Citation
South-east Asian linguistics

Material Information

Title:
South-east Asian linguistics essays in honour of Eugénie J. A. Henderson
Series Title:
Collected papers in Oriental and African studies
Creator:
Davidson, Jeremy H. C. S
Henderson, E. J. A ( Eugénie Jane Andrina ), 1914-1989
Place of Publication:
London
Publisher:
School of Oriental & African Studies, University of London
Publication Date:
Language:
English
Physical Description:
xi, 235p : ill, port ; 25cm

Subjects

Subjects / Keywords:
Languages -- Southeast Asia ( lcsh )
Genre:
festschrift ( marcgt )

Notes

Statement of Responsibility:
edited by Jeremy H. C. S. Davidson

Record Information

Source Institution:
SOAS, University of London
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
129321 ( aleph )
0-7286-0154-0 ( isbn )
Classification:
GM410 ( ddc )

Downloads

This item has the following downloads:


Full Text
SOUTH-EAST ASIAN LINGUISTICS
Essays in honour of
Eugenie J.A. Henderson
Edited by J.H.C.S. Davidson
School of Oriental and African Studies




Collected Papers in Oriental and African Studies
SOUTH-EAST ASIAN LINGUISTICS:
Essays in honour of Eugenie J.A. Henderson
Edited by
Jeremy H.C.S. Davidson
Lecturer in Vietnamese
School of Oriental and African Studies
SCHOOL OF ORIENTAL AND AFRICAN STUDIES
UNIVERSITY OF LONDON
1989


Published by the School of Oriental and African Studies
(University of London), Thornhaugh Street, Russell Square,
London WC1H OXG
(c) School of Oriental and African Studies, 1989
British Library Cataloguing in Publication Data
South-east Asian linguistics: essays in honour of
Eugenie J.A. Henderson. (Collected papers in
Oriental and African Studies)
1. Asian languages South-east
I. Davidson, Jeremy H.C.S. (Jeremy Hugh Chauncy Shane)
II. Henderson, Eugenie J.A. (Eugenie Jane Andrina,
1914-)
495
Printed in the United Kingdom
by Hobbs the Printers of Southampton
(2262/89)


CONTENTS
Page
CONTRIBUTORS vii
PREFACE ix
EDITOR'S NOTE x
EUGENIE J.A. HENDERSON: A BIOGRAPHICAL NOTE
- R.H. Robins 1
PUBLICATIONS OF EUGENIE J.A. HENDERSON
- Helen Cordell 5
mat ca: FROM "FISHES' EYES" TO "ANKLE BONES"
-- A VIETNAMESE CALQUE? Jeremy H.C.S. Davidson 11
SOME FEATURES OF MODERN KHMER LITERARY STYLE
- Judith M. Jacob 23
KHASI KINSHIP TERMINOLOGY *Lili Rabel-Heyman 43
ON PROSODIC RELATIONS BETWEEN FIJIAN BASES AND
VERBAL SUFFIXES G.B. Milner 59
A SIAMESE LETTER DATED 7 DECEMBER 1776
- Soren Egerod 89
TAI NAMES FOR THE OX William J. Gedney 111
FIRST AND LAST IN THAI, OR THE ORDER OF OPPOSITIONS
- *Mary R. Haas 129
LA TONOLOGIE DU LI DE HAINAN A.G. Haudricourt 133
PROTO-TAI *kh and *x- Li Fang-Kuei 143
UNCLES AND AUNTS: BURMESE KINSHIP AND GENDER
- David Bradley 147
THE BULGING MONOSYLLABLE, OR THE MORA THE MERRIER:
ECHO-VOWEL ADVERBIALIZATION IN LAHU
- James A. Matisoff 163


Fage
THE YAW DIALECT OF BURMESE John Okell 199
ORAL VOWELS AND NASALIZED VOWELS IN LEPCHA (RONG)
AS THE KEY TO A PUZZLING VARIATION IN SPELLING
- R.K. Sprigg 219
Frontispiece: Eugenie J.A. Henderson


CONTRIBUTORS
R.H. Robins, Professor Emeritus of General Linguistics, School
of Oriental and African Studies, University of London
Helen Cordell, Sub-Librarian, South East Asia and Pacific,
School of Oriental and African Studies
Jeremy H.C.S. Davidson, Lecturer in Vietnamese, School of
Oriental and African Studies, University of London
Judith M. Jacob (retired), formerly Senior Lecturer in Cambodian,
School of Oriental and African Studies, University of London
^Lili Rabel-Heymann, one-time Associate Professor of Linguistics,
University of Calgary, Alberta, Canada
G.B. Milner, Professor Emeritus of Austronesian Studies, School
of Oriental and African Studies, University of London
S^ren Egerod, Professor and Director, Scandinavian Institute of
Asian Studies, Copenhagen, Denmark
William J. Gedney, Professor Emeritus of Linguistics, University
of Michigan, Ann Arbor, USA
^Mary R. Haas, Professor Emeritus of Linguistics, University of
California, Berkeley, USA
A.G..Haudricourt, Directeur de Recherches honoraire au CNRS,
(Centre National de la Recherche Scientifique) Paris
Li Fang-Kuei, Emeritus Professor of Asian Languages, University
of Hawaii at Manoa, Honolulu, Hawaii, USA
David Bradley, Lecturer in Linguistics, La Trobe University,
Victoria, Australia
James A. Matisoff, Professor of Linguistics, University of
California, Berkeley, USA
John Okell, Lecturer in Burmese, School of Oriental and African
Studies, University of London
R.K. Sprigg (retired), formerly Reader in Phonetics,
School of Oriental and African Studies, University of London
vii




PREFACE
Eugenie Henderson is, as we all know, a most impressive
phonetician and linguist. Not only that, hut, as is obvious from
what follows in this collection of papers, she is also a most
informative and charming woman. Several years ago, I had the
idea of offering this token of awareness to her, and the response
was, as can readily be seen, high-powered.
I have, therefore, divided the contributions alphabet-
ically into language families (from Austro-Asiatic through
Austronesian and Tai to Tibeto-Burman) and within these sections,
once again alphabetically, by the surnames of the contributors,
in an attempt to express the range and equality of interest that
Professor Henderson has in South-East Asian languages.
There is no need for me as editor to discuss the topics
presented; their authors need no introduction and will, without
doubt, attract readers; the papers speak for themselves. With
that comment, I can only say that I, among many others, look
forward eagerly to Eugenie's work on Bwe Karen and all else.
Jeremy H.C.S. DAVIDSON
S.O.A.S. 1989
ix


EDITOR'S NOTE
Certain technical details found in this "book may require
clarification.
Romanizations: Languages -which are not normally written
in roman alphabets are romanized. The systems used are as follows:
Burmese is a transcription of the spoken language, not a translit-
eration of the script, and (where possible) follows the system
advanced by John Okell, A guide to the romanization of Burmese
(London: Luzac 1971, 66-7); Chinese is romanized into pinyvn,
unless indicated otherwise for the sake of clarity in particular
articles; while Khmer is written in accordance with the translit-
eration used by Saveros Lewitz, 1969* Bull. Eo. fr. Extr.-Orient
55, 163-9. Thai follows the phonetic systems described in the
next section. Variations may occur; naturally, names of places
and people may differ.
Transcriptions: Thai specialists who have contributed
papers may have evolved their own phonetic transcriptions (e.g.
Egerod, Gedney, Haas, Li) which vary somewhat from one to another.
Since all Thai specialists will be familiar with these
representations, I have left them as they are rather than choose
a 'standard1 form.
Bold face was not available for a wide range of IPA
symbols for type-setting in this volume. These are set in roman
and, for ease of reading, are underlined when they occur in the
running text.
Translations: Unless otherwise indicated, all transla-v
tions from languages presented are by the authors of the papers
in which they appear. Translations of entries do not always
follow their originals directly but may follow and 'reproduce* a
group, for the sake of linguistic connection.
All Chinese or nom characters are written by Jeremy
Davidson.
'k'k'k-k'k'k'k'k
I should like to thank all those who have helped to
make the task of editing the fifteen contributions easier. I also
wish to thank the Publications Committee of S.O.A.S. (especially
its three consecutive Chairmen, Professors JC. Wright,
R.H. Robins, and C. Shackle, and its Secretary Mr Martin Daly) for
their help and support. Finally, I owe a particular debt of
gratitude to the Editorial Secretary Miss Diana Matias and to the
x


staff of the S.O.A.S. Support Section, especially Mrs Joyce
Hutchinson, and to the photographer Mr Paul Fox for the
production of the present volume.
Jeremy H.C.S. DAVIDSON
The photograph on the cover is "by Hans Hinz and is reproduced
"by permission from Thai Painting published by Office du Livre,
Fribourg.
xi




