Generative phonology as applied to vietnamese dialects (a study based en middle vietnamese comparing the three major dialects of modern vietnamese) thesis of doctorate of linguistics

2 A modified Middle Vietnamese system is used asthe basis for the sound system, and the rules will pro-vide the information for the changes which have takenplace to produ~e the respectiv

FACULTY OF LETTERS

Univers it y of Saigon

V l1 1 f - t Q- 3 Zb

~ :T2-62f

GENERATIVE PHONOLOGY

as applied to VIETNAMESE DIALECTS

GENERATIVE PHONOLOGY

as Applied toVIETNAMESE DIALECTS:

A Study Based onMiddle Vietnamese,Comparing theThree Major Dialects

of Modern Vietnamese

Thesis submitted for Master's Degree

in the field of linguistics

PREFACE

-This thesis is written in partin 1 fulfillment ofthe requirements for a Master's degree in Linguisticswith the Department of English of the Faculty of Letters,Saigon University I trust that i t will also be of

value to all linguists in this area Much informationabout the theory used has been included in order to

make i t more readable for those not familiar with

generative phonology

I want to express my deep appreciation to Professor

D~ Kh~nh Hoan for his assistance in preparing and

sub-mitting the necessary forms for application and also

for arranging for the thesis committee

For assistance in some of the areas of the namese"dialects, I am grateful to Dr David Thomas

Viet-and Mr L~ Nggc Tr~ for their advice and correction

Dr Thomas also served as the advisor for this thesisand as such gave valuable assistance

Dr Kenneth Gregerson assisted in areas where

problems arose in the application of the

transforma-tional theory to the Vietnamese phonological system

I submit this thesis to the Faculty of Letters

with my highest regards

Barbara Friberg

Saigon, January 1972

TABLE OF CONTENTS

-PAGEI" Introduction 8 • • • • • • • • • • • • • 1

II Body

A Introduction to Transformational'

Theory • • 7

B Vietnamese Initial Consonants • 11

C Vietnamese Final Consonants • • • • 16

D. Vietnamese Vowels • • • • • • • 46

E Vietnamese Tones • • • • • • • • • 70III Summary and Application . 74

Notes 77

Bibliography • • • • • • • • • • • • • 81

i i

I Introduction

In the field of Vietnamese dialects, a variety of

,discussions can be found, from Cadiere's articles on

~high and low Vietnamese to the listings of Hoa and

Thompson and the word compilations and dictionary being

I

undertaken by the research center under Mr Tr~. But

to this point very little has been written about thephonological changes which are taking place across thedialects, and virtually nothing has been done in apply-ing the generative/transformational model to the Viet-namese dialects This discussion of ~he dialects willmake no attempts at being exhaustive - such a studywould require a myriad more research and material Atthe same time, trying to take into consideration asmany aspects of both the Vietnamese dialect studies andgenerat~ve phonology as possible, this paper will give

an over-all picture of what.is being done in the fieldand areas for further study Generative sound changeand spelling rules and feature notations will be used

as the terms for comparison These rules and featureswill primarily be those outlined by Chomsky and Halle

in SOUND PATTERNS OF ENGLISH, reinforcing their notion

of universal features with specific language data and

noting situations where this study indicates Q need toalter these proposed features Without manipulatingfeatures or data or producing 'ad hoc' rules, the sim-plicity measure of generality will be employed to dis-cover which rules most accurately and simply describethe data That is, rules will be formed (and featuresused) which ,qould cover the most general informationwithout redundancy or omission of actual sound patterns

of the language

2

A modified Middle Vietnamese system is used asthe basis for the sound system, and the rules will pro-vide the information for the changes which have takenplace to produ~e the respective sound systems of themajor, present-day dialects of Viet Nam There will

be only brief discussions of the many smaller dialects

-an area of study which could well serve as the basisfor some challe~ging and interesting research M~dng(and thus Proto~VietM~dng) will not be considered inthis discussion

The task of deciding which dialects are indeed themajor ones and the most generally used in problematic

