SUPRASEGMENTAL WMENA 3N
JAMAICAN CREOLE

Thedxfiui'thmofPhb.
MlCHlGAN STATE UNIVERSiTY
David L. Lawfan
E966

fii-iE—m?

This is to certify that the

thesis entitled

SUPRASEGMENTAL PHENOMENA IN JAMAICAN CREOLE

presented by

David Lloyd Lawton

has been accepted towards fulfillment
of the requirements for

_P_YL__IL_ degree in M

Major professor
Date Q47fifi /7Q

0-169

     
    

LIBRARY

Michigan State
University

 

“I E LL13; sawhflldi‘lhh

 

E? ATERML in 9.9.5-4.}

‘ wt +99%5 ‘

SUPRASEGMfiNTAL PHENOMENA IN

JAMAICAN CREOLE

RY

David L. Lawton

AN ABSTRACT OF A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of English

1963

ABSTRACT

SUPRASEGMENTAL PHENOMENA IN
JAMAICAN CREOLE

By David L. Lawton

In Jamaican Creole, unlike 'standard' dialects of English, tone
is lexical in a few minimal pairs. Tone levels are significant rather
than tonal contours within the tonemic system. The domain of tone is
the syllable, not the phrase. The tonemes are high-falling, high-
level, and low-level. Each utterance, whether it be a single lexical
item or a series of lexical items in linear sequence, is characterized
by one of two contours not primarily a part of the tonemic system.
These tone contours are stylistic variants and are characterized
in the one instance by'a fall from the utterance-final toneme, and
in the other by a rise from the utterance-final toneme. The style
contour (intoneme) basic to Jamaican Creole is a rising one. It
always occurs before final pause.

Attitude of speaker is indicated by total key shifts of entire
utterances or parts of utterances within the same context. This type
of intonation cannot be analyzed at present because there is no
methodology for identifying discrete elements in which total key
shifts occur; i.e. we cannot state the exclusive key distribution for

a particular 'attitude'.

David L. Lawton

The terminal contour for both questions and simple statements
is the same; but items ('words') ending with a lowblevel toneme have a
high-level toneme substitution in questions. Items with basic high-
level and high-falling tonemes keep the same pattern. The result is
that questions without question words can be identified only by con-
trast. Speakers of standard dialects of English are often much con-
fused by the apparent failure of speakers of Jamaican Creole to dis-
criminate between 'question' and 'statement'.

Stress is not phonemic in Jamaican Creole; it does not keep
apart minimal items. There is emphatic stress, however, which is
part of the metalinguistic system. Speakers oriented toward a stress
system often confuse tone with stress.

The implications of this study are that Jamaican Creole, as a
'dialect' of English can only be fitted into an overall pattern if we
are willing to include a dialect that has tone registers with some
intonation overlap, and if we include a stylistic tone contour as a
part of an intonation system. we must also be willing to include
within our overall pattern for English a dialect in which stress is
nonphonemic. Thus our analysis may point the way to an emphasis on
functional elements (suprasegmentals) of language that have not been
fully explored, and whose systematic presentation might offer an im-
plicit approach to the study of all creolized languages as separate

languages rather than as dialects of a 'matrix' language.

Cepyright by
David L. Lawton

1963

SUPRASEGMENTAL PHENOMENA IN

JAMAICAN CREOLE

BY

\

' i

David L." ‘Lawton

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of English

1963

AC KNOWLEDGMENT 3 AND IN TRODU CT ION

In attaining any goal, it is seldom that an individual succeeds
unaided. The deveIOpment and completion of this work is something I
have contemplated for many years, but the spark was lit in a class on
English sounds conducted by Professor William Heist in the summer of
1956.

A United States citizen, I had lived for ten years in Jamaica
from my eighth to my eighteenth year, returning to the States in 19h3.
I visited Jamaica again in 19h6 and in l9h8 to assess the possibilities
for research in Jamaican Creole, but it was not until the spring of
1959 that I was able to arouse sufficient interest in my preposed re-
search so that I could hOpe for some financial assistance.

In the spring of 1959, I attended a conference on creole lin-
guistics at the University College of the west Indies at Jamaica. The
conference was sponsored by the Rockefeller Foundation and attended by
well-known scholars in creolized languages, among whom were F.G. Cassidy
of the University of Wisconsin; Robert A. Hall, Jr., of Cornell
University; Albert Marckwardt of the University of Michigan; Douglas
Taylor of Dominica; John Berry of the School of Oriental and African
Studies, University of London; E. Bagby.Atwood of the University of
Texas; Jan Voorhoeve of the Bureau of Linguistic Research in Surinam
(University of Amsterdam); and R.B. LePage of the University College of
the west Indies, chairman of the conference.

In the summer of 1959 (subsequent to the conference) I returned
to Jamaica to carry out my field work with the help of a grant from the
Woodrow Wilson Foundation made possible through the recomnendation of

Professor R. B. Nye. During my three months in Jamaica, the University
ii

College allowed me full use of its facilities, which included an
excellent library and a Telefunken tape recorder. I travelled through-
out the whole island to check for significant tonal variation from
area to area before beginning my work with informants.

My informants were Mr. Arthur'Wild, Mr. Brooks, and Mr. Gentles,
all of Irish Town in the St. Andrew hills; Mr. Harris of Moore Town,
and.Hr. Lee of August Town, St. Andrew; Hr.'Wild was an old man of
about seventybfive, and had worked for several previous scholars, in-
cluding Professor DeCamp, as an informant. He was bilingual in Creole
and Standard English, and rather saphisticated linguistically for an
informant though unable to read and write. Mr. Brooks and Mr. Gentles
were about thirty years old, semiliterate, and had had infrequent
contact with speakers of noncreolized dialects of English. They were
semibilingual in Jamaican Creole and Standard English. Mr. Lee was my
principal informant. He was about thirty years old, semiliterate,
and had had the least contact of the five men with noncreolized dialects
of English. Unlike the others, he had never before been an informant.
All my informants were laborers.

During the research itself and since its termination, besides
Professors William Heist and Robert Geist, who have carefully read my I
work and made many timely suggestions, I have had encouragement and
assistance from several peOple. The generous and thoughtful criticisms
of Professor John Street, under whom I did my work in descriptive
linguistics, saved me from many pitfalls. Professor R.B. LePage of
the University College of the west Indies kindly extended to me the
facilities of the University College. Others who have been generous

with advice and encouragement are Professor F.G. Cassidy of the
iii

University of Wisconsin, whom I met in Jamaica while he was completing

Jamaica Talk and putting together his Dictionary of Jamaicanisms;

 

Professor David DeCamp of the University of Texas; Professor John
Berry of the School for African and Oriental Studies at the Univer-
sity of London; Mr. Charles Elliott of the University of Michigan, a
structural linguist and a close friend; and Professor Kenneth L. Pike
of the University of Michigan. Father A.H. Jones of the University
of London made a tonometer analysis of representative magnetic re-
cordings from my speech samples and generously supplied me with the
results.

iv

l.

2.

2.1

2.2

E]
//

LIST OF SYMBOLS; RELATED MATTERS

- phonetic brackets for cited data

. phonemic slant lines used for all '-emic' items

- any consonant, any vowel, respectively. The sequence may
be reversed or expanded; as for example, CCVC, CVCV, VVC,
VCV C , etc.

- hypothetical linguistic form

- high-falling toneme

- high-level tongue

- lowblevel toneme

. terminal contour (intoneme)

- becomes

- similar to

- equal to

- falling tone glide

' rising tone glide

- terminal pause

Contrastive units of pitch are TONEMES.

Tonemes may occur on a syllabic vowel or on a syllabic

consonant and are inherent in the lexical form of the word

and its syllable or syllables.

When the pitch of a syllable neither rises nor falls during

production within the limits of perception, the tone is

said to be a LEVEL TONEME.

Any language that has pitch (tone) contrasts between level

tonemes has a REGISTER SYSTEM.

V

2.3

Zeb

3.0

h.o

Tonal features of an accentual system.that indicate 'attitude
of speaker' without changing lexical meanings are INTONEMES.
The tonal members of an intoneme are allointonemes. Alloin-
tonemes are to a toneme what an allophone is to a phoneme.

If in two languages A and B, A has a fully tonal system and B
a fully intonal system, A is said to be a TONE language, and B
an INTONATION language. If language C shares some significant
features of A and some significant features of B, it is said
to have OVERLAP.

'When in any syntactic or morphological construction, one member
can be removed and have substituted for it a different item,

the undisturbed portion of the construction is the substitution

 

frame and the exchanged portion ... the substitution item}.

 

(Kenneth L. Pike, Tone Languages, 50.).

Since this is a scientific linguistic study, its stylistic

format generally follows the injunctions laid down in the LSA Style

Sheet, Language, (July~September 1962), 3833, part 2, 3-7, except

where the regulations of the University specify otherwise.

55
TABLE OF CONTENTS

Acknowledgments and Introduction

Symbols; related matters

1.0
2.0
2.1
2.2
2.3
3.0
h.0
h.1
h.2
4.3
h.h
5.0
6.0
7.0
8.0
9.0
10.0
11.0
12.0
13.0
1h.0

General problems of tonal analysis
Operational procedures; informants
Methodology and special problems
Attitudes toward tonal analysis
Assumptions

Phoneme list

Syntactic frames for tonal analysis
Type-one frames

Type-two frames

Type-three frames

Type ¢ frames

Tone vowels

Toneme distribution; minimal pairs
Tonemic domains

Basic forms

Stylistic variants

Toneme classes; patterning

Free discourse and key shifts
Examples of key shift

Style intonemes

Problems with stylistic variation

Conclusion

vii

Page

\O\OGDNU\

12
13

15
16
17
18
22
23
25
25
27
28
29
3o
32

Page
36
Appendices

53
Bibliography

viii

LIST OF APPENDICES

Appendix

16.0

17.0
17.1
17.2
17.3
18.0
19.0
20.0
21.0
22.0
23.0
2h.0
2h.1
2h.2
2h.3
25.0
26.0

27.0

'Diasystem vs. Overall Pattern: The Jamaican
Syllabic Nuclei'

Dialects H,M,L

Phonemic diasystems
Intrastructural aIIOphonic rules
Continued rules

Allophonic problems

DeCamp's consonant list
Consonant clusters

Vowel list; problems

Tone vowels

Vowel phonemes and allophonic distribution
Tonometer analysis

Tonometer analysis continued
Inferences drawn from analysis
Discussion of exception.(dgpi)
Stress

Explanation of text

Text from free discourse and translation

Page

37
38
38
38
39
39

h2
h6
h7
A7
50
so
so
51
52

O. The use of the term creole raises questions about its application
to a language or dialect1 and the essential linguistic features that
differentiate creoles from pidgins and artificial languages.

