] C/liq a. kamli *! *

b. kamni * c. kam|i *! * *

However, with the constraints posited thus far, we cannot account for why /palmok/

mok]. That , both *[palbok] d [palmok] are incorrectly selected as optimal

om the input /palmok/, as illustrated in the tableau in (23):

/pal + mok/ VflapV *PreV-l Ident- Max *Hom Max *Het

surfaces as [pal is an

fr

(23) /pal + mok/ [palmok] ‘ankles’

place [lateral] C/liq [nasal] C/liq a. palmok *

* b. palbok * c. panmok *! d. pallok *! * e. palnok *! *

In the tableau above, both candidates (a) and (b) are chosen as optimal since they violate only

the lowly ranked constraint *Het C/liq, while other candidates are ruled out by violating

highly ranked constraints. I would suggest that modifying /palmok/ to *[palbok] is

unacceptable since an underlying sonorant consonant in onset position is realized as an

obstruent. Thus, the following constraint can be employed to rule out candidate (b).

(24) Pres-ons [son]: Preserve identical values of the feature [sonorant] for consonants in onset position.

-122-

s never violated in Korean and thus is undominated. As

n the tablea below, *[ ok] cannot e selected as timal due to violation

[son].

/ Pres-ons [son] *Het C/liq

The constraint Pres-ons [son] i

illustrated i u palb b op its

of Pres-ons

(25) /pal + mok

a. palmok * b. palbok *! *

The analyses of lateral/non-coronal nasal sequences discussed above illustrate that

the asymmetrical pattern found in the sequences is due to the prohibition against a lateral in

prevocalic position. That is, the lateral in /VmlV/ and /VNlV/ is realized as [n] in order

not to have a lateral in prevocalic position, while the lateral in /VlmV/ surfaces unmodified

since it does not occur in prevocalic position. Recall that the /VlNV/ sequence is not

attested in Korean.

Recall also that in Korean /ln/ is realized as [ll], as we have seen in (17). But the

reverse order of the sounds, i.e. /nl/ also gives [ll], as in /non + li/ [nolli] ‘logic’. Unlike

in previous accounts where the change is motivated by the Syllable Contact Law, I assume,

as discussed above, that the change of /nl/ is motivated due to the prohibition against a

lateral in prevocalic position and low perceptibility of the sequence triggered by phonetic

is modified to [ll] rather

change a lateral into a nasal than to

change a lateral into a nasal to avoid having a lateral in onset position or two similar

consonants in a sequence. Through the interaction of the constraints established thus far,

we can analyze the modification of /nl/ into [ll], as shown in tableau (26):

similarity of the two segments involved. The observation that /nl/

than to [nn] illustrates that for Korean it is worse to

-123-

the tableau above, candidate (a), which is most faithful to the input, is ruled out by

e optimal output. We can summarize the

b. Crucial rankings: *Hom C/liq, Max [lateral] >> Max [nasal]

Ident-place >> Max [lateral]

/non + li/ VflapV *PreV-l Max [lateral] *Hom C/liq Max [nasal] (26) /non + li/ [nolli] ‘logic’

a. nonli *! * b. nonni *! c. nonti *! d. notli *! * * e. not|i *! * * * f. nol l i *! * * | |

[lat][lat] g. nol l i

[lat]

*

In

violating the highly ranked constraint *PreV-l. Candidates (b) and (c) are excluded due to

the violation of Max [lateral]. Candidate (d) cannot be chosen as optimal since it violates the

highly ranked constraint *PreV-l. Candidate (e) is excluded by violating the constraint

VflapV. In the case of candidate (f) with a non-geminate [ll], a lateral occurs in prevocalic

position and thus it is ruled out by violating the constraint *PreV-l. Candidate (g) with a

geminate lateral has no fatal violation of the highly ranked constraints although it violates

the constraint Max [nasal]. Thus, it is selected as th

constraints established for our analyses of nasal/lateral sequences in Korean as follows:

(27) a. Undominated constraints: VflapV, *PreV-l, Pres-ons [son]

*Hom C/liq >> *Het C/liq

We now turn to the pattern found when a nasal is adjacent to a stop consonant. In

Korean, stop plus nasal sequences are changed into nasal plus nasal sequences, e.g. /kuk +

mul/ [kuNmul] ‘soup-broth’. It is assumed in this study that the change is motivated by

-124-

in (28),

uNmul] can be correctly selected as the optimal output from the input /kuk + mul/:

l/ -b /kuk + mul/ *stop /___ son Pres-ons [son]

the low perceptibility of an unreleased stop before a sonorant consonant. As illustrated in

the tableau below, by adding the constraint *stopunrel/___son to the constraints

[k

(28) /kuk + mu [kuNmul] ‘soup roth’ unrel

a. [kukmul] *! b. [kukpul] *! c. [kuNmul]

The tableau above illustrates that the change of /kukmul/ is motivated to avoid having an

unreleased stop before a sonorant consonant. Among the constraints established thus far,

those th

a sonorant consonant and thus

stopunrel / ___ son is an undominated constraint.

that have been established thus far, we can also account for the

pattern fo in non-coronal s lateral sequences. Recall that in Korean,

surface as [N ] and [mn], respectively, e.g. /p´p + li/ [p´mni] ‘pinciple of

[ky´Nni] ‘separation’, while /lk/ and /lp/ surface unmodified, e.g.

like in previous accounts where the pattern is explained by

employing the notion of syllable contact. For example, in /p´pli/, /l/ occurs in prevocalic

at do not play a crucial role in the analysis are left out in the tableau above. It also

shows that it is worse to change a sonorant in onset position into a stop than to change a

stop in coda position into a sonorant consistent with the analysis of [palmok] in tableau

(25). In Korean, an unreleased stop never occurs before

*

With the constraints

asymmetrical und top/

/kl/ and /pl/ n

law’, /ky´k + li/

/kalpi/ [kalpi] ‘ribs’, /kal + ku/ [kalku] ‘desire’. This asymmetrical pattern is due to

the prohibition against a lateral in prevocalic position and the avoidance of an unreleased

stop before a sonorant, un

-125-

9) /p´p + li/ [p´mni] ‘principle of law’

VflapV Ident-lace

Max ]

*Hom q

position and thus cannot surface unmodified. The candidate *[p´pni] cannot be selected as

optimal due to the constraint *stopunrel/___ son, as illustrated in the tableau below:

(2

/ p´p + li / *stopunrel/___son *PreV-l p [lateral C/li

a. [p´pli] *! * b. [p´pni] *! * c. [p´lli] * d. [p´p|i] *! * * e. [p´mni] *

At this point, no evidence is found that there is any crucial ranking between *Hom C/liq and

the other constraints in tableau (29).

Unlike in a non-coronal stop plus lateral sequence, no modification occurs in a lateral

plus non-coronal stop sequence since /l/ does not occur in prevocalic position and the stop

is released into the following vowel. Thus, for example, [kalpi] is chosen as the optimal

output of the input /kalpi/, as shown in the tableau in (30):

(30) /kalpi/ [kalpi] ‘ribs’ /kalpi/ *stopunrel/___son *PreV-l Ident-place Max [lateral]

a. [kalpi] b. [kanpi] *! c. [kalli] *! d. [kampi] *! *

As can be seen from th tableau above, candidate (a), which is m aithfu o the input, does

y constr int while other candidates do. Thus, the input /kalpi/ surfaces as

alpi] without modification.

e ost f l t

not violate an a

[k

-126-

ers ‘t’ and ‘l’’ /tikÈt + liÈl/ *PreV-l VflapV Ident-

place Max [lateral]

*Hom C/liq

Max [nasal]

With the constraint rankings established, we can also provide an account of the

modification of /tl/ into [ll], not relying on the Syllable Contact Law, as shown in tableau

(31):

(31) /tikÈt + liÈl/ [tikÈlliÈl] ‘the lett

a. [tikÈtliÈl] *! * b. [tikÈtniÈl] *! c. [tik tti l] *! È È d. [tikÈt|iÈl] *! * *

d. [tikÈlliÈl]

Candidate (a), although it is most faithful to the input, cannot be selected as optimal since it

ons [son] place [lateral] C/liq ax

[nasal]

contains a lateral in onset position. Modifying /l/ in the /tl/ sequences into [n] or [t], as in

candidates (b) and (c), solves the problems that the /tl/ sequence has. However, the

modification runs into another problem. It incurs the violation of Max [lateral]. Candidate

(d) cannot be chosen as optimal since it violates the highly ranked constraint VflapV. Thus,

candidate (e) without any violation of the constraints is selected as the optimal output.

In the case of the /lt/ sequence in Korean, unlike the /tl/ sequence, it is not

modified since it does not violate any highly constraint motivating phonological change:

(32) /kalte/ [kalte] ‘reed’ /kalte/ Pres- *PreV-l VflapV Ident- Max *Hom M

a. [kalte] * b. [kante] *! c. [katte] *! d. [kalle] *!

-127-

hosen as optimal.

4) a. Undominated constraints: VflapV, *PreV-l, Pres-ons [son], *stopunrel/___son b. Ident-place >> Max [lateral] >> *Hom C/liq >> Max [nasal], *Het C/liq

n or

bstruent sequences, e.g. /kun + mantu/ [kummantu] ‘fried dumpling’, /sin-pal/

g it is

ssumed in this study that the change is motivated by the weak perceptibility of the coda

logical survey,

eterorganic sequences are more vulnerable to a phonological process than homorganic

sequenc n t hat th traints,

b. *Het NC: A homorganic sequence of a nasal and a consonant is prohibited.

c. *Het NC >> *Hom NC

In selecting candidate (a) as optimal, the ranking relation between Max [lateral] and *Hom

C/liq is crucial. As we can see from the tableau above, Max [lateral] is necessarily ranked

above *Hom C/liq. Otherwise, candidate (a) would not be c

(33) Max [lateral] >> *Hom C/liq

We can summarize the constraints established for our analyses of C/lateral sequences in

Korean as follows:

(3

Let us now consider the analysis of the patterning of nasal plus nasal or obstruent

sequences. In Korean, place assimilation is attested in heterorga ic nasal plus nasal

o

[simbal] ‘shoes(-foot)’, /kam + ki/ [kaN i] ‘cold’, etc. As discussed in chapter 3,

a

nasal and articulatory simplification. Recall that, according to the typo

h

es in asal plus nasal or obstruen sequences, and t e ranked cons

repeated in (35), were posited to account for the general cross-linguistic pattern:

(35) a. *Hom NC: A heterorganic sequence of a nasal and a consonant is prohibited.

-128-

6) *Het NC >> *Hom NC, Ident-place

lus nasal or obstruent sequence is modified by place

assimilation in Korean, it follows the place assimilation patterns below, which are also

attested in a heterorganic obstruent plus obstruent sequence.

(37) Korean place assimilation (Iverson & Lee 1995, Jun 1995)

/nm/ [mm] /kun + mantu/ [kummantu] ‘fried dumpling’

/tÓp/ [pp] /patÓ + pap/ [papp’ap] ‘rice with red beans’

b. A coronal assimilates to a following velar.

/tk/ [kk] /mit + ko/ [mikko] ‘believe and’

c. A labial assimilates to a following velar.

[´kk’o] ‘carry and’

d.

Nmul] ‘river water’ [toNdÈN] ‘equality’

/Np/ [Nb] /c´N + pi/ [c´Nbi] ‘maintenance’ /Nk/ [Ng] /caN + kun/ [caNgun] ‘general’

/kt/ [kt] /nok + ta/ [nokt’a] ‘melt (Declarative)’ /kp/ [kp] /puk + paN/ [pukp’aN] ‘the northward’

The observation that a heterorganic nasal plus nasal or obstruent sequence is modified by

place assimilation in Korean implies that the constraint Ident-place is dominated by *Het NC.

Thus, the following constraint hierarchy can be established:

(3

When a heterorganic nasal p

a. A coronal assimilates to a following labial.

/np/ [mp] /sin + pal/ [simbal] ‘shoes(-foot)’

/nk/ [Ng] /han + kaN/ [haNgaN] ‘Han river’

/mk/ [Ng] /kam + ki/ [kaNgi] ‘cold’ /pk/ [kk] /´p + ko/

A velar never assimilates to a following segment in place. /Nn/ [Nn] /caN + nal/ [caNnal] ‘a market day’ /Nm/ [Nm] /kaN + mul/ [ka

/Nt/ [Nd] /toN + tÈN/

-129-

lation, while a coronal or a labial does, we can see

that in Korean it is more important to keep a velar than to keep a coronal or a labial. Thus,

g rank constraint can be posited based on Jun (1995)’s analysis of Korean

lace assimilation:

nstraint triggering place assimilation is formulated as Weakening

ands articulatory effort conservation. In the present study, it is formulated as

ng er h rorga on-

nces a more li targets f phonolog al pr s than ho organic

quid s ences. I a assume t t place as ilation o rs in a he rorganic

lus obst ent sequen e to enhance the perceptibility of the coda obstruent and

to a following segment in

Given that a velar never undergoes assimi

the followin ed s

p

(38) a. Pres-Dor: Preserve a dorsal. b. Pres-Lab: Preserve a labial. c. Pres-Cor: Preserve a coronal. d. Pres-Dor >> *Het NC >>Pres-Lab, Pres-Cor, Ident-place, *Hom NC

In Jun’s analysis, the co

which dem

*Het NC to account for the cross-li uistic gen al pattern t at hete

oces

nic nasal/n

liquid seque re kely o ic se m

nasal/non-li equ l os ha sim ccu te

obstruent p ru c

to simplify articulation, just as in a heterorganic nasal plus consonant sequence. The

triggering constraint can be formulated as *Het OO which prohibits a heterorganic

obstruent plus obstruent sequence.

Let us first examine how assimilation of a coronal nasal

place is accounted for with the constraints in (38d).

-130-

Pr H Pres-Lab Pres-Cor Ip

(39) /han + kaN/ [haNkaN] ‘Han river’ /han+kaN/ es-Dor * et NC dent-

lace *Hom NC

a. hankaN *! b. hantaN *! * *

c. haNkaN * * *

As shown in the tableau above, the /nk/ sequence is unacceptable since it violates *Het

NC. If /nk/ is modified into [nt] in order not to violate *Het NC, the change triggers a

fatal violation of Pres-Dor. Thus, [Nk] is selected as the optimal output of /nk/.

Similarly, place assimilation of a labial nasal to a following velar can be analyzed as

follows:

(40) /kamki/ [kaNki] ‘cold’ /kamki/ Pres-Dor *Het NC Pres-Lab Pres-Cor Ident-

place *Hom NC

a. kamki *! b. kanki *! * * c. kampi *! *

d. kaNki * * *

In the tableau above, candidates (a) and (b) cannot be selected as optimal since they have a

critical violation of *Het NC. Although candidate (c) does not violate *Het NC, it is ruled

ut by violating Pres-Dor as the result of modifying a dorsal into a labial. Thus, candidate

ected as optimal.

Place assimilation is not attested in Korean l ar s fol

ow er, with e constrain establish thus ar, it ncorre ly pred ed

at place assimilation will occur in the sequence, as illustrated in the tableau in (41):

o

(d), which does not have a fatal violation, is sel

when a labia or vel nasal i lowed

by a coronal. H ev th ts ed f is i ct ict

th

-131-

/tÇamtÈlta/ Pres-Dor *Het NC Pres-Lab Pres-Cor Ident- *Hom NC(41) /tÇam + tÈlta/ [tÇamdÈlda] ‘fall asleep’

place a. tÇamtÈlta *! * b.tÇampÈlta * * * * c.tÇant lta * * * È

Candidate (a) is ruled out by *Het NC. Thus, candidates (b) and (c) are incorrectly selected

as optimal. To remedy this problem, we need to note that an onset is never modified by

place assimilation in Korean, and place assimilation does not occur before a coronal, e.g.

/nop+ta/ [nopt’a] ‘high (Declarative), /nok+ta/ [nokt’a] ‘melt (Declarative)’. To

account for these patterns, Jun (1995) posits the following undominated constraints:

r conson n po n. Pres (pl(__cor)): Preserve place cues for consonants before a coronal.

action of the two undominated constraints with the constraints in

now account for why place assimilation is not attested in Korean when a labial or velar

nasal is followed by a coronal: assimilation would violate the highly ranked constraints Pres

(pl(onset)) or Pres (pl(__cor)), as shown in the tableau in (43):

(43) /tÇam + tÈlta/ [tÇamdÈlda] ‘fall asleep’

Do or)) ss- Pres- Ident- *Hom

(42) Pres (pl(onset)): Preserve place cues fo ants i onset sitio

By the inter (38d), we can

/tÇamtÈlta/ Pres- Pres Pres- *Het Prer (pl(__c (pl(on et)) NC Lab Cor place NC

a. tÇamtÈlta * b. tÇampÈlta *! * * * c. tÇantÈlta *! * * *

-132-

us, candidate (a) is selected as optimal.

This analysis also accounts for the observation that a velar nasal in coda position

in (44),

assimilating a velar to a following consonant in place is unacceptable since the change

incurs a violation of Pres-Dor. In addition, assimilating a consonant in onset position to a

velar triggers a violation of Pres(pl(onset)). Thus, when a velar nasal is followed by a labial

or a coronal, place assimilation does not apply to the sequence.

(44) /tÇ´Npi/ [tÇ´Nbi] ‘maintenance’

Dor (pl(__cor)) (pl(onset)) NC s-

Lab Pres-Cor

Ident-place

*Hom NC

In the tableau above, candidate (a) violates *Het NC. However, when place assimilation

applies to avoid the violation of *Het NC, as in candidates (b) and (c), it incurs a fatal

violation of Pre-Dor or Pres(pl(__cor)). Th

never undergoes place assimilation in Korean. As illustrated in the tableau

/tÇ´Npi/ Pres- Pres Pres- *Het Pre

a. tÇ´Npi * b. tÇ´mpi *! * * c. tÇ´Nki *! * *

P ss ion is also attested in Korean when stop plus nasal sequlace a imilat ences are

m] which

s follo s:

5) /mat + my´nÈli/ [mammy´nÈ|i] ‘first daughter-in-law’

ace

modified to nasal plus nasal sequences. Thus, the /tm/ sequence is realized as [m

can be analyzed a w

(4 /mat-my´nÈli/ *stop / Pres- Pres- *Het Pres- Pres- Ident-unrel

__son Dor (pl(onset)) NC Lab Cor pl a. matmy´nÈli *! * b. manmy´nÈli *!

c. mammy´nÈli * * d. manny´nÈli *! *

-133-

>>Pres-Lab, Pres-Cor, Ident-place, *Hom NC

5.3 Summary

I have argued that the alternations of son/C clusters in Korean are the result of

segment contact and a segment specific positional restriction, rather than by syllable

contact. The motivation for phonological change in each type of Korean son/C clusters

proposed in the present study can be summarized as follows:

(47) Motivation for change in a Korean son/C cluster Cluster Motivation a. /nl/: restriction disallowing /l/ in prevocalic position avoidance of perceptually similar segments in a sequence b. /ln/: avoidance of perceptually similar segments in a sequence c. /ml/, /Nl/ /tl/, /pl/, /kl/: restriction disallowing /l/ in prevocalic position d. stop plus nasal: weak perceptibility of the stop e. heterorganic nasal plus nasal or obstruent sequence: weak perceptibility of the nasal, articulatory simplification

There is no crucial ranking relation between *stopunrel/__son and the set of the constraints

Pres-Dor, Pres-(pl(onset)). The constraints employed for the analysis of the patterning of

nasal/non-liquid sequences are as follows:

(46) *stopunrel/__son, Pres-Dor, Pres (pl(onset)), Pres (pl(__cor)) >> *Het NC

-134-

By motivating the modification of a nasal or obstruent plus lateral sequences in terms of a

restriction prohibiting a prevocalic lateral, a unified account can be provided with respect to

the change of /l/ in post-consonantal, word-initial and intervocalic position in Korean.

The segment contact account propos

can naturally extend to the account of phonological modifications of obstruent plus

obstruent sequences. That is, as in son/

obstruent plus obstruent sequences can be understood as being motivated by segment

contact

ed to explain the patterning of son/C clusters

C clusters, phonological processes involving

. For example, just as in heterorganic nasal plus nasal or obstruent sequences, place

assimilation occurs in heterorganic obstruent plus obstruent sequences in Korean. This can

be assumed to be motivated to enhance the perceptibility of the preconsonantal obstruent,

which is unreleased in Korean, and to simplify articulation.

-135-

erceptual salience of a son/C cluster. However, within the

liquid and liquid

C H A P T E R 6

CONCLUSION

It is argued in the present study that phonological processes involving son/C

clusters are the result of segment contact rather than syllable contact. Central to the

segment contact account is the view that low perceptual salience is a key factor motivating

phonological processes in son/C clusters. Low perceptual salience of a son/C cluster is

assumed to be triggered by (a) phonetic similarity of two segments in the cluster or (b) the

occurrence of a segment in contexts with less robust phonetic cues. In addition, it is argued

that a restriction disallowing a lateral in prevocalic position can motivate phonological

modifications in son/C clusters.

The segment contact account is superior to the syllable contact account for a

number of reasons. First, the segment contact account can explain why tautosyllabic son/C

clusters are targets of phonological processes. That is, it can account for phonological

processes involving both heterosyllabic and tautosyllabic son/C clusters in terms of the

same mechanism, i.e. low p

syllable contact account, some additional formal mechanism is needed. This is because a

syllable boundary is not crucial to the segment contact account while it is to the syllable

contact account. Second, the segment contact account can provide a unified account of the

same types of phonological modifications found, for example, in nasal plus

- 136 -

plus nasal sequences in languages such as such as Boraana Oromo, Korean, Leti and Toba

Batak. On the other hand, it has to be assumed within the syllable contact account that the

modification of nasal plus liquid sequences and that of liquid plus nasal sequences result

from d both

types of sequences in the languages. This is because liquid plus nasal sequences do not

predicted by the segment contact account,

ic C/liquid sequences imply those in

it

is incorrectly predicted that both homorganic and heterorganic sequences will be

he same in either violating or not violating the Syllable Contact Law. In

nguage, sonorant

nt

y the syllable contact account:

e more likely targets of phonological processes

an sonorant plus sonorant sequences. Finally, it is correctly predicted within the segment

ontact account that unreleased stop plus sonorant sequences are more likely to be

modified than released stop plus sonorant sequences due to the weak perceptibility of the

unreleased stop consonant.

This study is based on an extensive database of sound alternations found in son/C

clusters and thus provides a strong empirical base for cross-linguistic comparison and for the

study of phonotactic phenomena concerning son/C clusters. This study also lends robust

mpirical support to the segment contact approach to phonological change.

ifferent factors although the same types of phonological processes apply to

violate the Syllable Contact Law while nasal plus liquid sequences do. Third, as discussed in

chapter 3, the segment contact account correctly predicts the cross-linguistic patterning of

son/C clusters. According to the survey, as

phonological processes involving heterorgan

homorganic C/liquid sequences in a given language. Within the syllable contact account,

considered t

addition, the segment contact account correctly predicts that, in a given la

plus sonorant sequences will be more likely to be modified than obstruent plus sonora

sequences. The opposite pattern is incorrectly predicted b

obstruent plus sonorant sequences will b

th

c

e

- 137 -

A P P E N D I X A : A T Y P O L O G I C A L S U R V E Y O F L T E R A T I O N S I N S O N C C L U T E R S

pendix i es results from a typolog y of sound lternations in

n/C clusters in 31 different languages. The entry for each language is laid out as indicated

elow. Explanatory notes are in italics.

N meaccording to the Summer Institute of Linguistics Ethnologue (http://www.ethnologue.com)

Phonemic consonant inventory

P ter of n/ clu ters:Symbols used: - Shaded cells: for obstruent plus obstruent sequences, which are not discussed in this study. - blank cells: for a son/C cluster whose ph nologi l patter ng is n t ava ble

ples below since more than one pattern is attested. - u est Notes: any interesting facts or observations related to son/C clusters References: pertinent references on the language

A N / S

This ap nclud ical surve a

so

b

a (General, specific language families)

at ns so C s

o ca ni o ila- The downward arrow directs the reader to the exam ×: natt ed

138

Bilabial lveo r Pa tal

Choctaw (Muskogean, Western)

Phonemic consonant inventory: Labiodental Interdental A la latal Velar Glot

Stop p t k b Fricative f T s S h Affricate tS Nasa m l n Lateral l Glide w y - In some dialects, /Ò/ occurs instead f o /T/.

tS m Patterns of son/C clusters:

p b t k f T s S h n l w y p bl b bb t ll k km kn kl f ff T s S Sm Sn Sl h hl tS m ↓ ↓ ↓ ↓ ↓ mm ↓ ↓ ↓ ↓ ↓ mm n ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ nn l ↓ l lb hf~ff ss hS~SS hh ltS~

tStS lm nn ↓ ↓ ↓ ll lw yy hs~

w ww ~l y

Examples: /pl/ [bl] /tap-l-i/ [tabli] ‘to cut in two (passive)’

/atob-l-i/ [atobbi] ‘to pay’ assive)’

fl/ [ff] /lofli/ [loffi] ‘to pare’ [Sl] /baSli/ [baSli] ‘to gash’

li] ‘to finish’

i!nlih/ [akai ‡:lih] ‘indeed’ ll/ [ll] /hallo‡s/ [hallo‡s] ‘leech’

/kowli/ [ko:li] gloss not given

/bl/ [bb] /tl/ [ll] /akat-l-i/ [akalli] ‘to patch (p/kl/ [kl] /to!klo/ [to!kolo] ‘two’ //Sl/ /hl/ [hl] /tahli/ [taha

/ml/ [mm] /famli/ [fammi] ‘to spank’ /VnlV/ [v ‡:lV] /ak/ /Vwl/ [V:l] (in active forms of baSli-baSa verb types)

139

‘to open’ lp/ [Tp] or [Òp] depending on dialects when /l/ is a passive infix

[aTpisa] or [aÒpisa] ‘to set a date’ lp/ [pp] w is a p

-panSi/ [ppanSi] ‘hari (of the head)’ [lb] ittee!lbi/ [ittee!lbi] ‘lip’ [Tt] or [Òt] depending on dialects when /l/ is a passive infix

altaha/ [a Òtaha] ‘to be prepared’ [ht] or [tt] when /l/ is a prefix

tahtSi/ [htahtSi] or [ttahtSi] ‘shoulder’ k] or [Òk] depending on dialects when /l/ is a passive infix

l-ka a/ [a ama] or a] ‘to b stopperedkk] when /l/ is a prefix

/l-ki!Si/ [kki !Si] ‘chest’ [hfo!ni] or [ffo !ni] ‘bone’

/a- oha ffo /ls/ [hs] or [ss] /l-sana!htS hsan Si] [s ing/ [hS or [S S kba/ Sakba ] m’ /lh/ [hh] /l-hatip/ p’ / / [ht tS]

/l-t S SonkaS] or [tSonkaS] eart/lm/ [lm] /almo/ ‘ g re/ln/ [nn] /ta-l-na/ ‘ w ed (/l/ passive infix)

/l-noti/ ‘tooth’

/ly/ [y yi ‘foot’ /lw/ [l /a a/ ] ‘ y’ / / /k ! ma ‘ ew ’ /nn/ [n /o / !nna] ‘ w

[k /pakna/ ‘[km] /hokmi/ [hokomi] ‘to burn up’

rain

el en fol ortakes sho el

wh edes [k

s ents o ty

/wl/ [ww] /tiwli/ [tiwwi] / /alpisa/ / hen /l/ refix

/l /lb/ /lt/

/

/lt/

/ Ttaha] or [a

/l-/lk/ [T /lk/ [

/a- m Tk [aÒkam e ’

/lf/ [hf] or [ff] /l-fo !ni/ l-f / [a ha] ‘to be wound around’

i/ [ a!ht or sana!htSi] ‘w ’ lS/ ] S] /l- a [h ‘ar

[hhatip] ‘hiltS S] or [tS

Sonka / [ht ‘h ’ [almo] to be athe d’ [tanna] to be eav ’

= [nnoti]

(/l/ = prefix) /l-yi/

y]

w [y ]

[mm]

] lwaSa i !mmah/

[alwaSa to frsomm n [kani

[om h] m here

n’

n] n]

!nna to datop’

/kn/ [pakana]/km/ /Sn/ [Sn] /piSno/ [piSno] ‘we’ /Sm/ [Sm]

V/ V] i [pi/VnC [v ‡:C /p !nsalik/ ‡!:salikat] ‘that I see’ [o:ba] to/VmCV/ [v ‡:CV] /omba/ ‡ ‘ ’

Notes: - [k]: r eased wh (The release the form of a very rt vowel of the same quality as the vow

lowed by [b], [m], [n] [l].

ich prec

].)

References: Nickla , Thurston Dale. 1974. The Elem f Choctaw. Ph.D. dissertation. Universi

of Michigan.