,Thomp~on gives listings from specific towns and hoa

gives an oversimplified listing for South, Central and

North This discussion will involve the dialects thatare commonly considered as Northern, Central and South-ern with comments on the variant dialects in a few cases.The Central dialect covers a large and ~aried region,

so the basic discussion will include the speech of the

Da Nang dwellers, and notes given for the rural areasand other provinces in the area In a given region

dialect differences may be as much dependent on familybackground and education or a formal versus informal

situation as on geography One area may also containspeakers of all or many of the country's dialects

This is particularly true in Saigon, where people haverecently come from all areas of the country

For the following discussion, these dialect areasare defined: (1) Northern - considered that which iscommonly accepted in South Viet Nam as Northern and

usually spoken in Hanoi (noting here that.there areseveral variant dialects in the North today.) Many

of the speakers of Northern have come from North VietNam before or during 1954

are based on this dialect

"

Nguyen Dinh hoa's worksand serve as an accuratedescription (2) Central - As noted above, this is

the most diverse area, with a different dialect for

- 'J

-many of the provinces and population centers of the

Central region from IDla Trang to Quang Tr~. Cadiere'sdescription of Low Vietnamese from the Binh ~inh areawill be mentioned, as this seems descriptive of many

of the unusual sound changes of that region Generally,however, only the most commonly found sound changes

will be considered for the Central dialect

(1) Southern - This dialect is spoken by most of theVietnamese inha"bi tants in the delta, including Saigon,where this dialect will be found to be spoken by themajority of the people in other than formal, learnedsituations

These three dialect divisions should not be sidered as set boundaries, nor as a continuum of soundchange The Southern dialect seems to have developedfrom Middle Vietnamese separately; developing sound

con-patterns which are not seen in Northern or Central

Similarly, Northern and Central developed dialects

from Middle Vietnamese Similarities can be seen due

to lack of isolation of any of these areas The entirefield of dialect development is yet another area of

research which might well be undertaken by someone

The body of this thesis will be divided into five

sec-tions, each being considered separately, yet all as an

integrated picture of the application of the genarative

theory to the Vietnamese sound system as manifested in

the various dialects The parts will be presented in

this order: A An introduction and background material

of the generative theory B The initial consonants of

Vietnamese This section will include definitions and

clarifications of the various rule and feature n~ta­tions which will be used in the entire paper C F~nalConsonants These are placed before the vowels, be-

cause of their fewer variances and thus the ease of

handling the changes in the three dialects Also, i t

is proposed that diachronically the consonant changes

were affected first by the vowels which preceded them;

the vowel changes coming at a later period, some of

them resulting from the changes in the final consonants

D Vowels Although the vowel system must be presented

in Section C in order to understand the final consonant

changes, this discussion will be following that for

final consonants Since this is the area of greatest

divergence and inconsistency, i t will also include

com-me~ts on the diverse Central dia1edB and remarks on the

'5

-limitations of the feature system that has been presented

by Chomsky and Halle E Tones Although Chomsky andHalle do not consider this area in their presentation,

i t can readily be considered in the same phonologicalframework as the vowels and consonants, and simple fea-ture changes can be shown to describe the varying tonemanifestations of the dialects

The thesis will conclude with a summary of the

rules and their application to the Vietnamese language.Although the theory and basis of this thesis are those

of the generative phonological framework, much of themore familiar terminology will be used throughout thepaper in orde~ to facilitate the discussion and the

reading of i t

IIA THE GENERATIVE THEORY

In American linguistic circles, generativ~ logy is presently in a state of rapid development

phono-Every aspect of i t is being attacked, modified or pirically supported Much of the theory stems from thedistinctive feature system first set forth by Jakobson.(Jakobson and Halle:1956 ; Jakobson, Fant and Halle:195l):Halle then applied this system in his acclaimed work onRussian phonology (THE SOUND PATTERN OF RUSSIAN:1959)

em-He later shows the relationship between phonology andgrammar <"Phonology in Generative Grammar' :1962)

vlhile Halle concentrated on phonology Bach and Chomskyset down the bases for transformational grammar (Bach:i964 and Chomsky:1957 and 1965)