By definition, 'a pidgin is nobody's native language',2 and as
such is differentiated from an artificial language by being created
'under the pressure of practical circumstances in a bilingual
situation [rather than] by a scholar sitting quietly in his study'.3
A creolized language, on the other hand, is a language formed from
a pidgin; i.e. speakers of different 'first languages have taken a
pidgin as their common language and have raised children for whom
it is native'.h Such is the case with Jamaican Creole, which
derives from English in contact with'West African languages, of
which Twi seems to provide the largest number of noannglish words.
Jamaican Creole is spoken in Jamaica, the third largest of the

islands in the Greater Antilles, by the majority5 of the almost two

 

1Both language and dialect are defined as a 'collection of more or
less similar idiolects'; a 'dialect is Just the same thing with this
difference: when both terms are used in a single discussion, the
degree of similarity of the idiolects in a single dialect is presumed
to be greater than that of all the idiolects in the language'. See
C.F. Hockett, A Course in Modern Linguistics (New Ybrk: Macmillan Co.,
1958), 322. '_ '—

ZHockett, h23.

 

 

3Hockett, A22.
hHockett, h23.

SR.B. LePage and David DeCamp, Jamaican Creole: Creole Language
Studies (London: Macmillan Co., Ltd. 196UTrjfrll7-8.

2

6
million inhabitants, many of whom are semibilingual in Jamaican
Creole and Standard English, and descended from language communities

in which Twi, Mende, Mandinka, Ewe, Igbo, Efik, Yoruba, Kongo, and

possibly Swahili were spoken.7

Jamaican Creole is distinguished from Standard Jamaican English,
the dialect spoken by educated Jamaicans in the community, by three
specific linguistic features: morphology, syntax and suprasegmental

phenomena. Of the syntax, F.G. Cassidy writes that

the most striking differences between the folk speech
[Jamaican Creole] of Jamaica and the educated speech

are not in the sounds, still less in the vocabulary-
they are in the grammar, the functional patterns into
which the words fall. It is also in this respect that
the most fundamental influence from African backgrounds

is to be seen, and that is why Jamaican folk speech is
not a dialect in the same sense that the rural speech

of Devonshire or Lancashire . . . are dialects of English.
Those who would hold that Jamaican folk speech is not

to be considered a type of English at all, but a.new

and different language,will find their strongest arguments
here.

6The degree of mutual intelligibility between speakers of Jamaican
Creole who speak it from birth but understand Standard English
(Jamaican Standard), and those who have heard Jamaican Creole from
birth but speak Standard English has never been statistically es-
tablished. It seems more accurate to use 'semibilingual' than
'bilingual' in accordance with Hockett's criteria in A Course in
Modern Linguistics, 327. Hockett writes that 'Each speaks his own
personal variety of his own language, but has learned by experience
to understand the speech patterns of the others.‘

 

 

7R.B. LePage, 'General Outlines of Creole English Dialects in the
British Caribbean', Orbis: Bulletin International de Documentation
de Linguistique, 6:2, I957, 380.

 

8F.G. Cassidy, ngaica Talk, (London: Macmillan Co., Ltd., 1961), h9.

 

3

The suprasegmentals reveal an accentual system9 in which tone is

functionally different from that of Standard English}0

No structural
analysis of Jamaican Creole suprasegmentals is extant.

This study is an analysis of the suprasegmentals of Jamaican
Creole with special emphasis on tone and its role. So far, the
published work on Jamaican Creole has dealt only cursorily with
suprasegmental phenomena and not with their use in a structural frame.11
Opinions on the difficulty of making a tonal analysis vary. One scholar
suggests that the special services of an Africanist are needed;12
another feels that the study of tonal phenomena is so difficult that
'angels might well fear to tread'.13 Bibliographies are hardly help-

ful since relevant scholarship contains nothing of any consequence

on the suprasegmentals,and dates only from 1957.1h

 

9'Accentual system! refers either to that of a stress language ex-
emplified by English and Spanish or to that of a tone language ex-
emplified by Cantonese and Twi, in which 'most syllables in con-

text carry one or another of a small set of contrasting tonal contours,

and even.monosyllabic utterances show some of the contrasts'. See
Hockett, 100.

10Cassidy, 26-33.

llLePage and DeCamp, 137-9.
12LePage, 38h.

lBCassidy, 3o.

1hR.B. LePage's 'General Outlines of Creole English Dialects in the
British Caribbean' is the first scholarly publication to offer a
scientific sketch of Jamaican Creole phonology and grammar based on
English methodology; it contains nothing on the suprasegmentals and
is therefore not directly helpful to my analysis.

h
Of two studies recently published, Jamaican Creole by R.B. LePage and

 

David DeCamp deals largely with the ethnic makeup of Jamaica and the
historical linguistic milieu. A short section of the work lists
the segmentals, gives some tonal phonetic data, and barely touches

the grammar. F.G. Cassidy's Jamaica Talk, the other study, is directed.

 

largely at a lay audience. The first four chapters discuss historical
matters, pronunciation and grammar. The rest of the book treats
vocabulary items, their etymology and relevance to Jamaican life.
Professor Cassidy gives many phonological details, although naturally
enough, does not present them in the format of a structural grammar.

A detailed description of the phonology has been in preparation
for some time, but none of it has yet been published.15 Soon to be
published is an etymological dictionary of Jamaican Creole and a

sketch of Jamaican grammar.16 At the LSA conference in December 1961,

 

1SLePage mentions a detailed study of Jamaican Creole phonology by
V.L. McLoskey in his 'General Outlines . . .' The study still has
not been published so far as I know. It is not readily available to
anyone, nor do I know whether it is intended as a thesis or as a
separate scholarly publication.

16Information in a letter to me from Professor F.G. Cassidy in which he

says the dictionary will be published by the Cambridge University Press
(England), but not before l963-6h. The sketch of Jamaican grammar will
probably be published in Orbis by Mrs. Beryl Bailey as soon as it is
finished.

5
Professor DeCamp in a paper entitled Piasystem vs. Overall Pattern:

 

the Jamaican Syllabic Nuclei17 presented the syllabic nuclei accord-

 

ing to a dialect theory postulating three dialects of Jamaican English
and offering intrastructural rules for going from one dialect to
another. A complete review of work done on Jamaican Creole up to
1959, together with work in progress, is given by F.G. Cassidy in

18
American Speech.

 

1.0 The problems of tonal analysis in Jamaican Creole are similar to
those in any linguistic analysis in which the main concern is supra-
segmental phenomena. Of utmost importance are the selection of an
adequate corpus,l9 the method of gathering and collating, and the
determination of the structural elements of the dialect from the data.
The adequacy of a corpus does not rest on vast, random sampling,
but on a selective sampling based on the analysis of specific linguistic
phenomena-awhich, in this instance, is primarily tone. As H.A. Gleason

aptly puts it,

 

17In a letter dated February 23, 1962, Professor DeCamp tells me his

plans for publication of the paper are uncertain. See Appendix (16.0) for
extracts from the letter and the diagram of his preposed treatment

of syllabic nuclei.

18XXXIV(October 1959), 163-71.

19'The stock of recorded utterances constitutes the corpus of data,

and the analysis which is made of it is a compact description of the
distribution of elements within it'. See Zellig Harris, Structural
Linguistics, (hth ed.; Chicago: University of Chicago Press, 1960), 12.

 

 

6

A sufficiently large corpus of randomly gathered material
would seem to meet this need [an adequate corpus]. But there
are difficulties. Some grammatical features are quite in-
frequent. An unplanned corpus may have to be extremely
large to include adequate representation of these. Others
are very common; even a.moderate amount may exemplify these
far beyond what is required to establish or confirm an
analysis. Vast amounts of unneeded data can only complicate
recording, filing, searching, and all the other processes
involved in analysis . . . A linguist, therefore, strives

to obtain a corpus EBat will be representative in the
minimum total bulk.

The adequacy of the informants, like that of the corpus, lies
not so much in quantity as in selection. The criterion for selection
is not how'many informants the linguist can find, but the representative-
ness of their speech. If the validity of the corpus depended chiefly
on the number of informants, the analysis of important linguistic
phenomena in several languages would be invalid.21

The linguist, then, must determine whether or not the informant
is a native speaker. He must be able to elicit responses either by

speaking the informant's language or by having an informant who can

22
speak his. Obviously, when the linguist is confronted by a language he

 

20An Introduction to Descriptive Linguistics,(rev.3 New York: Holt,

Rinehart and Winston, 1961), I95.

 

21The analysis of 'southern drawl' in American speech was made largely
from the corpus provided by one representative speaker. See Kenneth

L. Pike, The Intonation of American English (7th printing; Ann Arbor:
University STIMichigan’Press, 1958), 105; so was the analysis of Swahili
by Zellig Harris. See Harris, 97. See also I. Richardson, The Role

of Tone in the Structure of Sukumu (London: School of Oriental and
African Studies, 1959), ii.

 

 

2Z'Phonemic analysis is carried out under varying conditions. The most
favorable ones are when the analyst can work through direct face-to-face
contact with a speaker of the language or informant--especially if the
analyst and informant both control some language, in addition to the
informant's native language that the analyst wants to study' . Hockett,
102.