EUGENIE J.A. HENDERSON: A BIOGRAPHICAL NOTE
R.H. Robins
This introduction is not the place for a biography or
even a full-length appreciation of Eugenie Henderson's scholarly
career. Although she formally retired on 3 September 1982from
her post in the School of Oriental and African Studies and in the
University of London as Professor of Phonetics, no one of her
many friends and colleagues, nor she herself, surely, regards her
retirement from teaching as the final conclusion of her active
participation in linguistic studies and research.. But the
contributions from the international world of scholarship to
this volume of studies, like those in the volume published in
1979 by Chulalongkorn University (Thongkum 1979)-, testify to the
esteem and affection in which she is held by the now large
community of her colleagues, former students, and friends.
Eugenie Henderson's involvement in the School of
Oriental and African Studies began in 19^2, when she was
appointed to a lectureship in phonetics. This appointment, like
those of a number of her contemporaries, was part of the immediate
response by the armed services to the requirements created by the
Japanese entry into the Second World War. The Department of
Phonetics and Linguistics, under the headship of Professor J.R.
Firth, was almost wholly given over between 19^2 and 19^5 to
Japanese language courses for Royal Air Force and Royal Navy
personnel, including some highly specialized courses in Japanese
phonetics for specific intelligence purposes.. Eugenie Henderson's
part in the development, administration and teaching of these
courses was probably second only to that of Firth himself.1
Some of its continuing framework of teaching in the
Department had its origin in the forced response of its staff to
wartime conditions (cf. Firth 19^5), and the impetus behind its
expansion was maintained after the war by the University of
London's implementation of the Scarbrough Report (.19^7), which
recommended a strong development of Oriental and African studies
linguistic, cultural and historical a development which,
naturally, was concentrated in the School dedicated to such
studies.
The post-war expansion of linguistic work in the
Department of Phonetics and Linguistics covered languages from
several different areas of the Far East, the Middle East, and
Africa. While still a postgraduate student under Professor
Daniel Jones at University College London in 1937, Eugenie
Henderson had begun the study of Thai, which laid the foundation
1+


of her lifelong connection with the languages, peoples and
cultures of South East-Asia.
Her studies in Thai were reinforced after 19^5 with
studies, mainly through informants in London, in other South-
East Asian languages, including Vietnamese, Cambodian, and Mon.
In 19^6 the Department of South-East Asia and the Islands was
re-established, and significantly enlarged soon afterwards, with
lectureships in Thai, Cambodian, Vietnamese, and Mon; during
their first years of appointment, the lecturers appointed to
these posts were attached to the Department of Phonetics and
Linguistics and came directly under the supervision of Eugenie
Henderson. The 'pro-seminar1 that she organized as a forum for
the presentation and discussion of their research became a model
for the later postgraduate seminar of the Department.
In 1953 she was appointed Reader and, in I96U,
Professor of Phonetics in the University of London.
Eugenie Henderson's involvement in South-East Asian
languages was further strengthened by her service, after the
retirement of Professor D.G.E. Hall, as Acting Head of the
Department of South East Asia and the Islands, from 1960 until
1966, during a crucial period in the consolidation of that
Department. At the same time she continued to play a full part
as a senior member of the Department of Phonetics and Linguistics
and, from 1966 until 1970, after the retirement of Professor
N.C. Scott, she was Head of this Department.
Throughout the years 19^5 to 1981 Eugenie Henderson
involved herself in the subject of general linguistics as a
whole. Since 195^- she has been a member of the Council of the
Philological Society, serving as the Society's Treasurer from
1965 to 197^, and from 1977 to 1980 she was Chairman of the
Linguistics Association of Great Britain.
Though her studies of the languages of South-East Asia
were mainly conducted in this country, she has made two fairly
extensive visits to that part of the world. In 195^ she visited
Burma as a Visiting Professor of Rangoon University and engaged
in fieldwork on Bwe Karen and Chin, this latter research resulting
in the publication of her book Tiddim Chin (Henderson 1965a),
while in 1975 she taught in Thailand, in a summer institute
organized by the Thailand Research Project in Bangkok.
Eugenie Henderson's publications have been mainly
concerned with the phonetics and phonology of South-East Asian
languages, in some cases breaking entirely new ground and also
applying the theory of prosodic phonology, developed by Firth
during the 19^0s and early 1950s, to different language material.
Her 19^8 'Prosodies in Siamese: a study in synthesis' (see
Henderson 1970a) has for long been recognized as one of the most
1+


thorough and revealing studies in prosodic analysis that we have,
and is regularly recommended to students embarking on this
aspect of phonological theory, especially since its republication
in 1970 in Palmer's Prosodic analysts.
A glance at her bibliography, however, shows that
South-East Asian languages by no means exhaust her fields of
interest. She has contributed papers on typological and
historical topics at international congresses (e.g. Henderson
1965b, Indo-Pacific linguistic studies Pt.jj), and has published
phonetic studies of Ossetic (e.g. Henderson 19^9) and. of other
Caucasian languages. In this latter area, her article 'Acoustic
features of certain consonants and consonant clusters in
Kabardian' (Henderson 1970b), though quite brief, has become
something of a classic in the literature of experimental phonetics.
British phoneticians look back with pride on the great
nineteenth-century phonetician and Anglist Henry Sweet, and Firth
himself often claimed that his prosodic theory was, in part, a
development of ideas latent in Sweet's own work. Sweet's
writings were voluminous, and in 1971 Eugenie Henderson performed
a valuable service to anyone concerned with the history of
phonetics and phonology by selecting and making available the
most important passages from his books and articles in a single
book, The indispensable foundation: a selection from the writings
of Henry Sweet (Henderson 1971).
It was also the happiest of circumstances that in the
month in which she retired Professor Henderson was honoured by
the world community of linguists in speaking as an invited
rapporteur, at the Plenary Session on Phonetics and Phonology,
on the present state and the prospects of this branch of
linguistics, at the Thirteenth International Congress of
Linguists in Tokyo in September 1982. In this year she was also
elected a Fellow of University College London and, in 198^,
President of the Philological Society.
NOTE
1. The exigencies of wartime language teaching left an indelible
and ultimately treasured impression on those of us who were
involved in it. I have a vivid personal memory from 19^
of Eugenie Henderson and myself shouting each other down
while endeavouring to conduct simultaneous classes in Japanese
pronunciation and Japanese grammar as we sheltered with our
students (all service personnel) in the corridor of our
departmental territory, to avoid the worst effects of
VergeItungswaffen T, affectionately known as 'doodlebugs'.
1+