I t is Chomsky and Halle's combined efforts in THESOUND PATTERN OF ENGLISH that serve to formalize some

of the generative phonological theory, but.even this

work is sketchy and incomplete in sections Postal

tried to defend the position of generative phonology

and systematic phonemics while devoting much space topointing out the fallacies of the autonomous (taxonomic)system (Postal:1968; Chafe:1970) Harm~ INTRODUCTION

7

-TO PHONOLOGICAL THEORY attempts to bring together thenotions of rUles, ordering, distinctive features, etc.,

in an introductory book on generative phonology though his terminology is not always comparable to that

Al-of Chomsky and Halle, this work serves as a thorough,understandable foundation for the theory

On the first page of this book Harms defines thearea which geuerative phonology covers:

" ••• those' components of generativegrammar which serve to provide the phonetic

representations of utterances in any given

language Thus i t includes the phonological

form of morphemes a~d morphs listed in the

lexicon - also those inserted via certain

rules, the morpheme-structure rule component

and the phonological rule component Ulti~

mately, however, the generative phonology

seeks to provide a general theory explaining

the competence of the native speaker in the

sounds.of his language."

" The fundamental unit of generativephonology is the distinctive feature, although

various nonphonological features, boundary

symbols, and syntactic bracketing also play

an important role The phoneme -

specifi-cally, the systematic phoneme - has clearly

-redundancies for any given sound in any given ment.

environ-The universal phonetic system is an area which isalso receiving much attention The defining of features

which will describe any sound in any language

consist-ently has been the goal of generative phonologists

Harms (1969) describes some of Jakobson's articulatoryfeatures Chomsky and Halle (1968) present a good basis(much expanded and altered by present-day transforma-tionalists) for these universal features

Quoting again from Harms (p 12) as he clarifiesthe systematic phonemic/phonetic issue:

"The primary aims of generative phonologyare to provide phonemic representations of

morphemes and a series of ordered rules tha~,

together with information about boundary

phenomena, (1) adequately express the

phono-logical generalizations of the language and

(2) at the same time determine the phonetic

form of all utterances in the language."

Simplicity is the criteria for the best rules - asimple rule coupled with one that covers the most generalarea of sounds Harms (p 11) also clarifies this

simplicity measure:

"The role of the simplicity metric is

to devise a method of counting the elements

present in the morphemes and rules in such a

way that the more highly valued phonological

9

-sys~em Cor a given language (in terms oC its

generality) will always be the more economical

one • Thus the simplicity metric Corces the

analysis in the direction oC greater

gener-ality."

These are only the basic issues considered Much

has been written about the rules, their order, notation,

and componants, and the theoretical Cunction oC the

distinctive Ceatures to eliminate redundant inCormation

as well as diachronically inaccurate distinctions The

authors and pUblications mentioned in the preceding

dis-cussion should be studied Cor a complete picture oC the

theory and application oC generative phonology

More speciCic Ceature descriptions and rule

nota-tions will be given in the Collowing section, as i t is

seen to apply in the Vietnamese sound system

lIB VIETNAMESE INITIAL CONSONANTS

When illustrating dialect differences, the z/y

distinction between the North and the South will often

be cited as the sound change; and i t is this that is

most noticeable to the non-native when first hearing thespeech of a Vietnamese These i~itial consonant changesare an obvious dialect variation Looking at the dis-tinctions between the three dialects from a generativeviewpoint, most of the changes occur in one area of soundproduction - voiced, lax obstruents - and these changescan be explained in terms of simple, general rules

Before looking at these distinctions, however, the

framework for the initial consonants must be given, andsome information about the rules and.features which

will be used to describe the changes

It is generally ac~epted that sounds can be divided

·3

into two major classes - obstruents and sonorants

According to Chomsky and Halle (1968:p 102) sonorantsare sounds produced with a vocal tract cavity config-uration that makes spontaneous voicing impossible

Technically this is a narrowing of the air passage to

a critical point, beyond which voicing cannot take

place Acoustic research has shown that vowels, glides,

11

-nasal consonants and liquids are sonorant, while stops,fricatives, and affricates have been found to be obstru-ents The application of this feature in Vietnameseresults in natural classes of sonorants, including thenasals, vowels, lateral and glides; and the obstruentsincluding all other consonants.