 

I'll ills-II lull"! ill! lull—ll l l '|.1.I. I111 illl

7

does not know, he must use his ingenuity in eliciting data until he

has mastered the rudiments of the language he is trying to analyze. In
general, 'any person past his first years of learning to talk, speaks
the language of his community as a 'native' if he has not been away from
the community for long periods9.23

2.0 For ten years I lived in Jamaica, where I attended primary'and
secondary school. During this period I gained an intimate knowledge

of Jamaican Creole, and participated fully in the social and linguistic
life of the community. This experience aided me greatly in the selection
of informants for this study. Often people who have lived for years in
Jamaica will maintain that they neither understand nor speak Jamaican
Creole, and even native-born middle class Jamaicans will deny emphatically
that they can speak it. All such assertions should be accepted with
reservations, since in Jamaica knowledge of Creole tends to be equated
with low social status. (It is quite true that some peOple cannot speak
Creole well, but they certainly understand it well enough to communicate
with servants and workmen.) If an individual claims, however, that he

is not a native speaker of Jamaican Creole, his claim may be valid,
since it is not necessary to be a native speaker to be able to speak

and understand a second language or dialect. Thus, in Jamaica the
selection of informants requires not only an adequate command of Creole,
but an understanding of the implications Jamaican Creole has for the
residents of the Island as an indicator of social level. My early

experiences equipped me to handle both aspects of the problem.

 

23Harris, 13-h.

8

2.1 One of the major difficulties of suprasegmental analysis in
Jamaican Creole has been the working out of a methodology for determin-
ing distinctive tones. Free conversation itself can serve only as a
check on minor phonetic features, since it does not provide a rigidly
controlled context. Such a context demands brevity and repetition of
similar utterances to avoid stylistic variations of speech that may be
confused with significant features. Unfortunately, few scholarly works
contain an explicit methodology.

Kenneth L. Pike's Tone Languages contains the most detailed and

 

clearly outlined procedures for the analysis of tone languages. His work,
of course, derives largely from analyses of Mexican Indian languages

which have proved to be tonal. Zellig Harris's Structural Linguistics

 

does not pretend to give detailed procedures for field work, nor provide
a detailed plan for obtaining data, but serves 'as a reminder in the
course of the original research, and as a form for checking or present-
ing the results . . .‘2h Although Harris's systematic methodolOgy
applies equally well to segmental and to suprasegmental phenomena, his
statements are oriented toward scholars familiar with mathematical

logic rather than to the Bloomfieldian linguistic tradition implicit

in Trager and Smith's Outline of Structure. Harris writes that it is

 

'more convenient to consider the elements as purely logical symbols,
upon which various Operations of mathematical logic can be performed'.25
For specific work with suprasegmental phenomena in Jamaican Creole,

neither the detailed and generally explicit procedures of Pike nor the

 

21‘Harris, 2.

25Harris, 18.

9

'mathematical logic' of Harris offers more than an outline. Each
language presents its own problems, and requires its own answers.
2.2 Quite apart from the problem of designing an adequate methodOIOgy
is the attitude of some scholars toward suprasegmental analysis. If
the analysis of suprasegmental phenomena generally has been complicated
by a lack of precedent, it has also been complicated by certain pre-
vailing attitudes toward such analyses:

The great majority of present-day linguists fall into

one or more of a number of overlapping types: those

who are convinced that tone cannot be analyzed, those

who are personally scared of tone and tone languages,

generally, those who are convinced that tone is merely

an unnecessary marginal feature in those languages where

it occurs, those who have no idea how to proceed with

tone analysisé those who take a simplistic view of the

whole matter.
Gleason goes on to point out that 'even experienced linguists commonly
know no more of the range of possibilities in tone systems than the
over-simple distinction between register and contour languages . . .
It is worth pondering that very little has been published on any phase
of field techniques in linguistics'. Of pertinence to Jamaican Creole
is his statement that 'tone and intonation often become seriously
intermeshed . . . But no other two phonological systems are as diffi-
cult to disentangle as are tone and intonation in some languages'.
2.3 I have assumed that suprasegmental features can be analyzed

following Pike's general methodology in his Tone Languages. Syntactic

substitution frames are used extensively for test items of the same

‘—

26HA. Gleason, review of The Role of Tone in the Structure of Sukumu,
by I. Richardson, Language(ApriI-June,“19617,‘37:2, 302.

10
form class varying in phonetic shape. The purpose of substitution

frames is to provide a constant against which to test the relative
pitch level of morphological forms. Not only are substitution frames
valuable for discovering significant tones, whether tonemic or in-
tonemic, but also for controlling significantly the changes of key
which frequently occur in free discourse.27 The checking of contras-
tive tones is also made simpler.

we can make a final check on the accuracy of frame analysis by
finding pairs of words which differ in relative pitch and meaning,
although sometimes these pairs may not be as definitive as frames in
identifying tonal contrasts. When members of the same form class are
drawn from the substitution list and grouped according to relative
height of pitch as low, mid, high, falling and rising, we can use the
working procedure in which a few dozen representative forms are
examined from each group. Identifying the relative height of tones
depends on finding a frame that has a stable high-level toneme or a
low-level toneme. ‘we assume that a frame of approximately ten syl-
lables will have one or more high syllables. much of the methodology
is implicit in the analysis dealing with the identification of sig-

nificant tones.

 

27"The general key or octave, or general 'pitch of the voice' may change
and force a change of all the syllables and tonemes in the phrase. At
one time the informant may speak in a high voice and at another time

in a low voice . . . The attempt to fix a certain auditory note for
establishing a comparison of absolute height with other tonemes won't
work. The change in key may make the conclusions false by the inves-
tigator's postulating too many tonal contrasts". Pike, 27.

11

3.0 The segmental phonemes of Jamaican Creole are taken from DeCamp's

list; /v/ and /h/ are omitted.28

/p/ /t/ /6/ /K/
/b/ /d/ /.‘J/ /g/
/r/ / 3/ /§/
/z/
/m/ /n/ /n/
/1/
/r/
/w,3/
The patterned arrangement of consonant phonemes follows Hockett.29 I

The vowel phonemes?o as I have analyzed them, are

/i/ /u/
/e/ / 0/
/a/

Diphthongs: /ie/, /uo/

/a;/, /0;J_/

 

28Appendix, 1 9.0.

29Hockett, 60.

3C’Appendix, 2 3.0 .

12

h.0 Frame-word sequences for identifying tonemes fall into the

following syntactic patterns:

.1 'WOrd sequence, test item, word sequence
.2 Word sequence, test item

.3 Test item, word sequence

.h Test item in isolation

An example of items from the substitution list is given in each syn-
tactic frame, and is fully representative of the phonetic shape of all
items of its form class. The items used are listed to the right of
each frame for easy identification. Tones are represented phonetically
by placement above or below the normal linear level of the items.
Falling tones are illustrated by a repetition of the vowel or syllabic
consonant wherein the falling tone occurs. The frame is given with

tonemes marked.

13

h.l Examples of test items are /bula, mata, sintin, ruoti, at, at,

pitieta, kasada, dalambat, talawa, kotaku/.

E dimaataafimi]

[cm uootiafimi]

[ di‘atarimi]

[ dipitieetaafimi]

[ dikasaedaafimi]

[didal‘mb‘starimfl

rdit‘l‘waar imi]

[ dikotakuafimi]

/dIbfiliafimi/ 'The bula is for me'
/dmataa:im.i/ 'The mortar is for me'
/dIsInt1n‘arimi/ 'The thing is for me'
/diru6ti§fimi/ 'The ruoti is for me'
/dIitif1mi/ 'The hat is for me'

/diitifimi/ 'The heart is for me'
/d'ipitiétaa£imi/ 'The potato is for me'
/dik§sidiifimi/ 'The casava is for me'
/did§12mbitarimv 'The dalambat is for me'
/dit;l;waafimi/ 'The strong [one] is for me'

/dIk5tikuafimi/ 'The basket is for me'

lb

h.2 Test items are the same as in h.l.

[afimidibula]
[‘arimidimaata]
[arimidiSintin]
[afimidiruooti]
rarimidiat]
[arimidiaatl

[afimidipitieeta]

kasa

[afimidi ads]

[afimididalambaat]

[afimiditalawa]

[afimidikotaku]

/arimidibfila/ 'The bula is for me'

/afimidiméta/ 'The mortar is for me'
/§fimidisIntin/ 'The thing is for me'
/éfimidiru6ti/ 'The ruoti is for me'
/a£imidiat/ 'The hat is for me'
/Arimidiét/

'The heart is for me'

/§fimidipitiété/ 'The potato is for
me'

/aflmidik§séda/ 'The casava is for me'

/af1mldidil§mbét/ 'The dalambat is for me'

/§£imiditaiawa/ 'The strong [onelis
for me'
/§fimidik8thkfi/ 'The basket is for me'

15

h.3 Test items are the same as in b.l.

[buladounde] /bfilhdounde/ 'The bula is down there'
[mastadounde] /mét5dohndé/ 'The mortar is down there'
faintindounde] /sintindoundé/ 'Something is down there'
[ruootidounde] /ru6tidoundé/ 'The ruoti is down there'
[atdounde] /5tdotnde/ 'The hat is down there'
[aatdounde] /atdoundé/ 'The heart is down there'
[pitietadounde] /pitiétadohnde/ 'The potato is down there'
[kasaadadounde] /k535dhdofinde/ 'The casava is down there'
[dalambaatdounde] /d§l§mbétdoundé/ 'The dalambat is down there'

[talawadounde] /t§l§wadoundé/ 'The strong [oneT is down there'

[kotakudounde ] /k5tékfidounde/ 'The basket is down there'

h.h ¢ Frames

rat]
[‘at]
[bit]
[biit]
rsut]
[suut1
[buotl
[liet]

[ mout]

E mail]

/=t/
/£t/
/bIt/
/b it/
/sat/
/sut/
/tu6t/
/liét/
/mo£t/

/mail/

16

'hat, hurt'

'heart'

'bit'

'beat, beet'

'soot'

'suit'

'boat'

'late'

'mouth'

'mile'

17
5.0 All phonetically long vowels carry high-falling tone in all
contexts. Short vowels carry a tone as high as the first element of
the tone on [eat] but there is no tone glide. The tone on short
vowels as in [at] is high-level. Hence we can analyze short vowels
as significantly'high,31 so that all monosyllabic free morphemes
with short vowels may be assigned a high toneme as in /it/, and all
free morphemes with phonetically long vowels a high-falling toneme as
in /at/. Complex nuclei also carry high-falling tones as in /budt/
and sometimes a low-level toneme, depending on the phonetic environ-
ment (h.3).