REFERENCES
Firth, J.R. 191+5. Wartime experiences in linguistic
training. Mod. long. 26, 38-1+6.
Henderson, E.J.A. 191+9. A phonetic study of Western
Ossetic (Digoron). Bull. Soh. Orient.
Afr. Stud. 13, 36-79.
................1965a. Tiddim Chin: a descriptive study
of two texts (London Orient. Ser. 15).
London: Oxford Univ. Press.
................1965b. The topography of certain phonetic
and morphological characteristics of
South East Asian languages-. In Indo-
Pacific linguistic studies Pt.II:
Descriptive linguistics (eds.) G.B. Milner
& E.J.A. Henderson. Amsterdam: North-
Holland, 1+00-31+.
................1970a. Prosodies in Siamese: a study in
synthesis. In Prosodic analysis (ed.)
F.R. Palmer. London: Oxford Univ. Press,
27-53.
................1970b.. Acoustic features of certain
consonants and consonant clusters in
Kabardian. Bull. Sch. Orient. Afr.
Stud. 33, 92-106.
................1971. (ed.) The indispensable foundation:
a selection from the writings of Henry
Sweet (Lang. & lang. learning 28).
London: Oxford Univ. Press.
Palmer, F.R. 1970. (ed.) Prosodic analysis. London:
Oxford Univ. Press.
Scarbrough Report 19^7. Report of the interdepartmental
commission of enquiry on Oriental
Slavonic3 East European and African
Studies. London: HMSO.
Thongkum, Theraphan L. 1979- (&t al. ed.) Studies in Tai and
Mon-Khmer phonetics and phonology: in
honour of Eug&nie J.A. Henderson.
Bangkok: Chulalongkorn Univ. Press.
1+


PUBLICATIONS OF EUGENIE J.A. HENDERSON
Helen Cordell
19^0 Specimen of Thai. Le maZtre phonitique 69, 11-12.
19^3 Specimen of Annamese (Tonkinese dialect), with notes on
pronunciation. Le maZtre phon&tique 79, 6-8.
19^+8 Notes on the syllable structure of Lushai. Bull. Sch.
Orient. Afr. Stud. 12, 713-25.
19^9 A phonetic study of Western Ossetic (Digoron). Bull. Seh.
Orient. Afr. Stud. 13, 36-79.
Prosodies in Siamese: a study in synthesis. Asia Major
n.s.l, 189-215. (Repr. with later observations and ideas
in Prosodic analysis, (ed.), F.R. Palmer. London: O.U.P.,
1970, 27-53.)
1950 & H.W. Bailey. Digoron word-list. Bull. Sch. Orient. Afr.
Stud. 13, 381-88.
1951 The phonology of loanwords in some South-East Asian
languages. Trans. Philol. Soc. 131-58. (Repr. in
Prosodic analysis, (ed.), F.R. Palmer. London: O.U.P.,
1970, 5U-81.
1952 The main features of Cambodian pronunciation. Bull. Sch.
Orient. Afr. Stud. 14, 1^9-7^-
1957 Colloquial Chin as a pronominalized language. Bull. Sch.
Orient. Afr. Stud. 20, 323-27.
1959 Karen Languages. Encyclopaedia Britannica. Chicago,
London.
The tones of the Tai dialect of Songkhla. Bull. Inst.
Hist. Philol. Academia Sinica 30, 233-35.
1961 Tonal exponents of pronominal concord in southern
Vietnamese. Indian linguistics 22, 86-97.
Tone and intonation in Western Bwe Karen. J. Burma Res.
Soc. Fiftieth Anniversary publication 1, Rangoon, 59-69.
1963 Notes on Teizang, a northern Chin dialect. Bull. Sch.
Orient. Afr. Stud. 26, 551-58.
Introduction to Linguistic comparison in South-East Asia
and the Pacific, (ed.) H.L. Shorto. London: Sch. Orient.
Afr. Stud. (Collected papers in Oriental and African
studies; 4), 1-6.
1+


1964 Marginalia to Siamese phonetic studies. In In honour of
Daniel Jones: papers contributed on the occasion of his
eightieth birthday..., (eds.), D. Abercrombie et al.
London: Longmans, 4-15-24.
1965 Indo-Pacific linguistic studies, (eds.), G.B. Milner &
E.J.A. Henderson. Amsterdam: North-Holland. Published
simultaneously as Lingua 14 & 15. Part 1: Historical
linguistics; Part 2: Descriptive linguistics. Papers
submitted to the Conference on linguistic problems of
the Indo-Pacific area, London, 5-8 Jan. 1965.
The domain of phonetics: an inaugural lecture delivered
on 5 May, 1965. London: Sch. Orient. Afr. Stud.
The articulation of final ?-nhf and f-chT in Vietnamese.
Proceedings of the Fifth International Congress of
Phonetic Sciences (Munster, 1964), (eds.), E. Zwirner &
W. Bethge. Basel: Karger, 348-52.
Final -k in Khasi: a secondary phonological pattern. In
Indo-Pacific linguistic studies, 1, (eds.), G.B. Milner
& E.J.A. Henderson. Amsterdam: North-Holland, 459-66.
Tiddim Chin: a descriptive analysis of two texts. London:
O.U.P. (London oriental series; 15).
The topography of certain phonetic and morphological
characteristics of South East Asian languages. In Indo-
Pacific linguistic studies, 2. (eds.), G.B. Milner &
E.J.A. Henderson. Amsterdam: North-Holland, 400-34.
1966 Khasi and the 1-clusters in seventeenth century Tonkinese:
a preliminary glance. In Essays offered to G.H. Luce in
honour of his seventy-fifth birthday, (eds.), Ba Shin
et al. Ascona: Artibus Asiae. 1, 139-50.
Towards a prosodic statement of Vietnamese syllable
structure. In In memory of J.R. Firth, (eds.), C.F.
Bazell et al. London: Longmans, 163-97.
1967 L'etat actuel des etudes ethnolinguistiques anglaises
sur l'Asie du Sud-Est. Rev. Ecole Nat. Langues Orient.
4, 85-104.
Grammar and tone in South-East Asian languages. Wiss.
Zeitschrift Karl-Marx Univ. Leipzig, 16, 171-78.
Vowel length and vowel quality in Khasi. Bull. Sch.
Orient. Afr. Stud. 30, 564-88.
1970 Acoustic features of certain consonants and consonant
clusters in Kabardian. Bull. Sch. Orient. Afr. Stud. 33,
92-106.
1971 The indispensable foundation: a selection from the.
writings of Henry Sweet, (ed.), E.J.A. Henderson. London:
O.U.P. (Language and language learning; 28).
6


1971 Structural organization of language: 1 phonology.
In Linguistics at large, N. Minnis (ed.), London:
Gollancz, 35-53.
1975 Phonetic description and phonological function: some
reflections upon hack unrounded vowels in Thai, Khmer
and Vietnamese. In Studies in Tai linguistics in honor
of William J. Gedney, (eds.), J.G. Harris and J.R.
Chamberlain. Bangkok: Central Institute of English
Language, Office of State Universities, 259-70.
1976 Khasi initial clusters. In Austroasiatic studies J,
(eds.), P.N. Jenner et al. Honolulu: Univ. Press of
Hawaii, 523-38.
Thai phonetics sixty years ago: gleanings from the
unpublished notes of Daniel Jones. In Tai linguistics
in honor of Fang-Kuei Li3 (eds.), T.W. Gething et al.
Bangkok: Chulalongkorn Univ. Press, 162-70.
Vestiges of morphology in modern standard Khasi. In
Austroasiatic studies3 J, (eds.), P.N. Jenner et al.
Honolulu: Univ. Press of Hawaii, 1+77-522.
Vestiges of morphology in some Tibeto-Burman languages.
In South-East Asian linguistic studies 2, (ed.), Nguyen
Dang Liem. Canberra: Australian Nat. Univ. (Pacific
linguistics. Series C; 42), 1-17.
1977 The larynx and language: a missing dimension? In The
larynx and languages (eds.), G. Fant and C. Scully.
Fhonetica 34, 256-63.
1978 Notes on Yes-or-No questions and allied matters in Karen
and Chin. In Spectrum: essays presented to Sutan Takdir
Alisjahbana on his seventieth birthday(ed.), S. Udin.
Jakarta: Dian Rakyat, 1+52-68.
1979 Bwe Karen as a two-tone language? an enquiry into the
interrelations of pitch, tone and initial consonant.
In South-East Asian linguistic studies, 3, (ed.), Nguyen
Dang Li£m. Canberra: Australian Nat. Univ. (Pacific
Linguistics. Series C; 45), 301-26.
1980 How can genetic-comparative classifications be based on
typological considerations? (Discussion). In Typology
and genetics of language^ (eds.), T. Thrane et al. Trav.
Cercle Lingu. Copenhague 20, 11+5-53.
1981 Towards a history of phonetics, (eds.), R.E. Asher &
E.J.A. Henderson. Edinburgh: Univ. Press.
Introduction to A comparative word-list of Old Burmese,
Chinese and Tibetan^ by G.H. Luce. London: Sch. Orient.
Afr. Stud., i-iv.
1982 Tonogenesis: some recent speculations and the development
of tone. Trans. Fhilol. Soc.3 1-21+.
1+