Traditionally five or more points of articulationhave to be referred to along with voicing, tenseness,etc., in order to discuss the Vietnamese initial conso-nants (these being labial, alveolar, retroflexed, pala-to-alveolar, velar, and glottal) Chomsky and Halle(1968:p 299) have posited three parameters which moreaccurately show distincive features for all the points

of articulation, and which eliminates a lot of fluous information about the sounds Using these para-meters allows for natural classes to be readily dis-cussed in terms of.actual distinctions and simplifiesthe feature system At this point these parameters will

super-be considered in detail for clarity

Coronal - noncoronal: Coronal sounds are produced

with the.blade of the tongue raised from its neutralposition This includes the traditional dental, alveo-lar, and palato-alveolar sounds Noncoronal sounds then

~re those produced with the blade of the tongue in theneutral position - these including the tradition labial

(lip articulated) and velar and uvular (tongue-root

articulated) sounds

Anterior - nonanterior: Anterior sounds are produced

with an obstruction (lips, tongue, etc.) that is located

in front or the palato-alveolar region of the mouth;

such as the labials, dentals, and alveolars

Nonante-rior sounds are produced without such an obstruction;

the palato-alveolars, velare, uvular or pharyngeal

sounds (Chomsky and Halle 1968:p 104)

By considering combinations of these two features,

the main points of articulation can be discussed in

terms of the following feature units:

labial

alveolar

alveolar

palato-velar,uvular

'I

' anterior-I:_+c o r o n a l .J

' "a n t e r i o r -

-; - coronal !

In many languages retroflexed consonants can be

considered as simply ~~~;J ,since they are not

- 11

-distinc~ from nonretroflexed palato-alveolar sounds.

In Vietnamese however, th ~s distinction is readily

noted, so a third cavity feature must be considered

Distributed - nondistributed~ This parameter's festation may vary from language to language, but is

mani-defined for its universal characteristics by Chomsky

and Halle (1968:p 112+) Distributed sounds are

pro-duced with a constriction that extends for a longer

distance along the direction of the air flow than the

nondistributed sounds Thus a retroflexed consonant

(represented by a dot under the consonant [!]) can bedistinguished from a nonretroflexed sound at the same

point of articulation: the retroflexed being r-disr]and the nonretroflexed [+disr]

This parameter proved more valuable in the

Vietnam-ese sound system than first appeared because of its

use-fu1ness in describing the lateral (1] At first the

lateral was problematic, because all descriptions give

i t as a sonorant, yet i t did not group with the glides,

nor could i t readily be seen to group with the rrare Ls ,

although this would be its natural class Some of the

phonetic descriptions give [1]

tinguishing i t from all other

simply as lateral,

dis-4

consonants but no reason

presents itself for distinguishing lateral from other

classes of sounds in this system The distributed

-nondistributed distinction which has already been pointed

out for the obstruents can be seen to extend to the

son-orants as well ; [I} bei:g considered as r -disr] and [n}

as (+4isr] Articulatorily the tongue-tip

constric-tion of the [1] is slight, while the constriction of

the [n] along the direction of the air flow is greater

High - nonhigh: This feature, which will become more

obvious in the vowel and tone systems,5 is introduced

at this point to show the distinction of the glottals

considered

from the velars Both [?] and [h] are

I ::~~ J ;distinguished from each other

by

tenseness (following paragraphs) Velars then have the

added feature of [+high]; this describing the tongue

root position in the mouth Chomsky and Halle (1968:

pp 10+) show these as glides, but in

function phonetically only as initial

Vietnamese they

6

consonants

Although the glottal stop is mentioned here, i t will

not have a bearing on the analysis, since i t is not

distinctive in this phonemic system, but occurs

pre 1'5

-dictably before all initial vowels Its placement in

MiddleVietnancse is due to its phonemic function today

in Mddng.