Two- and three-syllable free morphemes carry no significantly
different toneme pattern in isolation than they do in frames. The
frame items themselves remain relatively stable. Substitution items
placed in the frames either go tonally higher than adjacent frame
items, remain the same height, or drop slightly lower in tone before
final pause. Phonetic variation within frames is from two-thirds of a
semitone to three semitones, maximum variation. Commonly, variation
is about two-thirds of a semitone within short, sentence-type utterances.
Wherever a phonetically long vowel occurs, the tone spread is greater
and becomes [-‘i] rather than E- ]. Tone glides, however, are phonetic
since there are no occurrences of such glides in contrast. The best
treatment within a framework of significant tonal contrasts is to

regard glides as nontonemic within a level system. The consistent

 

31Pike , 63 e

18

occurrence of high-falling tone with long vowels enables us to des-
cribe vowel quantity when we are given the tonemic domain32 of the
syllable in question.
6.0 Looking at the following items, we can now observe the facts
supporting the statement in 5.0 about the function of tone with
phonetically long vowels. Alternate solutions are presented in the
Appendix.

[maatal > /méta/"mortar'

[mats] > /mata/ 'matter'

[tuutu] > /tfitu/ 'feces'

t - .
[ utu'l > /tutu/ 'privates'

[blita] > /bit‘a/ 'beater'

[bits] > [bita / 'bitter'
The two-syllable free morphemes given above fully represent the dis-
tribution of suprasegmental features with phonetically long and short
vowels. In each example, the initial syllable carries inherent,
audible stress (loudness), and the final syllable, no stress. Stress
is fully'predictable on initial syllables of all two-syllable free
morphemes. High-falling tone occurs with every long vowel whereas
high-level occurs with short vowels. Since high-falling tone occurs
with phonetically long vowels and complex nuclei, we can treat all

long vowels as purely phonetic. This treatment provides a.more con-

venient linguistic statement, since we can thus reduce the number of

 

32The 'domain' is simply the environment in which something occurs.
When we refer to tonemic domain in relation to 'syllable', we mean tone
environment. When we refer to syllabic domain in relation to 'tone',
we mean syllabic environment.

 

 

 

[Ii III. III] I [Ii l
. II‘ A I J llllluf ‘ \llI-I: I

l9

vowel phonemes to five.33 Complex vocalic nuclei cannot be similarly
reduced, of course, since we cannot predict the vowel glide from the
vowel nucleus. Stress has no relevance as a minimal differential
feature to keep otherwise identical items apart. Instead, we have
high-falling and high-level tones in contrast-/’/ and /‘/, respectively.
The assignment of phonetically long vowels to /'/ gives a more complete,
consistent, and simpler structural interpretation of the data than has
been given by LePage and DeCamp.3h

The evidence for a low toneme rests on the distribution of tone
within the frames I have used, and on the significant differences of
tone between otherwise similar items. It is certainly reasonable to
expect that where tone is contrastive, some tone will be significantly
high, others significantly low.35 The representative frame of a whole
set of frames /di___afimi/ gives /L-afimi/ consistently lower in tone
than the other members of the syntactic frame, and since substitution
items, if they are low toned, do not go perceptibly lower, a low toneme

can be assigned to each syllable of the sequence /¥-afimi/ >~/afimi/.

33!If we are confronted with two or more ways of identifying allophones

as phonemes, both or all of which equally well meet other criteria,
we should choose that alternative which yields the most symmetrical
portrayal of the system'. Hockett, 109.

31‘Appendix , 21.0 ff.

35MB , 63 e

20

Other evidence for a low toneme is

[mieribrOun] /miéribrofin/ 'Mary Brown'
[mieribroun] /miérlbroun/ 'Mary is brown'
[kyansOJ /ky?=\ngb/ 'can so '
[kyaango] /kyingb/ 'can't go'
[9191] /11/ 'yes'

[9191] /ii/ 'no'

Even though we have been fortunate enough to find two
minimal pairs and one subminimal pair, such items are not strictly
necessary for setting up significant featureso-in this instance,
tonemes. Criteria such as frames, for example, would prove adequate.
Conceivably, we may not be able to find.minimal pairs in a language;
it may even be theoretically possible for a given language to have
no truly minimal pairs. Gleason writes that it is 'conceivable that
extensive search might fail to uncover any minimal pairs for two
closely similar sounds. In some languages, minimal pairs are much
more difficult to find than is the case in English, so much so that
the analyst cannot afford to depend upon them. They are by no means
necessary, but merely the most definitive evidence when they can be
found. Other methods can, however, provide a quite reliable

analysis'.37

 

37Gleason, 52 Introduction to Descriptive Linguistics, 25.

 

 

21
In Jamaican Creole tonemes are clearly tied in with the
syllable nucleus, in most instances the peak vowel, though it is not
uncommon to have syllabic consonants as in /§m§di/ 'somebody', in
which /s/ is syllabic. with phonetically long vowels we have non-

38

phonemic contour overlap. No distinctive status need be given

noncontrastive tone glides since they are a constant and predictable

feature of a single component.39 In Jamaican Creole, then, there are

three tonemic levels /'/, /'/ and /‘/.

 

38'Any language whose tonemes are all to be interpreted in terms of an
interrelated system of levels is basically a register-tone t3 .. If,
however, glides develOp which, though composed of juxtaposed level
tonemes, have some characteristic of timing or placement that seems
different from the timing or placement of the separate level tonemes,

a contour overlap may be postulated for classification'. Pike, 12.

39This component is high tone. Another way of stating the same thing
is to say that if in every occurrence of.A,B,C we have Ay, By, Cy,

we may simply set up components A,B,C, since 2:13 predictable in

the environments given. See Harris, 52ff.

22

Basic tonemic domains appear in the following statement.

7.0

Tonemic Domain

/'/

/'/

/‘/

Syllable
/bét/
/r£-/
/ma'L-/
/tfi-/
/ou6-/
/sén/
/-naiz/

/mE-/
/b3t/
/bE-/
/pi-/
/k6-/ /-1o'1/
/t5-/ /-15-/
/ré'-/ /-k§-/

/-1§/
/-t§-/
/-s;1/

Item
/bét/
/riri/
finith/
/tfitu/
/cu6c6/

/ga'n/

/r§k5naiz/

finisfi/
/b5t/
/b"ilh/

smash

/t§15wa

'bath'
'whistle'
'mortar'
'feces'
'chayote'
'went'

'recognize'

'lift up'
'bat'
'cake'
'potato'
'basket'

'strong one'

/r§k§naiz/ 'recognize'

/b\'ilh/
new
/m58i/

‘cake'
'basket'

'lift up'

23
8.0 Since our main concern is not the writing of a complete grammar,
but an analysis and systemization of those suprasegmental phenomena that
lend themselves to analysis, two important statements should emerge
from the total data: a statement of basic patterning and a statement
of function.

A linguistic form may be given as 'basic' if its distinctive
features allow 'predictablilty in description'.’40 From this statement
it would seem that stable linguistic forms in isolation or in.context
would facilitate a descriptive summary of the tonemic patterning in
Jamaican Creole; i.e. we can say, given a and b'as significant
entities, if wherever 3 occurs, 1 occurs also, but not b, then 2
can be described in terms of y, or z in terms of 2. Our choice
depends on the greater or lesser complexity of the descriptive
statement. we can say very little about the patterning of b, for
example, unless we can state its distribution. 'Patterning' for
tones in some languages can be relevant either to the features of
grammatical relations defined, or to class markers; for example,
modifier-head sequences having high and low tones--*CVCV(modifier)+
*C§(head), or nouns with one tone pattern, adjectives with another,

and so on.

AQQVV: A-

'The criterion for deciding which should be taken as the 'basic'
or 'inherent' toneme or tonemic pattern is PREDICTABILITI IN DESCRIPTION.
Convenience of description usually dictates the choice of that tonemic
attern as basic which will most easil allow for the statement 2:
rules predicting how‘the toneme§_§iII %e pertuFBEd‘EIsewEere. Pike, 75.

 

 

2h

In Jamaican Creole tonemes are coterminous with syllables.hl
We refer to these syllables as having certain tonemic domains. Harris
points out the linguistic convenience of not setting up 'absolute
divisions, e.g. word and phrase, and then [saying] that various
relations cross these divisions . . . Instead, the domain of each
element, or each relation among elements, is indicated when the element
in question is set up'.ha Tonemic patterning in Jamaican Creole is
complicated by stylistic variationl‘3 which is bound up with tonal
features describable as intonation overlap before final pause. There
is no toneme change within the system of levels, but a change in the
direction of the terminal tone glide which is simply a stylistic

variation.

 

bl

Pike refers to tonemic syllabic patterning analogous to what happens
in Jamaican Creole in a general way: 'A number of EurOpean languages
have been described as utilizing pitch in the differentiation of
meaning of various lexical items, but with the placement of the pitch
limited to certain types of syllables [as in Swedish.and Norwegian] or
to specific places in the word. One must watch for the possibility of
significant pitch contrasts which are limited, for example, to stressed
syllables, or to long vowels . . . [Some languages] have some charac-
teristics which may eventually lead to the postulation of a new type
of tone language or to a slight modification of the present definition
of tone languages'. Pike, lh-S.

2 .
Harris, 15.

h
3Hockett suggests setting up 'style morphemes' for handling instances

of variant forms of the same lexeme, constructions, and sentence types,
but does not feel that this solution is entirely convincing.
Hockett, 278-9.