1983 Phonetics and phonology in the eighties: prospects and
problems. Proceedings of the Xlllth International
Congress of Linguists, Aug. 29 Sep. 4, 1982, Tokyo.
Tokyo, 209-19.
1985 Feature shuffling in Southeast Asian languages. In
Southeast Asian linguistic studies presented to Andre-G.
Haudricourt, (eds.), Suriya Ratanakul, David Thomas &
Suwilai Premsrirat. Bangkok: Mahidol Univ., 1-22.
Greenberg's 'universals' again: the case of Karen. In
Linguistics of the Sino-Tibetan area: the state of the
art, papers presented to Paul K. Benedict for his 71st
birthday, (eds.), G. Thurgood, J.A. Matisoff & D. Bradley.
Canberra: Australian Nat. Univ. (Pacific lingu. Series
C; 87), 138-to.
Patterns of baby language in Bwe Karen. Linguistics of
the Tibeto-Burman area, 8 (2), 55-59.
1986 Some hitherto unpublished material on Northern (Megyaw)
Hpun. In Contributions to Sino-Tibetan studies, (eds.),
J. McCoy & T. Bright. Leiden: Brill, (Cornell linguistic
contributions; 5) 101-3^.
1987 J.R. Firth in retrospect: a view from the eighties. In,
Language topics: essays in honour of Michael Ealliday,
(eds.), R. Steele and T. Threadgold. Amsterdam: John
Benjamins. 67-78.
A phonetic oddity in Thai. In, Lai su? Thai: essays
in honour of E.H.S. Simmonds, (ed.), J.H.C.S. Davidson.
London: Sch. Orient. Afr. Stud. (Collected papers in
Orient. Afr. Stud.), 52-62.
Forthcoming
J.R. Firth's 'First impressions' of Tibetan pronunciation.
In, Prosodic analysis and Asian linguistics: to honour
Richard Keith Sprigg, (eds.), David Bradley, Martine
Mazaudon, and Eugenie Henderson. Canberra: Australian
Nat. Univ. (Pacific ling.)
& R.H. Robins. John Rupert Firth and the London School.
In, Oxford International Encyclopaedia of linguistics,
(ed.),, W.O. Bright. New York: Oxford Univ. Pr.__,
Khasi clusters and Greenberg's universals. Papers from
the Helsingdr symposium on Austroasiatic linguistics and
literature, (ed.), J.O. Svantesson. Copenhagen: SIAS;
London: Curzon Pr.
One word or many? A problem for the lexicographer of
pre-literate languages. In, Proceedings of the 18th
International Conference of Sino-Tibetan Languages and
Linguistics. Bangkok.
1+


Forthcomings continued
Problems and pitfalls in the phonetic interpretation of
Khasi orthography. In, Austro-Asiatic linguistics:
essays in honour of H.L. Shorto, (ed.), J.H.C.S. Davidson,
London: Sch. Orient. Afr. Stud. (Collected papers in
Orient. Afr. Stud.)
1+




mljtt ca: FROM "FISHES' EYES" TO "ANKLEBONES":
A VIETNAMESE CALQUE?
Jeremy H.C.S. Davidson
'm£t c£ c£i gl1 (n6m= ^ ): "what
are fishes' eyes?",' asks Vietnamese (VN). The earliest record
so far discovered in answer is Alexandre de Rhodes' Diotionarium
entry (l651:i+56): 'm^t ca: artelho do yh: talus, i.', in which
the Latin talus means 'ankle, anklebone'. Legrand de la Liraye
(187^:1501)) translates it as 'cheville du pied', Taberd (1877:
271a) as '0ccuius piscis; malleoli', while Ravier (1880:706a)
notes 'malleolus #U.m&t ci (ndi. ch&n tay)' (=place of the feet/
legs]gj(| and hands/arms ). C&a (l897:86b) describes it as
'fleshed bone swelling out on both sides of the foot at the
point where it joins the leg'; Tif (l898:q.3, 3a4, p.51 =
p.216, 1.19) uses it cryptically to translate the Chinese
character jjjfchuk 'ankle' (>'anklebone'), hu/i in Modern Standard
Chinese (=putong hua/MSC). Again, Bonet (l899:^07a) records it
as 'cheville du pied; litt. oeil de poisson', Hue (1937:555a, 2)
repeating it as 'cheville du pied, cor, chevillon', while Tru
(1960:299b) defines it as 'heads of bones jutting out at the
neck of legs (=ankle) (qv. Tru 1970:11, 89^+a) and the HSi Khai
Tv' (1968:337b) tells us that" it is 'the head of a bone'budding
out near the neck of the leg' (=i.e. the ankle), hence 'the
anklebone'.2
3
This seems to be highly figurative language, a
descriptive pun (qqv. Emeneau 19^7; Hoa 1955; Davidson 1978:37ff.,
1986a:35^, n.37), for to find a fish's eye, probably rather dead
in appearance, looking up at you from the anklebone of a person's
leg, bulging out dull, then whitish as it stretches from where
it peaks to the darker skin surrounding it on legs that have
travailled in fields of paddy, sea-salt, or whatever else, in a
Vietnamese-style climate, will certainly provide yet another
superb and humorous expression of the impression intended, since
the Vietnamese have long associated their vital environment with
all the other natural symbolism that they envisage (e.g.
Davidson 1978).
Such visual association (qv. n.3),^ is frequently
complemented by word-play afforded by the opportunities of
homophony. An encounter is the Modern Mon (MM) /mot coiq/
'anklebone' (lit. 'eye of the elephant'; Shorto 1962:170b mbt
eoiq; ref. p.96b coiq^ 'elephant'), while in inscriptional
language one finds ciq^/ciq/ n.'elephant' (Shorto 1971:98) and
jun/jmn/ 'lower limb' {ibid.,125), where a mat2/mot/ n.'eye'
UbicZ.,285) plus jun would have provided us with the expected
1+


Mon-Khmer (MK) and Malayo-Polynesian (MP) 'eye of the leg1 for
an 1 anklebone1! This 'confusion1 may well have stemmed from,
or led to, folk-etymology and folk taxonomy an investigation
lying in waithut it further emphasizes the point already made
of the desire for the use of figurative language and especially
for descriptive punning.
Parts of the body are, however, often not identified
with any precision in basic and vernacular Vietnamese language,
that is, language which excludes or may not include the use of
the specific, pointed Han-Vi^t (HV) vocabulary found in the
literary language of the educated elite, and which is also 'pre-
modern', that is, prior to the introduction of western, scientific
and other specialist vocabularyA general area of the body may
be referred to, but not a specific part of it, by a single,
special word. So, chUln (van. chcfa, chtih) stands broadly for both
'leg' and 'foot', tay for 'arm' and 'hand', while together chjjn
tay means 'limbs1.8 When an anatomical feature does captivate
attention and is accorded a specific term, structural or func-
tional perception of it is a common reason for the formation of
the word or words describing it. So, many South-East Asian
languages have captured in their vocabularies' imagination the
concept of the ankle or wrist as the 'neck' of the leg or arm
and the anklebone as the'"eye" of the leg/foot' even though in
actuality it is not a part of the ankle itself but a swelling
out on both sides of the lower end of the two leg bones (tibia
and fibula) at the point at which they articulate with the tarsal
bones (qv. Cua 1987:86b).
Naturally, words for 'hands' and 'arms', 'feet' and
'legs' may well have changed in the language fa,mi lies that
inhabit South-East Asia, China, South Asia, Oceania and other
related areas as their languages developed from the early through
to the modern forms, but the phonetic similarity is even now
often noticeable. The same is also evident for the words for
the 'eye'9 (e.g. Shorto 1971:285, i.a.) since one finds links
between the Sino-Tibetan, Sino-Tai, Austroasiatic and Austro-
nesian forms in both their phonetics and the semantic functions-^
performed. So, as in Vietnamese we have an 'arm' tay, its
'neck' cff jfrfebecomes immediately recognizable as the 'wrist'
c£ tay, as does the leg's :'neck' become an 'ankle' co ch&n11
(e.g. Gouin 1957:l8lb). Other examples are numerous (qv. n.&
below) but, especially when we find the 'board, table' (b£tn
cf. van plank '^^L ) ^ helping to structure a 'hand' bhn ta/or
'foot' bhn ch£n,J-3 why do we not 'eye' the 'leg' (*mfet chstn) to
make an 'anklebone'? After all, we have already met the partially
successful Modern Mon attempt of mot coiq and the Malay mat a kaki.
Bahnar provides us with mat jong 'cheville du pie^' (Dourisbourne
1899:203; cf. Katu mani£tgmaniit, identified as m&t ca in the
Vietnamese equivalents but translated as 'ankle' in the English
(Costello 1971:27, 30) and probably from ma 'eye', cf. katam
(loc.cit.)), while Sedang gives us ma cheang 'ankle' (lit. 'eye
1+