The one ~ann er of articulation feature needed toshow distinctions within the class of obstruents is

tense - nontense This distinction corresponds quite

closely with Thomp$on's fortis - lenis distinctions for

initial consonants (196~:pp. 19,46) Because of thevariance in actual physical distinctions of the tense -

nontense sounds, no detailed explanation will be given,

but the distinctive function within the system is shown

i~ Chart A.7

The source feature of voicin~ will serve to tinguish the sounds in the tense - nontense classes

dis-from each other Among the sonorants, neither tenseness

nor voicing is a distinctive feature, since the points

of articulation and consonantality serve to distinguish

them sufficiently in this system

Chart A gives the sounds written with symbols

which have a phonemic significance in this system and

shows their distribution according to the features just

discussed Chart B gives corresponding Vietnamese

spell-ings for clarity

-i-In the discussion following the Middle Vietnamese

phonemes will be written between slashes - /n/i the

17

-sound changes for the dialects will be written phonet

-ica11y between square brackets - [n] , and the Vietnamese

spelling symbols will be without special notation - ~i.

In order to illustrate the distribution of th e

features in Vietnamese, Chart C wi l l give th e matrix

for the initial consonants with relevant featuros being

shown as + or - , and non-rel evant features left

Before showing how transformational rules can be

applied to the various Vietnamese dialect changes, i t is

necessary to discuss briefly the rule notation that will

be employed in the succeeding discussions Chomsky and

Halle give a very clear discussion of rules (1968;pp

330+), and this format will basically be used for this

paper along with the notations given by Harms (1968)

The rules have been generally presented in terms

symbolized by this formula:

where A and B represent distinctive feature units for

a given phonological system All units of.features

arc enclosed in square brackets - [ J ~ The

arrow st.ands for "is act.ualized as", and the diagonal

slash means "in t.he context of"; with X and Y ing respectively the l eft- an d right.-hand environment.s

represent-X and Y can be in t.erms of dist.inct.ive feature units orst.rings of units, or 'word' boundaries9 (which are indi-

cated by I), or they may ~ ) e unspecified, indicatingthat the environment is inconsequent.ial

The following guidelines have been adhered to when

writing the rules (and t.hus reading and applying them)

for Vietnamese dialect changes:

(1) Rules may eit.her add features or change them

B may express a sign change of any feat.ure in A; or B

19

-may add a feature which is not in A.

(2) Any feature which will involve a sign change

in B must be first given in A

(3) If n given feature is not necessary to tinguish the unit from others under discussion, i t neednot appear in A, but i t will be considered as having thesame value in B unless the presence of another fc~ture

dis-indicates a different value for the first feature

(4) Only features which have been designated asdistinctive for the phonological system can appear in

A, but in adding features to B, distinctive featureswhich were not used in the phonological system, but

which are re~lized by universal rules or filler rules,may be added

In the phonological system under discussion thefeature [distributed] is relevant only for [+antJ

+corsounds, so in discussing any of the other points of

articulation i t is not necessary to mention the feature[distributed) •

Although most filler rules will be omitted in thisthesis, a few examples of the more useful ones will bementioned here for clarity of discussion In this

system all consoanants are [-vocalic] and all vowels

are [ -consonantal); so these features need not be

indi-cated for initial distinctions, but can.be later added

10with filler rules such as the following

This does not eliminate those that are marked as

[:~~:~J which is the case with the y- and w-glides

Two more examples will show how the matrix given

for initial consonants can be filled in:

Two other notational devices will also be found in tht~

paper: curved braces [ ] and the variables o<fJ J

-Curved braces are used a~ an abbreviatory device, with

the meaning of either/or They can be used to give

alternatives in terms of single features, feature units,

or strings of units The variables are also an

abbre-viatory device; these used to indicate that eith~r a

+ or val~e can be considered as applicable for a

given feature Within a rule using the same variable

for different features indicates that the same value

must be considered for both features in a particular

application A negative variable (-~) indicates thatopposite values must be considered for the tWQ features -when Qne feature is + the other must be • and visaversa The use of such abbreviatory devices allows forgreater generality while achieving greater simplicitywhen describing a sound change of the system~ll