25
9.0 Terminal tone glides are in stylistic alternation for any minimal
free form, or any sequence of minimal free forms. We can have
[maataiw and [mnatafij for the same lexical item. Descriptively, it
is simpler to state these forms as the result of operational external
sandhi (stylistic variants). They differ in the direction of the
final tone contour glide. There is no change in tonemic level. we
represent both variants by the examples fbratat/ and /matat/ in

Mt

which /t/ stands for the style intoneme, in free variation with
the distribution [I] or [5] before final pause.

The selection of /t/ is arbitrary and based on the greater
frequency of occurrence of the rising terminal contour for 'statement'
lexemes.

10.0 The two main tonemic patterns in Jamaican Creole are /' ‘/, /m£tA/
and /' ‘/, /bfilh/. The patterns are determined by frequency of
occurence of two-syllable free morphemes, and may be referred to as
'pattern 1' (/’ ‘/) and 'pattern 2’ (/b ‘/). Three-syllable free
morphemes, by no means abundant in Creole, fall into three patterns
exemplified by /tanéhe, pitiéta, kotakfi, t513wi/. Exceedingly
interesting, however, is what happens when we have free morphemes in

construction. High-falling tone remains high-falling, but low-level

 

hh'Intonations [and hence, intonemes] are distributed over phrases,
rather than being completed on single syllables-othough a single
syllable may also constitute an entire phrase, and thus have an
intonation applied to it . . . Intonations frequently carry shades of
meaning . . . Intonations are extrinsic . . . parts of utterances'.
Pike, 15.

26
tone is perturbed to high-level when its domain is the second syllable

of a compound. At first, it would seem plausible to postulate a
definite and consistent pattern for compounds and modifier-head con-
structions, but the facts do not bear out such a solution. For example,
/gfiwé/'go away' has the same pattern as /b51i/ 'cake', but both forms
are different by class and function. Initial syllables, syllables with
phonetically long vowels, and final syllables of a sequence seem to
keep their basic tones; syllables with complex nuclei tend to do the
same. Syllables in positions other than those stated have perturbed

tones. Examples of perturbed tones are

/s‘i'bi/ but /s'ibiswit/ 'kind of orange'
/bfi1h/ " /b51§kie’k/ 'cake' (iterative form)
/gu§ba/ " /guéb3bbd/ 'guava bud'

/p£sh/ " /pasnbrofin/ 'type of banana'
/gfiwe/ " /guw§man/ 'Go away, man'
/m'a'ka/ " /makabakfi§/ 'type of fish'

Examples of nonperturbed tones are

/as/ and /ésplantn/ 'type of plaintain'

/pinuah/ " /pInwinm5kh/ 'type of them

/3§k£s/ " /5§kés£mpa/ 'saddle-basket for a jackass'
/kyan/ " /kyanplant/ 'can't plant'

Ina-III"!
ll!

 

27
11.0 Attempts at accurate statements about the function of tone in

MS

Jamaican Creole based solely on data from free discourse are ex-

tremely difficult to make at the present time, and simply result in
phonetic statements, and the postulating of far more levels than the
system can feasibly accommodate on a significant level.h6 Speakers

of Jamaican Creole often shift freely from one key to another in free
discourse without any warning.h7 An example of a possible key shift is
illustrated below. To the left of the page are numbers representing
'key' levels, '1' being the highest, with the other numbers in des-

cending order. Significant tone contrasts are marked with the toneme

symbols we have been using throughout the paper.

h5DeCamp states, after pointing out that 'intonation' has not been
analyzed in Jamaican Creole that 'calm conversation, including moderate
exclamations, may be analyzed into five relative pitch levels . . .
When extremely animated, a speaker may also use two or three higher
levels. All levels are much more widely spaced than in American
English.t Jamaican Creole, 138. See Appendix for discussion.

 

l‘6'The attempt to fix a certain auditory note for establishing a com-
parison of absolute height with other tonemes won't work. The change
in key may make the conclusions false by the investigator's postulating
too many tone contrasts.‘ Pike, 27.

l‘7'The general key, or octave, or general 'pitch' of the voice may
change and force—3 change of all the syllables and tonemes in the
phrase. At one time the informant may speak in a high voice, and at
another time in a low voice.’ Pike, 27.

28
12.0 Key Shift in Free Discourse
'You call that sugar and water.‘

Example of 'normal' statement

1.

 

2. / y a k a 1 d 5 t s u g

 

Example of fall to lower key, followed by rise to higher key
l./yfikél

 

 

 

tet/
2. d 5 t a a w a
30 sag
Example of abrupt shift of key

tet/

 

2. /yfikald§t

 

 

 

Illlllllil’ll’lll'tlll'i“

29
Key shifts such as those in the preceding diagram result when the
speaker becomes 'emotional', showing anger, joy, or a high degree of
excitement. The height of the voice is affected, but there is no
significant change in the normal tonemic arrangement.ha In the
first example, the utterance is normal and unemotional. The second
example begins on a relatively high key, slides lower, then rises.
The third example, using the same utterance, begins on a lower key,
but rises abruptly before termination. 'we cannot, however, attribute
any specific emotional 'meaning' to the second in comparison with the
third. we cannot analyze this type of intonation resulting from
key shifts because we cannot set up accurate controls, nor can we make
valid distributional statements.
13.0 A type of intonation which we can analyze is that used to dis-
tinguish certain types of questions from simple statements, one speaking
style fromianother. Questions with question words such as /§w3im§gb t/
'Where is he going?', have the same terminal contour pattern as simple
statements: /Imdofindbt/"He is down there'. Questions without question
words share the same terminal contour /'t /, but have /" / in the final
syllable raised to / ' / or / ’ /, thus: /Imdound5r/. But when / ‘ /
is raised to / ' / or / ' /, it is always before final pause. When
/ ' / precedes / ‘ /, / ’ / becomes / ‘ / in question form. Some

examples are

haPike, 22-5.

 

 

“r7",' 2*. _,.. -I: ~"

W W'-

 

I list. 1J1!
Ill; Ii i ii
4 II I

 

30

/nainalt1/ 'ninth night'
/nainaitt/ 'ninth night?‘
/demd'e'd3t / "They are there'

/d§md8d§?/ 'Are they there?‘

Put another way, we can say that for a question without question word,
we have a tone contour rising from / ' /, and from / ' /. For a
statement, not only is it possible to have a contour rising from / ‘ /
as the overwhelming majority of statements do, but also a contour
falling from a low toneme; for example, [bulafig. The lexemes [bula’]
and [bulafiJ have the same meaning and are stylistic variants by final
contour. These stylistic variants cause a good deal of trouble and,
for the speaker of Standard English, result in utter confusion since
he associates a rising tonal contour before final pause with 'question
intonation'; in Jamaican Creole there may be no question involved at
all. Final / ' t / and / ' t /"question' are unambiguously differen-
tiated from / ‘ t./ 'statement' by tonal contrast, but that /" t /
becomes /" ? /'when 'statement' becomes 'question', is a difference
likely to be unperceived by a stress-oriented listener.

lh.0 Although Hockett sees no sure solution for the problem of stylis-
tic variants, h9 there seems to be a real need for one in Jamaican
Creole to handle all the data, There are two courses open to us: to

treat stylistic variants of tone as belonging to the same dialect,SO

 

h9Hockett, 279.

50'If linguistic change results in groups of persons between which
communication is disturbed, these groups speak dialects of the language'.
Bloomfield (ed.). Sol Saporta, 32.

 

 

 

 

31

or to treat them as'shared features' 3 that is, given two languages
I and Y with X2 and Y2 occurring, we can set up z_as belonging some-
times to X and sometimes to Y. we could state the relation of g_to
both X and Y by providing intrastructural rules. For Jamaican Creole
this approach seems feasible but outside the sc0pe of this study,
since the whole grammar of the dialect would have to be analyzed
together with any other related dialect, or dialect in 'contact'.
‘weinreich refers to what we symbolize as z.as 'interference phenomena',5
which are particularly important in areas where bilingualism or semi-
bilingualism is common. Certainly, Jamaica is a rich field for examin-
ing languages in contact, with especial emphasis on lexemes whose dis-
tribution is the same but whose tonal domains differ. Hitherto, of
course, emphasis has been generally placed on segmental features with
reference to vocabulary items.52
In this study we suggest a shift of emphasis away from the lexico-
statistical approach to investigation of the systematic features of
suprasegmental phenomena within one creolized language--Jamaican Creole.

Our hepe is, of course, that similar studies can be made for all creol-

‘_

51'. . . Two or more languages will be said to be in CONTACT if they

are used alternately by the same persons. The language-using individuals
are thus the locus of the contact . . . Instances of deviation from the
norms of either language will occur in the speech of bilinguals as a
result of their familiarity with more than one language, i.e. as a
result of language contact, will be referred to as INTERFERENCE
phenomena'. Uriel Weinreich, Languages in Contact (New York: Linguistic
Circle of New York, l9Sh), l.

 

52Weinreich.

 

32

ized languages. For Jamaican Creole, however, we have a suprasegmental
system of three tonemes and one intoneme, no inherent phonemic stress,
and a complicated interlocking of stylistic variants. Emotional
categories--pain, anger, joy-—correspond to complete key shifts for
which there is as yet no apparatus for placing them within a systematic
framework.
15.0 Assuming Jamaican Creole to be a dialect53 of Englishsu poses
several related problems: the nature of an 'overall pattern' for
English if we accept the hypothesis of 'overall pattern'; the criterion
of mutual intelligibility for related 'dialects' of a single language;
and the linguistic apparatus by which we fit related 'dialects' into
a single language.

Generally, the idea of an 'overall pattern' for English has been

oriented toward lexical and segmental problems of relation, distinc-

tiveness, and patterning.55 Very little attention seems to have been

 

53Caught in the difficulty of differentiating 'dialect' from 'language',
‘weinreich writes: "It is immaterial whether . . . two systems are
'languages', 'dialects of the same language', or 'varieties of the same
dialect'. The greater the difference between the systems, i.e. the
more numerous the mutually exclusive forms and patterns in each, the
greater is the learning problem and the potential area of interference".