140

Pa

Chukchee (Chukotko-Kamchatkan, Chukot) Phonemic consonant inventory: Labial Alveolar latal Velar Uvular Glottal voiceless stop p t k q / voiceless tS affricate

voiceless Ò9 fricative

nasal m n N voiced stop r voiced (w) y Ø (w) continuant Patterns of son/C clusters:

p t k q / tS Ø r m n N Ò9 w y p mm mn mN t nm nn nN k Øm Øn NN Øy q / /n m / tS tSm tSn tSØ tSw × Ò9 Ò9m × Ò9Ø Ò9w Ò9y Ø wm Øn kw Øy r rm nn rN rw ty m mk Ò9 mr m m mØ Nn p nt Ò9 nr nm nn nwn n ny ntSN p q Ò9 Ø m Nk N ntS nr m Øn ØN mwn ny w wtS wr wn wt Ò9 w wyy yp Øt y/ r ym ywyk Ø Øy yq ØtS Ò9 yØ Ø Øn yN

Examples: /pm/ [mm] //´ q ma/ [/´ qamma] gloss not g

Ò9qep- k ’ Ò9 ap- Ò9 iven

f hides’ ’

’

cf. //eÒ9qep-´k/ [/e ´ ] ‘hammert/ t] /pn/ [mn] /r´pn-´ [r´mn´ ‘flesh side o

cf. /r´p´n/ [r´p´n] ‘flesh side of hidepN mN] p t- ´t/ / [ /Øa- N´ Ò9en/ [ØamN Ò9en] ‘having news’

´Ò9] cf. /p´N´Ò9/ [p´N ‘news’-Ò9en/ Ò9e/tm/ [nm] /Øa-tm´ [Øanm´ n] ‘he killed’

m kcf. /t´ -´k/ [t´m´ ] ‘kill’ /tn/ [nn] ´ nn´t] ‘teeth’ /r´tn- t/ [r´

t´ ´n]cf. /r´ n/ [r´t ‘tooth’ /tN/ [nN] /yanot-Nenqay/ [yanon-Nenqay] ‘a boy in front

/ cf. yanot/ [yanot] ‘in front’ /km/ [Øm] /m´k-mi:9Øer-e/ [m´Ømi:9Øere] ‘many guns’

141

/kn/ [Øn]

part’

cf. /aq´n/ [aq´n] ‘fishing pole’ ] /p´Ò9m-at-´k/ [p´Ò9mat´k] ‘to get dark’

/Øn/ ‘was superfluous’ ’

ry/ [ty] /kor-yo/ [kotyo] ‘purchase’ rn/ [nn] /kur-nin/ [kunnin] ‘he bought it’

cf. /kur-´k/ [kur´k] ‘to buy’ mk/ [mk] /n´-mk´-kin/ [n´mk´kin] ‘often’ m:9/ [m:9] /mim:9-´-qatSa/ [mim:9 qatSa] ‘place near the water’ mØ/ [mØ] /Ø´tam-Ø´rØ-´n/ [Ø´tam-Ø´rØ-´n]‘accuracy’ mr/ [mr] /mren-ti/ [mrenti] ‘mosquito guards’ mN/ [mN] /:9 mN´:9te:9-´k/ [:9 mN´:9te:9 k] ‘tell stories’ np/ [np] /Øa-n-pera-w-:9en/ [Øanperaw:9en] ‘decorated’ nt/ [nt] /Øe-nt-´-:9in/ [Øent´:9in] ‘he has cut off’ nn/ [nn] /Ø´nnik-qey/ [Ø´nnik-qey] ‘small animal’ nw/ [nw] /m´k-at´nw´:9-´n/ [m´kat´nw´:9 n] ‘having many Np/ [mp] /taN-pera-k/ [tamperak] ‘to look good’ Nk/ [Nk] /taN-kaw-kaw/ [taNkawkaw] ‘good sugar’ Nq/ [Nq] /n´-teN-qin/ [n´teNqin] ‘good’ NtS/ [ntS] /taN-tSottSot/ [tantSottSot] ‘good pillow’ N:9/ [n:9] /tSeN:9-´t/ [tSen:9-´t] ‘drawers’

cf. /tSeN´:9/ [tSeN´:9] ‘drawer’ Nr/ [nr] /wiNr-´t/ [winr-´t] ‘hoes’

cf. /wiN´r/ [wiN´r] ‘hoe’ Nm/ [Øm] /m´t:9´N-more/ [m´t:9´Ømore] ‘we five’

cf. /m´t:9´N-´n/ [m´t:9´N-´n] ‘five’ Nn/ [Øn] /enawr´N-nen/ [enawr´Ønen] ‘he gave it’

cf. /enawr´N-´k/ [enawr´N´k] ‘to give as a gift’ NN/ [ØN] /pet/iN-Ninqey/ [pet/iØNinqey] ‘boy with a cold’

cf. /pet/iN/ [pet/iN] ‘cold’ Nw/ [mw] /taN-waØ´rØ-´n/ [tamwaØ´rØ´n] ‘good life’ Ny/ [ny] /teN-y´:9qet-´k/ [teny´:9qet´k] ‘to sleep well’

cf. /teN-´:9/-´n/ [teN´:9/´n] ‘good’ y:9/ [Ø:9] /qey-:9 nanØet/ [qeØ:9 nanØet] ‘truth’

cf. /qey-we/ [qeywe] ‘correct’ yr/ [Ør] /w/ay-ran/ [w/aØran] ‘grass hut’

/m´k-nennet/ [m´Ønennet] ‘many otter’ /kN/ [NN] /te-pÒ9ek-N-´k/ [tepÒ9eNN´k] ‘make footwear’ cf. /pÒ9ek-´t/ [pÒ9ek´t] ‘footwear’ /ky/ [Øy] /m´k-yan/ [m´Øyan] ‘greater/qm/ [/m] /Neraq-m´t:9 N´n/ [Nera/m´t:9 N´n] ‘seven (2+5)’ cf. /Nireq/ [Nireq] ‘two’ /qn/ [/n] /aqn-´t/ [a/n´t] ‘fishing poles’ /Ò9m/ [Ò9m/Øm/ [wm] /etl9eØ-m´tq´m´t/ [etl9eØm´tq´m´t] ‘honey’ cf. /n-etl9eØ-qin/ [netl 9eØqin] ‘sweet’

[Øn] /Øe-Ønu-Ò9in/ [ØeØnuÒ9in] /rw/ [rw] /Øa-rw-´-len/ [Øarw´:9en] ‘he has split// //////////////

/

/ / / //

/ /

142

/ay-ma/ [Øaw/ayma] gloss not given ym/ [ym] /eymew-´k/ [eymew´k] ‘he approached’

k/ [tSaØna:9k´k] ‘to make tea’ cf.

/yN/ [yN] /tSiØ-meyNet-´k/ [tSiØmeyNet´ grow somewhat’ [Ø a ] seagull’

cf. ey´qey/ qey´qey] ‘nestling’

es:cz, 6 l c an s. SLS 16, 1:

Scot 8 C o M rphology. P D.

Davi . 1987. Diss de tion in Chu chi. ESCOL 87, 235-246

cf. /Øa- w//yn/ [Øn] /tSay-na:9k-´ /tSay/ [tSay] ‘tea’

k] ‘to /yy/ y] /q y-ya/yaq/ [qaØya/yaq ‘young /q [

Referencwi

Kensto Michael. 198

ussell. 19. The phono ogy of Chuk

ukchee Phhee conson

logy and t 79-96.

Krause, t R 0. Topics in h no o h.dissertation. University of Illinois at Urbana-Champaign.

Odden, d imilation as le k ’ .

143

Eg N -C g r

Phonemic consonant inventory: ab l Pa tal Velar Labio-

velar Glottal

D ( iger on o, Gu )

L ia Alveolar la

Stop W b W W dW k kW g gW k g p p b t t d p b Affr tS W

dZ icat tS dZ

WFrica ive h hW t f v s sWNasal m mW n ≠ N NW Lateral l lW Trill or flap

r

Glide w y Patterns of son/C clusters: p pW b t tW dW k kW g gW kp gb tS bW d tSW dZ dZW m mp mb m mk mkp mgb

bW

n n nt nd n S

n tSW

pW t

n dZW

≠ N Nk N Ng N Nkp Ngb

kW gW

l l lb l pW bW

r rp rb rk rkp rd w y f v s sW h hW m mW n ≠ N NW l lW r w y p b t d tS dZ k g gm gn gl gr kp gb f v s m m mr m ml

144

m y f v s sW h hW m W n ≠ N NW l lW r wn nf ns nl n nw

lW≠ N Nm Nn Nl Nr l lm lw r rf rs rm n ≠ rw w wr y Examples: /gm/ [gm] /nyUŸgma! [nyUŸgma!] ‘camel’ /

’

/gr/ [gr] ⁄]

] ‘small ne’ mk/ [mk] /yeŸmke!ni !/ [yeŸmke!ni !] ‘bat wing’

mkpe!g/] ‘disobedient person’ mgb/ [mgb] /tSaŸmgbEŸrEŸ/ [tSaŸmgbEŸrEŸ] ‘a broom’

m] /dZO!mmO!g´rI!/ [dZO!mmO!g´rI!] ‘to greet (evening)’ friend is there.’

mr/ [rm] /nOm + rI/ [nOrmI ] ‘scorpion’ npW/ [npW] /kUŸnpWE!r/ [kUŸnpWE!r] ‘soft thing’ nt/ [nt] /laŸnta!/ [laŸnta!] ‘blunderbuss’ nd/ [nd] /bø⁄ndi Ÿø⁄/ [bøŸndiø⁄] ‘anger’ ntS/ [ntS] /tSuŸntSu!leŸ/ [tSuŸntSu!leŸ] ‘shade of tree’ ntSW/ [ntSW] /nEŸntSWa!a!lIŸ/ [nEŸntSWa!a!lIŸ] ‘unstable person’ ndZW/ [ndZW] /daŸndZWE!/ [daŸndZWE!] ‘wall’ nf/ [nf] /kUŸnfuŸu!li !/ [kUŸnfuŸu!li !] ‘very black thing’ ns/ [ns] /ba!nsU!a!/ [ba!nsU!a!] ‘bravery’ nl/ [nl] /U!nla!/ [U!nla] ‘that one’ nlW/ [nlW] /kUŸnlWEŸE⁄/ [kUŸnlWEŸE⁄] ‘green thing’ nw/ [nw] /meŸnweŸe!/ [meŸnweŸe!] ‘father’s y. brother’ Nk/ [Nk] /ko!NkoŸoŸro !/ [ko!NkoŸoŸro !] ‘queen mother’ NkW/ [NkW] /kUŸNkWa!a!rI!/ [kUŸNkWa!a!rI!] ‘things’ Ng/ [Ng] /gaŸNgaŸN/ [gaŸNgaŸN] ‘a dance’ NgW/ [NgW] /kUŸNgWE!g´rI!/ [kUŸNgWE!g´rI!] ‘weak thing’ Nkp/ [Nkp] /be!Nkpo!Nø!/ [be!Nkpo!Nø!] ‘one by one’ Ngb/ [Ngb] /daŸNgbaŸla!/ [daŸNgbaŸla!] ‘walking stick’ Nm/ [Nm] /nIŸha!Nma!/ [nIŸha!Nma!] ‘waves’ Nn/ [Nn] /so!Nni !/ [so!Nni !] ‘to be seated’ Nl/ [Nl] /tO!NlI!/ [tO!NlI!] ‘to argue’ Nr/ [Nr] /gba ŸNre!/ [gbaŸNre!] ‘to be smart’

/gn/ [gn] /kpeŸgnE!n/ [kpeŸgnE!n] ‘powerful person/gl/ [gl] /glI!/ [glI] ‘to gargle’

/tuŸgre!/ [tuŸgre!] ‘mortar’ /mp/ [mp] /kUŸmpaŸla⁄/ [kUŸmpaŸla ‘needs winnowing’/mb/ [mb] /kOŸmba!/ [kOŸmba!] ‘t.z. ball’ /mbW/ [mbW] /bWe⁄mbWi !/ [bWe⁄mbWi ! sto//mkp/ [mkp] /lWeŸmkpe!g// [lWeŸ//mm/ [m/ml/ [ml] /n¶Ÿ da!m lo!!le!/ [n¶Ÿ da!m lo!!le!] ‘My//////////////////////

145

[/oŸlpWE!/] ‘liver’ lb/ [lb] /bi ŸlaŸlbi Ÿe!/ [bi ŸlaŸlbi Ÿe!] ‘orphan’

[bE!lbWa!] ‘any’ lm/ [lm !lm I!n] ‘yellow’

[lw Z ŸlwE!EŸ/ Ÿlw EŸ] ‘an or [rp] /kUŸr !tUŸ/ [kUŸrpO!O!t ] ‘coal pot’ [r I!/ ] ‘to be [rd] /kEŸr EŸ/ ] ‘very much’ [rk] /taŸrk n] ‘now’ [rkp] /kpi ! [kpi !rkpi !r black’

/ [maŸrfa!] ‘gun’ rs/ [rs] /daŸrsI‡!I‡!/ [daŸrsI‡!I‡!] ‘metal pot’

!/ [hø!rmi !/] ‘to yawn’ /rn/ [n] /dO nI/ [d I!] ou

cf. /dOr/ [d r !/] (sg.)’ / / [≠] /dO + ≠Ua/ [d ≠UŸa!] un r fo soup/rw/ [rw] /laŸrwIŸ/ [la wIŸ] /wr/ [rw] /sIwe + rI/ [s wI] o die’ - ther so /C clust s: info matio not av labl References: Crouch, M. 1994. The Phonology of Deg. ms. GILLBT, Ghana.

/lpW/ [lpW] /oŸlpWE!///lbW/ [lbW] /bE!lbWa!/ / ] /so I!n/ [so!lm/lw/ ] /aŸd Wa [aŸdZWa E! chid’ /rp/ pO!O UŸ/rb/ b] /wOŸrb [wOŸrbI! hot’ /rd/ d [kEŸrdEŸ/rk/ I!n/ [taŸrkI!/rkp/ rkpi !rkpi !r/ kpi !r] ‘very/rf/ [rf] /maŸrfa!//rm/ [rm] /hø!rmi

r + O!n ‘s p (pl.)’ OŸ ‘soup

r≠ r OŸ ‘h ge r ’ Ÿr ‘to catch’ (pl.) Ir ‘t

O n er r n ai e

146

ilabial lv Glottal

Finnish (Uralic, Finnic) Phonemic consonant inventory: B Labiodental A eolar Palatal Velar Stop p (b) t d k (g) Fricati e v (f) s h Nasal m n NLateral l Trill r Glide y √- /b, f/: used in loanwg, ords

d k s h m n N l r √ y Patterns of son/C clusters:

p t p t nn d k hn kr s ss sl h hrm hr m m~mp m nt mm n mp n Nk~ nn t~nn ng ny N N NN k~NN l l y ls lm ll lt~ll k~l~l ll r rt~rr r rm k~r-ry rr rr r√ √ y

Exa le

] g’ mp s:

hm/ [hm]

amman] ‘comb gen.’ /kampa/ [kampa] ‘comb’

yard’ nt/ t] or [nn]

/kanto/ [kanto] ‘stump’

/tn/ [nn] /a√atnee/ [a√annee ‘open, pot-3s/kr/ [kr] /√uokra/ [√uokra] ‘rent’ /kn/ [hn] /näk + nyt/ [nähnyt] ‘see, 2ptc’ /sl/ [sl] /kaisla/ [kaisla] ‘bulrush’ /sn/ [ss] /nousnut/ [noussut] ‘rise, 2ptc’

hr/ [hr] ///mp/ [mm] or [mp]

/kamp + an/ [k /mt/ [nt] /lum + ta/ [lunta] ‘snow, par.’ cf. /lumena/ [lumena] ‘snow, ess.’ /mm/ [mm] /mummi/ [mummi] ‘granny’ /np/ p] /menen pihalle/ [menem pihalle ]‘I am going into the [m

[n/

147

on/ [kannon] ‘stump, gen.’ nk/ [Nk] or [ng]

[ahdiNko] ‘difficulty’ nkan/ ] andle, gen’

[nn] r] [ny] [Nk] o

/laNka/ ] ‘thread, nom. sg.’ /laNk + an/ n] ‘thread, gen.’

[NN] [kaN-Nen] ‘bar-gen’

/kult + a/ [kulta] ‘gold’ an/ [kullan] ‘gold, gen.’

/lk/ [lk] or [l] / lko/ [pelko] fear/pelkon/ [pelko] fear gen

imp. 2sg.’ /lke/ [lke] or [lye]

/polke + n/ [polyen] lkea [polkea] s mp

/ls/ [ls] /lm/ [lm] /su:delma/ [su: elma kiss

/ / lnut/ [tu t] / llo/ [pallo] a

[rt] or [rr] / ta √i ] tre ’

/ t + a a/ √ir ta]/ [rk] or [r]

/ ärkä/ [märkä] wet /märkässä/ [märässä] ‘wet, ine.’

[rye] /särke + n/ [säryen] ‘break, 1sg.’ [nurmikko] lawn’

r√ah ah

- /k/ alternates with [h] before [t], [d], and [n].

arlsson, Fred. 1999. Finnish: An Essential Grammar. Routledge: London and New

/kant + /

/ahdinko/ /sa [sangan ‘h /nn/ /manner/ [manne ‘mainland’/ny/ /Nk/ r [NN] [laNka

[laNNa/NN/ /kaN-Nen/ /lt/ [lt] or [ll]

/kult +

pe ‘ ’ ‘ , .’

‘stamp, 1sg.’

/po / ‘to ta ’

d ] ‘ ’ /ln [ll]

[ll] tu llu ‘open, 2ptc’

/ll/ pa ‘b ll’ /rt/

√ir / [ [

rta ‘s am √ ri t ra ‘to stream’

/rk m ‘ ’

/rke/ /rm/ [rm] /nurmikko/ ‘/rn/ [rr] /purnut/ [purrut] ‘bite, 2ptc’ /rr/ [rr] /pur-ra/ [purra] ‘bite-linf’ /r√/ [r√] /ha ko/ [har√ ko] ‘sparse’ Notes:

References: K

York. Sulkala, Helena & Merja Karjalainen. 1992. Finnish. Routledge: London and New

York.

148

N

Lab l

Fur ( ilo-Saharan, Fur) Phonemic consonant inventory: ia Den./Alveolar Palatal Velar Stop b t d Ô k g Fricative f s (h) Nasal m n ≠ N Latera l l Tap r Glide w y - /h/: found only in two lexical items, in one of which it varies with [/] Patte s of son/ srn C clu ters:

d Ô r w y b t k g f s m n ≠ N l b t d Ô k g f s m m mt md mm ml b n mb nt~nd Nk ns nn NN ll rrnd ≠≠≠ N ~Ng Ng n mm nn NN ll rr nt ÔÔ Nk s l lt~ld lg lm lN ld lf~lb ll r rb rt~rd rg rm rn rN rl rr rÔ r≠ w y

Examples: /mb/ [mb /reŸmbe / [reŸmbe] ‘ladle’ ] mt/ [mt] /nuŸmti Ÿ/ [nu Ÿmti Ÿ] ‘Albizia amara’

nd] /C-umd-/ [C-und-] ‘pluck’ [lm] or [mm] /*C-uml/ [C-ulm] or [C-umm] ‘find’

nb/ [mb] /iŸn beŸleN/ [i ŸmbeŸleN] ‘he says this’ nt/ [nt] or [nd] /kaari-nta/ [kaarinta] or [kaarinda]

‘enclosure;fold’(pl.) nk/ [Nk] /naŸma i Ÿn kuŸa/ [naŸma i ŸN kuŸa] ‘then he said this’ nd/ [nd] /maŸndi Ÿn/ [maŸndi Ÿn] ‘meagre’

ns/ [ns] /unseŸl/ [unseŸl] ‘tear’ (tr. 1 sg. pt.) [nn] /kinnaŸ/ [kinnaŸ](< *kitaŸn)

‘younger brothers or sisters’ n≠/ [≠≠] /in ≠aŸmuroŸ/ [i≠≠aŸmuro] ‘this spider’

//md/ [/ml/ //

// //nn/ /

149

[ i ŸNNoŸoŸl] ‘this tuft of grass’ nl/ [ll] /baŸin-le/ [baŸille] ‘to/from the elder

brother’ nr/ [ /in ro ŸoŸ/ ‘this river’

[ / Ÿ Ÿ [ b ŸÔÔebbioŸ] ( too’ [ N ŸN ›p ŸNko›p] or [su ŸN › [Ng] /aNgir/ [aNgir] ‘elephant’

/Ns/ [ns] /aŸluN-siŸ/ aŸlunsi Ÿ] ‘I know (it)’ [mm] /duiN mu Ÿrta] ‘my horse’

[nn] /uuŸN niino/ ‘meat of cow’ ‘many’

[deelleŸeŸl] ‘his donkey’ /kuiN reŸito/ [kuirre Ÿito] ‘my fields’

lt/ [lt r [ [kol-taŸ] or [kol-da Ÿ] ‘months’ d/ /g Ÿ/ [g Ÿ] ‘argument’ g/ /d Ÿ≠ [daŸlgaŸ≠ ‘ am ’ f/ /dilfa Ÿn/ [dilfaŸn ] or [dilba ] tra sp m /ki a/ [ki ‘heart’ N/ N /fu i/ [fu Ni] ‘show’ (3 sg. pf.) l/ [ll] /ulloŸ/ [ulloŸ] ‘steal’ (1 sg pf) b/ /g Ÿ [g Ÿ ‘ orkt/ /k Ÿr-ta Ÿ/ [keŸweŸrtaŸ] or [keŸweŸrdaŸ] ‘porcup es

/rt/ /duŸrte/ [d ‘soft’g/ /zaŸrgaŸ≠ [z Ÿ≠ trap’

/rs/ /urse/ [urse] ‘penetrate’ (1 sg. pt.) /rm/ m] d [d Ÿ] ‘skin’ /rn/ fa neŸ/ [farneŸ] ‘female (animal)’ ≠/ ≠ b ≠o [bor≠oŸ] ‘jackaN/ rN ta N [ta Na] ‘ ot’/ [rl /k-oŸro Ÿ Ÿ/ [k-oŸrla ‘rounds’

rr/ [rr] /u Ÿrro ›/ [uŸrro ›] ‘drive out’ (1 sg pf)

i, 1990. elm .

/nN/ [NN] /i Ÿn NoŸoŸl/ /

/ rr] [na

[irro ŸoŸ] ntio]

im) too’ /Nt/ nt] /naŸN tiŸo ‘he seized (h

/NÔ/ ÔÔ] g] (1 ex.)

/naŸN Ôe/su ko

bio/ Ÿ [na ‘he threwgop] ‘hoe

it) sp.’ /Nk/ Nk] or [ / [su

/Ng/ [

/Nm/ ŸrtaŸ/

[duimmu Ÿ /Nn/ [uuŸnniino]

ŸNNa]

/NN/ [NN] /soŸNNa/ [soNl/ [ll] /deeN leŸeŸl/ /

/Nr/ [rr] / ] o ld] /kol-taŸ/ /l [ld] uŸlda uŸlda/l [lg] aŸlga / ] sw p/l [lf] or [lb] Ÿn ‘ p .’/l / [lm] Ÿlm Ÿlma] /l [l ] lN l /l /r [rb] u Ÿrbal / uŸrbal] st ’ /r [rt] or [rd] eŸwe in ’

[rt] uŸrte] /r [rg] / aŸrga ] ‘pigeon

[rs] [r / aŸrmaŸ/ aŸrma [rn] / r

/r [r ] / or Ÿ/ l’ /r

l [ ] / ŸraŸ- a/ Ÿr fo

/r ] l-a Ÿ] /

References: Jakob Angelika. A Fur Grammar. H ut Buske Verlag. Hamburg

150

Laryngeal

Kazakh (Turkic, Aralo-Caspian)

Phonemic consonant inventory: Labial Dental Alveolar Palatal Velar Stop p b t d k g Fricative (f v) s z (S) Z x V h Affricate (c) (t S dZ)Nasal m n N Latera l l Trill r Glide w y s - /v, f, c, tS, dZ, S, x, h/: occur only in loanwords Patterns of son/C clusters: p b t k g s z Z x Ø h m n N l rd w y p b t tt d dr k g s z z m Z x Ø h m mp mk b md md m n mb nt nd Ng nZ nb nd nd N Ng NZ Nn l lp ld lk lg lZ lm ld r rp rb rt rd rk rg rs rm rn rl w y yd yl E/

xamptl/ [tt] /at + lAr/ [attar] ‘horses’ ql/ [qt] /balIq + lAr/ [balIqtar] ‘fish’ zm/ [zm] /kyzmet/ [kyzmet] ‘service’ mp/ [mp] /kempir + ge/ [kempirge] ‘old woman, Dat.’ md/ [md] /atam+di+ki +n/ [atamdikin] ‘my father’s’ mk/ [mk] /keliwi mümkin/ [keliwimümkin]‘he might come’ mm/ [mb] /kymmat/ [kymbat] ‘expensive’ ml/ [md] /khanIm + lAr/ [khanImdar] ‘ladies’ nb/ [mb] /¨nan + ba/ [¨namba] ‘don’t believe’ nd/ [nd] /tüsindi/ [tüsündi] ‘he understood’ ng/ [Ng] /kön + gen/ [köNgen] ‘one who agreed’ nZ/ [nZ] /kel + gen + Zok/ [kelgenZok] ‘he did/has not arrived’

les:

///////////

151

‘don’t believe’ nn/ [nd] /minnet/ [mindet] ‘duty’

r/ [aywandar] ‘animals’ Ng/ [ /tö ömeN ‘lowe

/NZ/ [NZ] /eN + Zaksy/ NZaksy] ‘the [ u + n¨N/ N ka ‘form’

/ld/ [ld] /älde/ [älde] ‘some’ [lk] /xalk¨/ ¨ ‘people’ e [kelgen e did/has not arrived’

/lZ/ [lZ] /ol + Zerde/ rde] ‘that place’ [lm] /al + mA/ a] ‘don’t take’

ll/ [ld] /köl + lAr/ [köldär] ‘lakes’

another’ rt/ [rt] /art + ¨nda/ [art¨nda] ‘behind’

en öttim/ [köpirdenöttim] ‘I crossed the bridge’

e] ‘together’ rga ast

inst’

de] ‘sometimes’

/yl/ [yl] /ay + lAr/ [aylar] ‘moons, months’

References: Bekturova, A. Sh. & Sh. K. Bekturov. 1996. Manual of the Kazakh Language. Almaty

ges

/nm/ [nb] /sen + mA/ [senba]//nl/ [nd] /aywan + lA/ Ng] meN + gi/ [t

[egi] r’

best’ /Nn/ Nn] /k l¨N [kul¨ n¨N] ‘the colt’s’ /lp/ [lp] / lp¨/ [kalp¨]

/lk/ [xalk ] /lg/ [lg] /k l + gen + Zok/ Zok] ‘h

[olZe/lm/ [alm //rp/ [rp] /gurp¨/ [gurp¨] ‘custom’ /rb/ [rb] /bir + biri/ [birbiri] ‘one //rd/ [rd] /köpird/rk/ [rk] /kyrk¨/ [kyrk¨] ‘forty’ /rg/ [rg] /birge/ [birg /ot¨r + gan/ [ot¨ n] ‘sit-p .par’ /rs/ [rs] ‘aga /kars¨/ [kars¨]

/körmespin/ [körmes/rm/ [rm] pin] ‘I may not see’ /rn/ [rn] /murn¨/ [murn¨] ‘nose’ /rl/ [rl] /qiyar + lAr/ [qiyarlar] ‘cucumbers’ /yd/ [yd] /keyed/ [key

“Rauan”.Kara, Da!vid Som 02. Kazak. Langua of the World/Materials 417. Lincomfai. 20

Europa.

152

ba)

Bilab Interdentalpalatal

Velar Labio-velar

abio-palatal

Kikamba (Niger-Congo, Kikuyu-Kam Phonemic consonant inventory: ial Alveo- L

Stop k (g) (b) t (d) Fricative B D s (z) Affricate Liquid l Nasal n n

N m 1 N

Glide y w Á - /N/: an underlying moraic nasal phoneme which is unspecified for place of articulation

o n i arenth se o o nd un e yi ly, ved

atterns of son/C clusters: [mb] /ko-N-buly-a`/ [koombulya`] ‘to grab me’

grabe’

for me’

sure’

guar/Nz/ [nz] /ko-ko!Nz-aŸ/ [kokoÿo`nza`] ‘to fold’

climb

’ ntain’

Roberts-Kohno, Rosalind Ruth. 2000. Kikamba Phonology and Morphology. Ph.D.