of the leg/foot') and ma k6ng 'wrist' (Smith 1962:15-16) and we can
note en passant the Chrau j£ng 'leg, foot' (Thomas 1966:21+),^
unfortunately with no further relevant data. Going south in
present-day Vietnamese territory, we discover more support from
the Cham mdta takai (cf. takan takay, Moussay 1971:379b 'cheville
du pied'; cf. ibid. ,221+ mflta le 'cheville') with the other common
uses of the 'eye/face' word as 'node, joint', etc. The same
occurs in Nicobarese (Roepsforff 188U:78, 122, ll+9, mathou,
okloaka-la; olmat, etc.; cf. Man 1898-9:6, 178a).
This Vietnamese 'neck' ccf of the 'arm' or 'leg',
joining them to 'hands' or 'feet', finds many other Mon-Khmer
language equivalents, two directly apparent examples being Modern
Khmer ko: day 'wrist', ko: CYiq 'ankle' (Jacob 1971:1) and Modern
Mon ko?cl,Q 'ankle', ko? toa 'wrist' (Shorto 1962:53ab1^ Of
immediate interest, therefore, and suggestively cognate is the
Lao kho: thao 'ankle' (lit. joint of the foot) (Kerr 1972:129b;
cf. ibid.,ll8b kha:) which is supported by the Modern Standard
Thai (MST) khoo thaaw 'ankle; lit. joint of the foot' (not 'leg',
khaa, (n.b. Li 1977:213, §10.6.29 among other entries, e.g. §2.23,
1U.11.1+, etc.) with a secondary vernacular taa turn (thaaw)
'eye of the node (of the foot)' >'anklebone'. Thus, the possibil-
ity of 'link' cognates joint-neck-tibia-shin keeps springing
to mind. Indeed, there is an impression of 'joining, uniting,
linking' and of 'reaching, attaining, going to' (see GSR 675a-d
J^t j e. *gep/y^p/ho C=MSC he3) encouraged by the n6m character
for 'neck'tftfe which one also finds as Han-Viet, and in the MSC
compound ktLLou /v 1 skeleton' (= khS lau;*cf. MSC guge 'ibid.'
=HV c&t cach). Additionally, there is the Vietnamese
word for 'joint; physical articulation' khcfp vfefe (>'bone-joint1
= ru xufahg), which strongly suggests a relation between it and
the Vietnamese ^ 'unite, etc.', joining things together.
Reverting in this context to a possibly pre-sinitic
word (or a very early Chinese loan?) preserved in present-day
Chinese territory, one finds tj^j^ MSC jiao (GSR ll66p *k'og/
k'au/k'iao fj^ 'tibia' ; =?HV giao); its variant character hgg^
used for 'joint' and presented in Ningpo dialect as gap*
(Morrison 1876:20b), ngetu in Swatow (Gibson 1886:107c; cf. VN
ngau hdp 'join again' =?Tie-chiu/Hokkien) and its parallel
MSC jlAg e.g. jinggu 'shinbone' (>'tibia'; (cf. GSR
831k. *g'ieng/yieng-Zhing^leg, shank (Lunyii)'), as well as the
two pronunciations for 'neckMSC geng/jing (GSR 831n. *kieng/
kiang:/king, and *g'ieng/g'iang/k'ing7..(Tso)), all of which are
related, at least phonetically7 Hence the pronunciations for
the word for 'neck' in Vietnamese c£?, the Thai khoo, Lao kho:,
a large number of Mon-Khmer languages (qv. Shorto 1971:52), and
Archaic Chinese *k'og (GSR ll66p), thrust cognacy before us.1^
The image of an articulated (=joint), tube-like connection of
varying lengths between the torso and its ending appendages does,
of course, bring the possible linked, root-origins of such words
together, while the closeness in pronunciation of the 'leg' word
1+


in a range of South Eastern Chinese dialects (qv. Bauer 1987),
e.g. kha, keuk (=HV ciftfc, also read khiicfe, giving rise to gffc
'"base, foot (of a tree)1 LnSm^II (Tru 1960:i+6a)) reconstructed
as Archaic Chinese *kiak (GSR 776g-h), forcefully implies a pre-
sinitic origin of the word that ends up as the Chinese character
}(MSC jue, jiao).
But here the Vietnamese mat ca 'anklebone' remains
distinctly in the forefront of one's mind. If the base word
ch&n 'leg, foot' which qualified other associated ideas giving
us words like 'ankle', 'foot', 'toe', 'toenail', 'heel', etc.,
(qv. n.8 below) fulfils its purpose, then why does an 'anklebone'
suddenly turn from the '"eye" of the leg' into a 'fish's eye'?
Naturally, in a Vietnamese language situation such as this, one
looks to the Chinese reservoir of vocabulary for help in solving
the problem. In Archaic and in Ancient Chinese one does not find
the 'anklebone' separately listed, one finds the 'ankle' j^L
(GSR 351 j. *g'lwar/ywa:/hua ankle (Li.) = MSC hua, hu^i), the Han-
Vi£t reading for which is hoa (Thieu
Chifu 1966: 660a)17 but which
is intriguingly identified as an 'anklebone'. (ibid.), and clearly
recorded as such in the Tu> (l898:q.3, 3a4=p.51 ^ Ajf
=p.216, 1.19 as : 'khoa mat ca' ) while the MSC gives^us hu£i 'tor
'ankle' and huaizigu for 'anklebone' (cf. i.a. Maclay
1929: 1090 ku£-galik jjffffl? Other HV 'pedal' terms flourish
too. Meanwhile, the MSu^Snkle' occurs as jiaoyangu ffijl BftL-Jf
(var. j iaohuai, jueyan) while the Ningpo kyiah tsang jSfe
'ankle' (cf. kyi£h gao fliti 'ankle-joint', Morrison l876?20b)
relates directly to a Hakkla kiok tsang 'heel', and the Hakka
'ankle' kiok muk ^ (cf. kiok ngin flfip fjg^, Maclver 1926:
288b) connects well with the Shanghai kyak mok. Yet, focusing
on the 'eye',as the centre of attention in this 'joint', this
'link', strikingly important is the awareness of the '"eye" of
the "leg", foot', in Southern, and in particular in South Eastern
Chinese dialects. There is Cantonese keuk ngaan (kwut) Ajj^
' ankle (bone)' (qv. MSC, above; cf. Maclver 1926:3^b, Hakka khTfcr
kwut 'the anklebone'), Lungtu khaa muk, Fuzhou ka ngu mek
(-'the leg Cjoint]'s ox's eye') (Maclay 1929:1021)the
Swatow kha-mak 'anklebone' (and variants, Lechler 1.883:9) and the
Amoy 'ankle' k'a bak (liter, kiok bok, kha-bak, kib7 bak; cf. C.
Douglas 1899: 10a, 257b, kha-bak, 'ankle', R. kiok,col, kioh).
The thought came to my mind of a word of pre-sinitic origin
preserved in what are termed South Eastern Chinese dialects and
then spurred into a caique by the creative imagination of the
Vietnamese because it rhymed descriptively with the phonetic and
thence the visual awareness of a 'fish's "eye"' (mat <*kha
mak). And, almost instinctively one senses that the Hokkien
CHkn.=Fuzhou] dialects offer the greatest suggestions. Among
them, I think that Tie-chiu C=ChaozhouH enlivens us. In it we
find mak (Sj eye (Goddard l883:103a)19 and W 'foot' (ibid.,
63a), the tone, pitch and cpntour of whose pronunciation conform
well with the Vietnamese mat ca, strengthening the argument since
it was the speakers of Tie-chiu who made up the largest Chinese
lU