Sh'As for the English part of the vocabulary, it is overwhelmingly the
largest: of the more than four thousand Jamaicanisms that this book
[Jamaica Talk] records, the non-English element, including mixed com-
pounds, is less than ten per cent; a good ninety per cent are of
English origin, though often altered in form'. Cassidy, 397.

55Hockett, 33h-7.

 

33

paid to the function of suprasegmental phenomena within dialects of
English except in a cursory way.56 Hockett even suggests the possibility
that French and Italian may be fitted into essentially the same set;57
i.e. that they may be related to the same overall pattern synchronically.
But if we fit Jamaican Creole into an overall pattern of English, we
must either attempt to discount tonemic and intonemic overlapping, or

we must explain how tonemes can be functional in one dialect of

English but not in others. And what happens to segmental grammar

where utterances have no segmental structure of predication as in
/miéribrofin/’Mary is brown', and other utterances where alternate
constructions of the same morphemic elements have the same semantic
distribution: /3fimIditin/ : /ditin5fim1/ 'The thing belongs to me'?

To try to fit Jamaican Creole into an overall English pattern might
prove to be more complicated than the pattern itself, and merely

obscure the relation of 'dialect' to 'common core'é8 If we should
succeed, however, in fitting Jamaican Creole into this 'overall
pattern', granted its desirability, we might then conceivably have no
grounds for rejecting such purely tonal creolized dialects as

Saramaccan, "the language of a people descended from escaped slaves . . .
living along the upper reaches of the Surinam River, in the interior of

Surinam. The language is related to the 'Negro English' of the coast

 

56See Pike, The Intonation of American English (7th printing. Ann
, O

Arbor: University of‘Michigan Press,‘l9

57Hockett, 336.

 

8
5 Hockett, 33h.

.IL

 

 

 

 

 

3h

59

(Sranan) and to the other Bush-Negro languages, Aucaan and Matawari".

We should also have to discard the criterion of mutual intelligi-
bility, since Jamaican Creole, at any rate, is only partly intelligible
to speakers of a 'standard' English dialect—-who find the process
something like identifying a few French words here and there in a
large context. The interference with mutual intelligibility seems to
result from tonemic and intonemic overlap plus difference in syntax,
rather than from difference in lexicon.60

The linguistic apparatus for fitting Jamaican Creole into the
'common core' would have to be a series of clearly stated intrastruc-
tural or rewrite rules for relating Jamaican Creole to the overall
pattern. Possibly, the necessary transformation rules can be worked
out, and may provide an interesting problem which a computer system
can attempt to solve. It does appear, nevertheless, that much would
be lost in an attempt to relate Jamaican Creole to an 'overall pattern'.
Sameness would be emphasized where total structures are only tenuously
similar and essential differences would be obscured.

The type of analysis attempted in this paper and the particular
emphasis on suprasegmental phenomena suggest that Jamaican Creole and

other creolized languages may better be classified, not as a part of

 

59

J. Voorhoeve, 'An orthography for Saramaccan', werd, XV(December,
1959), No.3, b360

60John Lotz writes: 'In Scandinavia . . . Danish, Norwegian, and
Swedish, though mutually intelligible, are regarded as different
languages because of their political status-~whereas Lappish, which
includes a number of nonmutually intelligible language communities,
is usually counted as one language'. Frontiers of Knowledge, as
reprinted in fsycholinguistics, 2. ..

 

 

35

some central language-- English, French, or Spanish-~with which they

share, or may share, only certain significant features, but as

separate systems of communication, suprasegmentally unique. It is in
the function and pattern of tone (and stress where applicable), that
creolized languages differ most--Jamaican Creole in the overlapping of
tonal levels (tonemes) with tonal, final contours (intonemes).
Considering related languages from the systematic structuring of their
suprasegmentals might tell us a good deal more about why languages with
similar lexicons fail to be mutually intelligible, and why lexicon is
not necessarily a reliable and certain index for determining that

languages are related.

 

W“
-_

 

....

all!“ “‘5." ' '

 

36

APPENDIX

16.0 David DeCamp's paper, 'Diasystem vs. Overall Pattern: The
Jamaican Syllabic Nuclei', read at the December 28, 1961, LSA meeting

is particularly interesting from the viewpoint of the theoretical
presentation of the segmentals of Jamaican English, although it neglects
to point out what bearing the suprasegmentals might have on vowel
distribution and the total vowel inventory. My use of the term 'creole'
assumes a narrower and more rigid distinction than he makes. The gist
of his thesis is that

There are some sets of dialects of English which can-

not be incorporated into any overall pattern yet prOposed.
I reject these alternatives: 1. declare dialect M to be
'tcn-English'; 2. declare all Jamaican to be non-English;
3. expand the overall pattern to twelve vowels, or fourteen,
or twenty-six, or whatever it takes to cover all varieties;
and h. scrap the overall pattern in favor of something
like weinreich's diasystems. I reject these and suggest
retaining the old three-times-three as a sort of 'kernel',
with transformation-like conversion rules to account for
the varieties like dialects M and L. These rules are
tricky and involve levelqmixing, but at least some of

the logic involved has been worked out by Chomsky in the
Third Texas Conference and in his Logical Structures of
Linguistic Theory. Dialect M is the result‘of Dialect
H'plus conversion rule number one (applied after the
allophonic rules, so that it applies to a.mixed string
containing both phonemes and allophones). Dialecg L is

the result of H plus all of the conversion rules. 1

 

 

 

61Excerpt from a letter to me from David DeCamp at the University of
Texas, February 23, 1962.

David DeCamp

37

DIASYSTEM vs; OVERALL PATTERN:

The Jamaican Syllabic Nuclei

The University of Texas

December 28,1961

 

 

 

 

 

 

 

 

 

 

 

17.0
Dialect H:
/1/ /u/ /1h/ /uh/ /1y/ /uw/
bit book beer tour beat boot
/e/ /e/ /eh/ /eh/ /oh/ /ey/ /ow/
bet bud bear bird tore bait toe
/c/ /o/ /-h/ /oh/ /ay/ /oy/ /W/
pat pot lard lord bite boil bout
Dialect M:
/i/ /u/ /ih/ /uh/ /1y/ /uw/
bit book |beer 2 tour beat boot
/5“ /b‘h/
bay toe
/e/ /e/ /eh/ /eh/ /oh/
bet bud bear bird tore
E/ae/ /o/ /ah/ /oh/ /ty/ /ay/ /aw/
pat pot lard lord bite boil bout
IDialect L:
_/1/ /u/ /1h/ /uh/ /iy/ /uw/
,bit book beer tour beat boot
, bay toe
/e/ /o/ bear tore
ibet bird
1 bud
/a/ /ah/ /a.v/ /aw/
pot lard bite bout
pat lord boil
62

' raised

 

 

 

38

17.1 PHONEMIC DIASYSTEMS:

D1 3 H,M,L//i ~11 seHL /e/R$ L /a/ ~HsL/O/ //

/e‘~e/ /e~ 3 ~ 0 ~ 9/ /o" ~ 0/
H M

 

 

 

M ,M
D2 :H,M,I//e~o~g,L /i/~L/a/ ~H,L /1_1/
/ie“/ /ae~a~a~e/
M H,“ M /u ~ O‘/
17.2 INTRASTRUCTURAL AILOPHONIC RULES: '

/e/ + /y/-—) /e/ * /y/ Generalized:63

/e/ + /h/—» MK + me /P/ + /P1/ Mp1 + /p1/p
9

/0/ + /W/—-)/0/w * /W/o

/o/ + /h/——>/o/h + No

1- [z :1 ——» m

. /1/
2 [Mr /n/ -+ /u/ + /h/

3. /
/?/} H/a/
h. /ah/——) /8/

S. /J/ /a/ in instressed [sic] syllables
—) /o/ in stressed syllables

 

63The symbols under 'Generalized' seem to be a rule statement for
the phonemic intrastructural transformations' under 17.2.

39
18.0 Although our primary concern is not with a complete listing of

phonemes 229 their a110phones, a brief discussion of phonetic variations
among segmental data would not be amiss. It would not be amiss be-
cause there are phones in Jamaican Creole that do not occur in any
variety of English which we can call 'standard'. LePage's statement
that the consonant system.is 'the same as that of RP'6h obscures the
phonetic facts important to the assignment of allophones to their
phonemes. DeCamp's approach is to state a 'complete' phonemic list of
items for one informant with the comment that "there is no single
Jamaican English dialect. Jamaican idiolects range in a continuous
spectrum from the speech of the English expatriates to the 'bush talk'
of isolated villages".65 I doubt whether DeCamp would still maintain
this position for a structured description. Certainly, there is a
good deal of phonetic variation among speakers of different dialects
of the same language, but the variations are within the same dialect.
Further, DeCamp lists only one allophone for each consonant phoneme:66
19,0 /p/rp1, /+./ m, / as / m, /k/ m, /b/ m.

/d/ Edi, /3/ [3], /g/ Eel, /f/ [f]. /s/ [8].

/§/ [5], /h/ [h]. /v/ [v], /z/ [z], /m/ [m].

/n/ Inl. /n/ rm, /1/ [17. /r/ [a]. /W/ [W].

/J/ [31.

 

6hLePage, orbis, 383.
65LePage and DeCamp, 135-7.

66LePage and DeCamp, 137.

hO
My list (3.0) does not include /v/ or /h/. DeCamp's listing of

/v/ results from his grouping all the dialects of Jamaican English
together under the name of Jamaican Creole. Instead of /v/, my corpus
has [bl as in [Iqbal 'lover'. A letter from DeCamp agrees with my
finding not only [b1 but [gl, a bilabial and a velar fricative, res-
pectively,67 neither of which is mentioned in DeCamp or LePage.

To /b/ I assign rbl because there is no /v/ in Jamaican Creole. Both
[b1 and [b] are phonetically similar and in free variation. Similarly,
both [g1 and [31 are phonetically similar and in free variation.