- C nso ants n p e s: n t f u d rl ng deri P/Nb/ cf. [kobulya`] ‘to ’ /Nt/ [nd] /ko-N-ta!l-aŸ/ [koondaÿlaÿ] ‘to count m

cf. [kotaÿlaÿ] ‘to count’ /Nd/ [nd] /ko-liNd-aŸ/ [kolii ŸndaŸ] ‘to cover’ /Nk/ [ng] /ko-N-kon-a`/ [koongona`] ‘to hit me’ cf. [kokonaŸ] ‘to hit’ /Ng/ [ng] /ko-bi !Ng-aŸ/ [kobi ÿi`nga`] ‘to shut’ /ND/ [nD] /ko-N-Di!m-eC-aŸ/ [koonDi !me.aŸ] ‘to measure cf. [koDi ÿmaÿ] ‘to mea/Ns/ [ns] /ko-N-suNg-a`/ [koonzuu`nga`] ‘to guard me’ cf. [kosuu`nga]` ‘to d’

/NdZ/ [ndZ] /ko-loNdZ-aŸ/ [kolooŸndZaŸ] ‘to chase away’ /Nl/ [nd] /ko-N-le!es-y-aŸ/ [koondeÿe`sya`] ‘to cause me to a mountain cf. /ko-le!es-aŸ/ [koleÿeŸsaŸ] ‘to climb a mou Notes: - Kikamba is a typical Bantu language with CV syllable structure.

References:

dissertation. The Ohio State University.

153

Labial Alveolar Palatal Velar

Kimatuumbi (Niger-Congo, Bantu) Phonemic consonant inventory:

Stop p d c k b t g Fricative f v s z Affricate tS dZ Nasal m n ≠ N Liquid l, r Glide w y - /v, z, f, r/: appear only in loanwords Patterns of son/C clusters: p b t d c k g f v s z tS dZ m n ≠ N l r w y m mp mm nt nn Nk NN ≠≠ mm nn ≠≠ nn NNw ≠≠ n mb mb nd nd Ng Ng ≠dZ ≠dZ m n ≠ N nd Ngw ≠dZ Examples:

mpa'andi ä'/ ‘planter’ /mp/ [mp] /muä-pa'andi ä'/ [-baa/mb/ [mm] /muä 'nda/ [mmaa'nda] ‘slave’

mt/ [nt] /kuä-muä-te'lekya/ [ku äuänte'lekya] ‘to cook for him’ [nn] /muä-da'laa'nzi ä/ [nna'laa'nzi ä] ‘bitter orange’

cf. [mi äda'laa'nzi ä] ‘bitter oranges’

cf. [mi ägu'äuända] ‘fields’ /mtS/ [≠tS] /mu ä-tSo'omi 'ä/ [≠tSo'omi 'ä] ‘reader’ mdZ/ [≠≠] /muä-dZi äi ängi 'ä/ [mu ä-dZi äi ängi] or [≠≠i äi ängi] ‘you should

enter’ mm/ [mm] /muä-ma'ti 'ä/ [mma'ti 'ä] ‘platerer’ mn/ [nn] /ne'muäli 'ä/ [nne'muäli 'ä] ‘dancer’ m≠/ [≠≠] /mu ä-≠a!gi !ä/ [≠≠a!gi !ä] ‘scratcher’

/mN/ [NN] /muä-Na'andi 'ä/ [NNa'andi 'ä] ‘player’ ml/ [nn] /muä-la'alo/ [nna'alo] ‘fool’

cf. [ala'alo] ‘fools’ mw/ [NNw] /muä-wa!atSi 'ä/ [NNwa!atSi 'ä] ‘thinker’

cf. [awa!atSi 'ä] ‘thinkers’ my/ [≠≠] / mu ä-yi!imbi !ä/ [≠≠i !imbi !ä] ‘singer’

cf. [ayi'!imbi !ä] ‘singers’ np/ [mb] /n-bala!ai !/ [mbala!ai !] ‘bald heads’

cf. [lupala!ai !] ‘bald head’

cf. [abaa'nda] ‘slaves’ //md/ /mk/ [Nk] /muä-ku!Nuu!ndi ä/ [Nku!Nuu!ndi ä] ‘filterer’ /mg/ / muä-gu'äuända/ [NNu'äuända] ‘field’ [NN]

/

///

/ / / /

154

[mba!u] ‘ribs’ nt/ [nd] /n-teela!/ [ndeela!] ‘pieces of wood’

‘piece of wood’ nd/

/nk/ [Ng] /n-ki !li/ [Ngi !li] ‘palms’ ‘palm’

[Ng] -go !i/ Ngo!i] ‘ropes’ /ntS/ dZ] /n-tSapi 'äi ätSa' ≠dZ api Sa'] ‘clean, Cl.9 deverbal

adj i 'äi ächa'] ‘clean, Cl. 11’

-dZii !Ngya! !Ngya!] ‘entered (pl.)’ nm/ [m] /niä-ma≠a/ [ma≠a] or [ni äma≠a] ‘I know’

'oli äte/ [no 'oli äte] or [ni ä-no'oli äte] ‘I sharpened’ n≠/ [≠] examples not given

e not iven nd !la/ !la

f a ‘p r Ngw] /n-wIk !/ [Ng kI!lya]! ‘covered )’

cf. [luwI lya!] ‘covered g.)/ ≠ /n- I!ma/ [≠d ma] ‘poles’

f. yI! !] ‘p

dden, David. 1996. The Phonology and Morphology of Kimatuumbi. Clarendon Press. xfo .

/nb/ [mb] /n-ba!u/ / cf. [lute!ela!] / [nd]

cf. [luki !li]/ng/ /n [

[≠ / [ 'äi ät.’

cf. [luätSap/ndZ/ [≠dZ] /n / [≠dZii//nn/ [n] /ni ä-no//nN/ [N] exampl s g /nl/

[ ] /n-laac

[ndaa ] ‘peppers’

/ . l [lu a!la]

I!lyaeppe ’

/nw [ wI (pl. kI! (s ’

/ny [ dZ] y ZI! c [lu ma ole’

Notes: References: O

O rd

155

Klamath (Penutian, Klamath-Modoc)

onemic conson nt inven : Labial Alveolar Palato-

lvPalatal Velar Uvular Glottal

Ph a tory

a eolar Stop p p’ t t’ d k k’ g q q’ G b / Affricate tS tS’ dZ Fricative s h Nasal m9 m n9 n n’ m’ Lateral l l 9 l’ Glide w9 w w’ y9 y y’ - C’: a glottalized consonant

/ p’ t’ tS’ s Patterns of son/C clusters:

p t tS k q b d dZ g G k’ q’ m mt mtS mq mb md mdZ mG m/ m t’ mtS’ mk’ mk p’ m n mp nt ntS nq mb nd ndZ ng nG n/ nq’ nnk nt s’ ntS’ nk’ m 9 m 9 m 9q’ t’ m 9k’m’ m’tS m’/ n’ n’tS l lp lt lt lq lb ldZ lg lG l lp ’ ltS’ lk’ lq’ S lk ld / ’ lt ls l9 l9t l9b l9/ l9q’ w wt wt wq wb wdZ wg wG w/ wp’ wt’ wtS’ wk’ wq’ wsS wk wdy yt ytS yq yb ydZ yG yp’ yt’ ytS’ yk’ yk yd y/ yq’ ysw9 w9dZ w w9tS’ w9k’ 9p’ y 9 y 9q’

h m n9 m’ n’ l l9 l’ w y w9 y 9 w’ y’ n

p pn pm’ pl pl9 pl’ pw pw’pn’ py’ t tm tn tn9 tm’ tn’ tl tl9 tl’ tw ty 9 tw’ ty’ tS tSm tSn9 tSm’ tSl tSl9 tSl’ tSw tSy 9 tSw’tSn tSn’ tSy’ k kn km’ kl kl9 kl’ kw9 kw’kn’ ky’ kw q qm qn qm’ ql9 qw9 qy 9 qwqn’ qw qy qy’ ql ’b b bn l bw by d dm dl dw dy g gm gl gw gy gn G G Gl n Gw Gy p’ P’l p’n p’w p’y t’ t’l t’l9 t’w t’y tS’ t tS’n ’w tS’y S’l tSk’ k k’n ’l k’w k’y q’ q’m q’l q’w q’y q’n s sm sn sm’ sl sl9 sl’ sw sy 9 sw’ sn’m h mm mn ml my m mn’ my’n nh nn n ll lh nw9 ny 9 nw’ ny’ n’ nw ny n9 n9w9 l lh lm ln lm’ ll lh l/ lw9 ly 9 lw’ lln’ lw ly y’ l9 l9w9 l9w’ l9w l9y w wn wm’ wl wl9 wl’ ww wy 9 wy’ wn’

156

w y y’ h m n n9 m’ n’ l l9 l’ w9 y 9 w’ y yh yn ym’ yl yl9 yl’ yw’ yn’ yw yy w w9m’ wl9 wl’ 9

Examples:

/ I] /pn’/ [pn’] /ton’Ipn’I [ton’Ipn’ ‘five’ /pl/ [pl] /d´pl´l/ [d´pl´l] ‘loon’

obj.

wn

’

e

/ rt’ ’

´/

’ /tSw/ [tSw] //ItSw´/ [/ItSw´] havings’

/bl/ [bl] //ebli/ [/ebli] ‘brings a long back’

/tm/ [tm] /s´tm´blI/ [s´tm´blI] ‘asks back’ /tn/ [tn] /hI/e!tna/ [hI/Itna] ‘right around there’

] er /tn’/ [tn’] /do/e!tn’I/ [do/otn’I ‘one from right ov there’

/t’l/ [t’l] /gEt’leGI/ [gEt’leGI] ‘goes over a in’ mounta

/t’l9/ [t’l9] /nqot’l 9 / [nqot’l 9 ] ‘scorches on a surface’

’´/ /dm/ [dm] /sEdmEttS [sEdmEttS’´] ‘washes one’s o (cloth, hair)’

/dw/ [dw] /qbEdw´t/ [qbEdw´t] ‘taste!’ /kn/ [kn] /mId´nsIknI/ [mId´nsIknI] ‘from your place’

/kl 9/ [kl 9] //Ikl 9 / [/Ikl 9 ] long obj. ‘puts a down on

/kl/ [kl] /w´klo/ [w´klo] ‘buckskin sack’ /kl’/ [kl’] /sbokl’I/ [sbokl’I] ‘sweats in a

sweathouse’ /kw/ [kw] /dokw´/ [dokw´] place name

/ ] /k’n/ [k’n] /tS’w´k’n´ [tS’w´k’n´ ‘cottontail rabbit’ /k’l/ [k’l] /ntSik’lGa/ [ntSik’lGa] ‘drips down’ /k’w/ [k’w] //Inok’wa/ [/Inok’wa] ne cross’ ‘lets someo/gm/ [gm] /dZIgmIn/ [dZIgmIn] ‘iron, nail’

’ /gn/ [gn] /lognyGI/ [lognIGI] ‘pl. stand above /gl/ [gl] /wdoglgi/ [wdoglgi] ‘comes to eat

someone’ /qm/ [qm] /d´qmIl/ [d´qmIl] ‘hat’

gniz/qm’/ [qm’] /q´yqm’´/ [q´yqm’´] ‘does not reco someone’

wl/qn/ [qn] /q’aqneg [q’aqnegw´l] ‘is dirty on top’ shi/qn’/ [qn’] /qaqn’ul’s/ [qaqn’ul’s] ‘elkhide armor

on/ql/ [ql] /m´ql´Gs/ [m´ql´qs] ‘Indian, pers /q l 9/ [q l9] /boq l 9 / [boq l 9 ] ‘has a cataract’

/Gl/ [Gl] /dZEGlE/ [dZEGlE] ‘blood’ /tSn’/ [tSn’] //owItSn’ [/owItSn’´] ‘go along shoulder to

shoulder’ /tSl 9/ [tSl 9] /pEtSl 9y/ [pEtSl 9I] oot inside‘puts a f

‘makes s

157

´/ ´] /tS’w/ [tS’w] //IltS’wy/ [/IltS’wy] objs. right

/ ] ’´/ sation’ / Z] /

k

/ml/ [ml] /wdoml´mn´/ [wdoml´mn´] s/

´/ ] /nt/ [nt] /bIbl´nt´nn´/ [bIbl´nt´nn´] on both

´/ ] /nt’/ [nt’] /getant’it/ [getant’it]

/ a] s’

ne in

/nG/ [nG] /manGas/ [manGas] ‘housefly’

e’

/lp/ [lp] /GElp´/ [GElp´] ‘(animal) is hot’

/ a] /lk’/ [lk’] /gobalk’ItS’kanga/ [gobalk’ItS’kanga] in

/tS’n/ [tS’n] /GEtS’not [GEtS’not ‘grows up with’ ‘brings pl. up to’

/sn/ [sn] /tS’osnI/ [tS’osnI] ‘forever’ /sl/ [sl] /tSEslEyn/ [tSEslEyn] ‘plant’ /sw/ [sw] /d´sw´/ [d´sw´] ‘puts a hand into

water’ /mb/ [mb] //´mbo/ [/´mbo] ‘water’ /mt/ [mt] /qemt´nt/ [qemt´nt] ‘at the back’ /md/ [md] //´md´/ [/´md´] ‘digging stick’/mk’ [mk’ /s´q’´mk [s´q’´mk’´] ‘denies an accu/m// [m/] //am/a/ [/am/a] ‘plays with a doll’ /mdZ [md /p/omdZip [p/omdZip] ‘cousin’ /mn/ [mn] //´mn´lin´ll´/ [/´mn´lin´ll´] ‘goes along the ban

crying’ ‘swims behind’

/m’// [m’/] //om’/´m’ [/om’/´m’s] ‘so-so’ /m9k’/ [m9k’] //´m9k’´/ [/´m9k’´] ‘maybe’ /np/ [mp] /honptSI/ [homptSI] ‘one like that’/nb/ [mb] /honbatn [hombat´ ‘flies ashore’

‘all around sides’

/nd/ [nd] /sot´nd´nG [sot´nd´nG´ ‘wires’ ‘over there’

/nk/ [nk] //anko/ [/anko] ‘wood’ /nk’/ [nk’] /nqEnk’wa [nqEnk’w ‘yells acros/ng/ [ng] /d´ng´tS/ [d´ng´tS] ‘palate’ /nq/ [nq] /d´nqn’I/ [d´nqn’I] ‘how many times’/nq’/ [nq’] /nqEnq’ya/ [nqEnq’ya] ‘shouts at someo

the road’

/n// [n/] /s/Ew´n/I/ [s/Ew´n/I] ‘give pl. objs.!’ /ndZ/ [ndZ] //IndZIl/ [/IndZIl] ‘angel’ /ns/ [ns] /mId´nsIknI/ [mId´nsIknI] ‘from your plac/nh/ [nh] /penhI/ [penhI] ‘be naked’ /nn/ [nn] /ndannok’/ [ndannok] ‘all three’ /nl/ [ll] /honlina/ [hollina] ‘flies along the bank’ /nl 9/ [lh] /honl 9y/ [holhi] ‘flies into’ /nw/ [nw] /bonw´/ [bonw´] ‘drinks’ /n’tS/ [n’tS] //on’tSe/ [/on’tSe] ‘later’

/ld/ [ld] /loldm/ [loldam] ‘winter’ /lk/ [lk] /gobalkanga [gobalkang ‘waves around’

‘soars around circles’

158

/lq’/ [lq’] //Is´lq’I/ [/Is´lq’I] lG/ [lG] //ElG´/ [/ElG´] ‘names’

] /d´l/ak/ [d´l/ak] ‘truly’

/lm/ r’ ln/ [ln] /inb´talniN/ [imb´talniN] ‘payment’ ll/ [ll] /p’alla/ [p’alla] ‘steals’ l l 9/ [lh] /pal l 9a/ [palha] ‘dries on’ ll’/ [l/] /re!ryall’i/ [yalyal/i] ‘clear’ lw9/ [lw9] /tS’´y´lw9 sg´/ [tS’´y´lw9 sg´] ‘backs away from’ lw/ [lw] /bEl´/ [bEl´] ‘puts the tongue into

liquid’ lw’/ [lw’] //œlw’´/ [/œlw’´] ‘(stars) twinkle’ ly/ [ly] //´ly´ns/ [/´ly´ns] ‘onion(s)’ l 9t’/ [l 9t’] /m’oq’´ l 9t’´/ [m’oq’´ l 9t’´] ‘gets wet’ l 9q’/ [l 9q’] //I l 9q’´/ [/I l 9q’´] ‘removes hair or

feathers’ wk’/ [wk’] /tS’lIwk’I/ [tS’lIwk’I] ‘close it!’ wb/ [wb] /niwbanI/ [niwbanI] ‘drives a herd to the

end’ wd/ [wd] /gEwd´nt/ [gEwd´nt] ‘in my’ wg/ [wg] /blEwgI/ [blEwgI] ‘has a lot’ wtS’/ [wtS’] /q´wtS’I/ [q´wtS’I] ‘wokas juice’ wn/ [wn] /nt’Iwneg´/ [nt’Iwneg´] ‘falls into a hole’ wn’/ [wn’] //Iw9awn’aa/ [/Iw9awn’a] ‘is propped up

against’ wm’/ [wm’] //iwm’em’i/ [/iwm’em’I] ‘heckleberry season’ wl/ [wl] /hIwliw´/ [hIwliw´] ‘drapes off the edge’ wl 9/ [wl9] /sn’Ewl 9Is/ [sn’Ewl 9Is] ‘cold’ w9m’/ [w9m’] //Iw9m’atS’´/ [/Iw9m’atS’´] ‘is on the end’ w9l 9/ [w9l 9] //Iw9l 9I/ [/Iw9l 9I] ‘is inside’ w9y9/ [w9y9] //Iw9y9eni’a/ [/Iw9y9en’a] ‘is inside in’ w9y’/ [w9y’] /gow9y’´sGs/ [gow9y’´sqs] ‘venereal disease’ yp’/ [yp’] /sw´yp’ek’´/ [sw´yp’ek’´] ‘strins the eyes’ yt’/ [yt’] /bl´yt’it/ [bl´yt’it] ‘above’ yd/ [yd] /bl´yd´lk’ni/ [bl´yd´lk’ni] ‘God (one-from-

above)’ yG/ [yG] //oyGI/ [/oyGI] ‘puts a long obj. over’ yl/ [yl] /snEylig´/ [snEylig´] ‘camps on the bank

of a stream’ yl’/ [yl’] /snEyl’G´/ [snEyl’G´] ‘hearth’

/lg/ [lg] /tSElg´/ [tSElg´] ‘eats the roasted skin of birds’ ‘corn’

//l// [l//ltS/ [ltS] /k’ul’altS´/ [k’ul’altS´] ‘goes to gather

Valeriana’ /ltS’/ [ltS’] /d´ltS’I/ [d´ltS’I] ‘arrow’

[lm] /snelma/ [snelma] ‘makes it thunde//////

////

//

/////

//////////

//

/

159

/yl 9/ [yl9] //Eyl 9y/ [/Eyl 9I] ‘puts the head inside’ yw/ [yw] //Eyw´/ [/Eyw´] ‘puts the head into

water’

ences: . A. R a Dictionary niversity of Cali nia Publicatio in

ornia Press: Berkeley and Los Angeles. . A. R a . niversity of Calif nia Publicatio in

inguis U C er eley and Los A geles.

/

ReferBarker, M

Linguistics. Vol. 31. University of Calif. 1963. Klam th . U for ns

Barker, ML

. 1964. Klam th Grammariversity of

U or nstics. Vol. 32. n alifornia Press: B k n

160

abial

Koasati (Muskogean, Eastern) Phonemic consonant inventory: bil dental Palato-alveolar velar glottalStop p b t c k / Fricative F Ò s h Nasal m n Latera l l Glide w y Patterns of son/C clusters:

p b c k s t / F Ò h m n l w y p pl b bb t tm tn tl c cl k km kn ↓ / F Fn FF Ò ↓ s sm sn sl h hm hn hl m mp mb ↓ ↓ ↓ F Ò s mm n ↓ n ↓ ww yy n np nb nc nk nÒ ns nh nn nnnt l lp lb lt lc lk lF nÒ lh lm nn ll lw ly w ww y yl

Examples:

[bb] g many

look t something’

/tn/ [tn] / !:ka] / li!:c] ’s mouth

/km/ [km] /i !kmot/ [i !kmot] id not gather it’

r [k´lr [to !k r [to !glon] ‘two’

/Fl/ [FF] /siF-li-n/ [si !FFin]

/pl/ [pl] /hipli !/ [hipli !] ‘snow’ /bl/ /Òob-li-n/ [Òo!bbin] ‘to hole somethin

times’ /tm/ [tm] /matmici!ksin/ [matmici!ksin] ‘to raise the head and

a/matna:Òi !:ka [matna:Òi ‘radio’

/tl/ [tl] /alotlo!:kan/ [alotlo !:kan] ‘to be full (pl.)’ /cl/ [cl] /ico:Òawicli!:c [ico:Òawic ‘to cause someone

to pucker’ ‘They d

/kn/ [kn] /acokna!/ [acokna!] ‘pick [tool]’ /kl/ [kl] o ] or [gl]

/to!klon/ [to!klon] o ´lon] o/Fn/ [Fn] /ho !Fnan/ [ho !Fnan] ‘to smell something’

‘to buck’

161

ki]

/tiÒ-li-n/ [ti!ÒÒin] ‘to push something’ / eyelids’

/sn/ [sn] /aka!snon/ [aka!snon] ‘to be hungry [plural]’

on/ ] ’ n/ elling’

/ ] /am-pacokko!:ka/ [ampacokko!:ka]‘my chair’

/mt/ [mt] /aho!mta!kko‡/ [aho!mta!kko‡] ‘I do not moan.’ n am- s

ci!n/ i !n] stick your tongue out at him!’

in am- sses/

mk/ [nk] in am-series of possessive prefixes

/am-kasto!/ [an-kasto!] ‘my flea’ ‡F] /am-Fo!:si/ [a‡-Fo!:si ] ‘my bird’

/Vms/ ‘my saw’ mh/ [mh] /asaki!mho ‡/ [asaki!mho ‡] ‘It [rice;beans] is not

overcooked.’ Vmh/ [V‡h] in am-series of possessive prefixes

/am-holikFa!/ [a‡holikFa!] ‘my shirt’ mm/ [mm] /Falammi !/ [Falammi] ‘north’ /am-ma!:ma/ [amma!:ma] ‘my mother’ mn/ [nn] /am-na:sinca!:ka/ [an-na:sinca!:ka] ‘my book’ ml/ [mm] /taFilam-li-n/ [taFila!mmin] ‘to overturn [once]’ ml/ [ll] in am-series of possessive prefixes

/am-layki!/ [allayki !] ‘my dung’ mw/ [ww] /am-wilo!/ [awwilo!] ‘my shoes’ my/ [yy ] /am-ya!:ya/ [ayya!:ya ] ‘my elder sibling’ np/ [np] /coskanpaci !/ [coskanpaci !] ‘goldeneye or bufflehead

duck’ nb/ [nb] /iha!:ni banboki !/ [iha!:ni banboki !]‘rolling land’ nt/ [nt] /konti!/ [konti !] ‘greenbriar’ nc/ [nc] /sanco!/ [sanco !] ‘sand’ nk/ [nk] /li !nkan/ [li !nkan] ‘to swallow two or more

things or times’

/Òl/ [Òl] /ittiÒlapa!tki/ [ittiÒlapa!t ‘copperhead [snake]’ /Òl/ [ÒÒ] in some clusters

/sm/ [sm] /mismi !hlin [mismi !hlin] ‘to flutter the

/sl/ [sl] /slo!nkasi/ [slo!nkasi] ‘switch’ /hm/ [hm] /okso !hm [okso!hmon ‘to be slushy/hn/ [hn] /Fotohna!:ci [Fotohna!:cin] ‘to be good-sm/hl/ [hl] /cowahla!/ [cowahla!] ‘cedar tree’ /mp/ [mp] /oho!mpan [oho !mpan ‘to dine’

/mb/ [mb] /lambi !/ [lambi !] ‘caterpillar’ /am-baya!:na/ [ambaya!:na] ‘my stomach’

/mt/ [nt ] i series of posse sive prefixes /am-towa!/ [antowa!] ‘my onion’

/mc/ [mc] /icoklam [icoklamc ‘Do not

/mc/ [nc] series of po sive prefixes /am-cico!:ba [ancico!:ba] ‘my horse’ /mk/ [mk] /komko!/ [komko!] ‘amphiuma’ /

/VmF/ [V/VmÒ/ [V‡Ò ] /am-ÒakFo!:si/ [a‡-ÒakFo!:si] ‘my brother’

[V‡s] /am-sikkaci!/ [a‡-sikkaci !] /

/

/ ///

///

////

162

[o!nÒat] ‘We two arrived over there’ ns/ [ns] /itto insa!:wa/ [itto insa!:wa] ‘branch’

[hinhi !:can] ‘to expect someone’ nn/ [nn balinn

[nn] examples not given [lp] /hacokp [hacokpalpa!] ‘butterfly’

/lb/ [lb] /olbi !/ /lt/ [lt] no [ilto!hnon]

[lc] /alcika!:n lcika!:na] [lk] /na:silka!/ [na:silka!] ‘that which is said’

(1plneg(1A)-SHOOT:AND:HIT-neg:comp-phr:term) [olFa!] ‘a sprout’

/lÒ/ [nÒ] /kil-Òa!:h-o-V‡/ inÒa ‡] ‘We do not shoot it.’ /lh/ /asi !lhan/ [asi!lhan] ‘to a/lm/ [lm] /malmahka!/ [malmahka!] igh g’ /ln/ /i:t-il-noca!/ i:tin ] ‘bedroom’ /ll/ /patilli!/ pati azal/lw/ [lw] /hassolwa!/ [hassolwa!] ‘spanish moss’ /ly/ [ly] / lyokc [alyokca!] ‘makings [archaic]’ /wl/ /kaw-li-n/ [ka!wwin] ‘to snap many things’ /yl/ [yl li!:cin/ o w l so ing’ References:

ball, eoffr D. 1 1. Ko ati G mm tud in th Anth olog of North erican Indians. University of Nebraska Press.

/nÒ/ [nÒ] /o!nÒat/ //nh/ [nh] /hinhi !:can/ / ] / a!/ [balinna !] ‘need’ /nl/ /lp/ alpa!/

[olbi !] ‘hominy’ /ilto!h n/ ‘to work’

/lc/ a/ [a ‘yarn; untrue tale’ /lk/ /lF/ [lF] /olFa!/

[k !:ho [lh] sk’

in‘l tn [nn] [ll]

[[

noca!lli!] ‘ ea’

a a!/ [ww]

] /Òoy [Òoyli!:cin] ‘t hir meth

Kim G ey 99 as ra ar. S ies e rop y Am

163

ola Laryngeal

Korean (Language Isolate)

Phonemic consonant inventory: Bilabial Alve r Palatal Velar Stop p pÓ p’ t tÓ t’ k kÓ k’ Fricative s s’ h Affricate tÇ tÇÓ tÇ’ Nasal m n N Liquid l Glide w y Patterns of son/C clusters:

p k s tÇ m n N l t hp mm mn × mn t mm nn × ll k Nm Nn × Nn s Nm Nn × × h × × tÇ × × m mb md mg ms mh mtÇ mm mn × mn n mb nd mb ns nh ntÇ mm nn × ll N × Nn Nb Nd Ng Ns Nh NtÇ Nm Nn l × ll lp lt lk ls lh ltÇ lm ll

Examples: pm/ [mm] /pap + mas/ [pammat] ‘appetite’ (<‘ rice’ + ‘taste’) pn/ [mn] /sip + ny´n/ [simny´n] ‘ten years’ pl/ [mn] /p´p + li/ [p´mni] ‘pinciple of law’ tm/ [mm] /mat + my´nÈli/ [mammy´nÈ|i] ‘first daughter-in-law’ tn/ [nn] /mat + nui/ [mannui] ‘first sister’ tl/ [ll] /tikÈt + liÈl/ [tikÈlliÈl] ‘the letters ‘t’ and ‘l’’ km/ [Nm] /kuk + mul/ [kuNmul] ‘soup-broth’ kn/ [Nn] /n´kn´k-/ [n´Nn´k-] ‘enough’ kl/ [Nn] /ky´k + li/ [ky´Nni] ‘separation’ mp/ [mb] /kim + pap/ [kimbap] ‘maki’ (< ‘laver’ + ‘rice’) mt/ [md] /tÇam + tÈlta/ [tÇamdÈlda] ‘fall asleep’ mk/ [Ng] /kam + ki/ [kaNgi] ‘cold’ mm/ [mm] /kam + mul/ [kammul] ‘juice from a persimmon’ mn/ [mn] /pam + namu/ [pamnamu] ‘chestnut tree’

/ml/ [mn] /sam + lyu/ [samnyu] ‘third rate’ np/ [mb] /sin + pal/ [simbal] ‘shoes(-foot)’ nt/ [nd] /kun + te/ [kunde] ‘army’ nk/ [Ng] /han + kaN/ [haNgaN] ‘Han river’ nm/ [mm] /kun + mantu/ [kummantu] ‘fried dumpling’

//////////////

////

164

nnun] ‘big eyes’ nl/ [ll] /non + li/ [nolli] ‘logic’

[toNdÈN] ‘equality’ Np/ [Nb] /tÇ´N + pi/ bi] nance’

] /tÇaN + kun/ [tÇaNgun] ‘ge eral ’ [Nm] a/ a] ‘rain season’

n] /tÇaN + nal/ [tÇaN day’ [Nn] /y´N + lak/ [y´Nnak] ‘downfall’

/lp/ [lp] /kalpi/ [kalpi] ‘ribs’ [lt] ‘reed’ [lk] /kal + ku/ ] ‘desire’ [lm] / [kalmaN] ‘desire’

ln/ [ll] /pul + nÈN/ [pullÈN] ‘incapability’

References: Davis, Stuart and Seung-Hoon Shin. 1999. The syllable contact constraint in Korean:

n Op imality-The retic nalysis. Jour l of E st A an Li uisticIverson, G. and Kee-Ho Kim. 1987. On word-initial avoidance in Korean. Harvard Studies in

orean Lingui ics. Vol. 2, 377-393. Cambridge, Mass.: Harvard University. Iverson, G. and Hyang-Sook, Sohn. 1994. Liquid representation in Korean. In Kim-

enau You g-Ke (ed.) Theoretical Issues in Korean Linguistics. 7 -100.Iverson, G. K. and S.-S. Lee. 1995. Variation as optimality in Korean cluster reduction.

oceedi gs of Eastern tates onfer ce On ingu tics 9

/nn/ [nn] /kÓÈn + nun/ [kÓÈ//Nt/ [Nd] /toN + tÈN/ / [tÇ´N ‘mainte/Nk/ [Ng n/Nm/ /Nn/ [N/Nl/

/tÇaN + m [tÇaNmnal] ‘market

/lt/ /kalte/ [kalte] /lk/ [kalku/lm/ /kal + maN/

A t o a na a si ng s 8, 285-312.