The major Vietnamese dialects show a remarkable

correlatio~ in sound changes which involve the initialcon.onants What appears to be completely unrelated

can be shown to be a change, addition or shifting of

one or two features

Rule I

The most obvious sound change comes witb the duction of the Middle Vietnamese Idl and Ill. The

pro-following rules and discussion will show how these

changes have taken place by the feature changes and

their correlations Widely used today in the Northerndialect is the pronunciation of Id/, I~/, and Ir/ -

d, gi, r - all as the same [z) ~12 These three

conso-nants form a class of nontense, voiced coronals For

Northern this shift from t~e pronunciation in Middle

Vietnamese can be shown by adding the feature [+strident]13and shifting all three sounds i~to the [+anterior]

position, as is ahown in Rule I Distributionally thesenow take the position of the Id/, but because of the

addition of the feature [+strident], they become a [z]

Rule I: Northern [z] shift - (d, gi, r ~ [z] )

+cor

"'ant+vd-tns,xd i s r

)

[

+a n t J+s~ri+dJ.sr

I #

,

Ly (1968:p 20) discusses these sounds for a dialect

of Hanoi that is still heard today In this case the

I~I remains as i t is in Middle Vietnamese, gi - I~I

becomes [z] and Id I becomes [~]. In this case (+stri]

is also added and variables are used to show the

oppo-site values for [distributed] and [anterior]

Rule Ia: ( /d ,~ [~J i /zI ~ [z])

+corC(ant+vd+disr-tns

-Note, in this case, i f in A the feature unit toins [+ant] than B must contain [-disr] ; if A contains[-ant], B must be specified as [+disr].

con-RUle II

The Central shift of this class of sounds is seen

in d and gi becoming a y-glide These Northern andCentral changes from Middle Vietnamese can reasonably

be considered as separate events, and not as one suIting from the other I~ / remains unchanged in thiscase The feature unit [-vocl1 presupposes a change

re consJ

in B to [+sono] from [-sono] in A

Rule II: Central [y] shift - t d, gi~ [y])

Rule I:J:I

-sono+cor+disr-tns+vd

dia-tion is that Southern evolved separately The d and

gi as well.as the labial !v' all become a y-glide inmany areas But there are a number of areas where the

Iv l undergoes a different type of change than the Idl

and I l l Initially in this case the labial becomes

palatalized so that a [vY] or even a [bY] will be heard.For the Southern shift, then, Rule II will be applied

as i t is in Central, but Rule III sho~ the possibility

of Ivl going to either a y-glide or a palatalized ial · The curved braces in B indicate that either the

lab-feature [+high] is added to the feature unit of A or

the obstruent in A becomes a glide

Rule III: South IvI shift - (Iv1 + Iy] .Iv Y], rbY] )

-sono+ant-cor+vd-tns

)

[

- c on s]

-vocl+cor

4

In order to illustrate these three rules as they

are applied to the Viutnomese dialects, exanplos with

the same vowel which shows no dialect change can be

given and their initial consonant changes shown AI~houghthe phonetic symbols written to this point will never

be written on any level of analysis (-the rules will

suffice ·t o show the sound changes-) they nre written

here in order to show the pronunciation variations

from dialect to dialect

25

-Rules I - III will be illustrated using these Vietnamesewords:

pro-nunciations are evidenced:

(i) Iza.! ) [ zo]

Iwords' are mani:fested as :follows:

in the Southern dialect:

(iV~ Iva I )

Any other examples which have the initial consonantBpel~ing qf d, gi, r or v will result in like sound change

Ruie IV

The second large shif: in Northern nppears in thevoiceless palato-alveolar and retroflexed obstruents.The x Is/ becomes [s] and the tr It I becomes [t5]

a reversal of positions in the system Usually theseare considered as separate, unrelated phenomena, butwith the application of the generative model this re-versal can be shown with one rule which changes thevalue of [anterior], drops the retroflexion and addsthe feature of stridency (for preciseness of distinc-

[a]. Again retroflexion is dropped, stridency addedand the positions of [anterior] exchanged

27

-Rule V: Southern/Central Spirant Flip:

+cor-tns-vd(J(ant-<Xdisr

-\~ -::::~:;l

+strjj

Simple examples sbow the application oC Rules IVand V Cor the dialects Considering these Vietnameseexamples:

(v) xa(vi) sa(vii) tra

'Car''to Call''to investigate'

Application oC Rule IV produces the Northern ciations as Collows:

pronun-(v) laa l -7 rsa]

(vi) /~a/ ~ r~a] 17(vii) I~a/ ~ rtsa]