20
population contributing to the 'early' Vietnamese vocabulary,
just as it appears to be Swatow speakers who influenced Thai
(Egerod 1959).
For here, it seems, we do have a Chinese-Vietnamese
caique created by the Vietnamese who, hearing this South Eastern
'Chinese' word *k'a ( association with their word c^ 'fish' and, preferring the fun21
behind a 'fish's "eye"', mat ca,22 to the neutrality of such
words as might otherwise have been their special terms (e.g.
*go chin 'hillock of the leg/foot', *mat chan 'eye of the leg/
foot'), transposed the Tie-chiu k'a m^k ('foot's eye' >'ankle-
bone )23 into their own language, reordering the word as mak k'a
to fit their syntax, to create ^ casque, so that the 'ankle-
bones' became 'fishes' eyes', mat ca
NOTES
1. Immediatelyy we are alerted to an unusual etymology.
Normally, mat .'eyes' are classified as animate by con ^L,
so that con mat ca would mean 'the eye of fishes' even
though the syntactic rhythm feels heavily clumsy. But here,
it is classified as inanimate by cai, suggesting a variant
meaning. What is the reason? (qqv. ThierLi CtuJu 1966:660a
IHan-Vil t/(HV= Sino-VietnameseD gjg. hoa, (or khoa).
1. cai mat ca chan 'the "anklebon^fc" of the legs'; cf. Anh
1957:510b Loa, '"fishes' eyes" on the two sides of the
ankle', i.e. anklebones).
2. Listings of dictionary entries could continue, but they are
not being included as they are meaningfully repetitive, if
not identical; e.g. Hung 1955:4l4a; Thanh Nghi 1967:885a:
'balls of bone jutting out on both sides of the ankle';
Hoa 1967:270a, etc.
3. So, too, is the Palaung ra-ngys jung 'ankle' (lit. 'joint
of the legM (Milne 1931:18)7with 'anklebone' (loc.cit.)
a ka-ang i-ar, that is, 'bone of the fowl, hen' (?cockspur.
CboneJ; cf. VN cufa 'cockspur'; xub^Lg ~ 'astragalus').
4. Note, interestingly, that in Chinese we have the word 'eye',
MSC yan, reconstructed as GSR 4l67. *ngen/ngan:/yen eye (Yi)
which is also a 'loan for *ngan/ngan:/en^protrude as a knob
(Chouli)' (loc.cit.). The HV is nhan/nhan (qqv. de Rhodes
1651:54$, cf. 456; Tip-^c 1898: q.3, 5a3-4, p.52 =p.2l8, 1.5
'nhan mat; Davidson i975:597, no.487). In passing, 'faces'
and 'eyes' do, understandably, go together (e.g. Shorto
1971:285; Tu>mb 1898: q.3, la5, P.50^-^3 =p.215, 1.9
' dien mat')Tj^/TQ
15


5. For instance, there is truiig caffi S^v? 'eggs of fishes1 (qv.
Bonet 1899:34la trang c£ 'oeuf de poisson'), whence comes
'spawn1 and then, because of the visual similarity, 'black-
head, comedo' (uSa 1967:516a; Hue 1937:1088b; cf. Gouin 1957:
l4lb). This is possibly a play on words, on cluing 'symptom
of an illness' (nSm/HV ; MSC zheng) plus c£ 'fishes^,
whence 'small white face pimples'. Another example is hon
dalj&tfj (lit. 'balls/stones of the genitals'>) 'testicles'
(Hfea 1967:84a; Ku£ 1937:191b) paralleled by trAig d^i (lit.
'eggs of the genitals'>) 'sperm' (Bonet l899:34la; cf. Gouin
1957:l48lb^, i.a.). Note ciMypS) fto urinate^' (aj-so sometimes
written^.in n&n). Similarly, one finds go ma
literally ^mound/knoll of the cheek', so 'cheekbone' (Bonet
1899:230b/ joi^e, pommette'; Hue 1937:3l4a; Gouin 1957:482b),
but why mat ca for 'anklebone'?
6. The play on words could also have given us mot coiq1 'eye of
the elephant' and mot coiq^ 'eye of the ridgepole of the
house' (Shorto 1962:96b). Compare the Archaic Chinese use
of (MSC jisLO 'tibia') for 'the tapering end of the
spofe of awheel (Chouli)', (GSR ll66p.).
7. Nowadays, the scientific and medical vocabulary of Vietnamese
is as detailed in its anatomical and other descriptions and
recognitions (e.g. Thanh Nghi 1967:1532a; Hoa 1967:565a-b)
as its main new source language, French; in fact, it is
probably made richer by ready recourse to Chinese wherever
that proved necessary. Thus Vietnamese has a very rich
vocabulary and usually treats technical-type terms in three
ways:
(i) identifying them by direct translation into VN e.g.
'astragalus' = xiic?ng ciia 'cockspur bone' ;
(ii) by use of Chinese terms in the HV pronunciation of
their characters, although a degree of Vietnamese word-
ordering may; be introduced e.g. 'haemophilia' = benh huyeft
huh (iii) by transliteration from the Frenche.g.'xanthin'=
xang-tin.
8. qqv. Shorto 1971:125; Davidson 1975:597, esp. nos. 474,
489-92. So, paralleling one another, are such compounds as:
ban ('table/board') chan 'foot' b\n tay 'hand'
ngon ('toe/finger') chan 'toe' - ngon tay 'finger'
mong ('nail/claw') chAn 'toenail' mong tay 'fingernail'
c£? ('neck') chan 'ankle' c tay 'wrist'
\ 7
(Hoa 1967:58a, 407b; also Bonet 1899:407a; i.a.; cf. cang
'paw, leg (of animal)'). Nonetheless, there are independent
words for special parts of the body, e.g. gffi 'knee', versus
1+


khuyu/cui tay 'elbow' ; bap chan 'calf (of the leg)' ; cfai
'thigh'; g£t chan 'heel' (cf. g£n g6t 'tendon, Achilles
heel...') ; nSm tay 'fist'. And, although there are specific
'finger' .words, e.g. ngon tay ci.1 'thumb', ngon tay t£t
'little finger', one can easily manufacture parallels like
'big toe', 'little toe', etc., by substituting chan for tay.
And so on. The versatility is immense, as is readily shown
by the Tif-Brib.
9. See, e.g. Bauer 1987 on the 'leg' in South Eastern Chinese
dialects and Tibeto-Burmese; and, in general, Davidson 1975:
I, 296ff. & nn.; 1986b:63, & references).
10. The use of the 'eye' as a node of the bamboo (e.g. Bonet
loc.cit., mat tre 'noeud de bambou'; Hue 1937:555a2, ibid.',
Gouin 1957:801b. ff.) and sometimes as the joint of a part
of the body and so on, is also well observed throughout
South-East Asian languages (see i.a., Shorto 1962:170a ff.;
1971:284ff. ; Moussay 1971:22*0.
o
11. qv. de Rhodes (1651:126) 'c&^tay: collo da mao: collum
manui', Cqv. also p.715 ^ ; ch£n: collo do pb: collum
pedis'. (Hu£ 1937:101a; cf. Gouin 1957:l8lb).
12. In Malay, one apparently circles the leg to form an ankle
(Wilkinson 1932:1, 338a, gelang, cf. 290 a; **96a kaki) but
the 'anklebone' is still the mata kaki 'eye of the leg' (loc.
eyb. & II, llUa, mata). The thought that the Vietnamese
mat ca might embalm an MP variant of this is farfetched
(even given the *proto-language forms being proposed at
present) because of the distinctly Mon-Khmer word for 'leg'
in Vietnamese.
*k
13. In Bahnar, I understand that / pa: 13/ is a possible
reconstruction. Might this imply a MK k- 'body' prefix
which is now lost in modern Vietnamese?
1*+. Pacoh deals where relevant with bones but does not seem to
have a special term for the 'anklebone'. (Watson 1979:382;
cf. parreat (ati).).
15. Note that h- /h-/ and kh- /X/ variants are common in
Vietnamese, e.g. n.l above.
16. Relevant, too, are the semantic similarities and, to a
certain extent, the various reconstructed readings of
MSC jia, jia (#GSR 630a *kap/kap/kia be on both sides of
(ShT) ; support (Shu};.-., loan for 6301. all around (Shu)
C = jxfc jia,xial and especially ?/ tsiep/tsie C=MSC xieJ
encompass, embrace (.Shi); all around (Chouli).); n.b. GSR
630k, too.
1+