K st

R d, n y , 7

Pr n S C en L is 4.

165

ic consonant inventory: palatal velar

Kuman (Trans-New Guinea, Chimbu) Phonem bilabial alveolar Voiceless stop p t k Voiced prenasalized

nd ng

stop

mb

Fricative s Nasal m n Liquid |, l K Glide w y Patterns of son/C clusters:

p t k mb nd ng s m n | l K w y p p| t k

mb nd nd | ng s m mn n ↓ | l lt K Kk K mb m ↓ tn w y

Examples:

|/ [p|] /p|o/ [p|o] ‘hear, singular’ nd|/ [nd|] /nd|a-n/ [nd|an] ‘mouth-your sg.’

[lt] /olto/ [olto] ‘long’ Kk/ [Kk] /yunaKka/ [yunaKka] ‘I will bring’ Kmb/ [mb] when ‘mb’ is an initial consonant of a subject-suffx

/p|aKmbuka/ [p|ambuka] ‘hear, 3 sg. future’ Kn/ [tn] /kaK-na/ [katna] ‘leg, 1 sg.’ Kn/ [|] in three nouns

/agiK-na/ [agi|a] ‘brother, 1 sg.’ cf. /agiK-n/ [agitn] ‘brother, 2 sg.’ Km/ [m] when ‘m’ is an initial consonant of a subject-suffx

/p|aKmnga/ [p|mnga] ‘hear, 1 pl. future’ |n/ [tn] /mbi|-no/ [mbitno] ‘head, non-sg.’ mn/ [mn] /amniKo/ [amniKo] ‘milk’ nn#/ [n] /ndan-n/ [dan] ‘belly, 2 sg.’

/p//lt/ //

//

/

///

166

-no] ‘belly, non-sg.’

/r, l, K/: do not occur word-initially , s, w, y, l/: do no word- ally

- Only /m, n, K, l, k/ can occur in coda position.

es: man ‘Liquids’. Languages and Linguistics in Melanesia

Trefry, D. 1969. A Comparative Study of Kuman and Pauaian. Pacific Linguistics. Series B. No. 13.

/VnnV/ [VnnV] /ndan-no/ [ndan Notes: - - /p, t t occur fin

eferencR

Lynch, John. 1983. On the Ku14 (1-2), 98 12. -1

167

Phonemic consonant inventory: Labial Dental Alveolar Velar

Leti (Austronesian, Southwest Maluku)

Stop p t d k Continuant v s Sonorant n m l, r Patterns of son/C clusters:

p d v s m t k n l r p pn pl pr t tm ↓ t l 9 tr d nn ll dr k km kn kl kr v mm vn vl vr s sm sn sl sr m md mv ms mm mnmt ml mr n mp nt nv ns nm nn ll nr l lt lm ll lk r rk rv rs rm rn rl

Examples: /pn/ [pn]

nura)

ss’ v a-den

leti/ ’

/kn/ [kn] /knu!ni/ [knu!ni] ‘pillow’

/pni !nu/ [pni !nu] ‘fool’ all reptile /pl/ [pl] /pli !lki/ [pli!lki] k.o. sm

(Emoia cya/pr/ [pr] /m-dupra/ [mdupra] ‘you flee’ /tm/ [tm] /tmu!ra/ [tmu!ra] ‘tin’ /tOtmemi/ [tOtmemi] ‘mirror’

e/ /tn/ [tn] /utu + n [utnu] ‘his flea’ n n/tn/ [t ] /tninne/ [t i !n:I] ‘his ear’

n dne /iatne/ [/yat I] ‘his ban r([t ] = oiced int tal nasal stop)

] /tl/ [tl]9 /pO!tla/ [pO!t l 9a ‘bottle’ /tlE:sa/ [t l9E!:sa] ‘we (inc.) read’ /tr/ [tr] /motru/ [motru] ‘motorboat’ /dn/ [nn] /lOdan/ [lOnna] ‘rattan’ /dl/ [ll] /dudal/ [dulla] ‘horn’ /kuda + [kulleti] ‘Letinese horse’ /dr/ [dr] /dre:sa/ [dre:sa] ‘Andreas

n/km/ [km] /kmiaratne/ [kmjarat I] ‘its narrowness’ /kakma/ [kakma] ‘cloud’

/riEkna/ [ryEkna] ‘let’s say’ /kl/ [kl] /kla!si/ [kla!si] ‘glass’

168

/kr/ [kr] /krait/ [krait] ‘to be slow’ /sikri/ [sikri] ‘Sikir’

u/

/vl/ [vl] /vlira/ [vlira] ‘weaving rod’ ’

/ [sm] rne/ I] /ssisma/ [ssisma] ‘sepia’

/ [sn] /snu!rna/ [snu!rna] ‘thread’ l] /slE!tna/ [slE!tna] ‘strait’

/mt/ [mtaut] ‘to be afraid’ /md/ [md] /koni + mderi/ [konimderi] ‘grasshopper + Mderi’

/ms/ [msyai] ‘you (pl.) rise’ /mm/ [mm] /mmE!i/ [mmE!i] ‘table’ /mn/ [mn] /mniniru/ [mniniru] ‘soft’ /ne!mnu/ [ne!mnu] ‘he drinks’ /ml/ [ml] /mlili !lu / [mlili !lu] ‘sour’ /mr/ [mr] /mrenu/ [mrenu] ‘Mrenu’ /np/ [mp] /n+pura/ [mpura] ‘he is wild’

/kerna + papra/ [kerampapra] ‘Babarese turtle shell (medicine against poisoning)’

/nt/ [nt] /pEnta/ [pE!nta] ‘grass’ /ntEna/ [ntE!na] ‘he tastes’ /nd/ [nd] /dudduandi/ [duÊÊwandi] ‘now’ /nv/ [nv] /nvEra/ [nvEra] ‘he opens’ /nanvetrao/ [nanvEtrao] ‘he eats maize’ /ns/ [ns] /ionsu/ [/yonsu] ‘rainbow’ /nm/ [nm] /nmo!ri/ [nmo !ri] ‘he lives’ /nn/ [nn] /nnE!i/ [nnE!i] ‘sign’ /nl/ [ll] /na + losir/ [llosir] ‘3sg, to follow’ /na + leti/ [lleti] ‘Alety (clan)’ /nr/ [nr] /maranrOmna/ [maranrOmna] ‘marna of Roma’ /lt/ [lt] /tultulu + enu/ [tultulwenu] ‘hat + turtle’ /ld/ [ld] /laldie/ [laldiyI] ‘the sulphur once’ /lk/ [lk] /lki !rna/ [lki!rna] ‘game of marbles’ /lm/ [lm] /mvialmi/ [mbj’almi] ‘you return’ /ln/ [ll] /vulan/ [vulla] ‘moon’ /Ela + ne/ [Elle] ‘sister, poss.’ /rt/ [rt] /kartu/ [kartu] ‘thorn’ /rd/ [rd] /rdutarlEta/ [rdutarlEta] ‘they wall a village’ /rk/ [rk] /pOrka/ [pO!rka] k.o. pole

rs/ [rs] /lirsi/ [li !rsi] ‘tornado’

/vm/ [mm] /vavi + m [vammu] ‘your pig’ /vn/ [vn] /vnutna/ [vnutna] ‘iron’

/ai + vlakar/ [aivlakar] ‘wood + crossed/vr/ [vr] /vrO:na/ [vrO:na] ‘axe’ /sm /smuaka [smwa!karn ‘his smartness’

/sn/sl/ [s/sr/ [sr] /mOsra/ [mOsra] ‘sword’

[mt] /mtaut/

/mv/ [mv] /mvuaka/ [mvwaka] ‘you (sg.) ask’ [ms] /msiai/

/

169

[vOrvOra] ‘silverish’ rm/ [rm] /rmo !ri/ [rmo !ri] ‘they live’

[nErma] ‘he is aware’ rn/ /r s

k o te v / ] o

[rl] /r lE! w ll a village’ /r fo nd a village

References:

r & Aone van Engelenhoven. 1997. Nonmoraic

tics 36.1, 65-101. spoken in Tutukei).

/rv/ [rv] /vOrvOra// /nErma/ / [rn] nani/ [rna!ni] ‘they wim’ / eŸrna/ [keŸrna] ‘Chel nia imbrica L.’ / u:ra + -ne [ruarne ‘his m untain’ /rl/ dutarlE!ta/

ta/ [rdutar ta] ‘they

ya

lE! [rlE!ta] ‘the u ’

Hume, Elizabeth, Jennifer MulleGeminates in Leti. Phonology 14.3.

Hume, Elizabeth V. 1997. Vowel Preservation in Leti. Oceanic Linguisan Engelenhoven, Aone. 1995. A description of the Leti language (asV

Ph.D. dissertation, Leiden.

170

icoalveolar Retroflexed Palatal Velar

Mara (Australian, Maran) Phonemic consonant inventory: Bilabial Interdental ApStop b d1 d ˇ Ô gNasal m n1 n ¯ ≠ NLateral l 1 l Rhotic | Glide w ® y - /d1, n1, l 1/: occur only prevocalically, only in nouns, almost all being flora-fauna terms. - /n1/ occurs only stem-initially. - /N/ can occur in onset position, i.e. Narug ‘curlew’ - No instance with d-final /m/ except wor verb-particle n‡im ‘to camp’

n ¯ ≠ N l1 l | ® y w Patterns of son/C clusters: b d 1 d ˇ Ô g m n1

b × × bn b¯ b≠ N × × × × × × ×d 1 × × × × × × × × × × × × ×d ↓ × dn d¯ d≠ dN × × ↓ × × nwˇ ↓ × ˇn ˇ¯ ˇ≠ ˇN × × × × ¯y ×˙ Ôm × Ôn Ô¯ Ô≠ ≠N × × × × × × ×g ↓ × gn g¯ g≠ × × × N × × Ny ×m mb × × mÔ mg × ×mˇ × m¯ m≠ mN × × × × × × ×n1 × × × × × × × × × × × × × × × × × × ×n nb nd nd × n ng nm n n nNar × × nw1 Ô × ↓ ↓ ≠ N × × ↓ ¯ b × Ô g m × ¯ ≠ ¯N¯ × ¯ˇ ¯ ¯ ¯ × ¯ × × × × × ¯y × ≠ ≠b × ≠ˇ ≠g ≠m × ≠¯ ≠ × × × ≠N × × × × ≠y ≠wN Nb × × Nˇ NÔ Ng Nm × × ↓ N≠ gN × × N × N Ny Nwl1 × × × × × × × × × × × × × × × × × × ×l lb ld lˇ lÔ lg lm l l¯ l≠ lN × × l × × × l® ly lw b × × ˇ Ô g m × × ≠ N × × × × y w| |b × d × |Ô |g |m × ↓ × |≠ × | × × |® |y |w| |N ® × × × × × × ®m × ® × ®≠ ®N × × × × × ®y ®wy × × × × × y × × y¯ × × × × × m yN × × × ×w × × × × × × wm × w¯ × × × w × wN × × × × Examples: /bn/ [bn] /uab + ni/ [uabni] ‘juvenile euro, purposive’

et’ /

/b¯/ [b¯] no example given /b≠/ [b≠] no example given /bN/ [N] /Nayab + Nayab/ [NayaNayab] Nayab ‘to be qui/dm/ [dm] /gad + min [gadmin] ‘skin-and-bones’

171

/

tuck’ /dw/ [nw] /waad + waad/ [waanwaad] wal ¢ad ‘(sky) to clear’

/ ] d penis! (swear word)’

/ˇn/ [ˇn] /≠awaˇ + ni/ [≠awaˇni] ‘hawk sp., purposive’

/dm/ [nm] /mad + mad [manmad] mad ‘to mix’ /dn/ [dn] /wambigud + ni/ [wambigudni] ‘plant sp., purposive’/d¯/ [d¯] no example given /d≠/ [d≠] no example given /dN/ [dN] no example given /d / [] (in CVCV roots) /a|ad + a|ad/ [a|aa|ad] a|ad ‘to fall into hole’ /d/ [n] (in the only CVC roots)

/ud + ud/ [unud] ud ‘to get s

/ˇm/ [ˇm] /Ôigiˇ + min [Ôigiˇmin ‘crooke/ˇm/ [¯m] /miˇ + miˇ/ [mi¯miˇ] miˇ ‘to level out lumps’

/ˇ¯/ [ˇ¯

/ˇN/ [ˇN] no example given /ˇ/ [] (in CVCV roots) /agaˇ + agaˇ/ [agaagaˇ] agaˇ ‘to strip off

ark) from’

/Ôn/ [Ôn] /NaˇiÔ + ni/ [NaˇiÔni] ‘hibiscus, purposive’

] no example given /ˇ≠/ [ˇ≠] no example given

(b/ˇy/ [¯y] /yaˇ + yaˇ/ [ya yaˇ] yaˇ ‘to wait’ ¯/Ôm/ [Ôm] /biÔ + min/ [biÔmin] ‘big-eyed’

/Ô¯/ [Ô¯

/

/gn/ [gn] /ÔiÔug + ni/ [ÔiÔugni] ‘hare-wallaby, purposive’

ative’

fork’ /

/mb/ [mb] /bambal/ [bambal] ‘sea-sponge’ p’

es) to break’ + o’

/m¯/ [m¯] /≠im+¯a+ganÔi/ [≠im¯aganÔi] ed out’ to camp’ ut’

/nb/ [nb] /bulanba/ [bulanba] ‘senior (wife)’

] no example given /Ô≠/ [Ô≠] no example given /ÔN/ [ÔN] no example given /ÔN/ [≠N] /NarÔ + NarÔ [Nar≠NarÔ] NarÔ ‘to carve’ /gm/ [gm] /wag + min/ [wagmin] ‘black’

//gm/ [Nm] /mug + mug [muNmug] mug ‘to forget’

/g¯/ [g¯] no example given /g≠/ [g≠] /¯a-ÔiÔug-≠u(|) / [¯aÔiÔug≠u(|)] ‘hare-wallaby,

allative/loc/gN/ [gN] /wugNan/ [wugNan] ‘myrtle’ /g/ [N] /u|g + u|g/ [u|Nu|g] u|g ‘to be aloft in tree /gy/ [Ny] /yi|ig + yi|ig [yi|iNyi|ig] yi|ig ‘to twitch’

/mbul/ [mbul] ‘to load u/mˇ/ [mˇ] /ˇum + ˇum/ [ˇumˇum] from ˇum ‘(wav/mÔ/ [mÔ] /¯a-ubam Ôani/ [¯aubamÔani] ‘from the juvenile eur/mg/ [mg] /≠im+gu+ganÔi/ [≠imguganÔI] ‘he camped out’

‘we camp/m≠/ [m≠] /≠im + ≠im/ [≠im≠im] from ≠im ‘/mN/ [mN] /≠im+Na+ganÔi/ [≠imNaganÔi] ‘I camped o

172

/nd/ [nd] /bandiyan/ [bandiyan] ‘king brown snake’ ] il (bird)’

/nÔ/ [nÔ] /binÔa/ [binÔa] ‘hair’ a] p.’

the fire’ / a|Ôaa er’ (Oblique)

/ ] i|i|i/ ue) /

forward’ a] / ue)

et stuck’

¯a lost’ u¯bil il u¯bili]‘cluster of stars’

/nd 1/ [nd1] /aland 1a|bura/ [aland 1a|bura ‘chestnut ra

/ng/ [ng] /baingama/ [baingam ‘kangaroo s/nm/ [nm] /badbanma/ [badbanma] ‘midriff’ /nn/ [n] /ˇiwan + niwi-nbuni/ [ˇiwaniwinbuni] ‘we fell into/#nn [nNan] /n+Na+n Na/ [yana|ÔaaNa] ‘wife’s moth/n¯/ [n] /¯ad-¯ad/ [¯an¯ad] (Nasalization) [¯anad] ¯ad ‘to run’ /#n¯ [nNa¯ /n+Na+¯ [≠a¯i|i|i] ‘small ones’ (Obliq/n≠/ [n≠] /n+≠iyin [n≠iyin ‘name’ ]/nN ] /mun+Na-nbuni/ [munNan ‘I was / [nN buni] leaning/#n/ [#nN /n+Na+uga [≠auga] ‘island’ (Obliq/n/ [n] /un+ud/ [unud] from ud ‘to g/#nr/ [#nNar] /n+®u|u/ [nNa®u|u] ‘section’ /nw/ [nw] /wa¯an+wa n/ [wa¯anwa¯an] from wa¯an ‘to be/¯b/ [¯b] /bu¯bilib i/ [bu¯b ib/¯ˇ / [¯ˇ ] /ba¯ˇiÔ/ [ba¯ˇ¯

iÔ] ‘kidney’ / Ô ¯/ [ Ô ¯Ôi/ ¯Ôi] ‘wattle

/

to it’

] /bu|uwa [bu|uwa sp.’ /¯g/ [¯g/ /ba¯giba/ [ba¯giba] ‘fat turtle’ /¯m/ [¯m] /ba¯mi|a [ba¯mi|a] ‘tree sp.’ /¯n/ [¯] /ma¯+niwi+niNai/ [ma¯iwiniNai] ‘we adhered/¯≠ / [¯≠ ] /Ôa¯≠ in/ [Ôa¯≠in] ‘bloodwood’ /¯N/ [¯N] /wa¯Nu/ [wa¯N/¯y/ [¯y] ait for’

|i/ /≠m/ [≠m] /gi≠mugi≠mu / [gi≠mugi≠mu] ‘sleepy cod’ /≠¯/ [≠¯] /ˇa≠+¯a+ÔuÔu≠i/ [ˇa≠¯aÔuÔu≠i] ‘we sliced it’

+ÔuÔu u≠i] it’ /≠/ [≠] /u≠+u≠/ [u≠u≠] from u≠ ‘to coil’

up’ /≠w/ [≠w] /wi≠+wiÔ/ [wi≠+wiÔ] from wiÔ ‘to suck honey’ /Nb/ [Nb] /ÔawuNbuN/ [ÔawuNbuN] ‘short’

/ aN] ’

/Ng/ [Ng] |/ ction’ ¯iN iN] ut

a≠ ‘we hit it’ /N¯/ [g¯]

u] ‘black plum’ /ya¯+yaˇ/ [ya¯yaˇ] from yaˇ ‘to w

/≠b/ [≠b] /bi≠binÔaa/ [bi≠binÔaa] ‘bat’ / /≠ˇ/ [≠ˇ] /ˇa≠+ˇa≠ [ˇa≠ˇa≠] from ˇa≠ ‘to slice’

/≠g/ [≠g] /mabi≠ga [mabi≠ga|i] ‘dugong harpooner’

/≠N/ [≠N] /ˇa≠+Na ≠i/ [ˇa≠NaÔuÔ ‘I sliced

/≠y/ [≠y] /yi|i≠+yi|iÔ/ [yi|i≠yi|iÔ] from yi|iÔ ‘to pick

/Nˇ/ [Nˇ] /miˇalaNˇalaN [miˇalaNˇal ‘darter’ /NÔ/ [NÔ] /ÔuNÔuN/ [ÔuNÔuN] ‘to suck

/balaNga [balaNga|] ‘river jun/Nm/ [Nm] /mi¯iN+mi / [mi¯iNmi¯ from mi¯iN ‘to flick o

tongue’ /N¯/ [N¯] /®aN+¯a+≠i/ [®aN¯ i]

/®aN+¯ang+a≠i/ [®ag¯anga≠i] ‘he hit me’

173

/N≠/ [N≠] /NN/ [gN] /raN+Na+≠i/ [ragNa≠i] ‘I hit him’ /N/ [N] /uN+uN/ [uNuN] from uN ‘to swell up’

/ /Ny/ [Ny] /yi|iN+yi|ig/ [yi|iNyi|ig] from yi|ig ‘to move’

’ /lˇ/ [lˇ] /ˇilˇil/ [ˇilˇil] ‘feathered stick’ /lÔ/ [lÔ] /gulÔi/ [gulÔi] ‘grinding stone’ /lg/ [lg] /bilg/ [bilg] ‘to be weak’

/lm/ [lm] ¯ ] /ln/ [l] /bal+niwi+Ôa≠i/ [baliwiÔa≠I] ‘we wrote it’ /l¯/ [l¯] i|i/ | I] /l≠/ [l≠] /≠il iluna/ [≠il iluna] ‘trevally fish’

’

on’ /ly/

[yulway]

ˇ ˇ ˇrtles’

sp.’

/ son

/baya/ [baya] ‘male’

/gu|b/ |d/ [|d] /ma|da/ [ma|da] ‘feather’ |ˇ/ [|ˇ] /gambi|+ˇi/ [gambi|ˇi] ‘my daughter’s son’ |Ô/ [|Ô] /mu|Ôi/ [mu|Ôi] ‘hand’

/wa|Ô/ [wa|Ô] ‘to get’ /|g/ [|g] /Ni|gNi|g/ [Ni|gNi|g] ‘chestnut rail (bird)’ |m/ [|m] /Na|ma/ [Na|ma] ‘armpit’ |n/ [n] (n=inflectalbe verb stem initial C.)

/wulumbu|-naÔi/ [wulumbu|naÔi]‘they saw them’ /|n/ [|](n=prefix or suffix initial C.)

/≠iN≠iN / [≠iN≠iN] ‘sharp’

/N®/ [N] /®aN+®aN [®aNaN] raN ‘to hit’

/Nw/ [Nw] /wiN+wiN/ [wiNwiN] from wiN ‘to rest’ /lb/ [lb] /balba/ [balba] ‘river’ /ld/ [ld] /guldu|u/ [guldu|u] ‘bonefish

/balgi≠/ [balgi≠] ‘salty’ /balma¯a/ [balma a ‘hat’

/wul+¯i| [wul¯i i| ‘small ones’ ≠ ≠

/lN/ [lN] /≠ilN/ [≠ilN] ‘to be soakingl giÔin/ n] / / [l] /wul+i| [wuli|giÔi ‘experts’

/l + a a ] ok®/ [l®] /®al ® l/ [®al® l from ral ‘to ho [ly] /bi|iilyi/ [bi|iilyi] ‘parrot sp.’

/lw/ [lw] /yulway/ ‘stringybark’ /b/ [b] /ubab / [ubab] ‘juvenile euro’ / / [ ] / ilNgaˇiNga/ [ˇilNgaˇiNga] ‘vine /Ô/ [Ô] /waÔa/ [waÔa] ‘dugong and tu

sp.’

/g/ [g] /bagan/ [bagan] ‘mangrove/waÔag/ [waÔag] ‘mud’

/m/ [m] /mbimu|/ [mbimu|] ‘grub sp.’ /n/ [] /ba+niwi+ini/ [baiwiini] ‘we sneaked along’ /≠/ [≠] /wu≠in/ [wu≠in] ‘genital cover’ /N/ [N] /buNu/ [buNu] ‘ashes’

i /l/ [] /waNgu+l [waNgui] ‘sister’s daughter’s (vocative)’

/y/ [y] /w/ [w] /yawar/ [yawar] ‘empty-handed’ /|b/ [|b] /gu|b+gu|b/ [gu|bgu|b] ‘stone edge’

[gu|b] ‘to cut with stone edge’ ///

//

174

u|iwibiliNanÔi] ‘we like it’ |≠/ [|≠] /mu|≠mu|≠/ [mu|≠mu|≠] ‘bubbles’

[wi|≠] ‘to graze’ |N/ |N] t u

/a|Nai / [a|Na ‘to scratch’ /a|+a|/ [a|a om a| ‘to cut up’

/|®/ |®] /wu|+ra [wu|®ambi] ‘together ual)’ |y] /wu|ya/ [wu|ya] h sp.’

/ga|wi®i/ [ga|wi i] ‘dog’ /®b/ [®b] /ma®ba®b [ma®b ®ba] ‘wood bits’

[®ˇ] /ˇama®+ˇama®/ [ˇama®ˇama®] from d ¢amar ‘to warm up’ /®Ô/ [®Ô] /ga®Ôi|/ [ga®Ôi|] ‘python sp.’

/ba®Ô/ [ba®Ô] ‘to wag tail’ g/ [ g] /ga gunÔa/ [ga®gunÔa] ‘night-heron’

[bi®g] ‘to clap boomerangs’ / / [®m] /ba® [ba®munu] ‘goanna sp.’ / ®] Ôa®+niwi Ôu≠ [Ôa®iwi ≠i] we/ / [®≠ ®≠/ [gu®≠]

/ga®≠ |imba/ [g ®≠i| a] ig c wd/ / [®N /Nab +Nabar/ [N ba®N r] om/ [®y] /Ôa®+ya+Ôu u≠i/ [Ôa®yaÔu ≠i]

®w] /Naba®+wala+ma/ [Naba®walama] ‘they got sick’ /ym/ [ym] /¯alayma|yi/ [¯alayma|yi] ‘soon’

/yN/ [yN] /mayNu/ [mayNu] ‘ochre’ /wm/ [wm] /iwmaiw+min/ [iwmaiwmin] ‘crooked, zigzagging’

I+ini/ ] me out’ ini/ ]

/w/ [w] /awaw/ [awaw] ‘other side’

References: Heath, Jeffrey. 1981. Basic Materials in Mara: Grammar, Texts and Dictionary.

Department of Linguistics, Research School of Pacific Studies, The Australian

/bu|+niwi+biliNanÔi/ [b/ /wi|≠/ / [|N] /ya|N/ [ya ‘to spli p’

i] /|/ [|] |] fr

[ mbi/ (D/|y/ [ ‘fis/|w/ [|w] ®

a/ a/®ˇ/

/® ® ®/bi®g/

®m munu/ ®n/ [ / +Ôu i/ ÔuÔu ‘ pulled it’ ®≠

] /gu imb

‘to have an itch’ ‘bi a ro ’

®N ] a® a aba fr Nabar ‘to be sick’ ®y/ Ô Ôu ‘he was dragging it this

way’ /®w/ [

/y¯/ [y¯] /ˇay+¯iwi+ma/ [ˇay¯iwima] ‘we fled’

/w¯/ [w¯] /baw+¯ [baw¯iini ‘you ca/wN/ [wN] /baw+Na+ [bawNaini ‘I came out’

National University.