As for /h/, we are faced with setting up a 'floating' phoneme
or not setting it up at all for Jamaican Creole, since it may or may
not be present before vowels and is often absent where we would ex-
pect to find it, as in [aaba] 'harbor'. Its distribution cannot be
accurately stated. It may or may not precede vowels and there are no
minimal forms in which [h] is distinctive. To set up /h/ for Jamaican
Creole, violates the linguistic facts.

Neither /a/'nor /C/ is part of the phoneme inventory of Jamaican
Creole.68 If we were to include /v/, which occurs in Standard Jamaican
English, then we would logically have to include /5/'and /C/, which
also are members of the phoneme inventory of Standard. we cannot whim-
sically include one phoneme and exclude two others by applying one set

of criteria to /v/ and another to /8/'and /C/.69

 

67DeCamp's letter

68Cassidy, 36; LePage and DeCamp: 137-

69Missing from Jamaican Creole, /8/ and /D/ are also missing from
virtually 'all English-based pidgins'. Robert A. Hall, Jr., 'Pidgin
Languages', Scientific American, 200:2. 12h-3h.

 

 

 

hl

20.0 Jamaican Creole shows further variation from Standard English

70
in consonant patterning. In cluster /ny/, /n/ has the a110phone [h/;

71 Initial clusters

consonants are palatalized before front vowels.
/skr, sp, spr, st, str/ from Standard become /kr, p, pr, t, tr/ as in
[kraani] 'scrawny', [pwail] 'spoil', [prinfl] 'springe', Etik] 'stick',
[tranga] 'stronger'. Before nasals, /s/ is often syllabic as in

[smadi] 'somebody', but is sometimes followed by a.vowel as in [Siniekl
'snake'.72 In alternation before pause are nasals and nasalized vowels.

For example, [iml ~'[I] 'he, him', [ an] 'can't'.
21.0 The vowel phonemes present problems that previous studies do not
seem to have resolved.73 LePage lists the vowels and their allophones

in the following order:7h

/a/ short, Open, central
hfllmgommcmud
/ii/ corresponds to RP [13

/i/ " " " [I]
/nu/ " .... [u]
/u/ " " " [U]
/e/ " " n [a]

7OLePage and DeCamp, 137.
71LePage, 383.

720aasiqy, hoz.

73DeCamp, 135-6; LePage, 382-3.

7hLePage, 382-3.(LePage uses IPA transcription.).

llll'"l

 

 

 

142
Rising (stress) diphthongs: /ie/, fies/ 'face'

/uo/, /kuot/ 'coat'
Falling (stress) diphthongs: /ai/, /tail/ 'tile'

/ou/, /mout/ 'mouth'

DeCamp omits /aa/ from his list of vowel phonemes that are otherwise
identical with LePage's, but writes that '/a/ regularly occurs both

short and long'.75 It is not clear whether 'short and long' indicate

a phonemic distinction or merely a phonetic one.76

22.0 We have three clearly possible ways of interpreting the data
relevant to a structural description of the 'long vowels' of Jamaican
Creole- We can say that all instances of 'long vowels' may be re-
Presenbcd as sequences of two short vowels-33, 3919 ii; that 'long
Vowels' may be accounted for by assigning phonemic status to length
(durationL-ah, uh, 11.3.3 that 'long vowels' are conditioned by tone and

have no phonemic value .

*—
‘—

7630 Siortsema in 'Problems of Phonemic Interpretation, II: Long Vowels
in ‘ TOne Language', Lin (1959), VIII, h2-6h, discusses the problem

°f 1°I'lg vowels in YoruBa, and points out that 'long vowels occur', as

'L linguistic statement, must be carefully delimited before the term

'long vowel' can have amr real significance. 'What is to be considered

a long vowel in a tone language? Should it have one and the same level
tone throughout or can it have a gliding tone? Should it be in one
”11.33): or spread out over two syllables? If it can cover two syllables,
would these two syllables be within one word or can we still speak of

‘ 1°98 vowel' when its second part belongs to the next part of the text
”min-d?!

 

 

 

1:3

If we argue for a, 31‘}, ii as sequences of short vowels, we

would have to show, for example, that the second vowel of the first
syllable of Lana, M, {Lid}, 'mortar, boot, whistle' is structurally
necessary. Short vowel sequences would provide a good way to analyze
the data provided the initial vowel of each pair occurred with one
ton. , and the following vowel with a variable one. Let us assume by
way' of supposition *[ ‘a, nu, 11'] and *[a‘, nu, ii] occurring with
lexically different morphemes. Obviously, we would have no choice but
to set up _a_a_, 23, E as short vowel sequences to account for the con-
trasts between *[‘a] and -I-[a"l; {“11} and *[uu]; 44:11] and *[11].
But Jamaican Creole, unlike Sarsmaccen (another creolized language of
the Caribbean area), has no contrasts which would Justify setting up
'long vowels' as a sequence of short vowels. 'Long vowels' do not
Cecur across morphemic Junctures, and the camponent of length always
carries falling tone as in [maata], [bunti], [“111]. To set up */aa,
1m, 311/ as significant units would have no explicit or implicit
'tructurei value.
our second option is to assign phonemic status to vowel length
3° that we have */neht.a, buhti, fihfi/ in which 2 represents length
(dur‘tion). That falling tone is a consistent feature of the second
°°mp°n9nt of a 'long vowel' may appear to lend this thesis some sup-
port. Items such as [than] 'fresh', [buil] 'boil' show tone per-
turb“‘-‘-ion on vocalic glides. The first example above carries a high-
1°‘"1 tone, the second, a high-falling tone. Since tone is variable
with honeyllabic vocoids of complex nuclei, we would have a length
phat-11.1“. for 'long vowels' and would have to distinguish them from

vocnic glides with tone perturbation. Our structural statement would

 

 

 

 

M:

be extremely complex and would reveal neither an accurate nor an

adequate picture of Jamaican Creole. The attempt to establish a
lengt‘ta phoneme implicitly assumes tone to be phonetic or part of
the intonational framework.

If tone were phonetic, we would expect forms such as *[mahta,
buhti, fihfi] to represent adequately the significant linguistic
features of Creole. An informant should respond favorably to vowel
length as '-emic ' , and vowel length should include all the phonetic
features necessary for identification of the cited forms. Unfortunately,
vowel length with high-falling tone removed and stress substituted
results in unintelligibility. The informants consistently failed to
identify the form as Creole. Stress was then placed on the final
syllablc with the same result--unintelligibility. Speakers of a
'standard' dialect of English almost invariably substitute stress for
Pitch in Creole and are therefore not understood by native speakers.

A length phoneme would certainly not be helpful to someone learning
0’9013 , since length is not significant. If we persisted in setting

up ‘1 length phoneme, we would be dangerously close to an implicit
StabIHent that Creole has neither significant tone nor significant
pitch; That informants consistently failed to identify length (duration)
as distinctive where tests were made suggests that tone must be con-
”(1°de significant, £123 vowel length. A length phoneme does not make
P°33ible a satisfactory structural statement, since it does not accom-
mod‘t. all the pertinent linguistic data relevant to vocalic nuclei.

If we are 'in accord with the criteria of completeness, consis-

 

 

 

 

hS
tency, and simplicity”,7 we must postulate a reasonable structural

interpretation for the analysis and synthesis of all meaningful data.
We find that stress in Creole is no substitute for tone, and that
'long vowels' cannot be satisfactorily analyzed as vowel plus length
without obscuring tone flmction. Our third possible structural inter-
pretation of the phonetic data is to set up tone as '-emic' and
vow]. length as '-etic'. All 'long vowels' have high-falling tone

as in [maata, bu’utzl, fiifi]; usually, high-falling tone occurs with
complex nuclei, sometimes high-level, and sometimes low-level tone.

By analyzing tone as '-emic' and 'long vowels' as '-etic', a satisfac-
tory if not unique structural interpretation is given. A non-Creole
Speaker given a form such as /bwana/ and told to place high-falling
tone on the vowel of the first syllable will reproduce the utterance
Clearly enough for it to be identified as Creole by a native speaker.
Th. voicel also becomes phonetically long. The treatment of Fas, uu, ii]
33 ‘ll—Ophones of /a,u,i/ conditioned by tone seems to offer the best
“mt-11rd description of the data for conciseness, simplicity and com-

Pl'hncss. The following vowel distribution is shown with apprOpriate
t'°n°"\i<': domains .

 

77
Archibald A. Hill, Introduction to Linguistic Structures (Harcourt,
Bragg. and Company, 19335 32-3.

 

 

 

 

"I'nll'-’ I

 

23.0 Phoneme

4/1/

/h/

/a/

/0/
/o/
/1e/
/uo/

/ai/

/ou/

ho
Allophone
E 111

r1]
[nu]
[u]
[n]
[a]
[:3
[Ad
tie]
[go]
[all

[on]

Tonemic Domain

/ ./ /i‘ir1/ 'whistle'
/“/; /‘/

/'/ Anili/ 'coolie'
/ ‘/, / '/
{sfi‘s‘zfiiié slaw
/ 7. / '/

/./ does not occur
/./ does not occur
/‘/, / ‘/. / '/

/'/, / ‘/, / '/

/‘/. / ‘/, / "/

/'/, / ‘/, / '/

 

 

 

b7
2h.0 It is worth our while to consider the mechanical tonometer

analysis presented below because of the inferences we can draw. At
the same time, we must keep in mind the limited function of a machine
ane1.ysis.78 Father A.H. Jones of the University of London has kindly

provided me with a statement which I reproduce.
Zhel The reading scale on the tonometer is arranged thus:

a. The whole numbers represent pitches which are a
semitone apart. The lower the figure, the lower
m. Thus '12 10' means that in this two-
syllable word, the first syllable is two semi-
tones (l2-10) highdr than the second one: the
word thus has this tone-form: ’ -

b. The fractions are fractions of a semitone. Thus
in sentence 1, 'naata' falls through 2% semitones.

c. Sliding tones are marked thus: 12%,- 1‘\10 means
'slide down from 12% to 10' .

l‘tae language on the whole appears to have a remarkably small
Overall range of pitch-much less than English, a feature Professor
Bcrry and I have noticed EIso in Liberian Creole.