17. But loa in Anh (1957:510b; Tru 1970:278a4) which implies an
original *kl- initial consonant cluster and a shift from it
to h- and kh- variants, cf. n.15 above.
18. n.b. gie mek ciu flk $ (lit. 'foot-eye-ball1, Adam
1891/1905: 11; cf. Maclay 1929:1222) which brings to mind
the Burmese khie-myak-ci, transcribed as chi-mye?-si 'foot-
eye-ball' ='eyeball' (note Adam 1905:106 mek-ciu-ci).
19. From the Chinese in which there is MSC mu 'eye', recon-
structed *miok/mjuk/mu (GSR 1036a-c.), comes the H^n-Viet /
muc_ (de Rhodes lb51:488; muc, vide mouc; 483 mouc, con mat:
oihos: oculi; TtfEih l898:*q.3, laU,*p.50 =P-215,
Z.7 muc_ m^t). En passant, we may note that final -t_ and -c_
/k/ are often interchanged, especially in southern Vietnamese
dialects.
20. A large Tie-chiu refugee migration one which preceded the
late nineteenth century influx of Cantonese who then also
provided a vast source of vernacular, material culture
vocabulary is known to have taken place during the Ming
dynasty, and Tie-chiu loans to, and pronunciations of Chinese
characters in, Vietnamese as distinct from the Han-Vi^t
forms one normally encounters are numerous. Many of such
words are readily found in Cua (1897: e.g. 374a thao; 380a
thau; 402a tho; 433b tia 'daddy' Ccf. Egerod 1959:'no. 159H;
451a toa Ccf. Bonet 1899:323bD) and in Tru (1970: e.g. II,
17a mang, etc.) but not all dictionaries are as detailedly
precise. En passant, no Tie-chiu words have, it seems been
recorded in de Rhodes' Dictionarium, a compilation principally
of the northern dialect.
21. As seen in the more risque or vulgar puns like (jfu-me for
(Paul) Doumer and -cua 'crab's arse' for the much
despised 'discours'.
22. And its irony. -Note the idiom (thanh ngu) : ngu* muc h£n ch4u
^ ^IL) 1 fishes 1 eyes may be taken for pearls',
meaning that true and false are confused. Its main applica-
tion in both China and Vi^t
Nam was to people who used fake
materials to make counterfeit, 'real'-looking, goods, thus
deceiving others. This was usually jewelry for women and
girls.
23. Perhaps this also refers obliquely to notice of a lot of
Chinese migrant labour?
1+


REFERENCES
Adam, T.B.
Anh, -B^o duy,
Bauer, R.S.
Bonet, J.
Carstairs Douglas
Costello, Nancy A.
? \
Cua, P. Huynh tinh,
Davidson, J.H.C.S.
1905.(from 1891). An English-Chinese
dictionary of the Fooohow dialect.
Shanghai & Foochow: Lacy & Wilson,
Methodist Publ. Ho.
/ ^ a ? t/ j. n
1957. Han-Viet tid-ctien giSn-ySu
Saigon: Tru&fag-thi.
1987. fLegf in Southeastern Chinese
dialects and Tiheto-Burman root *pey
Tleg!. Ling. Tibeto-Burman Area 10(l),
169-7^.
1899. Dictionnaire annamite-francais
ilangue officielle et langue vulgaire).
Paris: Imp. nat., E. Leroux., 2 pts.
1899. Chinese-English dictionary of the
vernacular or spoken language of Amoy.
London: Publ. Office Presbyt. Church of
England (new ed.).
1971. NgiF-viing Katu: Katu vocabulary
(T?i-s^ch ngon-ngu? dan-tSc thi£?u-s£f
V let-Nam, vol.5), si.: Trung-tam hoc-lieu,
B6"Gidio-duc.
1897. -Bqi-Nam qu&c-dm td-vi. Saigon:
Imp. Rey, Curiol, & Cie.
1975. A new version of the Chinese-
Vietnamese vocabulary of the Ming dynasty.
I: Bull. Sch. Orient. Afr. Stud. 38(2),
296-315; II: op.cit. 38(3), 586-608.
1978. Images of ecstasy: a Vietnamese
response to Nature. In: Natural symbols
in South East Asia (ed.) G.B. Milner.
London: Sch. Orient. Afr. Stud., 29-5^.
1986a. Collaborateur versus Abstentioniste
(Tilong versus Tri): a political polemic
in poetic dialogue during the French
acquisition of southern Vi|t-Nam. Bull.
Sch. Orient. Afr. Stud., 49(2), 321-63.
1+


Davidson, J.H.C.S.
Dourisbourne, P.X.
Egerod, S.
Emeneau, M.B.
Gibson, J.C.
Goddard, J.
Gouin, E.
GSR
Hoa, Nguyen cfinh,
Hoi Khai Tr%
Hu£, G.
Hung, L^ van,
Jacob, Judith
1986b. Another source for information on
Ayutthaya Thai. In: Lai Su? Thai: essays
in honour of E.H.S. Simmonds (ed.)
J.H.C.S. Davidson. London: Sch. Orient.
Afr. Stud., 63-72.
18 89. Diotionnaire bahnar-frangais.
Hong Kong: Imp. Soc. Miss. Etr.
1959- Swatow loan words in Siamese.
Acta Orient. 23, 137-56.
19^7. Homonyms and puns in Annamese.
Language 23, 239-^.
1886. A Swatow index to the syllabic
dictionary of Chinese by S. Wells Williams,
Litt.D. & to the Dictionary of the
vernacular of Amoy by Car stairs Douglas.
Swatow: Eng. Presbyt. Miss. Press.
1883. A Chinese and English vocabulary
in the Tiechiu dialect. Shanghai: Amer.
Presbyt. Miss. Press (SOAS W?2.2b).
1957. Dictionnaire vietnamien-chinois-
frangais. Saigon: Imp. Ext.-Orient.
1957. B. Karlgren, Grammata serica recensa
(= GSR). Bull. Mus. Far East. Antiq. 29.
1955. Double puns in Vietnamese: a case
of 'linguistics play'. Word 11, 237-^.
1967. Vietnamese-English dictionary.
Rutland, Vermont: Tuttle.
1968. Viot-Nam tii <£iin ^ drafted by Ban
v&n hoc H8i Khai Tri ti£n dub kh&i th&o).
Saigon: Mac-Lam.
1937. Dictionnaire annamite-chinois-
francais: Tif-cti&n Vitzt-Hoa-Phdcp. si.:
Imp. Trung Hoa.
1955. Vietnamese-English dictionary.
Paris: Eds. Europe-Asie.
197^. A concise Cambodian-English
dictionary. London: Oxford Univ. Press.
1+