175

Moroccan Arabic (Afro-Asiatic, Arabic)

lab r lottal

Phonemic consonant inventory: bilabial iodental apical palatal vela uvular pharyngeal gvl stop t t∞ k q ? vd stop b b d d g ∞ ∞vl fric. f s s S X © h ∞vd fric. z z∞ Z Â ? nasal m m∞ n lateral l l∞ flap-trill r r∞ semi-vowels

w y

Patterns of son/C clusters:

t k q / b d g m n l r w y m mn ml mr n nt nq nb nd ng nm nl~ll nr~r

r

l lm ln~nn

lr~rr

r rm rn rl w y

xamples: mn/ [mn] /mn´?/ [mn´?] ‘prevent’ ml/ [ml] /mli©/ [mli©] ‘fine, good’

/Srahom + lek/ [Srahomlek] ‘he bought them for you (sg.)’

mr ¢/ [mr ¢] /mr ¢id¢/ [mr ¢id¢] ‘sick’ nt¢/ [nt¢] /nt¢ q/ [nt¢ q] ‘pronounce’ nt/ [nt] /q´nt/ [q´nt] ‘corner’ nq/ [nq] /nqi/ [nqi] ‘clean’ nb/ [nb] /ta-n-bus/ [tanbus] ‘I kiss’ nd/ [nd] /ta-n-dir/ [tandir] ‘I do’ nd ¢/ [nd¢] /nd ¢´?/ [nd ¢´?] ‘chew’ ng/ [ng] /ta-n-gul/ [tangul] ‘I say’ nm/ [nm] /nmla/ [nmla] ‘ant’ nl/ [nl] or [ll] /n-l?´b/ [nl?´b] or [ll?´b] ‘I play’

/ban-l-i/ [banli] or [balli] ‘I thought’ /nr ¢/ [nr ¢] or [r ¢r ¢] /n-r ¢mi/ [nr ¢mi] or [r ¢r ¢mi] ‘I throw’ /lm/ [lm] /lm´?/ [lm´?] ‘shine’ /ln/ [ln] or [nn] /l-nif-i/ [lnifi] or [nnifi] ‘to my nose’

/mal-na/ [malna] or [manna] ‘what’s the matter with us’

E//

//////////

176

[lrusiya] or [rrusiya] ‘to Russia’ r ¢m/ [r ¢m] /r ¢mi/ [r ¢mi] ‘throw’

[r ¢n´b] ‘hare’ r ¢l/ / ] e got

ences

ich rd S. 1962 Short Referen Grammar of Moroccan Arabic. Georgetown iv rsity Press ashington, D. .

/lr/ [lr] or [rr] /l-rusiy-a/ //r ¢n/ [r ¢n] /r ¢n´b/ / [r ¢l] t¢ayr ¢-l-u/ [t¢ayr ¢lu ‘h fed up’

Refer : Harrell, R a . A ce

Un e : W C

177

Palataar Glottal

Mutsun (Penutian, Costanoan) Phonemic consonant inventory: Bilabial Dental/ Retroflex Alveo-palatal/ Vel

Alveolar l/Palatalized Stop p t Ê c k / Affricate ts tS Fricative s S h Nasal n ≠ m Lateral l l y Tap or Trill

r

Glides w y Patterns of son/C clusters: p t Ê c k ts s S h m ≠ l l tS n y r w y p pl pr pw py t tm tn tl tly tr tw ty Ê Ê n Êl Êr m Ê Êw Êy c cl k km kn kl kr kw ky ts tsytS tSl tSr tSw tSys sm sn sl sr sw sy S Sm n Sl Sr Sy Sh hm hl hr hw hy m m mt mc km ms mS mh mmp mtS ml mr n np~ nt nÊ nc nk~ ns nS mh nm nn

mp k ntS nw ny

N≠ l lp lt lÊ lk lts ls lS lm ll lw ly ly lym r rp rt rÊ rk rts rs rS rh rm S rw rtw wp wt wk ws wm wn wl wr ww wyy yp yt yÊ yc yk ytS ys yS ym yn yr yw yy yl

tar)’

track’ a’

’ /tw/ [tw] /hitwiy/ [hitwiy] ‘clean (it)!’

Examples: /pl/ [pl] /kapla/ [kapla] ‘embraces’ (verb) /pr/ [pr] /kiprihte/ [kiprihte] ‘twisted’ /py/ [py] /tSupyuhte/ [tSupyuhte] ‘puffed up (from wind)’ /pw/ [pw] /hapwey/ [hapwey] ‘shoo (them) away!’ /tm/ [tm] /tSi:ritmin/ [tSi:ritmin] ‘lark’ /tn/ [tn] /hutna/ [hutna] ‘pounds (in a mor/tl/ [tl] /matla/ [matla] ‘is face down, on belly’/tly/ [tly] /pitlyan/ [pitlyan] ‘gopher dirt, gopher/tr/ [tr] /hotro/ [hotro] ‘sticks finger into vagin/ty/ [ty] /hotyon/ [hotyon] ‘tied (something) up

178

Êman/ [luÊman] ‘got wet’ /Ên/ [Ên] //eÊnempiy/ [/eÊnempiy] ‘put (him) to bed!’

in/ d one’

kloSmin/ [tokloSmin] ‘one having syphilitic lumps under ear’

k/ fellows’

/kw/ [kw] /hakwana/ [hakwana] ‘goes to gather (black)

/tSm/ [tSm] //awnitSmine/ [/awnitSmine] ‘tortoiese (objective case)’

g)’

/sn/ [sn] /tSisnan/ [tSisnan] ‘elder (tree, flower)’ ’ r’

es’ alth)’

/ uns much’ /

/Sl/ [Sl] /huSlu/ [huSlu] ‘sucks’ )’

e

!’ /mp/ [mp] /SuSpumpi/ [SuSpumpi] ‘flashes light (at someone)’

/mk/ [km] /rukka + mak/ [rukkakma] ‘house, nom. thematic pl.’ ematic pl.’

/mS/ [mS] //amSi/ [/amSi] ‘so that, in order that’ min/ min] oolish one’

eads’ ne’

/Êm/ [Êm] /lu

/cl/ [cl] /kacluSm [kacluSmin] ‘thick-legge/km/ [km] //onyekma/ [/onyekma] ‘companions’ /kn/ [kn] /Êuknu/ [Êuknu] ‘signals with eye(s)’ /kl/ [kl] /to

/kr/ [kr] /hotokros/ [hotokros] ‘cocoon rattle’ /ky/ [ky] /hokyoma [hokyomak] ‘ “sarnosos”, paltry

mussels’ /tsy/ [tsy] /mutsyaSte/ [mutsyaSte] ‘wrinkled (of rump)’

/tSl/ [tSl] /wetSlohte/ [wetSlohte] ‘big-mouthed’ /tSr/ [tSr/ /potSroSte/ [potSroSte] ‘having sores’ /tSy/ [tSy] /litSye/ [litSye] ‘is standing’ /tSw/ [tSw] /SatSwen/ [SatSwen] ‘opened (somethin/sm/ [sm] /hismen/ [hismen] ‘sun’

/sl/ [sl] /haslinit/ [haslinit] ‘fear me!/sr/ [sr] /tSisreSte/ [tSisreSte] ‘in bad humo/sy/ [sy] /misya/ [misya] ‘pretty on/sw/ [sw] /yiswanin/ [yiswanin] ‘got worse (in he/Sm/ [Sm] /ÊokceSmin [ÊokceSmin] ‘runner, one who r/Sn/ [Sn] /woSnonin [woSnonin] ‘choked’

/Sr/ [Sr] /taSri/ [taSri] ‘is firm, stiff (cloth/Sy/ [Sy] /miSya/ [miSya] ‘good ones’ /hm/ [hm] /Êaskuhmin/ [Êaskuhmin] ‘red one’ /hl/ [hl] /kehleSmin/ [kehleSmin] ‘hoarse one’ /hr/ [hr] /hihronin/ [hihronin] ‘slipped’ /hy/ [hy] /Êuhyin/ [Êuhyin] ‘is getting to be late in th

morning’ /hw/ [hw] /yihwiy/ [yihwiy] ‘shake (it)

/mt/ [mt] /namti/ [namti] ‘hears’

cf. /ru:k + mak/ [ru:kmak] ‘string, nom. th/mc/ [mc] //amcaSmin/ [/amcaSmin] ‘big eater’ /ms/ [ms] /Êumsan/ [Êumsan] ‘enjoys the taste’

/mh/ [mh] /mamhaS [mamhaS ‘crazy, f/mm/ [mm] /yumme/ [yumme] ‘cheats, lies, misl/mn/ [mn] /hamnaSmin/ [hamnaSmin] ‘crazy, foolish o/mr/ [mr] /homron/ [homron] ‘big strainer’

179

/ kes out’

/ /

[Nk /tursin + ka/ [tursinka] or [tursiNka] ‘I am (getting) cold’

/ ’

nh/ [nh] /hinhan/ [hinhan] ‘how much?’ [nm] /minmuy/ [minmuy] ‘below, under’

nn/ [nn] /mennen/ [mennen] ‘grandmother’ w] /hinwa/ [hinwa] ‘when?’

/lt/ /tSaltaSmin/ [tSaltaSmin] ‘one who urinates much’ lÊ/ [lÊ] /yolÊo/ [yolÊo] ‘is big-eared’ lk/ [lk] /halkaninak/ [halkaninak] ‘he became blind’ lts/ [lts] /peltsey/ [peltsey] ‘close your eyes!’ ls/ [ls] /halsaSmin/ [halsaSmin] ‘liar’ lS/ [lS] /helSo/ [helSo] ‘is soft’ lm/ [lm] /helmo/ [helmo] ‘is soft’ ll/ [ll] /Êallan/ [Êallan] ‘it’s hot’ ly/ [ly] /kulyan/ [kulyan] ‘blackbird’ lw/ [lw] /lilwinin/ [lilwinin] ‘amused (one)self’ lym/ [lym] /pelymo/ [pelymo] ‘is soft’ rp/ [rp] /tSirpiy/ [tSirpiy] ‘shout!’ rt/ [rt] /hu:purte/ [hu:purte] ‘having boils’ rk/ [rk] /tSorkon/ [tSorkon] ‘gets dry’ rts/ [rts] /murtsumak/ [murtsumak] ‘black ones’ rtS/ [rtS] /hartSuhte/ [hartSuhte] ‘is lacking a piece’ rs/ [rs] /yerseksi/ [yerseksi] ‘is torn’ rS/ [rS] /murSun/ [murSun] ‘has a molar toothache’ rh/ [rh] /marhaSmin/ [marhaSmin] ‘leafy one’ rm/ [rm] /mukurma/ [mukurma] ‘woman’ rw/ [rw] /parwes/ [parwes] ‘five’ yp/ [yp] /koyponin/ [koyponin] ‘got scared’ yt/ [yt] /loytohte/ [loytohte] ‘loose’ yÊ/ [yÊ] /layÊaSmin/ [layÊaSmin] ‘long one’ yc/ [yc] /hayceSmin/ [hayceSmin] ‘big-mouthed one’ yk/ [yk] /poykonin/ [poykonin] ‘suddenly got scared’ ys/ [ys] /haysa/ [haysa] ‘they’ yS/ [yS] //oySo/ [/oySo] ‘again’ ym/ [ym] /huyman/ [huyman] ‘gets weak’ yn/ [yn] //ammayni/ [/ammayni] ‘comes to eat’ yl/ [yl] /tsaylak/ [tsaylak] ‘he’s (lying) face up’

/np/ [np] or [mp] //itStSonpi [/itStSonpi] or [/itStSompi] ‘ta/nt/ [nt] /hinte/ [hinte] ‘does’ /nÊ/ [nÊ] /ranÊaSmin [ranÊaSmin] ‘long-necked one’ /nc/ [nc] /hinceSmin [hinceSmin] ‘wanderer, rover’ /nk/ [nk] or ]

/ns/ [ns] /hinseksi [hinseksi] ‘keeps quiet/nS/ [nS] /me:nSa/ [me:nSa] ‘you alone’ //nm/ //nw/ [n/lp/ [lp] /tsayyalpu/ [tsayyalpu] ‘puts oneself face up’

[lt] //////////////////////////////

180

[weyrohte] ‘big-mouthed’ yw/ [yw] /Êaywen/ [Êaywen] ‘mush’

sottowtak/ [sottowtak] ‘at the fire’ wk/ [wk] kun/ [law ‘swa ithout

[w sen iwsen] [wm] /nossowmak/ [nossowmak] ‘spirits, [wl] //awli/ sour’ [w / ‘sits, lives’ [wy] en ‘antelope’

:

/yr/ [yr] /weyrohte/ //wt/ [wt] // /law kun] llows w chewing’ /ws/ s] /hiw / [h ‘likes, wants’ /wm/ souls’ /wl/ [/awli] ‘is/wr/ r] /Êawra

/tiwy [Êawra]

iwyen] /wy/ / [t

otesN R fere cese n : Okrand, Marc. 1977. Mutsun Grammar. Doctoral dissertation. University of California,

Berkeley.

181

aryngeal

Oromo (Afro-Asiatic, Cushitic) Consonant inventory: Labial Alveolar Palatal Velar LStop p’ b t t’ d Î k k’ g / Fricative f s S h Affricate tS tS’ dZ Nasal m ≠ n Liquid l, r Glide w y

e s o : Patt rn f son/C clusters p’ b t t’ d k k’ g f s Î S t/ S tS’ dZ

m mb mt mt’ ms m/ n nb nt nt’ nd nk nk’ ng n/ nf ns nS nÎ ntS’ ndZ≠ l lp’ lb lt’ ld lk lg lf ls lt l/ ltS r rp’ rb rt’ rk rg r/ rf rs rtrt S’ rdZw y

m n l r w ≠ y

t’ nÎ k’ k’n g gn s sr m mm mn ml n nl ≠ l lm ll r rm rn rl rr w y

Examples:

n +/t’ / [nÎ] /fit’ na [finÎa] ‘we finish’

/

/e’

] uu/ Suu] ‘happiness’

n] /ga!mna!/ [ga!mna!] ‘wise person’

/k’n/ [k’n] /do ⁄ok’na!/ [do⁄ok’na!] ‘stingy’ /gn/ [gn] /Îa!gna!/ [Îa!gna!] ‘body’

s’ /sr/ [sr] /mi Ÿsra!atStSo!o [mi Ÿsra!atStSo!o] ‘good new ] beetl/mb/ [mb] /buŸmbi !i/ [bu Ÿmbi !i ‘dung

/mt/ [mt] /gaŸmta!a/ [gaŸmta!a] ‘unity’ /mt’/ [mt’] /laŸmt’ii / [laŸmt’ii] ‘leprosy’

/m// [m/] /Îi Ÿm/u!u/ [Îi Ÿm/u!u] ‘leak’ / !/] ‘hunter’ /ms/ [ms] /aŸda!msa!/ [aŸda!msa

mm/ [mm /gammatS [gammat//mn/ [m/ml/ [ml] /yimlaa/ [yimlaa](from Arabic) ‘wholesale’ /nb/ [nb] /baŸlaŸnba!lle!e/ [baŸlaŸnba!lle!e] ‘wood frame’

182

/nt/ [nt] /gaŸtaŸnta!raŸ/ [gaŸtaŸnta!raŸ] ‘I walk zig-zag’

/nd/ oat’

/nk/ /nk’/ ’u!u/ [haŸnk’a!ak’u!u] ‘egg’

ng/ [ng] /a!ango!/ [a!ango!] ‘force’ n// [n/] /t’aŸn/a!tStSa!/ [t’aŸn/a!tStSa!] ‘sinews’ nf/ [nf] /aŸlaŸnfa!ÎÎaŸ/ [aŸlaŸnfa!ÎÎaŸ] ‘I chew’ ns/ [ns] /baŸalaŸnso!ofi !i/ [baŸalaŸnso!ofi !i] ‘kind of tree’

/nS/ [nS] /i ŸnSi Ÿla!ala!/ [i ŸnSi Ÿla!ala!] ‘dill’ ntS’/ [ntS’] /a!antS’a!/ [a!antS’a!] ‘resin ndZ/ [ndZ] /ti ŸndZi !i/ [ti ŸndZi !i] ‘border’ nl/ [nl] /cinaa + lafa/ [cinlafa] ‘hemisphere’ lp’/ [lp’] /saŸlp’a!a/ [saŸlp’a!a] ‘light’ lb/ [lb] /aŸlbe!e/ [aŸlbe!e] ‘knife’ lt/ [lt] /aŸla!ltu!u/ [aŸla!ltu!u] ‘kind of tree’ lt’/ [lt’] /goŸlt’a!a/ [goŸlt’a!a] ‘useless meat’ ld/ [ld] /baŸaldi !i/ [baŸaldi !i] ‘pail’ lk/ [lk] /haŸlka!n/ [haŸlka!n] ‘night’

[lg] /aŸlge!e/ [aŸlge!e] ‘corn silk’ l// [l/] /baŸl/o!o/ [baŸl/o!o] ‘juncture’ lf/ [lf] /u!lfa!/ [u!lfa!] ‘pregnant’ ls/ [ls] /fi Ÿlsa!ta!a/ [fi Ÿlsa!ta!a] ‘preparation’ ltS/ [ltS] /ba⁄altSaŸ/ [ba⁄altSaŸ] ‘I hand over sth.’ lm/ [lm] /e!lma!/ [e!lma!] ‘milk’ ll/ [ll] /kaŸlle!e/ [kaŸlle!e] ‘child’s garment’ rp’/ [rp’] /hi Ÿrp’o!o/ [hi Ÿrp’o!o] ‘one-fourth’ rb/ [rb] /a!rba!/ [a!rba!] ‘elephant’ rt/ [rt] /aŸma!rti !i/ [aŸma!rti !i] ‘ring’ rt’/ [rt’] /k’uŸrt’u!mmii/ [k’uŸrt’u!mmii] ‘fish’ rk/ [rk] /ha!rka!/ [ha!rka!] or [a !rka!] ‘hand’ rg/ [rg] /a!rgaŸ/ [a!rgaŸ] ‘I see’ r// [r/] /bi Ÿr/a!a/ [bi Ÿr/a!a] ‘season, spring’ rf/ [rf] /aŸrfa!asa!/ [aŸrfa!asa!] ‘season’ rs/ [rs] /faŸrso!o/ [faŸrso!o] ‘Ethiopian beer’ rtS’/ [rtS’] /baŸrtS’u!ma!/ [baŸrtS’u!ma!] ‘chair’ rdZ/ [rdZ] /aŸrdZa!a/ [aŸrdZa!a] ‘kind’ rm/ [rm] /birma !du!u/ [birma !du!u] ‘free person’ rn/ [rn] /murna/ [murna] ‘section, unit’ rl/ [rl] /barlamee/ [barlamee] ‘biennial’ rr/ [rr] /haŸrre!e/ [haŸrre!e] ‘donkey’

eferences: itima, Tamene. 2000. A Dictionary of Oromo Technical Terms: Oromo-English.

/nt’/ [nt’] /fi Ÿrint’i !it’a!a/ [fi Ÿrint’i!it’a!a] ‘restive’ [nd] /haŸnda⁄k’ii/ [haŸnda⁄k’ii] ‘rainc

/nÎ/ [nÎ] /haŸnÎu!ura!a/ [haŸnÎu!ura!a] ‘umbilical cord’ [nk] /aŸkaŸnkaŸre!essa⁄/ [aŸkaŸnkaŸre!essa!] ‘leech’ [nk’] /haŸnk’a!ak

////

//////////lg/ ///////////////////// RB

183

o·ln. loret-Romanyach, Maria-Rosa. 1988. Gemination and Vowel Length in Oromo

issertation. Indiana University. h Nominal Plural in the Cushitic Languages.

Vero·ffentlichungen der Institute für Afrikanistik und Agyptologie der iversität ien.

Ru·diger Ko·ppe Verlag. KL

Morphophonology. PhD. Daborski, Andrzej e MorpZ . 1986. Th ology of

Un Wien. W

184

ila Velar

Ponapean (Micronesian) Phonemic consonant inventory: B bial Dental Alveolar Stop pp tw k Fricative s Affricate t°S Nasal m mw n N Liquid l, r - /t°S/ etroflexed nal affric: r coro ate - /r/: alveolar tap or alveolar trill

p pw t k s t °S m mw n N l r Patterns of son/C clusters:

p mm mwmw pVn pw t t(V)l k k(V)m NN kl s )r s(V)l s(Vt °S )r t °S(V)n t °S(V)l t°S(Vm ↓ ↓ °S mt m wmw

mw n mp mwpw ↓ Nk ↓ ↓ mm mwmw rr NN llN ↓ N(V)l l nd l(V)k ↓ nt °S lVn l(V) N r r(V)p ↓ ns nt °S nn ll

+ / of yam’ /pm / [m m ] /kehp + m ot°S/ wot/pwmw/ [mwmw] wo ut of breath’ /pn/ [pVn] /tihp + niki/ [tipaniki] ‘sin’ + ‘to possess’

erve as female servant’

given /kN/ [NN ] /e saik Nget°S/ [esaiNNet°S] ‘He’s not yet out of

breath’

et°Sine’ k’

t crab’ yant’

#mp/ [(V)mp] /mpe/ [mpE] or [impE] ‘beside it’

Examples: /pm/ [mm] /kehp + meir kelik [kehmmeirkelik] ‘variety

w w w w [kehmwm °S] ‘variety of yam’ /mwopw + m pw/ [mwomwmwop] ‘o

/tl/ [t(V)l] /lit + lituwih/ [litilituwih] lituwih ‘s

/km/ [k(V)m] /pwihk + men/ [pwihk(i)men] no gloss

/kl/ [kl] p/ /ak + lapala [aklapalap] ‘self-assertive’+ l s/ [lusulus ] lus ‘jump’ /sl/ [s(V)l] /lus u

/sr/ [s(V)r] /res + rese/ [resirese] rese ‘saw’ ‘sell’ /t°Sn/ [t°S(V)n] /net°Snet°S/ [n t°S] net°S

/t°Sl/ [t°S(V)l] /let°S + let°S/ [let°Selet t°S] let°S ‘flickt°Soro ar/t°Sr/ [t°S(V)r] /rot°S + rot°S/ [ro t°S] ro t°S ‘d

p onu/mp/ [m ] /emp/ [emp] ‘coc /mpei/ [mpei] ‘bou

/

185

wpw] /lim + pwoat°S / [limwpwoat°S] ‘five (oblong things)’

/mt°S / m + t°S aman/ [t°Samat°Saman] t°Saman ‘to remember’

/np/ [nampar] ‘trade wind season’ npw/ [mwpw] /nanpwungara/ [namwpwuNara] ‘between them’ nt/ [nt] /tuntune/ [tuntune] tune ‘tie gogether’ #nt/ [(V)nt] /nta/ [nta] ~ [inta] ‘to say’

/nk/ [Nk ] /nankep/ [naNkEp] ‘inlet’ ns/ [ns] /kens/ [kens] ‘yaws’

/sinsinom/ [sinsinom] sinom ‘sink in’ #ns/ [(V)ns] /nsen/ [nsEn] or [insEn] ‘will’ nt°S/ [nt°S] /nt°SiN/ [nt°SiN] ‘to write’

/t°Sent°Senek/ [t°Sent°Senek] t°Senek ‘hung up’ #nt°S/ [(V)nt°S] /nt°Sa/ [nt°Sa] ~ [int°Sa] ‘blood’ nm/ [mm] /nanmadau/ [nammadaw] ‘ocean, beyond the

reef’ /nmw/ [mwmw] /nanmwoalehdi/ [namwmwoalEhdi] ‘to rest’ /nN/ [NN ] /pahn NeteNete/ [pahNNEtENEtE] ‘roof of the mouth’ /nl/ [ll] /nanleN/ [nallEN] ‘heaven’

nr/ [rr] /nanras/ [narras] ‘ground level of a feasthouse’

/Nk/ [Nk] /Nked/ [Nked] ‘to roof’ /kaNkaN/ [kaNkaN] kaN ‘eat’ #Nk/ [(V)Nk] /Nkapwan/ [Nkapwan] or [iNkapwan] ‘a while ago’ Nl/ [N(V)l] /loaN + loaNe/ [loaN(i)loaNe] loaNe ‘to pass across’ lt/ [nd] /til + til/ [tintil] til ‘penetrate’ lk/ [l(V)k] /kil + kiles/ [kil(i)kiles] no gloss given

/ls/ [ns] in a reduplicated word /sel + sel/ [sensel] sel ‘tied’

ls/ [lVs] /mwesel + saN/ [mweselisaN] ‘leave from’ lt°S/ [nt°S] /t°Sal + t°Sal/ [t°Sant°Sal] t°Sal ‘click, tsk’ ln/ [lVn] /kihl + niki/ [kiliniki] ‘skin’ + ‘to possess’ lN/ [l(V)N] /Nal + Nalis/ [Nal(i)Nalis] no gloss given

[r(V)p] /pir + pirap/ [pir(i)pirap] no gloss given /rt/ [nt] in a reduplicated word

/tar + tar/ [tanter] tar ‘strike, of a fish’ rt/ [rVt] in a word

/taur + ti/ [tauriti] ‘climb downward’ rs/ [ns] /sar + sar/ [sansar] sar ‘fade’

/rt°S/ [nt°S] /t°Sir + t°Sir/ [t°Sint°Sir] t°Sir ‘narrowing’ rn/ [nn] /nur + nur/ [nunnur] nur ‘contract’ rl/ [ll] /lir + lirohro/ [lillirohro] ‘protective’

/mpw/ [m/#mpw/ [(V)mwpw] /mpwer/ [mwpwEr] or [umwpwEr] ‘twin’

[mVt°S] /t°S a/mmw/ [mwmw] /lim + mwut°S/ [limwmwut°S] ‘five (piles)’

[mp] /nanpar/ ///

/ //

//

/

////

/////rp/

/

/

//

186

References: oodman, Beverly D. 1995. Features in Ponapean Phonology. Doctoral dissertation.

y. Rehg, Kenneth L. & Damian G. Sohl. 1981. Ponapean Reference Grammar. University

Press of Hawaii: Honolulu.

GCornell Universit

187

Labial Alveolar Palatal Velar geal Glottal

Rendille (Afro-Asiatic, Cushitic) Consonant inventory:

PharynStop b t d g ˇ k Fricative f s x © h Affrica e tÇ dÛ t Nasal m ≠ N n Approximant w l, y ® Patterns of son/C clusters: b t d k g f s x © tÇ m n ≠ N w l ˇ ® y

b ®b t nn d dn ®d ˇ nn k g g ®g Vn f ®f s sm ↓ : x xm xn © h tÇ dÛ m nt m© ®m n nd nˇ nn ns ≠ N w l lb ld lˇ lk l lm ll ss ↓ lx ©® ® t ®g ® ®tÇ m ~®® b ® ®ˇ ®s ®n ®l f ®y yn Examples:

e/ ’ /dn/ [dn] /©a!adne/ [©a!adne] ‘we shaved’ d®/ [®d] /u!du® + e/ [u!®de] ‘I slept’ ˇn/ [nn] /©i ! + n + e/ [©i !inne] ‘we tied’ gn/ [gin] /e!®eg + n + e/ [e!®gine] ‘we sent’ g®/ [®g] /a!ga® + e/ [a!®ge] ‘I saw’ f®/ [®f] /ˇafa!® + o/ [ˇa®fo!] ‘cloths’ sm/ [sm] /gi !sme/ [gi !sme] ‘share’

/b®/ [®b] /aba!® + o/ [a®bo !] ‘mother’ /tn/ [nn] /©a!t + n + [©a!nne] ‘we stole

//////

188

n] /wa!®as + n + e/ [wo!®sine] ‘we mixed’

/xm/ ’ xn/ [ta!xna$n] ‘push, verbal noun’ mt/ [nt] /tu!m+t+e/ [tu!nte] ‘you ground’ m©/ [m©] /u!m©e/ [u!m©e] ‘begin, I/he’ m®/ [®m] /©a!ma®+e/ [©a!®me] ‘I shivered’ nd/ [nd] /manda:n/ [manda:n] ‘male twin’ nˇ/ [nˇ] /ginna !anˇa/ [ginna!anˇa] ‘sing! (pl.)’ ns/ [ns] /amansa! e/ [amansa! e] ‘I yawned (perfect)’ nn/ [nn] /xanna ! / [xanna ! ] ‘be ill’ lb/ [lb] /geleb+o!/ [gelbo] ‘evenings’ lt/ [lt] /galta!am/ [galta!am] ‘big girl’ lt/ [ss] /ˇe!l+t+a/ [ˇe!ssa] ‘you will give birth’ ld/ [ld] /ilda!wa/ [ilda!wa] ‘medicine’ lˇ/ [lˇ] /fi !lˇ/ [fi !lˇ] ‘comb (own hair)’

/©alalˇu!uˇe/ [©alalˇu!uˇe] ‘I chewd (perfect)’ lk/ [lk] /kulku!lˇ/ [kulku!lˇ] ‘decorate self’ ls/ [ss] lx/ [lx] /bo !lxe/ [bo !lxe] ‘feel pain, I/ he’ l©/ [l©] /yil©o!/ [yil©o!] ‘charcoals’ lm/ [lm] /adalme/ [adalme] ‘be ignorant, I/he’

/olm+o!/ [olmo!] ‘tears’ ln/ [ll] /ˇe!l+n+e/ [ˇe!lle] ‘we gave birth’ ®b/ [®b] /ga®bo !/ [ga®bo !] ‘shoulder blades’ ®t/ [®t] /is+le!+i !®ten/ [isle!i !®ten] ‘they went together’ ®ˇ/ [®ˇ] /fi !®ˇe/ [fi !®ˇe] ‘escape, I/he’ ®g/ [®g] /ˇa!®ge/ [ˇa!®ge] ‘be full, I/he’ ®f/ [®f] /yi!®fe/ [yi !®fe] ‘plait, I/he’ ®s/ [®s] /wo!®sa/ [wo!®sa] ‘mix! (pl.)’ ®tÇ/ [®tÇ] /ka!®tÇa/ [ka!®tÇa] ‘cook! (pl.)’ ®m/ [®m] /ka®am+o/ [ka®mo] ‘calabash (pl.)’ ®n/ [®n] or [®®] /fu !®+n+e/ [fu !®ne] or [fu !®®e] ‘we opened’ ®l/ [®l] /bi !®lab/ [bi !®lab] ‘sword’ yn/ [yn] /ka!yne/ [ka!yne] ‘we tried’

/sn/ [si/sn/ [sn] /na!sne/ [na!sne] ‘we rested’ /sl/ […] /a!l+is+le!+n+ita©/ [a!li…e!nita©] ‘we became friends

(together)’ [ta!xme] ‘push,verbal noun [xm] /ta!xme/

[xn] /ta!xna$n/ ///////////// ///// ////////////

189

im, Ronald J. 1981. Morphophonemics of the Verb in Rendille. Afroasiatic Linguistics.

o ural in shitic L . Vero·ffentlichungen der Institute für Afrikanistik und Agyptologie der

rs ien

References: S Vol. 8. Issue 1. 1-33.

aborski, Andr . The Z zej. 1986 Morphology of N minal Pl the Cu anguages

Unive

ität Wien. W .