I presume that you want to know the pitch-behaviour of the
key words (naata, bula, et.) in the various contexts of the
different frames, and I have only taken the pitches of these

key-words.

These key-words preserve the same itches in all the contexts,

with a few exceptions noted below. ave therefore given you
K ..a

 

78.0“. might think that the tonemes could be analyzed by instruments,
inasmuch as precise frequencies can be determined from instrumental
"cords. These measurements are a great help for describing the physical
nature. of the tonemes after they have been otherwise discovered, but do
not c=<>ntribute greatly to their analysis since (1) it is the relative
Plrh of tonemes which is significant, rather than their absolute pitch,
3: ( 2) tonemes change under various conditions so that the intervals
covnot remain fixed in such a way that they may be mechanically dis-
"1 -‘t‘ed; instruments merely record gross fluctuation rather than analy-
f‘gg it in terms of deviations of units within a system‘. Pike, Tone
guages, 21.

 

_- 7'
L

 

 

 

h8

<>nly the standard reading, taking just one sentence, and
<>ccasionaIIy cohfirthE It by noting another. But I worked
<>n the tonometer through all the frames.

The exceptions are these:

a. di naata daun de.
Hire, Ehe firsE—syllable of naata is lowered by
nearly two semitones: but the parent form of naata,
i.e. a falling pitch, is still preserved. On
listening to the tape it sounds as it in the frame
'di (ke 0rd) daun de‘, in every case the keyword
loses its first high tone and that both syllables
are spoken on a level tone, on the same pitch as
'di'. BUT THIS IS NOT SO. In every case the key-
word contains a fall in pitch, though in 'naata'
and in 'do i' the fall is smaller than it Ii'ii
the other rames for these words.

b. Do i daun de.
The firsf syllable of 'dopi' starts 1/3 of a semi-
tone lower than in the ther frames for 'dopi'.

 

 

0
e

a fi mi di dopi.

N053 that the pitch of 'do i' is inverted.

I wonder if your speaker thought he was asking a
question here? This is the only case in your-list
of such inversion, with a medium first syllable
and a rise and a fall on the second one.

The tonometer Readings
a. NAATA

 

12% 10
mi waen naa - ta

{‘10 2/3 10 2/3 10 lo 10 2/3 10 2/3
di nae ta daun de

b. EULA
{ 10 2/3 10
fi mi di bu — la

10 2/3 10
bu — la daun de

 

 

 

h9

c. BUBU
"""' 10 2/3 10
fi mi d5. bu - bu

 

di bubu a fi mi
di EuBu daun de BUBU has same reading as above

d. DOPI

 

10 10 2/3 101/3
efimidido .11 ”“

 

10 2/3 10
di do - pi s :1 mi
10 2/3 10
Ill di do - pi daiya

 

10 1/3 7 10 2/3 10 2/3
do - pi daun de

 

e. TAAWA etc. (The recording is especially poor for
these words)

1 10
as» w. m

12% m 10 2/3
man I!

13 'vwfik 10 2/3

 

wen ts
12% 10 . . . . (probably; as near as I can
di gee - ta e.fi mi see)
13% 10

we yu get d1 man - te tram

{12% M10 2/3
lee - ts

 

 

 

SO

2h.2 The conclusions we can draw from the tonometer analysis largely

support the auditory analysis which puts forth an argument for a
stmc tural classification of the phonetic tonal data into three tonemes.
The facts seem to be in favor of a relatively small tone variation
within any utterance; but any utterance may be arbitrarily raised or
lowered in total key with astonishing frequency. The tonometer analysis
tells us little about stylistic tone contours for a number of possible
rte-Leone: there might have been enough loss in the treble range of the
taped material, especially at the end of utterances, to obscure slightly
rising and falling contours of tone 3 the tonometer readings are rigor-
Oully applicable to substitution items rather than to final frame items 3
more importantly, the frames were designed to exclude stylistic contours
which commonly occur in free discourse.
2h-3 The exception to the tonal distribution mentioned in 2h.l (/dopi/),
PrObflbly derives from the informant's mimicking the investigator's ques-
tion intonation. The precise circumstances cannot be given. That
/d°P1/ represents the only instance of tone reversal in all the test
frames would seem to support the hypothesis that the informant may have
been responding to the investigator's question by mimicking hill. There
18 aCLao the possibility that although the frames were designed to
exclIJlde stylistic variation, a final rising stylistic contour intruded.
25'0 Inherent word stress is nonphonemic. Minimal fr": forms like
[pit-16th, k'o'taka, talawi/ 'potato, basket, strong', and pairs like
Anit‘a/ : /m'a'ti/ 'mortar','matter' have inherent phonetic stress in
”‘9 domain /'/, / - /, / ‘ / of three-syllable free forms, but initially

on two-syllable free forms regardless of tone differences. In three-

 

 

 

 

"a‘lll" -

 

51
syllable free forms, stress occurs medially when preceded by / ‘ / and

in the domain / . / as in /bITIéta/; finally, preceded by / ‘ / and in
the domain / ' / as in /kataKfi/; initially, in the domain / ' / and
followed by /' / as in /TXl§wi/. we can speak of one degree of phonetic
stress in comparison with no stress, but no phonemic stress.

Separate from inherent phonetic stress is stylistic, emphatic
stress which cammonly occurs at the syntactic level. In sentence-type
utterances, phonetically long vowels and complex nuclei commonly carry
emphatic stress as in /ru623pln3tipl?/ 'A rose apple is not a star apple'.
Emphatic, stylistic stress here falls on the penult, but not on the
penult in /3rimiaimstht5whdit/ 'The tawa belongs to Mr. Martin', where
emphatic stress occurs on the third syllable from the end in the
domain /fl/.

26.0 The short selection which follows is taken from free discourse
and illustrates same stylistic features of story-telling apart from
the distribution of tonemes and pause boundaries; for example, phrasal
redundancy, and the repetition of ’you know', an attention-getting
construction. The style intoneme coincides with short breath-group
pauses; the double-cross with longer terminal breath-group pauses.
Story-telling in Jamaican Creole can call for frequent pause breaks,
or it can continue over several sentence-type utterances without any

audible pause.

 

 

 

 

52
27.0 /hIst§5n£nsinou t imwozo T Imwozo ? waizmin ? waizyunuo f #

slnsibl f # wSmisEImsEnsibl ? iisfins ? # Imibibiguot t # guot ?
5:11 # ksmlai'k‘. 1 kodkya’r 1 imznlnwair 1 ,4! iniiya'd 1 #
Imwintopli‘ T dItrikaniwaif ? siwidiguot ? nofi ? imss f

waif 1 Erilsik 1 yfinou 1# irflaik 1,1; Ewigaidzkt‘. 1# hugzedguéb‘a 1,11
batodtditfifia 1;; Insolmibtdtizk 1 wimisttfititk 1 titiak 1#
d‘itlguibi 1 baromltda‘. 1#/

Mr. Anancy, now, he was a, he was a wise man. Wise, you know.
Sensible. When I say sensible, he had sense. He had a baby goat.
Goat child. Something like a cow-calf. He and his wife were in
their yard. He wanted to play a trick on his wife. She was
with the geat, now. He said, 'Wife, I feel sick, you know.

I feel sick. I want to go to the doctor.‘ He got two guavas
that pushed out his two jaws. He said he had toothache. When
I say toothache, toothache. His two gusvas pushed out his two

JNUe

 

 

 

1.

2.

3.

h.

9.

 

53

LIST OF REFERENCES

Books

Cassidy, F.G. Jamaica Talk. London: Macmillan Co. Ltd. 1961.

 

Gleason, H.A. An Introduction to Descriptive Linguistics. New York:
Holt, Rinehart and Winston, 1961.

Harris, Zellig. Structural Li¥§uistics. Chicago: University of
Chlcago ss, .

Hill, Archibald A. Introduction to Linguistic Structures. Harcourt,
Brace and Company,’1958.

 

 

Hockett, C.F. A Course in Modern Linguistics. New York: Macmillan
Cempany, I958.

LePage, R.B. and DeCamp, David. Jamaican Creole: Creole Language
Studies. London: Macmillan Company, Ltd. 1960.

 

Pike, Kenneth L. The Intonation of American En 1ish. Ann Arbor:
UniversityfofrflichiganPress,19% .
. Tone Languages: A Technique for Determining the
Number and Type of Pitch Contrasts in a Language,
with Studies in Tonemic Substitution and Fusion.
Ann Arbor:‘University of Michigan P;33., I957.

 

 

 

 

Richardson, I. The Role of Tone in the Structure of Sukumu. London:
School of African and Oriental Studies, 19§-.

10. Saporta, Sol (ed.). Psycholinguistics. New YOrk: Holt, Rinehart

 

and Winston, T961.

 

 

 

10.

11.

12.

13.

15.

Cassidy, F.G.

Gleason, H.A.

Hall, Robert,

LePage, R.B.

Siertsema, B.

Voorhoeve, J.

American, 200 3 2 e 12’4’3’4 e

Sh

ARTICLES AND PERIODICALS

'English Language Studies in the Caribbean',
American Speech, XXXIV(0ctober, 1959), 163-71.

 

Review of the R61e of Tone in the Structure of
Sukumu by I. Ribhardson,LanguagETApril-June, 1961),
37:2. 302.

Robert.A. Jr. 'Pidgin Languages', Scientific

 

'General Outlines of Creole English Dialects in the
British Caribbean', Orbis: Bulletin International
de Documentation Linguistigue. Louvain: Centre
International de Dialectologie Ginerale. 6:2, 195?.

 

'Problens of Phonemic Interpretation, II: Long
Vowels in a Tone Language', Lingua(1959) VIII, h2-6h.

'An Orthography for Saramaccan’, Word(December, 1959).
1533a 1136s

 

 

 

 

 

‘3’!
1:, .. .

 

W5!“ -. v; V l???"

 

 

 

111131111

31293

NSTATE UN

 

11111111111111111

003