Rule V's application results in these phonetic

pro-ductions Cor the Southern and Central dialects:

(v) !sa l - 4 [ sa]

(vi) l~al -4 rsa]

(vii) / ~ a I ~ r~a]

consonants when they occur

initially before a w-glide In the Da Nang area the Ihl

alone is dropped, as this first part of the rule shows:

Rule VIa: h-drop (pre-w-glide)

r- a n tj

-cor-hi

null / #

-[

- VOCl ]-cons-cor

In the Saigon area the Ikl will often be dropped

in this same position, but the other velars will still

be heard before the w-glide

Rule VIb: k-drop (qu- before w-glide)

-sono+hi-ant-cor+tns-vd

~ null [ -VOCl~-c on s

-cor

In the rural arens, outside of the major cities ofthe Central region and Saigon, there is a tendency todrop or change all the valars - Igl and 19/ are dropped

as well as the /hI and Ikl which have alrendy been

29

-mentioned, and the Ikh( combines with the w-glide to

form a different sound A more general rule, but madecomplex by the ommission of the same dropping tendency

with the lkh/, is written using double braces to show

the alternating features, within the class of

obstru-ents and between obstruobstru-ents and sonorants

Rule VI: Southern/Central back consonant drop

_ _ ) null [ -VOCl]-cons

-cor

, 18

Cadi ere shows several other changes in the back

consonants when they precede the w-glide In some cases

they become [h) and the glide remains as shown below~

IgwVI /ltwVI

This change, however, has not been regularized in all

forms in·which i t is found, so no general rule can yet

be formed

It is surprising that the Ikhl docs not follow thepattern of the other back consonants, but indeed i t

undergoes n change completely distinct from the otherBounds in its class both in the outlying Central regionsand in the South The juxtaposition of this velar.stopand the labial-glide results in a labial fricative

h

Ik wayl -} [fay] khoai 'potato'

Rule VIc: Velar to labial pre-w-glide

)

-sono-tns-vd+ant-cor-hi

/ # _ [+vocl]

Rules VI and VIc accurately show w'hat is happeningphonetically, but they fail to show the relationship

between the two Rule VIc must show repetition of ma~y

of the features for the initial consonant for clarity.Harms (l968:pp 68+) gives an alternative form for

altering more than one segment with a rule - the

sit-uation which occurs in Rule VIc In this case a ent format can be employed Rather than the normal:

differ-ab ~ cd I e j

a structural description and change can be given which

11

-show the various segments undergoing different changes-.

The general format can be given as:

Harms also discusses a special convention which

has been proposed for handling certain situations in

which there are exceptions to a general rule (1968 :

pp 71+) This involves the introduction of A i featurewhich eliminates the given segment from undergoing

another rule - such a feature as r'- n e x t rule] •

Consi-dering all of Rule VI, this sort of convention wouldsimplify the rules as.well as express the generalization

which is clearly seen

Rule VIc would be ordered before Rule VI, and could

[

- c on s]

-vocl-cor

These are just some notational devices which give

examples of the possible conventions which can be

em-ployed to economize in the descriptions Many more

notations will not be considered, but in a thorough

study of any area of Vietnamese phonology, these could

be employed to facilitate the description and accuracy

of the anqlysis

33

-This w-glide is seen to be written in Vietnamese

as either an 0 or u the 0 occuring before t he

[.-tns] ' a l an d I ~ I an d the u writte n before the [+tns]

Ie I , I i I , an d 1/\/ • . (After lk / they bo-th appear as ucoupled with a q-) The following examples show the

application of Rules VI and VIc for Central and Southern

dialects:

( ix) khoa 'speciality' (khue 'woman - !)

(xi) ng o a 'to boast' (nguy 'to be dangerous' )

Rule VIcx is applied first; i t specifying that the

/kh/ i t describes will no t be subject to Rule VI, which

To complete the discussion of initial consonants,

a few points of minor shifts need to be mentioned A

uniform shift in all th r e e dialects from Middle

) [ +continuant]

There is also a uniform dropping of the

rctroflex-ion for the I ~ / and a regionally nonuniform dropping

of i t for the I~/.

')5