190

Phonemic consonant inventory: Bilabial Dental/Alveolar Palatal Velar Glottal

Selayarese (Austronesian, South Sulawesi)

Stop p b t d k g / ÔPrenasalized Stop

mb nd ≠Ô Ng

Fricative s h Nasal m n ≠ N Liquid l, ® Patterns of son/C clusters:

p b t d Ô k g / s h m n ≠ N l ® p b t d Ô k g / /m /n /≠ /l /® s h m mp mb mm n nt nd ns nn n®≠ ≠Ô N Ng Nh NN N mp mb d ≠Ô Nk Ng nt n ns Nh mm ≠≠ ll n®l ® ®®

Examples: //m/ [/m] //ma!: u/ [/ma!:® ®u] ‘share a husband’

‡//

// /l/ [/l] /lambe/la!mbe®e/ [lambe/la!mbe®e] ‘somewhat long’

/®a!mmasa] ‘somewhat dirty’ !:lo] ‘ten’

b] /mba!≠na‡N/ [mba!≠na‡N] ‘stretch’

® /da!mme‡// [da!mme‡/] ‘female (animal)’

//a/ma!:e [/a/ma!:e‡/] ‘breathe’ !kâ//n/ [/n] //ni !kka/ [/ni ka] ‘marry’ (intr.)

/so!/na‡/ [so!/na‡] ‘dream’ //≠/ [/≠] /ta/≠u‡!/≠u [ta/≠u‡!/≠u/] ‘stumble’ ///®/ [/®] /amma/®a!mmasa/ [amma/mp/ [mp] /sampu!:lo/ [sampu/mb/ [m /ba!m:baN/ [ba!m:baN] ‘sick, feverish’ /mm/ [mma!: i] ‘stop’ (intr.) [mm] /mma!:®i/

191

[binto !:eN] ‘star’ nd/ [nd] /nda!:tala/ [nda!:tala] ‘pursue’

//o!n:daN/ [/o!n:daN] ‘invite’ /ns/ [ns] /ansu!:lu [ su /] ‘get out’ (intr.) /nn/ [nn] /nno!:/no ‡so/ [nno !:/no ‡so] ‘shake liquid’ (intr.) /ba a‡N/ [ba!nna ] ‘thread’ / / [n®] /n® / [n i !:o] ‘hold’ /raÔ ®a!: N/ [r Ôin® ÔiN] ‘rather ilig / / [≠Ô] /≠Ô ‡/ [≠Ôa!:m ‘w rk’ ntr /Ôa!≠ ÔaN/ [Ô !≠:Ôa ] ‘see’

/ [Ng] /Ng ®a/ [N !:® ‘shout, cheer on’ //a !:o [/ Nge N] ‘sway or shake’ (tr.) /Nh/ [Nh] /Nh [Nha!:u] ‘kiss’ (intr.) //a !kkuN/ [/ Nhu âkuN ‘p nis

/ ma NaN [maNN ] d/Np/ p [pekampeka o ct

/ b/ b b a am N [bambambamba or ho

cf. [bambaN] ‘hot’ [nt] /tun®uNtun®uN/ [tun®untun®uN] ‘hit lightly’

] aNN / a ]

’

a/ lo] [lla

n rd-

n e & H . 198 e ic s, 15.

/nt/ [nt] /binto!:eN/ /

/ // /an !:lu

!nn ‡Nn® i !:o ®

in Ôi a a!: d ent’≠Ô a!:ma a‡] o (i .)

: a N/Ng o!: go a]

Nge N/ a !:oa!:u/ Nhu a !k ] u

rh (intr.)

’ ’

/ NN [NN m

//pekaNpekaN/

N / aN ‘ti e [ ] N] ‘hook-like bje ’

cf. [pekaN] ‘hook’ N [m ] / amb Nb ba / N] ‘s t of t’

/Nt/ cf. [tunruN] ‘hit’ /Nd/ [nd] no example given/NÔ/ [≠Ô] no example given/Nk/ [Nk] no example given/Ng/ [Ng] no example given/Ns/ [ns] no example given /Nh/ [Nh] no example given/Nm/ [mm /maNNaNm aN [maNN mmaNNaN ‘sort of tired’ cf. [maNNaN] ‘tired’ /N≠/ [≠≠] no example given/Nl/ [ll] /luNaNluNaN/ [luNaNluNaN] ‘small pillow cf. [luNaN] ‘pillow’ /N®/ [n®] /annaN ®up [annan®upa] ‘6 kinds’ /ll/ [ll] /ba!llo/ [ba!l ‘beautiful’ /lla!:®i/ !:®i] ‘run’ ) /rr/ [rr] /bo !®®o/ [bo !®®o] ‘conceited’

(intr.

Notes:

only [N ly. - Amo g nasals, ] occurs wo final References:

, M 5. T ceanMithu ariann asan Basri he Phonology of S layarese. OLinguistic 210-255.

192

d q ≠ l r

Sidamo (Afro-Asiatic, Cushitic)

Patterns of son/C clusters: b t t’ k g tS dZ f s S m n

b mb t nt t’ d nd k Nk g Ng q tS dZ f s ns S m mb nt mm mm n nt n ns nS nn t’ nd ≠ ≠≠ l lb lt lt’ ld lk lg lq ltS ldZ lf ls lS lm ll ll r rb rt rt’ rk rt rtS’ rf rs rS rm rr rd rg rq S rr

Examples:

n ] / [/b / [mb /hab + nemm ham !mmo] ‘we forget’ n m / nto eaten’

o beo !omi/ d’

/ k’

mb/ [mb] /afimbe:limma/ [afimbe:limma] ‘ignorance’ t] /um-tino/ [untino] ‘she has dug’

/mn/ [mm] /um-noommo/ [ummoommo] ‘we have dug’ /nt/ -to/ [dZanto] ‘let her hit’

n’

fter an animal

dy a gift’

nn/ [nn] /adde:nna/ [adde:nna] ‘incisor (tooth)’ ≠≠/ [≠≠] /adi≠≠e/ [adi≠≠e] ‘thoughtfulness’ lb/ [lb] /albillitte/ [albillitte] ‘in the future’ lt/ [lt] /ofol-tu/ [ofol-tu] ‘she sat’ lt’/ [lt’] /bilt’a/ [bilt’a] ‘to be proud’ ld/ [ld] /ba:lde/ [ba:lde] ‘pail’ lk/ [lk] /balka ya:/ [balka ya:] ‘to break open’ lg/ [lg] /alga/ [alga] ‘throne’

/tn/ [nt] /it + oom [i mo] ‘we have+ /dn/ [nd] /gud nonn [gundo!nni] ‘they finishe

/kn/ [Nk] /duk + nanni/ [duNka!nni] ‘they carry’ /gn/ [Ng] /ag + nummo [aNgummo] ‘we dran/sn/ [ns] /has + nemmo/ [hanse!mmo] ‘we look for’ //mt/ [n/mm/ [mm] /yuummo/ [yuummo] ‘say, 1st sg.’

[nt] /dZan/nt’/ [nt’] /bint’e/ [bint’e] ‘top of a mountai/nd/ [nd] /bande:la/ [bande:la] ‘to roam’ /ns/ [ns] /adansa/ [adansa] ‘to go a

to kill it’ /nS/ [nS] /afanSa / [afanSa] ‘to give somebo

////////

193

[ba:lqe] ‘small insects living on the skin of fowls’

[golfa] ‘vision, divination’ ls/ [ls] agulsis a] o hinder’

[lS] Sa/ [adu:lSS] /ba:qult [ba:qultSo] ‘one pumpkin’ Z] a [goldZa] ‘wild boar’

/lm/ [lm] [galma] h] e!mmo/ [fulle!mmo]

[akkalle] ‘old age’ /arba/ ‘Sidamo astrology’

[rt] ti / [murtini] ‘you (pl.) cut’ [rt’] e:/ [birt’e:] ‘the top, outside’

/rd/ [rd] /birdi:ffata/ [birdi:ffata] ‘to stretch to reach a high place’

[rk] /harkum [harkumma:wa] ‘to be syphilitic’ [arge] ‘an object that has been

rrowed/ [rq [birqe] ‘ rare n l in

[rf ar :sa [a a: ] ‘red’ / [rs [a r sa ‘to cause to leave’ / S ad wa a/ [a aw rS ‘war-cry, shout of boasting’ S/ So/ [g :rt ‘an old person’ S’/ [rt S’ima/ [b rtS m ] ‘ rcum isi ’ / [rm] ba i / [barmille] ‘ nk f r w er

[rr !nni/ [marro!nni] ‘they went’ [rr ad rre [a ur ] ‘cat’

fer nces: Gasparini, Armido. 1983. Sidamo-English Dictionary. Bologna, Italy: E.M.I. Hudson, Grover. 1975. Suppletion in the Representation of Alternations. PhD

dissertation. UCLA. d , Gro 1 95 Phono gy of Eth op n angu es. In oh ld m ( ,

Handbook f n lo ic Theo . Oxf rd Bl k e 2 .ty: A Structural Account of Sonority

Relationships. Phonology 9. 61-99. Vennemann, Theo. 1988. Preference Laws for Syllable Structure. Berlin: Mouton de

Gruyter.

/lq/ [lq] /ba:lqe/

/lf/ [lf] /golfa/ / / a/ [agulsis

a] ‘incessant ‘t

/lS/ /adu:l prayer’ /ltS/ [lt

[ldSo/

/ldZ/ /goldZ/galma/

/ ‘a big‘we go out’

ouse’ /ln/ [ll /ful-n

/akkalle/ /ll/ [ll] [rb] /rb/ [arba]

/rt/ /mur-/birt’

ni/rt’/

rk/

ma:wa / /

/rg/ [rg] /arge/ bo ’ /rq ] /birqe/ a , u usua th g’ /rf/ ] / fa / rf sa/rs ] /agursisa/ gu si ] /rS [r

[rt] / a rS d a a]

/rt S] /ge:rt e So] /rt S’] /bart a ’i a ci c on/rm / rm lle ta o at ’ /rn/ ] /mar-no /rr/ ] / u / d re Re e

Hu son ver. 9 . lo i ia L ag J n Go s ith ed.) o Pho o g al ry o : ac w ll. 78 -797

Rice, Keren. 1992. On Deriving Sonori

194

Somali (Afro-Asia tic)

em onantStop Af Liquid

tic, Cushi Phon ic cons inventory: Fricative fricate Nasal Glide Bilabi b f m al Labio-dental Dental t1 d1 Alveolar s l, r n Post-alveolar ˇ S Palato-alveolar dZPalatal y Velar g k w Glotta h l / Uvular q X Pharyngeal © ? Patterns of son/C clusters: b t k g q f s S h m y w 1 d 1 ˇ / X © ? dZ n l r

b bn br t 1 d 1 d 1n d 1l ˇ ˇm k g gn gl / q ql f s sl S h X © ? ? ?l m

dZ m mb md 1 mm ml mr n nt 1 nd 1 k nd nn nl n n ns r Z nml lb ↓ ld 1 k ↓ ll ly l lm r rd 1 k rg rn rb rf rr ry r rs rm rwy w Examples: /bn/ [bn] i !m ©iibni !mo] ‘being a friend,

’

/d1n/ [d1n] [lad 1nahay] ‘I feel fine’

/saa©iib + n o/ [saa friendship

/br/ [br] /abri:q/ [abri:q] ‘tea-pot’ /la d 1nahay/

cf. lad1a!n ‘good’ /ˇm/ [ˇm] /he¢ m-o/ [he¢ mo] ‘intestines’

195

n ’ ay/

/gl/ [gl] /nuglaa / [nuglaa] not hardy’ o] oman’

/?l/ [?l] /dZe?laa/ [dZe?laa] ng’ ]

] /ml/ [ml] /kalumle/ [kalumle] ’

mr/ [mr] /nimra!d1/ [nimra!d1] ‘numeral’ /nt1/ [nt1] /ma!ant1a/ [ma!ant1a] ‘today’ /nd/ [nd] /mind1i !/ [mind1i !] ‘knife’

k/ [ni !nkii] ‘the man’

nd’

/nn/ ‘houses’ nl/ [nl] /uurkumma!anle/ [uurkumma!anle]‘clairvoyant’ nr/ [rr] /bar + nay/ [barray] ‘(we) taught’ lb/ [lb] /dZilbo/ [dZilbo] ‘knees’ cf. dZi !lib ‘knee’ lt1/ [lt1] /galt1ibba!©/ [galt1ibba!©] ‘become assimilated’ lt1/ [S] (in feminine noun inflection and verb conjugation)

/me!el + t1a/ [me!eSa] ‘the place’ /qo!sol + t1ay/ [qosoSay] ‘(she) laughed’

ld1/ [ld1] / ba:ld 1 + È/ [ba:ld1È] ‘m. bucket’ lk/ [lk] /ilk + o/ [ilko] ‘teeth’

cf. [ilik] ‘tooth’ lm/ [lm] /maalmo/ [maalmo] ‘days’

/ln/ [ll] /d1il + nay/ [d1illay] ‘(we) killed’ ln/ [lm] /kulan + ay/ [kulmay] ‘(he) encountered’ ll/ [ll] /kelli!/ [kelli!] ‘kidney’ ly/ [ly] /wil + ya:l/ [wilya:l] ‘m. young men’ rb/ [rb] /irba !d1/ [irba!d1] ‘needle’

/garb + o/ [garbo] ‘shoulders’ cf. [garab] ‘shoulder’

rt1/ [rt1] /hart1È/ [hart1È] ‘m. man’ rd 1/ [rd 1] /fard 1 + o/ [fard 1o] ‘horses’

cf. [faras] ‘horse’ /rk/ [rk] /ma!rkab/ [ma!rkab] ‘ship’

/hark + o/ [harko] ‘skins’

cf. he¢ a¢n ‘intesti e /d1l/ [d1l] /had 1al + [had 1lay] ‘I talked’

‘delicate, /gn/ [gn] /naag + ni!mo/ [naagni !m ‘being a w/ql/ [ql] /boqla!ad1/ [boqla!ad1] ‘hundredth’ /sl/ [sl] /?uslaa/ [?uslaa] ‘he was heavy’/?m/ [?m] /ga?mo/ [ga?mo] ‘hands’

‘(in a state of) liki/mb/ [mb] /lambarro/ [lambarro ‘numbers’ /md1/ [md1] /lam + d1a:/ [lamd1a:] ‘men’ /mm/ [mm /ˇuˇummo/ [ˇuˇummo] ‘hands’

‘fisherman/

/n [nk] /ni !nkii/ /ns/ [ns] /garo ! + sii/ [garansi Ÿi] ‘to make someone

understa/ndZ/ / [d1andZi !re] ‘ambassador’ /nm/ [nm] /d1unmay/ [d1unmay] ‘got organized’

[ndZ] /d1andZi !re

[nn] /min + ne/ [mine] /// //

//

/

////

//

196

rg/ [rg] /nirgö/ [nirgö] ‘young female

camels’ rf/ /©arfo/ ©arfo] ‘letters’

[rs] /aarso!/ rso!] ‘take /irmaanaa/ [irmaanaa] ‘ /askar + ni !mo/ [aska ‘being a soldier,

military service’ /rr/ [rr] / [wa!rran] ‘ ll news of

[rw] / -le/ [garwa:nle] ‘ unter’ [ry] /guryanno!qod ry nno !qod1] ‘returning home of

animals’

Saeed, John Ibrahim. 1993. Somali Reference Grammar. 2nd ed dp unwoo Pr ss: Kensington, Maryland, U.S.A.

Zabo ki, A dr j. 1 86. he orphol gy N minal Plural in the Cushitic Languages. Vero·ffentlichungen der Institute für Afrikanistik und Agyptologie der Universität Wien. Wien.

Index. dp Dunwoody Press: Kensington, Maryland, U.S.A.

cf. [harag] ‘skin’ /

//rs/

[rf]

[ [aa revenge’

/rm/ [rm] in milk’ /rn/ [rn] rni !mo]

wa!rran/arwa:n

te ’ /rw/ g

1/ h

/ry/ [gu a

References:

. D dy e

rs n ze 9 T M o of o

Zorc, R. David & Madina M. Osman. 1993. Somali-English Dictionary with English

197

nemic con onant y: Labial Dental Alveolar Palato- Palatal Velar Glottal

Swedish (Indo-European, Germanic) Pho s inventor

alveolar Stop p t d k b g Fricative f s S Ç Û v h Nas n al m N Lateral l Trill r Glide y Patterns of son/C clusters:

t k g v s S Ç Û h m n p b d f N l r y p b t d k g gn f v s S Ç Û h m mb mn ml mr n mp mb Nk Ng nl nr N Nt Nn Nl l lÛ lm ln r Ê ß rÛ rm rn y

Examples:

gn/ [gn] /mogna/ [mogna] ‘ripen’ mb/ [mb] /bomb/ [bomb] ‘bomb’ mn/ [mn] /hamn/ [hamn] ‘harbour’ /nœmna/ [nœmna] ‘mention’ ml/ [ml] /ramla/ [ramla] ‘fall’ mr/ [mr] /hamra/ [hamra] ‘hammer’ np/ [mp] /anpassa/ [am:pas:a] ‘to adapt’ nb/ [mb] /inbealning/ [im:b´tA:lniN] ‘payment’ nk/ [Nk] /ankomma/ [aN:kOm:a] ‘to arrive’ ng/ [Ng] /ango/ [aN:go] ‘to concern’ nl/ [nl] /vanlig/ [vanlig] ‘usual’ nr/ [nr] /inre/ [inre] ‘innter’

/// ////////

198

[tråNt] ‘narrow’ Nn/ [Nn] /vaNn/ [vaNn] ‘carriage’

[reNna] ‘rain’ /viN er’

[lÛ l [helÛ] [lm] /malm/ [ln o loud’

/mo [Ê] /bor [boÊa] [Í] r /

rÛ/ [rÛ] /arÛ/ [arÛ] ‘angry’ ll/ [formell] ‘formal’

/rn/ rn] /torn/ [torn] ‘t er [varna] ‘warn’

] / / [kä ‘vessel’ /s la/ [so ‘murmur’

ference llberg, Sta n. 197 . Grap onom Rules in Phonology: Studies in the

Expression Component of Swedish. ACTA Universitatis Gothoburgensis.

dissertation. Institute of Oriental Languages, University of Stockholm.

/Nt/ [Nt] /tråNt/ / /reNna/ /Nl/ [Nl] la/ [viNla] ‘stagg/lÛ/ ] /he Û/ ‘holidays’ /lm/ [malm] ‘ore’ /ln/ ] /m ln/ [moln]

[molnig] ‘c

lnig/ ‘cloudy’ /rt/ ta/ ‘away’ /rs/ /ko sa [koÍa] ‘cross’ //rm/ [rm] /forme

[ ow ’ /varna/ /rl/ [ kärl ]

or a]

RH

e s:e ffa 4 h ic

Pyun, Kwang-soo. 1987. Korean-Swedish Interlanguage Phonology. Doctoral

199

Phonemic consonant inventory: Labial Alveo-dental Alveo-palatal V tal

Tatar (Altaic, Turkic)

elar GlotStop p b t d g / kAffrica e ts tS dZt Fricative f v s z S Z x h Nasal m n N Liquid l, r Glide w y Patterns of son/C clusters:

p b t k g / ts dZ f v s z S d tS Z x h m mn mS n mp mb nd Nk Ng ntS N l lb lt lk lg ld r rt rk rg r rs rd f

m n l r w y N t tn tl k kl f fr z zn zm S S Sl m m mn mn n nn ny m m N Nn l lm ll r rm rn rl y yl

Examples: tn/ [tn] /at + nEN/ [atnïN] ‘of the horse’ tl/ [tl] /at + lÄr/ [atlar] ‘horses’ kl/ [kl] /balIk + lÄr/ [balIklar] ‘fish’ fr/ [fr] /yafrak/ [yafraq] ‘leaf’ zm/ [zm] /yaz + mÄ/ [yazma] ‘not to write’ zn/ [zn] /küz + nEN/ [küzneN] ‘of the eye’ Sm/ [Sm] /biS-meN/ [biSmeN] ‘2000’ Sl/ [Sl] /eSlä/ [eSlä] ‘work!’ md/ [mn] /uram + dÄn/ [uramnan] ‘from the street’ mS/ [mS] /yomSak/ [yomSaq] ‘soft’ mn/ [mn] /küze + m + nEN/ [küzemneN] ‘of my eye’ ml/ [mn] /khanIm + lÄr/ [khanImnar] ‘ladies’ np/ [mp] /un + pot/ [umpot] ‘ten puds’ (pud = Russian

weight = 40 pounds)

/////////////

200

[umbiS] ‘fifteen’ (lit. ten-five) nd/ [nd] /anda-monda/ [andamonda] ‘here and there’

[xatïNkïz] ‘women, woman-falk’ o irl)

[N a n [n SïgïS/ Sïg (l e) [mm] /min + mE/ [mime] [n h van + lÄr/ a ar] ‘animal [ny] k ‘north’ (lit. night-side) [N ] /kü + N + nEN/ Nn ] ‘of thy eye’

lb il baw/ w] ‘girdle, belt’ [balta-belän] ‘with the ax’

dÄ/ [awïlda] ‘in the village’ /lk/ [lk] [ kew ‘to wave’ /lg/ [lg] /bil / [bilg´] ‘sign’ /lm/ [lm] /alma + lÄr/ [almalar] ‘apples/ll/ [ll] /kül + lÄr/ [küllär] ‘ g/rt/ [rt] /korït + ï / [kortïN ‘thy worm’ /rd/ [rd] /Sähär + dÄn/ [S ‘ + re lat al s ffix/rk/ [rk] /kurïk + a/ [kurka] ‘he is afraid’

[rg] /siN r + / [s arga] ‘to thee’ /rf/ [rf] /xäref + e/ [xärfe] ‘his letter’ /rs/ [rs] N [ ar ] ‘thou shalt write’

/ ] b m [ m ‘do not go’ /rn/ ] Sähär + nE/ [ är ‘ city’ /rl/ [rl] Sähär + lÄr/ [ ärlä r] ‘cities’ /yl/ ay lÄ [aylar] ‘moon, months’

References: Poppe, Nicholas. 1968. Tatar Manual. Indiana University Publications: Bloomington.

/nb/ [mb] /un + biS/ //nk/ [Nk] /xatïn-kïz/

(lit. w ‘one who has burnt’

man-g/ng/ g] /y

] /kön + tn + gan/ [yaNga

[könt]

/ntS/ tS ïS] ‘east’ ‘me?’

it. sun-ris/nm/ /nl/ n] /k

/tön + yak/ ay [khayv nn s’

/ny/ [tönya ] /Nn/ n ze [küze

[bilbaeN

/lb/ [lt/ [lt]

] /b /balta + belän/

+ /

/ld/ [ld] /awïl +/selek + ew/ sel ]

g´’

villa es’ N ]

ähärdän] city pu - er ion u ’

/rg/ a gÄ iN

/yazarsï/

/ Ä/

y zbara s

a] ïN

/rm [rm ar + [rn / Säh ne] the

/ Säh [yl] / + r/

201

Toba Batak (Austronesian, Sundic)

lottal Phonemic consonant inventory: Bilabial Alveodental Alveolar Alveopalatal Velar GStop p b t d k g Affric e tS at dZFricative s h Nasal m n N Lateral l Trill r semivowels w y Patterns of son/C clusters:

p b t d k g tS s h m n N l r w y dZ p km kn N kl kr k b t km kn kN kl kr d k km N kkn k l kr g tS dZ s sm sn sN sr h m pp bb pt md pk Ng ptS pp mn N ml mr ps mm mn pp bb tt kk Ng ↓ ss kk mm nn NN ll rr dd N N N S N N N b kt ↓ g kt ks kk Nm n N nl r kp Nd l l lb lt lg ls lh lm ll N ll p ld l r rp rb rt rk rg rs rh rm rn N rr rd r ry w y

/pm/ [km] /raŸp + mo!dom/ [raŸkmo!dom] ‘sleep together’

/pr/ [kr] /rap + ro!/ [rakro!] ‘come together’ me’

time’

Examples:

/pn/ [kn] /a!lap + na/ [a!lakna] ‘get the one’ /pN/ [kN] /rap + No!t/ [rakNo!t] ‘get up together’ /pl/ [kl] /rap + la!o/ [rakla!o] ‘go together’

/tm/ [km] /huhu!t + ma!rsak/ [huhu!kma!rsak] ‘sad at the same tio/tn/ [kn] /huhu!t + no!ln l/ [huhu!kno!lnol] ‘stare at the same

/tN/ [kN] /huhu!t + No!t/ [huhu!kNo!t] ‘rise at the same time’ la ] /tl/ [kl] /huhu!t + la!0/ [huhu!k !o ‘go at the same time’

t /tr/ [kr] /yu! + roha!na/ [yu!kroha!na] ‘he is exasperated’ k/ m/ [km] /didok + ma + maddok + au/

[didokmamaddokau] ‘(he) said to me…’ /kn/ [kn] /halak + na + modom + i/

202

[halaknamodomi] ‘the sleeping man’

/kr/ [kr] /rE!krEk/ [rE!krEk] ‘shelf’ /sm/ [sm] /pO!sma/ [pO!sma] ‘a boy’s name’ /sn/ [sn] /a!sna/ [a!sna] ‘a girl’s name’ /sN/ [sN] /NO!sNOs/ [NO!sNOs] ‘worrying’

/mp/ [pp] /o!nom + pE/ [o!noppE] ‘just six’ /mb/ [bb] /laŸm + balga!/ [laŸbbalga!] ‘become bigger’ /mt/ [pt] bu!at] y’

/mk/ [pk] /laŸm + koto!r/ [laŸpkoto!r] ‘(become) more dirty’ /mg/ [Ng] /laŸm + gogo!/ [laŸNgogo] ‘become stronger’

e!t] lently’ !tik] ’

/mh/ [pp] no example given /

giv purna a/ a!ma] ’

/nb/ [bb] /soŸNon + ba!tu/ [soŸNobba!tu] ‘like a stone’

/nd/ [dd] /soŸNon + do!lok/ [soŸNoddo!lok] ‘like a mountain’ ku!li/

/pa!ntSi/ [pa!ntSi] ‘pan’

/ntS/ [ttS] across a word tSu!tSi/ i] ‘laundry ap’

/ns/ [ss] /siŸan + su!NE/ [siŸassu!NE] ‘from the river’ e’

/nm/ [mm] /soŸNon + ma!nuk/ [soŸNomma!nuk] ‘like a chicken’

/nN/ [NN] /hoŸlan + Nalu!kna/ [hoŸlanNNalu!kna]‘only the misery’

nr/ [rr] /sian + ru!ra/ [siŸarru!ra] ‘from the valley’ p] /daN + po!la/ [dakpo!la] ‘not necessary’

Nb/ t/

/Nd/ [Nd] /da!NdaN/ [da!NdaN] ‘pay heavily’ /Nk/ [Nk] within a word

/ba!Nku/ [ba!Nku] ‘bench’

/kN/ [kN] no example given/kl/ [kl] /la!klak/ [la!klak] ‘rind’

/sr/ [sr] no example given

/hoho!m + tabu!at/ [hoho!pta ‘let’s take (it) stealthil/md/ [md] no example given

/mtS/ [ptS] /hoho!m + tSe!t/ [hoho!ptS ‘paint (it) si/ms/ [ps] /holo!m + sao!tik/ [holo!psao ‘somewhat dark

/mm [mm] /u!mma/ [u!mma] ‘kiss’ /mn/ [mn] no example given /mN/ [mN] no example given /ml/ [ml] no example given /mr/ [mr] no example en /np/ [pp] /buŸlan + !m [buŸlappurn ‘full moon

/nt/ [tt] /siŸlan + to!ru/ [siŸlatto!ru] ‘from below’

/nk/ [kk] /soŸNon + [soŸNokku!li] ‘like a cooly’ /ntS/ [ntS] within a word

/saŸbun + [saŸbuttSu!tS so

/nh/ [kk] /siŸah + ku!ta/ [siŸakku!ta] ‘from the villag/ng/ [Ng] /soŸNon + ge!a/ [soŸNoNge!a] ‘like a worm’

/nn/ [nn] /ni!nna/ [ni!nna] ‘he says’

/nl/ [ll] /láOn + láOn/ [làOlláOn] ‘eventually’ //Np/ [k/ [Nb] /buNbu!N/ [buNbu!N] ‘wide open, vertically’ /N [kt] /daN + tabo!/ [daktabo!] ‘not delicious’

203

oundary /daN + koto!r/ [dakkoto!r] ‘not dirty’

ta!Nga/ [ta!Nga] ‘ladder’ NtS/ /daN + tSu!kk ktSu!kkup] ‘not enough’

/Ns/ [ks] /daN + sa!Na/ sa!Na] ‘not enough time’ [kk] /daN + kEbb [dakkE !N] ‘not

/Nm/ [Nm] /ma!NmaN/ N] ‘vow’ /Nn/ [Nn] /no!NnoN/

/nu!NNa/ [nu!NNa ‘already’ [Nl] (one occurrence)

/la!NlaN/ [la!NlaN] ni habited p e’ Nl/ [nl] across a word boundary

[ri!NriN] ‘joint’ /lp/ /talp lpe!] ‘on the brink’ /l ] /tu!lbas/ ‘r lenis/lt/ [lt] /palta!k/ [palta!k] ‘clearly visible’ /ld/ [ld] /da!ldal/ [da!ldal] ‘bee’ /lg/ [lg] /balg [balga!] ‘big’ /ls ] /sO!ls !lsOl] ‘repentance’ lh /pu!lha/ [pu!lha] ‘break dam to drain’

[lm] /yO!l lma] ‘man, generic’ /ln/ [ll] /bál + na/ álla] /l [lN] /talNa [talNa!] /l [ll] /tu!lla [tu!llaN] /rp/ [rp] /ga!rp / [ga!rpu] /rb/ [rb] /su!rb / [su!rbu] /r [rt] /a!rta/ [a!rta] ‘treasure’ /rd/ [rd] /marda!lan/ [marda!la ‘w lk, vb.’ /rk/ [rk] /park kas/ [parka!kas] ‘tool’ /rg/ [rg] /gO!rg / rga] ‘carve’ /rs/ [rs] /a!rsak/ sak] ‘sorrow’ /rh/ [rh] /sO!rh / [s rha] ‘spinning wheel’ /rm/ [rm] /sO!rm n/ !rmin] ‘reflect’ /rn/ [rn] /burna!N/ [burna!N up’ /r [rN] /si!rNo / [si!rNom] ‘lost, unfortunate’ / [ll] /pO⁄r + lak/ llak] ‘garden’ /r [rr] /pi! rik] tle, n.’ /ry/ [ry] /bu ry kind’ References: N ban, .J. 19 . A G mar oba ak. P ic Lin istics S ries D o. 3 Materials in Languages of Indonesia, No. 6. W.A.L. Stokhof (ed.). Department

of L nguistic , Resea h Scho l of P Stud s. The Australi Nation l University

/Nk/ [kk] across a word b /Ng/ [Ng] // [ktS] up/

[da [dak

/Nh/ a!N/ bba spread’ [ma!Nma

[no!NnoN] ‘immerse’ /NN/ [NN] ] /Nl/

‘u n lac//Nr/ [Nr] /ri!NriN/

[lp] e!/ [ta b/ [lb [tu!lbas] ep h’

a!/ / /

[ls [lh]

Ol/ [sO//lm/ ma/

[yO![b ‘his ball’

N/ l/

!/ ‘uneven’ ‘stab’ N/

u ‘fork’ ‘burn’ u

t/ n] a

a!a [gO!

[a!r

a O![sO

i

] ‘puffed N/ m

rl/ [pO⁄[pi!rr/ rrik/ ‘whis

ryu!/ [bu u!] ‘good,

aba P.W 81 ram of T -Bat acif gu e . N 7.

i s rc o acific ie an a

204

Turkish (Altaic, Turkic)

Dal

Glottal Consonant inventory: Bilabial Labio- ental- Palato- Palatal Velar

dental veolar alveolar Plosi p ve b t d k g (/) Fricative f z S v s Z Affricate tS dZ Nasal m n Liquid l, | Glide y h Patterns of son/C clusters:

p k g f v b t d s z m mp mt mg mf mb ms n nb~ nt nd Nk Ng

mb ns nz

l lt lk lg lf lv lb ls lz | |p |t |k |g ||b |d |s v

S tS d m n l | Z Z y h

p pm p| b b| t tm tl t| d dl d| k km kl k| g f fn v vm vl ↓ s sm sn | sl s z zm zl z| S Sm Sn Sl Z Zm tS tSm tSl dZ m mtS Z mm mn ml m| md mh n ntS ndZ nm nn nl~ n|

nn l ltS lm ↓ ll ly lh | |S |tS |m ↓ | |dZ y |h y yl h hm hl ↓

205

Examples: /pm/ [pm] /tSapmak/ [tSapmak] ‘gallop’

mak/ [atmak] ‘throw’ /tl/ [tl]

/dl/ [dl] /adliye/ [adliye] ‘justice’

/tSakma/ [tSakma] ‘flash of lightning’ lE|/

/vl/ [vl] /devlet/ [devlet] ‘state’ om.’

/ [sm]

|/

/ ’

man/ nt’ /tSm/ [tSm] /atSma/ [atSma] ‘inauguration’

/ Acc.’ p’

/mt/ [mt] /emtia/ [emtia] ‘goods’

[Sindi]‘snow’ ]

lice’ amak/

/mh/ [mh] /hamhalat/ [hamhalat] ] / r water’

/

/p|/ [p|] /tSap|az/ [tSap|az] ‘crosswise’ /b|/ [b|] /geb|e/ [geb|e] ‘hair-cloth glove’ /tm/ [tm] /at

/at + lE|/ [atla|] ‘horses’ /t|/ [t|] /kat|a/ [kat|a] ‘drop’

/d|/ [d|] /id|ak/ [id|ak] ‘perception’ /km/ [km] /kl/ [kl] /bal¨k + [bal¨kla|] ‘fish’ /k|/ [k|] /ak|an/ [ak|an] ‘contemporary’ /fn/ [fn] /defne/ [defne] ‘laurel’ /vm/ [vm] /tSavmak/ [tSavmak] ‘deviate’

/v|#/ [vV|] /dev|/ [devi|] ‘period, N/v|/ [v|] /dev| + i/ [dev|i] ‘period + Acc.’ /sm /asma/ [asma] ‘suspended’ /sn/ [sn] /esna/ [esna] ‘course’ /sl/ [sl] /aslen/ [aslen] ‘originally’ /s|/ [s|] /k¨s|ak/ [k¨s|ak] ‘mare’ /zm/ [zm] /azmak/ [azmak] ‘go astray’ /zl/ [zl] /k¨z + lE [k¨zla|] ‘girls’ /z|/ [z|] /az|ak/ [az|ak] ‘rare’ /Sm/ [Sm] /antlaSma [antlaSma] ‘treaty’ /Sn/ [Sn] /aSn¨/ [aSn¨] ‘ancient/Sl/ [Sl] /aSl¨k/ [aSl¨k] ‘provisions’ /Zm/ [Zm] /amenaZ [amenaZman] ‘manageme

/tSl/ [tSl] /atSl¨k/ [atSl¨k] ‘hunger’ /hm/ [hm] /ahmak/ [ahmak] ‘foolish’ /hl/ [hl] /ihlal/ [ihlal] ‘spoiling’ /h|#/ [hV|] /Seh|/ [Sehi|] ‘city + Acc.’ /h|/ [h|] /Seh| + i [Seh|i] ‘city +/mp/ [mp] /ampul/ [ampul] ‘bulb, lam/mb/ [mb] /bombok/ [bombok] ‘utterly useless’

/md/ [md] or [nd] /Simdi/ [Simdi] or

/mg/ [mg] /damgadZ¨/ [damgadZ¨ ‘stamper’ /mf/ [mf] /hemfiil/ [hemfiil] ‘accomp/ms/ [ms] /az¨ms [az¨msamak] ‘disdain stg.’

‘coarse’ /mtS/ [mtS /tSamtSak [tSamtSak] ‘wooden vessel fo/mdZ [mdZ] /amdZa / [amdZa ] ‘paternal uncle’

206

e’

se in temperature’

/bin+baS+¨/ [binbaS¨] or [bimbaS¨] ‘major’

/ng/ [Ng] [ns] /densiz/ [densiz] ‘peevish’

ntS/ [ntS] /kazantS/ [kazantS] ‘gain’ ndZ] /alt¨ndZ¨/ [alt¨ndZ¨] ‘sixth’

/nn/ [anne] ‘mother’

lt/ [lt] /alt¨/ [alt¨] ‘six’ ld/ [ld] /ald¨r¨S/ [ald¨r¨S] ‘attention’ lk/ [lk] /alk¨m/ [alk¨m] ‘rainbow’ lg/ [lg] /alg¨n/ [alg¨n] ‘pale’ lf/ [lf] /alfenit/ [alfenit] ‘silver alloy (for cutlery)’ lv/ [lv] /dZilve/ [dZilve] ‘grace’ ls/ [ls] /s¨|¨ls¨klam/ [s¨|¨ls¨klam]‘sopping wet’ lz/ [lz] /kolza/ [kolza] ‘rape’ ltS/ [ltS] /altSak/ [altSak] ‘low’ lh/ [lh] /bilhassa/ [bilhassa] ‘especially’ lm/ [lm] /ilm-i:/ [ilmi:] ‘scientific’ ln#/ [lVn] /aln/ [al¨n] ‘forehead, Nom.’ ln/ [ln] or [nn] /aln + ¨/ [aln¨] or [ann¨] ‘forehead + Acc.’

/ll/ [ll] /elli/ [elli] ‘fifty’ |p/ [|p] /a|pa/ [a|pa] ‘barley’ |b/ [|b] /be|bat/ [be|bat] ‘destroyed’ |t/ [|t] /aya|t¨/ [aya|t¨] ‘corruption’ |d/ [|d] /ka|deS/ [ka|deS] ‘brother, sister’

/|k/ [|k] /ku·|k/ [ku·|k] ‘fur’ /fa|k + a/ [fa|ka] ‘difference + Dat.’

|g/ [|g] /ka|ga/ [ka|ga] ‘crow’ |v/ [|v] /ki|ve/ [ki|ve] ‘sponsor of boy being

circumcised’ |s/ [|s] /a|sa/ [a|sa] ‘building plot’ |S/ [|S] /a|Sin/ [a|Sin] ‘Tk. yard, yardstick’

/mm/ [mm] /amme/ [amme] ‘everyone’ /mn/ [mn] /memnuniyetle / [memnuniyetle]‘with pleasur/ml/ [ml] /han¨m + lE|/ [han¨mla|] ‘ladies’ /m|/ [m|] /cem|e/ [cem|e] ‘gradual ri/nb/ [nb] or [mb]

/nt/ [nt] /as¨nt¨/ [as¨nt¨] ‘putting off’ /nd/ [nd] /bu|un + da/ [bu|unda] ‘nose + Loc.’ /nk/ [Nk] /sanki/ [saNki] ‘as though’

/ns/ /nz/ [nz] /benze|/ [benze|] ‘similar’ //ndZ/ [/nm/ [nm] /abanmak/ [abanmak] ‘lean over/against’

[nn] /anne/ /nl/ [nl] or [nn] /hayvan + lE|/ [hayvanla|] or [hayvanna|] ‘animals’ [onla|] or [onna|] ‘they’ /n|/ [n|] /tan|¨tSa/ [tan|¨tSa] ‘godde

/onla|/ss’

/lb/ [lb] /dZilbent/ [dZilbent] ‘portfolio’ /////////////

////

//

//

207

[bu|han] ‘evidence’ |tS/ [|tS] /ge|tSek/ [ge|tSek] ‘true’

n/ [b¨ld¨|dZ¨n] ‘common quail’ |m/ [|m] mek / [e| ‘to a

[|Vn / [bu|un] ‘|n] /bu|n + a/ [ ‘nose + Dat.’

/|l/ [|l] + [ ‘/|y/ [|y] /he|ye|delik/ [ ‘ od’s omnipresence’

[yl] /ay + lE [ e|] ‘moons, months’

Notes: eral, wo ve vocabulary do with [d , f, Z, l, m, n, r, z].

s: 85. The Oxford Turkish-English Dictionary. 3rd ed.

. ert h, K ren S art av . 20 0. T rki +/ Cl te ho ol y. CW P 2 ,

Oh St e U ive ity.rn lt, Ja lin. 997 Tu ish Ro led : Lo don nd ew o .

r Universit Pr s, E y H us Lo o

/|h/ [|h] /bu|han/ //|dZ/ [|dZ] /b¨ld¨|dZ¨/ /e| mek] ttain’ /|n#/ ] /bu|n nose, Nom.’ /|n/ [ bu|.na]

/Sehi| lE|/ Sehi|le|] cities’ he|ye|delik] G

/yl/ |/ ayl

- In gen rds of the nati n’t begin Z

eferenceRAlderson, A. D. & Fahir Iz. 19

Oxford niversit PreU y ssBa sc a & tu D is 0 u c C l/ us r P n og M O 000

io at n rs Ko fi k 1 . rk . ut ge n a N Y rkLewis, G. L. 1975. Turkish Grammar. Oxfo d y es l o e, nd n W.

208

Udi (North Caucasian, Lezgian)

Phonemic consonant inventory: Stop Affricate Fricative Nasal Rhotic

Bilab p b mial p’ Labiodental f v Alveodental t t’ d ts s n ts’ z r Palato lveolar t ’ dÛ Ç a t Ç Ç ÛPalatal tS tS dZ S ’Lateral l Velar k k’ g Uvular q q’ Â X Laryngeal h Patterns of son/C clusters:

p b t t’ d k g ts’ tS tS’ s s’ z  m n l r p pl t tt’ d dd dd q ql~Xl S Sn Sl m mp mb mtS mtS’ mn n nt’ ng l lp lk ↓ lm ll r rb rt rd rts’ rtS’ rz rm rr rr

Examples:

pl/ [pl] /plœkœn/ [plœkœn] ‘stairs, ladder’ tn/ [tt’] /tutna/ [tutt’a] ‘mulberry (gen.)’ dn/ [dd] /zid-na/ [zidda] ‘iron (gen.)’ dl/ [dd] /xodluÂ/ [xodduÂ] ‘woods’ ql/ [ql] or [Xl] /qoqla/ [qoqla] or [qoXla] ‘egg’ Sn/ [Sn/ /aS-ne-b-sa/ [aSnebsa] ‘(s)he works’ Sl/ [Sl] /aSl-axun/ [aSlaxun] ‘work-abl.’ mp/ [mp] /Sampesun/ [Sampesun] ‘to slaughter’ mb/ [mb] /zombesun/ [zombesun] ‘to teach’ mtS/ [mtS] /tSœmtSœ/ [tSœmtSä] ‘ladle’ mtS’/ [mtS’] /amtS’i/ [amtS’I] ‘empty’ mn/ [mn] /aXum-nu-p-e/ [aXumnupe] ‘laugh-2sg.-LV-PERF’ nt’/ [nt’] /bunt’/ [bunt’] ‘riot’ ng/ [ng] /dœng/ [dœng] ‘idiot’ (Persian ‘deng’)

/rl/ [rr] /saturla/ [saturra] ‘one-legged’ lp/ [lp] /k’al-p-e/ [k’alpe] ‘read-LV-PERF’

/lk/ [lk] /tülki/ [tülki] ‘fox’ lÂ/ [ÂwÂ] or [vÂ]

/dZolÂa/ [dZOÂwÂa] or [dZOvÂa] ‘group of people’

//////////////

/

/

209

[xalx] ‘people’ (Arabic h ¢alq) lm/ [lm] /elmux/ [elmux] ‘spirit, soul’

/ [pulmuÂon] ‘with the eyes’ (eye-pl-erg) ln/ [ /k’alnexa/ [k’allex ‘(s)he calls, reads’

/rt/ [rt] /tS [tS’urt] ‘cocoon’ martad/ [martad] ‘larg

[rb] /ser-b-esa/ [serbesa] ‘to build (present tense)’ /rd/ [ œ rd] ‘pain’ (Persian dard)

[ se [serzub il 1sg-LV-PRE ’ [rts’] /p’ ‘dung’

[ S’] /pa [partS’] ‘water jug’ [ Ø œt/ œt] ‘greeting’

rn/ [rr] /purnexa/ [purrexa] ‘(s)he flies’

N tes: - Initial /r/ is prohibited. - Initial consonantal clusters are not allowed except for recent loan words. - Initial /n/ is rare. - Initial /l/ is extremely rare before vowels ther an / . (m y cas in q stio

ar loanw rds.) References: Schulze, Wolfgang. 2001. An online grammar of Udi Ludw -Ma milia -

Universita·t Mu chen http: www lrz-m enche .de/~ sc e/ud hal )

/lx/ [lx] /xalx/ / /pul-muÂ-on/ ll] a]

’urt/ / et dish’ /rb/

rd] /d rd/ [dœ /rz/ rz] / r-zu-b-sa/ sa] ‘bu d- S/rts’/ irts’/ [p’irts’] /rtS’/ rt rtS’/ /rm/ rm] / hrm [Øhrm/

o

o th a/ an es ue n e o

. u

ign

xiw

nsul·n . ( // . h z in t.htm

210

Phonemic consonant inventory: Labial Alveo- Alveo- Velar Uvular Glottal

Uyghur (Altaic, Turkic)

dental palatal Stop p b t d k g q G / Affricate (ts) tS dZ Fricative f s z S Z x (v) h Nasal m n NLateral l Trill r Glide y w Patterns of son/C clusters:

t G r p b d k g q s m n l p pl t tm tl k kl q ql G Xn s sm sl z zl tS tSl h hm m ml mr n mp mb nd N Ng –q –G mm k ll N Nl l lp lt ld lg ls lm ll r rk rG rm rn rl y yl

Examples: /pl/ /äswa:p+lar/ [äswaplar] ‘tools’

kl/ [kl] /tSötSäg+lik/ [tSötSäklik] ‘person of Chughuchaq’ ql/ [ql] /beliq + lAr/ [beliqlär] ‘fish’ Gn/ [Xn] /tamaGniN/ [tamaXniN] ‘of (the) food’ sm/ [sm] /ismim/ [ismim] ‘my name’ (loanword from

Farsi) /sl/ [sl] /äslik/ [äslik] ‘a ~ the memory’ zl/ [zl] /qIz + lAr/ [qIzlar] ‘the girls’ tSl/ [tS] /kätSlik / [kätSlik] ‘nightly’ hm/ [hm]

/np/ [mp] /qalGan pul/ [qalGampul] ‘(the) remaining money’

[pl] /tm/ [tm] /kätma/ [kätma] ‘go, negative

l/ /t [tl] /at + lAr/ [atlar] ‘horses’ ////

///

211

ä/ [Sämbä] ‘([on] a ~ the) Saturday’ nd/ [nd] /mändä/ [mändä] ‘at my place’

[bügüNki] ‘today’s’ /ng/ [ /kö än] ing d

[ –qaldimq] ‘I remained’ [ ] /na –Ga] ‘to (the [mm] /kälgänmän/ n ‘I have come’ [ a ‘animals’

/ml/ [ml] /kh ‘ladies’ [mr] /amraq/ [amraq] ‘found’

/Nl/ [Nl] /SinyaN+liq/ [SinyaNliq] ‘(person) of Xinjiang’ lp/ [lp] /yälpügütS/ [yälpügütS] ‘a ~ the fan’

S/ [olturuS] ‘a ~ the meeting’ /ld/ [ld /tIl+ n/ ld ng (a at )’

[lg] /käl + gän/ älgän] avi come, m ’ /ls/ [ls] /wälsipid/ älsipit] icyc ’(?/lm/ [lm /tIl+maq/ ilmaq] the) ise ing /ll/ [ll] /köl + lAr/ ‘lakes’ /rk/ [rk /u+lar+kim+lär/ lar mla ‘Who are they?’

/ [rG] /bar + Gu + m/ arG m] y w nti to o’ ( u = ‘wanting to’, m = ‘my’)

/ [rm] /bar+ma+N/ [barmaN] plea ) d no go! (ma = ‘not’, N ‘d ) / /ularniN/ [ularniN] ‘their’

/ [r /shähär + lAr/ h lär ti/yl/ [yl] Ar/ yla o ’

- re ns an l / an r op on y d lete resulting in the lengthening of a preceding vowel

: F. 1991 r. Univ y o d

hi 971. ( ussian M ral

/nb/ [mb] /Sänb//nk/ [Nk] /bügünki/

Ng] rüngän/ [körüNg ‘hav appeare/nq/ –q] /män qaldim/ [mä/nG/ –G nGa/ [na ) bread’ /nm/ [kälgämmä ] /nl/ ll] /h yvan + lAr/ [hayvallar]

Imlar] anIm + lAr/ [khan /mr/

//lt/ [lt] /olturu

] dI [ti in] ‘to ue bl ive/lg/ [k ‘h ng co ing

[w ‘b le ) ] [t ‘( d ct ’ [köllär] ] [u ki r]

/rG [b u ‘m a ng g G

/rm ‘( se o t ’ = o’

/rn [rn] /rl l] [s ähär ] ‘ci es’

/ay + l [a r] ‘m ons

Notes: P co on ta l/ d / /: ti all e d

ReferencesHahn, Reinhard . Spoken Uyghu ersit f Washington Press: Seattle an

London. Nadz p, E. N. 1 translated fr R by D. . Segal) rn Uigur. om Mode Cent Department of Oriental Literature.Nauka Publishing House: Moscow.

212

Yakut (Altaic, Turkic)

Labial Dento-alveolar Palatal Velar Uvular Glottal

Phonemic consonant inventory:

Stop p b t k d g Fricative s X Â h Affricate tS dZ Nasal m n ≠ N Liquid l, r Glide y - /r/: a trill; a flap intervocalically

r y Patterns of son/C clusters:

p b t d k g s X  h tS dZ m n ≠ N l p pt b t tt d k kt g s st X Xt  h tS dZ m mt mtS mm mn mn n nt nn N nn Nk Ng mm N≠ N Nt NN Nm Nn NN Nnl ↓ lg lg ltS ldZ ll r rb rd rg rs rdZ rn ↓ y yd

Examples: /pl/ [pt] /Iskaap + lAr/ [Iskaaptar] ‘cabinets’ tl/ [tt] /at + lAr/ [attar] ‘horses’ kl/ [kt] /balIk + lAr/ [balIktar] ‘fishes’ sl/ [st] /muos + lAr/ [muostar] ‘horns’ Xl/ [Xt] /tuoX + lAr/ [tuoXtar] ‘what things’ mt/ [mt] /olom + tAn/ [olomton] ‘ford, ablative’ mtS/ [mtS] /emtSit/ [emtSit] ‘doctor, physician’ mm/ [mm] /tammaX/ [tammaX] ‘drop, drip’ mn/ [mn] /umnas/ [umnas] ‘stem, stalk’ ml/ [mn] /olom + lAr/ [olomnor] ‘fords’ nm/ [mm] /suun + mappIt/ [summappIt] ‘we do not wash’ nt/ [nt] /oyun + tAn/ [oyuntan] ‘shaman, ablative’ nk/ [Nk] /ün + kür/ [üNkür] ‘sloping, aslant’

////////////

213

/nn/ [nn] /munnuk/ [munnuk] ‘corner’ nN/ [NN] /oron + No/ [oroNNo] ‘on the bed’

[oronnor] ‘beds’ Nt/ [ /XatIN + tAn/ n] ch, ablative’

[ ] /tii KA/ ‘to the squirrel’ /Nm/ [Nm] /toN ‘frozen’

[Nn] /toNno/ [toNno] ‘it oze’ [Nn] /Xat ‘bi hes’

‘son, boy (ablative)’ VAr/ [uollaaVar] ‘than a son’

+ lar/ [illIlar] ‘they took’ [küölge] ‘t e lake

/lg/ /dZilga/ dZilga] ‘in a year’ S/ S S / bultSut] ‘hunter’ Z/ Z /X Z yI/ Xal aay ] ‘slope’

[ll] /küöl + lAr/ [küö er] ‘lakes’ [rb] /süürbe/ süü e] ‘tw nty’

/rd/ [rd] /k rde tuk/ kird eXtuk] ‘dirtily, filthily’ / [rg] /I rga Iarga] ‘in difficulty, in a

heavy…(+noun)’ / rs] / su n Xorsunnuk] ‘bravely’ Z/ [rdZ] /tahIr a/ [tah dZa] ‘outside’ / rn /udarnikk [uda i ‘you are a

sh ckw rker’ /rl/

/kötör + lAr/ ‘birds’ /rl/ hen r/ s af x-fi al a /l is fix-

/ahIIr + lar/ hII ar] ‘t eat’ cf. [ahaa] ‘to eat’, [ah

[yd /ubay + lAr/ uba dar ‘elder brothers’

Krueger, John R. 1962. Yakut manual. Bloomington, IN: IU Publications.

//nl/ [nn] /oron + lAr// Nt] [XatINta ‘bir/Nk/ NN N + [tiiNNe]

mut/ [toNmut] /Nn/ fr/Nl/ IN + lAr/

n/

[XatINnar] [uoltan]

rc/lt/ [lt] /uol + tAlt/ [ll] /uol + TAA/

/ld/ [ll] /il + dI/lk/ [lg] /küöl + KA/ o th ’

[lg] [/lt/ld

[lt ] /bult ut [ [ld ] ald aa [ dZ I

/ll/ ll /rb/ [ rb e

i eX [ e/rg a / [

/rs [

Xor n uk/ [/rd dZ Ir/rn [ ] In/

rn kkIn]

o o [rd] when /r/ is stem-final and /l/ is affix-initial

[kötördör] [ll] w / i fi n nd / af initial

[a ll heyIIr] ‘he eats’

/yl/ ] [ y ] References:

214

Yindjibarndi (Australian, Pama-Nyungan)

Phonemic consonant inventory: bilabial interdental alveolar palatal velar retroflex

Stop p t1T t Ô k Ê Nasal m n1 n ≠ N ¯Lateral l Glide w y1 | or r y ® - /r/: realized as a flap or a trill

atterns of son/C clusters: p t1T t Ô k Ê m n1 n ≠ N ¯ l w y 1 | y ®

P

p t1T t tm tn1 t≠ Ô Ôm Ôn1 Ô≠ Ô k Ê Êm Ên1 Ê≠ Ê

m mp n1 n np nt1T nt nÔ nk nm nn1 n≠ nN nd ≠ ≠p ≠t1T ≠Ô ≠k ≠m ≠n1 ≠≠ ≠Ô N Nk ¯ ¯p ¯t1T ¯t ¯Ô ¯k ¯m ¯n1 ¯≠ ¯N ¯ˇ l ly w y 1 | |p |k |m |≠ |N | |w |y 1 |y y yk ® ®p

Examples:

nl/ [nd] /maygan + la/ [mayganda] ‘my gum tree’ (locative)

¯l/ [¯Î] /yu|u¯ + la/ [yu|un¯Îa] ‘hair’ (locative) µl/ [µÔ] /yuÎa|iµ + lu/ [yuÎa|iµÔu] (proper name,

instrumental) ˇl/ [ˇ] /ku¯ˇ + la/ [ku¯Tata] ‘daughter’ (locative) Ôl/ [Ô] /kaNkaÔ + la/ [kaNkaÔa] ‘slack’ (locative) |l/ [|] /maTa\ +la/ [maTa\a] ‘red ochre’ (locative)

Examples of the other consonant clusters are not given.

otes: - /l/ and /r/ are prohibited in word- initial position.

/

//

///

- N

215

eferences: ndi Language. Canberra: The Australian National

RWordick, F.J.F. 1982. The Yindjibar

University.

216

Zoque (Mixe-Zoque)

o ic n in en ry Labial Aveolar Alveo-

palatal Velar Glottal

Ph nem consona t v to :

Palatal

Stop p t d Ô k g / b cFricative f s S h Affricate ts tS dZ Nasal m n ≠ N Liquid l, r Glide w y - /r/: realized as a flap or a trill Patterns of son/C clusters: p b t d c Ô g / f s S ts tS k h dZ

m mp mb × d mÔ mg /m mf × mh mts × × m × n nb nt nd nÔ × ng /n × ns nh nts ntS× ≠ × ≠b × × × × × × ≠dZ × × × N Nb × NÔ N Ng /N × Ns Nh Nts × × Nd k l lp × × × lk × /l ls × × ld r rp rb rt rg /r rs × rtS × rk × w × × wt wd × × w × × s × y × × yt yc × × × yh × × m n ≠ l r w y N

p × × pl pr pw py× ×b bl br by × t tn × × tr ty × × tw d × dr × × c Ô k × kn × × kl kr kw kyg gl gr gw × / f × × × fl fr × × × s sm sn × sl sr sw sy × S × × S l × Sy ≠ × S × h × hn × × × h hy× wts tsm tsn × × × × × tswtS tSn × × × × × × ×dZ m × mn × m My wn nm nn × n Ny w

217

N w m n ≠ l r y ≠ × × × × × N Nm Nn N≠ Nw Ny l lm × × × Lyr rm rn rl rw Ry w wm × × × Wy y ym yn × × × Examples:

hku/ hku] /pl/ [pl] //yaplastatsø [/yaplastatsø ‘he crushed it’ a a ’ /pr/ [pr] //y probatsø u/ [/y probatsøhku] ‘he approved it

hk

øhk øhku] ’

u

hku

it’

u/ ound’

/pw/ [pw] /kipwa/a/ [kipwa/a] ‘he already fought’ /py/ [py] /kipyahu/ [kipyahu] ‘they fought’ /bl/ [bl] /dyablo/ [dyablo] ‘devil’ /br/ [br] /mbalabra/ [mbalabra] ‘my word’ /by/ [by] / abyo !N/ [ abyo!N] ‘airplane’ / //tr/ [tr] //yatraNkats u/ [/yatraNkats ‘he propped it’ /tw/ [tw] /petwa/a/ [petwa/a] ‘he already swept/ty/ [ty] /setyembre/ [setyembre] ‘September’ /dr/ [dr] /padre/ [padre] ‘priest’

] /tsm/ [tsm /tøtsmen/ [tøtsmen] ‘toothache’ /tsn/ [tsn] //itsnapya/ [/itsnapya] ‘he bats his eyes’ /tsw/ [tsw] /tsitswa/a/ [tsitswa/a] ‘it already tore’ /km/ [km] /kowatøkmø/ [kowatøkmø] ‘Tuxtla Gutie!rrez’ /kn/ [kn] //yoknøktsahku/ [/yoknøktsahku] t as an ‘he left i

inheritance’ hk hku/kl/ [kl] /reklamatsø / [reklamatsø ] ‘he claimed it’

re o re o/kr/ [kr] /sek t / [sek t ] ‘secret’ g lo / l/ [gl] /sig / [siglo] ‘age’

/gr/ [gr] /pagre/ [pagre] ‘priest’ ø g a elf/gw/ [gw] //igwalats / [/i w latsøhku] ‘he made hims

equal’ tS tS /itS apya/ [ itS a/ ≠/ [ ≠] / ≠ / ≠ pya] ‘he bats his eyes’

li/fl/ [fl] //aflihido/ [/af hido] ‘worried’ f /fr/ [fr] /ko re/ [kofre] ‘chest’

/sm/ [sm] /k sm / [k sm ] ‘above’ ø ø ø ø/sn/ [sn] /kasnamb n/ [kasnamb n] ‘ladino’ ø ø

/ / [/sl/ [sl] / eslabo!N /eslabo!N] ‘fire striker’ e e/sr/ [sr] //isra !l/ [/isra !l] ‘Israel’

S s / [ s / y cooked/sw/ [sw] / o wa a/ So wa a] ‘she alread s/sy/ [sy] /nasyo!N/ [na yo !N] ‘nation’

Sl/ Sl] k Sla e/ Sla/ [ / a /an [ka /ane] ‘bread’ S S S S/ ≠/ [ ≠] /yo yoS≠ay [yo yoS≠ayu] ‘he tinkered ar

n/hn/ [hn] /kuhnayu/ [kuh ayu] ‘he knelt’ /hy/ [hy] /tuhyahu/ [tuhyahu] ‘they shot’

218

a/op ] / speak’ tu/

]

/ nayu ersed’ a/ ’

/ come’ a/ k’

’

u/ u] inging’ i/

ved’ /

behalf’ y/ y] e

/nw/ [nw] /minwa/a/ [minwa/a] ‘he already came’ /≠Ô/ [≠Ô] /ha≠Ôu/nømis/ [ha≠Ôu/nømis] ‘you did not read it’

ømis/ it’ /Nb/ [Nb] /maNba/ [maNba] ‘he goes’ /Nd/ [Nd] /tiNdiN/ [tiNdiN] ‘thick’

] //øNÔo/pya/ [/øNÔo/pya] ‘he is sleepy’ ticle

/Ng/ /tu] ‘he also went’ N// [/N] /kaN + /is/ [ka/Nis] ‘of the jaguar’ NdZ/ [NdZ] //aNdZi/u/ [/aNdZi/u] ‘goatee’ Nts/ [Nts] //aNtsoNu/ [/aNtsoNu] ‘he answered’ Ns/ [Ns] //aNsis/ [/aNsis] ‘lips’ Nh/ [Nh] /myaNhayu/ [myaNhayu] ‘he went for him’ Nm/ [Nm] //a/yaNmø/ [/a/yaNmø] ‘Ixhuata !n’ Nn/ [Nn] //aNnaka/ [/aNnaka] ‘mouth’ N≠/ [N≠] //aN≠a/ [/aN≠a] ‘left’

/Nw/ [Nw] //aNway/ [/aNway] ‘beard’ /Ny/ [Ny] /maNyahu/ [maNyahu] ‘they went’

/mp/ [mp] /≠empe/ [≠empe] ‘ever’ /mb/ [mb] /tsamba/ [tsamba] ‘he speaks’ /md/ [md] /tsamdøhka/opya/ [tsamdøhk ya ‘he meditates’ /mÔ/ [mÔ] /tsamÔo/yu [tsamÔo/yu] ‘he wanted to /mg/ [mg] /nømge/ [nømge/tu] ‘he also said’ /m// [/m] /kom + /aNø/ [ko/maNø] ‘to the post’ /mts/ [mts /tsamtsamnakyuy/ [tsamtsamnakyuy] ‘conversation’ /mf/ [mf] /pa!mfilo/ [pa!mfilo] ‘Pa!nfilo’/mh/ [mh] /tSamhayu [tSamhayu] ‘he told him’ /mn/ [mn] /tsamtsamnayu/ [tsamtsam ] ‘he conv/mw/ [mw] /tsamwa/ [tsamwa/a] ‘he already spoke/my/ [my] /nømyahu/ [nømyahu] ‘they said’ /nb/ [nb] /minba/ [minba] ‘he comes’ /nt/ [nt] /kwanto/ [kwanto] ‘how much’ /nd/ [nd] /mindø/u [mindø/u] ‘he intended to /nÔ/ [nÔ] /kenÔo/py [kenÔo/pya] ‘he wants to loo/ng/ [ng] /kenge/tu/ [kenge/tu] ‘he also looked/n// [/n] /pøn + /is/ [pø/nis] ‘of the man’ /nts/ [nts] /wantsø/y [wantsø/y ‘he quit s/ntS/ [ntS] /rantSo/ [rantSo] ‘ranch’ /ndZ/ [ndZ] /pøndZøk [pøndZøki] ‘image of a man’ /ns/ [ns] /winsa/u/ [winsa/u] ‘he revi/nh/ [nh] /kyenhayu [kyenhayu] ‘he looked on his

/nm/ [nm] /winma/Ngu [winma/Ngu ‘Huimango (plac name)’

/nn/ [nn] /masannø// [masannø/] ‘holy water’

/≠dZ/ [≠dZ] /ha≠dZøk [ha≠dZøkømis] ‘you did not do

/NÔ/ [NÔ/Nk/ [Nk] //øNkø/ [/øNkø] emphatic par

[Ng] /maNge/tu/ [maNge////////

219

/lp/ [lp] /kyulpa/ [kyulpa] ‘his blame’ /lk/ [lk] /balko!N/ [balko!N] ‘balcony’ /l// [/l] /perol + /is/ [pero/lis] ‘of the copper kettle’ /ls/ [ls] /kalseti!N/ [kalseti !N] ‘socks’ /ld/ [ld] /kaldu/ [kaldu] ‘broth’ /lm/ [lm] /kalmatsøhku/ [kalmatsøhku] ‘it calmed’ /ly/ [ly] /tSelyo/ [tSelyo] ‘Cliserio’ /rp/ [rp] /kwerpo/ [kwerpo] ‘body’ /rb/ [rb] //arbol/ [/arbol] ‘Christmas tree’

k/ orke/ [porke] ‘because’ /rg/ [rg] //alargatsøhku/ [/alargatsøhku] ‘it lengthened’

/rtS/ /martSa/ [martSa] ‘march’ /rs/ [rs] /marso/ [marso] ‘March’

/rn/ [rn] //ernesto/ [/ernesto] ‘Ernest’ rl/ [rl] /karlos/ [karlos] ‘Carl’

y/

/ws/ n’

/yt/ ment’

Wonderly, William L. 1951. Zoque II: Phonemics and Morphology. International

/rt/ [rt] /pwerta/ [pwerta] ‘door’ /r [rk] /p

/r// [/r] /lugar + /oyh/ [luga/royh] ‘at the place’ [rtS]

/rm/ [rm] //armas/ [/armas] ‘firearm’

//rw/ [rw] /sirwela/ [sirwela] ‘prune’ /r [ry] /koryentepø/ [koryentepø] ‘common’ /wt/ [wt] /flawta/ [flawta] ‘harmonica’ /wd/ [wd] /dewda/ [dewda] ‘debt’

[ws] /kawsa/ [kawsa] ‘legal accusatio/wm/ [wm] /rewma/ [rewma] ‘common cold’ /wy/ lready ran’ [wy] /powya/a/ [powya/a] ‘he a

[yt] /pleyto/ [pleyto] ‘argu/yc/ [yc] /kuycøm/ [kuycøm] ‘avocado’ /yh/ [yh] /hyuyhayu/ [hyuyhayu] ‘he bought for him’ /ym/ [ym] /teymø/ [teymø] ‘yonder’ /yn/ /reyna/ [reyna] ‘queen’ [yn] References:

Journal of American Linguistics 17, 105-123

220

lderso ry. 3rd ed. Oxford

very, oicing contrasts. Ph.D. dissertation. University of

Baertsch, Karen and Stuart Davis. 2000. Turkic C+/l/ cluster phonology. Paper presented

Baertsc h.D. dissertation. Indiana University, Bloomington.

cations in

Barker, M. A. R. 1964. Klamath grammar. University of California Publications in Linguistics.

Bekturova, A. Sh. and Sh. K. Bekturov. 1996. Manual of the Kazakh language. Almaty “Rauan”.

ger Ko·ppe Verlag.

nese and

Clemennd

on.

Crouch, M. 1994. The phonology of Deg. ms. Ghana: GILLBT.

B I B L I O G R A P H Y

A n, A. D. and Fahir Iz. 1985. The Oxford Turkish-English dictiona University Press.

Peter. 1996. The representation of vAToronto.

at MCWOP.

h, Karen. ierarchy. P

2002. An Optimality Theoretic approach to syllable structure: the split margin h

Barker, M. A. R. 1963. Klamath Dictionary. University of California Publi

Linguistics. Vol. 31. University of California Press: Berkeley and Los Angeles.

Vol. 32. Berkeley and Los Angeles: University of California Press.

Bitima, Tamene. 2000. A dictionary of Oromo Technical Terms: Oromo-English. Ko·ln: Ru·di

ho, Young-mee Yu. 1990. Parameters of consonantal assimilation in JapaC

Korean. Japanese/Korean Linguistics 1, 143-156. s, G.N. 1990. The role of the sonority cycle in core syllabification. In J. Kingston t

and M. E. Beckman (eds.), Papers in Laboratory Phonology I: Between the Grammar aversity Press, 283-333. Physics of Speech. Cambridge Uni

Côté, Marie-Hélène. 2000. Consonant cluster phonotactics: a perceptual approach. Ph.D. dissertati Massachusetts Institute of Technology. Côté, Marie-Hélène. 2003. Syntagmatic contrast in consonant deletion. ms.

221

Experimental Psychology 8, 153-162. Davis, Stuart and Seung-Hoon Shin. 1999. The syllable contact constraint in Korean:

Dickey, Laura Walsh. 1997. The phonology of liquids. Ph.D. dissertation. University of

Espy-Wilson, C. Y., S. E. Boyce, M. Jackson, S. Narayanan & A. Alwan. 2000. Acoustic

ts.

ujimura, O., M. J. Macchi, and L. A. Streeter. 1978. Perception of stop consonants with

.M.I.

Goodm University.

Heath, Jeffrey. 1981. Basic materials in Mara: grammar, texts and dictionary. Department of

rsity.

eath, Jeffrey. 1987. Ablaut and ambiguity: phonology of a Moroccan Arabic dialect. Albany: State

Hellberg, Staffan. 1974. Graphonomic rules in phonology: Studies in the Expression Component of

cademic Press.

Hudson, Grover. 1975. Suppletion in the representation of alternations. Ph.D. dissertation. UCLA.

Crowder, R. G. 1982. Decay of auditory memory in vowel discrimination. Journal of

An Optimality-Theoretic analysis. Journal of East Asian Linguistics 8, 285-312.

Massachusetts, Amherst.

modeling of American English /r/. JASA 108 (1), 343-356. Frisch, S., M. Broe, and J. Pierrehumbert. 1995. The role of similarity in phonotactic constrain

ms. Northwestern University.

Frisch, Stefan. 1996. Similarity and frequency in phonology. Ph.D. dissertation. Northwestern University.

Fconflicting transitional cues: A Cross-linguistic Study. Language and Speech 21, 337-346.

Gasparini, Armido. 1983. Sidamo-English dictionary. Bologna, Italy: E

an, Beverly D. 1995. Features in Ponapean phonology. Ph.D. dissertation. Cornell

Hahn, Reinhard F. 1991. Spoken Uyghur. Seattle and London: University of Washington

Press.

Harrell, Richard S. 1962. A short reference grammar of Moroccan Arabic. Washington, D. C: Georgetown University Press:

Hayes, Bruce. 1986. Inalterability in CV phonology. Language 62, 321-351.

Linguistics, Research School of Pacific Studies, The Australian National Unive

HUniversity of New York Press.

Swedish. ACTA Universitatis Gothoburgensis. Hooper, Joan B. 1976. An introduction to natural generative phonology. New York: A

222

),

1-402.

Hume, Elizabeth. 1998. The role of perceptibility in consonant/consonant metathesis.

Proceedings of the West Coast Conference on Formal Linguistics 17, 293-307.

.

Hume, Elizabeth and Keith Johnson. 2001. A model of the interplay of speech perception and phonology. In Elizabeth Hume & Keith Johnson (eds.), The role of speech perception

ech

n. Paper presented at the Second International Conference on Contrast in honology, University of Toronto.

Hume,

eption in shaping phonological assimilation rules. Language and Speech 35, 59-72.

Iverson, G. and Kee-Ho Kim. 1987. On word-initial avoidance in Korean. Harvard Studies in

Korean Linguistics. Vol. 2, 377-393. Cambridge, Mass.: Harvard University.

Iverson, G. and Hyang-Sook, Sohn. 1994. Liquid representation in Korean. In Kim-

Iverson, G. K. and S.-S. Lee. 1995. Variation as optimality in Korean cluster reduction. guistics 94.

Jakobi, Angelika. 1990. A Fur grammar. Hamburg: Helmut Buske Verlag.

hnstone, T. M. 1975. The modern Southern Arabian languages. Afroasiatic Linguistics 1,

1-29.

Hudson, Grover. 1995. Phonology of ethiopian languages. In John Goldsmith (ed.), Handbook of Phonological Theory. Oxford: Blackwell. 782-797.

Humes, L. E. 1985. Psychoacoustic foundations of clinical audiology, In Katz (ed.Handbook of Clinical Audiology. 94-115.

Hume, Elizabeth, Jennifer Muller and Aone van Engelenhoven. 1997. Nonmoraic geminates in Leti. Phonology 14. 3, 37

Hume, Elizabeth. 1997. Vowel preservation in Leti. Oceanic Linguistics 36. 1, 65-101.

Hume, Elizabeth. 2001. Metathesis: formal and functional considerations. In E. Hume, N

Smith & J. van de Weijer (eds.), Surface syllable structure and segment sequencing. Leiden: Holland Institute of Linguistics.

in phonology. Academic Press. 3-26

Hume, Elizabeth and Keith Johnson. 2002. The impact of phonological contrast on speperceptioP Elizabeh and Georgios Tserdanelis. 2002. Markedness of consonant place reconsidered: the case of Sri Lankan Protuguese creole. ms. Ohio State University.

Hura, S. L., Lindblom, B., and Diehl, R. 1992. On the role of perc

Renaud, Young-Key (ed.), Theoretical Issues in Korean Linguistics. 77-100.

Proceedings of Eastern States Conference On Lin

Johnson, Keith. 1997. Acoustic and auditory phonetics. Blackwell Publishers Ltd.

Jo

223

ality alifornia, Los Angeles.

articulatory Constraints. Proceedings of WCCFL 14, 221-237.

Japanese/Korean Linguistics 9, 114-124.

Karlsso h: an essential grammar. London and New York: Routledge.

Kimball, Geoffrey D. 1991. Koasati grammar. Studies in the Anthropology of North American Indians. University of Nebraska Press.

Kiparsky, Paul. 1985. Some consequences of lexical phonology. Phonology Yearkbook 2,

Kohler, K. 1990. Segmental reduction in connected speech: Phonological facts and phonetic explanations. In W.J. Hardcastle & A. Marchal (eds.), Speech Production and

luwer Academic Publishers. 69-92

Krause, Scott Russell. 1980. Topics in Chukchee phonology and morphology. Ph.D. dissertation.

Kruege Ladefoged, Peter. 1982. A course in phonetics. 2nd ed. Harcourt Brace Jovanovich, Inc.

bridge, Massachusetts: Blackwell Publishers Inc.

Levin, J eometry. ms.

Jun, Jongho. 1995a. Perceptual and articulatory factors in place assimilation: an OptimTheoretic approach. Ph.D. dissertation. University of C

Jun, Jongho. 1995b. Place assimilation as the result of conflicting perceptual and

Jun, Jongho. 1996. Place assimilation is not the result of gestural overlap: evidence from Korean and English. Phonology 13.

Kang, Hyunsook and Seo-Hwa Hahn. Canalization before a liquid in Yonbyon dialect of Korean.

Kara, Da!vid Somfai. 2002. Kazak. languages of the world/materials 417. Lincom Europa. n, Fred. 1999. Finnis

Kawasaki, Haruko. 1982. An acoustic basis for universal constraints on sound sequences. Ph.D.

dissertation. University of California, Berkeley.

85-138.

Speech Modeling. Dordrecht: K

Kornfilt, Jaklin. 1997. Turkish. London and New York: Routledge.

University of Illinois at Urbana-Champaign. r, John R. 1962. Yakut manual. Bloomington, IN: IU Publications.

Ladefoged, Peter and Ian Maddieson. 1996. The sounds of the world’s languages. Cam

uliette. 1988. A place for lateral in the feature g Lewis, G. L. 1975. Turkish grammar. Oxford University Press, Ely House, London W.

224

r of Peter Ladefoged. Academic Press. 157-168.

hD. Dissertation. Indiana University.

Lomba 105.

1-2,

McCarthy, John and Alan Prince. 1993. Prosodic morphology I: constraint interaction and satisfaction.

r

cCarthy, John and Alan Prince. 1995. Faithfulness and reduplicative identity. In Jill

Mielke, Jeff. 2003. The interplay of speech perception and phonology: experimental

ceanic Linguistics

15, 210-255.

Muller,ity.

of Indonesia, No. 6. W.A.L. Stokhof (ed.). Department f Linguistics, Research School of Pacific Studies. The Australian National

Nadzhip, E. N. 1971. (translated from Russian by D. M. Segal) Modern Uigur. Central

Newton ive interpretation of dialect: A study of modern Greek Phonology.

ambridge: Cambridge University Press.

Newma kuts language of California. New York: Viking Fund Publication in nthropology, no.2.

Lindau, Mona. 1985. The story of /r/. In Victoria A. Fromkin (ed.), Phonetic linguistics: essays in hono

Lloret-Romanyach, Maria-Rosa. 1988. Gemination and vowel Length in Oromo Morphophonology. P

rdi, Linda. 1995. Why place and voice are different: constraint-specific alternationsin Optimality Theory. Rutgers Optimality Archive #

Lynch, John. 1983. On the Kuman ‘liquids’. Languages and Linguistics in Melanesia 14, 98-112.

Report no. RuCCS-TR-3. New Brunswick, NJ: Rutgers University Center foCognitive Science.

MBeckman, Laura Walsh Dickey, and Suzanne Urbanczyk (eds.), University of Massachusetts Occasional Papers in Linguistics 18, 249-284.

evidence from Turkish. Phonetica 60, 3, 208-229.

Mithun, Marianne and Hasan Basri. 1985. The phonology of Selayarese. O

Jennifer S. 2001. The phonology and phonetics of word-initial geminates. Ph.D. dissertation. The Ohio State Univers

Murray, R. W and T. Vennemann. 1983. Sound change and syllable structure in Germanic phonology. Language 59, 3, 514-528.

Nababan, P.W.J. 1981. A grammar of Toba-Batak. Pacific Linguistics Series D. No. 37. Materials in LanguagesoUniversity.

Department of Oriental Literature.Nauka Publishing House: Moscow.

, B. 1972. The generatC

n, Stanley. 1944. YoA

225

Press.

hala, John J. 1981. The listener as a source of sound change. In Masek, C. S., R. A. avior:

Chicago Linguistic Society. Chicago: CLS. 178-203.

Ohala, aboratory Phonology I: Between the Grammar and Physics of Speech. Kingston, John &

hala, John. 1992. Alternatives to the sonority hierarchy for explaining segmental sequential constraints. Chicago Linguistics Society: Papers from the Parasession on the

en the grammar and the physics of speech. Cambridge: Cambridge University Press. 87-94.

Okrand mmar. Ph.D. dissertation. University of California, Berkeley.

rince, Alan S. and Paul Smolensky. 1993. Optimality Theory: constraints interaction in generative

er for Cognitive Science.

yun, Kwang-soo. 1987. Korean-Swedish interlanguage phonology. Ph.D. dissertation. Institute of

Rehg, Kenneth L. and Damian G. Sohl. 1981. Ponapean reference grammar. Honolulu:

ice, Keren and Peter Avery. 1991. On the relationship between laterality and coronality.

ice, Keren. 1992. On deriving sonority: a structural account of sonority relationships.

Phonology 9. 61-99.

Roberts-Kohno, Rosalind Ruth. 2000. Kikamba phonology and morphology. Ph.D. dissertation. The Ohio State University.

Nicklas, Thurston Dale. 1974. The elements of Choctaw. Ph.D. dissertation. University of Michigan.

Odden, David. 1996. The phonology and morphology of Kimatuumbi. Oxford: Clarendon

OHendrik, & M. F. Miller (eds.), Papers from the Parasession on Language and Beh

John. 1990. The phonetics and phonology aspects of assimilation. Papers in LMary E. Beckman (eds.). Cambridge University Press. 285-275.

O

Syllable. Chicago: CLS. 319-338.

Ohala, John. 1993. The perceptual basis of some sound patterns. In D. A. Connell & A. Arvaniti (Eds.), Papers in laboratory phonology, Vol. 1: Betwe

, Marc. 1977. Mutsun gra

Poppe, Nicholas. 1963. Tatar manual. Bloomington, IN: IU Publications.

Pgrammar. Report no. RuCCS-TR-2. New Brunswick, NJ: Rutgers University Cent

P

Oriental Languages, University of Stockholm.

Reed, C. E. 1947. A survey of pennsylvania German phonology. Modern Language Quarterly 8,276-289.

University Press of Hawaii.

RPhonetics and Phonology 2. The Special Status of Coronals: Internal and ExternalEvidence. 101-124.

R

226

97-425.

Saeed, John Ibrahim. 1993. Somali reference grammar. 2 ed. Kensington, Maryland: Dunwoody Press.

Sapir, J. D. 1965. A grammar of Diola-Fogny. London: Cambridge University Press.

.lrz-muenchen.de/~wschulze/udinhalt.htm)

rnal of the Phonetic Society of Japan. Vol. 5. No. 3. 49-63.

st

iences (ICSS). 224-235.

. 315-327.

,

ngeles.

Steridade, Donca. 1995b. Positional neutralization. ms. University of California, Los Angeles.

ngeles.

Steriade, Donca. 2001. Directional asymmetries in place assimilation: a perceptual account. In E. Hume & K. Johnson (eds.) The Role of Perception in Phonology. New York: Academic Press. 219-250.

Steriade, Donca. 2002. The phonology of perceptibility effects: the P-map and its consequences for constraint organization. ms.

Steriade, Donca. 2003. Knowledge of perceptual similarity and its phonological uses:

evidence from half-rhymes. Proceedings of the 15th International Congress of Phonetic Sciences. 363-366.

Stroomer, Harry. 1995. A grammar of Boraana Oromo (Kenya): phonology, morphology, vocabularies.

Köln: Rüdiger Köppe Verlag.

Rose, Sharon. 2000. Epenthesis positioning and syllable contact in Chaha. Phonology 17, 3

nd

Schulze, Wolfgang. 2001. An online grammar of Udi. Ludwig-Maximilians-

Universita·t Mu·nchen. (http://www

Seo, Misun. 2001a. A perception-based study of sonorant assimilation in Korean. OSUWPL 55, 43-69.

Seo, Misun. 2001b. A cross-linguistic study of the perception of sonorant sequences. Jou

Seo, Misun. 2002. The realizations of the Korean liquid in onset position. Proceedings of the 1 International Conference on Speech Sc

Seo, Misun. 2003. A perception-based study of sonorant assimilation. Japanese/KoreanLinguistics 11

Sim, Ronald J., 1981. Morphophonemics of the verb in Rendille. Afroasiatic Linguistics 8, 1

1-33.

Steridade, Donca. 1995a. Licensing retroflexion. ms. University of California, Los A

Steriade, Donca. 1997. Licensing laryngeal features. ms. University of California, Los A

227

Sulkala, Helena and Merja Karjalainen. 1992. Finnish. London and New York: Routledge.

Trefry, D. 1969. A comparative study of Kuman and Pauaian. Pacific Linguistics. Series B. No. 13.

Tserdanelis, Gerogios. 2001. Manner dissimilation directionality in Greek. OSUWPL 55, 172-199. Van Engelenhoven, Aone. 1995. A description of the Leti language as spoken in Tutukei. Ph.D.

dissertation. University of Leiden.

Vennemann, Theo. 1988. Preference laws for syllable structure. Berlin, New York, Amsterdam: Mouton de Gruyter.

Winters, Steve. 2003. Empirical investigations into the perceptual and articulatory origins of cross-

linguistic asymmetries in place assimilation. Ph.D. dissertation. The Ohio State University.

Wonderly, William L. 1951. Zoque II: Phonemics and morphology. International Journal of American Linguistics 17, 105-123.

Wordick, F. J. F. 1982. The Yindjibarndi language. Canberra: The Australian National University.

Wright, Richard A. 1996. Consonant clusters and cue preservation in Tsou. Ph.D. dissertation.

University of California, Los Angeles.

Zaborski, Andrzej. 1986. The morphology of nominal plural in the Cushitic languages. Vero·ffentlichungen der Institute für Afrikanistik und Agyptologie der Universität Wien. Wien.

Zorc, R. David and Madina M. Osman. 1993. Somali-English dictionary with English index. Kensington, Maryland: Dunwoody Press.