Kerr, A.D.
Lechler, R.
Legrand de la Liraye,
Le R.P.
Li Fang-Kuei
Maclver, D.
Maclay, R.S. &
C.C. Baldwin
Man, E.H.
Milne, Leslie
Morrison, W.T.
Moussay, G.
Ravier, M.H.
Rhodes, A. de,
Roepsforff, F.A. de,
1972. Lao-English dictionary. Washington:
Consortium Press, 2 vols.
1883. English-Chinese vocabulary of the
vernacular or spoken language of Swatow.
Swatow: Eng. Presbyt. Miss. Press.
187^. Dictionnaire elementaire annamite-
frangais. Pari s: Challamel.
1977. A handbook of comparative Tai.
Honolulu: Univ. Press Hawaii.
1926. (M.C. Mackenzie, rev.) A Chinese-
English dictionary: Hakka dialect as
spoken in Kwang-tung province. Shanghai:
Presbyt. Miss. Press.
1929. (S.H. Leger, rev. & enlarged)
Dictionary of the Foochow dialect.
Shanghai: Presbyt. Miss. Press.
1888-89. A dictionary of the central
Nicobarese language. New Delhi:
Sanskaran Prakashak., rep. 1975.
1931. A dictionary of English-Palaung
and Falaung-English. Rangoon: Supdt.,
Govt. Printing & Stationery.
1876. An Anglo-Chinese vocabulary of the
Ningpo dialect. Shanghai: Amer. Presbyt.
Miss. Press.
1971. Dictionnaire c&m-vietnamien-
franqais: Tif-dien CJiam-ViQt-Phczp. Phan
Rang: Centre culturel cam: Trung-t&m
v&L-hoa Cham.
1880. Dictionarium latino-annamiticum.
Ninh Phu: ex typis missionis tunquini
occidentalis.
1651. Dictionarium annamiticum3 lusitanum3
et latinum ope Sacrae congregationis de
propaganda fide in lucem editum. Roma.
188*+. A dictionary of the Nancoury
dialect of the Nicobarese language in
two parts: Nicobarese English and
English Nicobarese. Calcutta: Home
Dept. Press.
1+


Shorto, H.L.
Smith, K.D.
Taberd, Rev.
Thanh Nghi
ThiSu Chuli
Thongs, D. &
Th& Sang Luc
Tru, Le ngoc,
Tu>-Bi$b
Watson, R., Watson,
Sandra & Cubuat
Wilkinson, R.J.
1962. A dictionary of modem spoken Mon.
London: Oxford Univ. Press.
1971. A dictionary of the Mon inscriptions
from the sixth to the sixteenth centuries
(London Orient. Ser. 24). London: Oxford
Univ. Press.
1962. Ngti'-vyhg Sedang: Sedang vocabulary
(TS-sach ng&n-ngu* d&n-t^c thi&u-sft' Vi#t-
Nam, vol. 2). Saigon: Vien chuy§n-khao
ngu-hoc trung-tam thiidng-ngu? Kontum, Bo-
Giao-duc.

I877. (rev. J.S. Theurel) Dictionarium
annamitico-latinum. Ninh Phu: ex typis
missionis tunquini occidentalis.
1967. Vi£t-Nam tan tif-iien, minh hoa.
/
Saigon: Khai Tri.
1966. Ean-Vi&t tu-diin
Saigon: Huhg Long.
1966. Ng&vuhg Chrau: Chrau vocabulary:
Chrau-Vi£t-English (Ti?-sach ngdn-ngCiP
d&n-t6c thi^u-stf'Vi£t-Nam, vol. l).
Saigon: BS Gidfo-duc.
1960. Viet-ngu>chvnh-ta tu?-vi. Saigon:
Triicfrig-thi, 2nd ed.
19-70 (et al-J Viet-Nam tu? diin. Saigon:
Khai Tri", 2 vols.'
1971. (from 1898^. (Ctibn Ching-ho ed. ,
etc.) TuHtiib thanh-chlf tifrhoc gidi-
ngh-Ca-oa ffl g &
Hong Kong: Chinese Univ. Hong KongJ./Vfc
1979,. Ncfh Pacoh-Yoan-Anh: Nguhvifng
Pac6h-Vi&t-Arih: Pacoh dictionary: Pacoh-
Vietnamese-English (T^-s^ch ng&i-ngu>
d4n-t&c thieu-s£f Viet-Nam, vol. 25, pt.l).
Huntington Beach, Calif.: Summer Inst.
Ling.
1932. A Malay-English dictionary
{romanised). Mytilene: Salavopolous &
Kinderlis, 2 vols.
1+


SOME FEATURES OF MODERN KHMER LITERARY STYLE
Judith M. Jacob
The literary or formal style under review here is that
of modern prose, particularly that to be found in journals,
speeches and novels. Interest in this aspect of the Khmer
language was stimulated by the realization that certain assumptions
are made about it without any attempt at justification. For
example, one feels, especially when translating, that it is an
unnecessarily verbose style, containing far more words than can
be represented in the translation. Is this really so, or is it
a false impression due to language differences? One also tends
to say, without any specific parallels in mind, that modern Khmer
journalese is clearly imitative of its French equivalent. This
paper attempts to summarize the various linguistic differences
between the consciously formal modern style and the spoken, or
informal. Material has been collected for the purpose from
articles (post-1930) on literature and religion, particularly
those in the journal Karrfoujasuriycl, from newspapers (1950-72)
and from novels (1938-71). I am very much aware that there may
be many constructions, associable with style, which have been
missed and I present this analysis as a precursor of others.
The increase in the output of prose writing as compared
with poetry in the last few decades, during which, for example,
both the novel and the printed newspaper in Khmer have developed
from scratch, has caused prose writing to be held in higher
esteem than it traditionally was in Cambodia.1 Authors of novels
and writers of articles were trying consciously to raise the
standard of prose style in the 1950s and 1960s in response to the.
general wave of nationalistic feeling. Interesting changes in
style which thus took place in post-war Cambodia, especially
since independence in 1953, may be observed by comparison with
the factual and simple narratives of the Chronicles, composed in
the nineteenth century,2 or of the first written versions of
folktales3? or of the esteemed early twentieth century writer
Suttantaprija In. The analysis of these developments might be
presented in many different ways. I have decided to arrange the
various points in four sections which suggest my own idea of their
possible origin. These are:
1. Features which seem to be present in order to clarify long,
involved sentences;
2. A feature which suggests conscious or unconscious imitation
of Thai;
3. Features which suggest conscious or unconscious imitation of
French;
1+


4. Features which seem to reflect the desire to embellish,
using the traditional devices of Khmer literature.
1. Features which seem to be present in order to clarify long,
involved sentences
In the 1950s and 1960s, when upwards of 3000 new
technical terms based on Sanskrit and Pali were invented so as
to avoid using French loanwords any longer, the introduction of
this unfamiliar vocabulary added to the need to explain some old
loans to a general reading public and had some effects on the
style of written Khmer:
(i) In the early years of the use of the new vocabulary it
was common practice to place two words, one Khmer, one Indian,
side by side. Thus /smo :m yi:ecok/^ (Khm. + Skt.) 'beggar',
/syksa: so:t-rian/(Skt. + Khm.) 'education', /prate:h-ci:at
mi:atophu:m/ (old.. + new borrowing) 'one's native country', all
occur in the newspapers of the 1950s.
(ii) The unwieldy sentences produced by use of the new words
in complex clauses led to the excessive use of the literary
particles / nvu/ and ney/ which are used only with restraint in
earlier prose writing such as that of the Chronicles, the folk-
tales, or the writing of Suttantaprija In.
/nvu/ (spelt nuv)* In Khmer, the object, when it is
expressed, normally follows the verb immediately, e.g. /mv:l
kon/ 'see a film'. In an informal or colloquial style, no particle
connects the object to the verb. The use of /nvu/ as an indicator
that the object of the verb follows is a literary device which is
useful when, as in journalese especially, several phrases occur
between the verb and its object and it is helpful to know that
the next word, or words, will be the object, e.g.
/coh phsa:y pi: lv:k knoq sa:;PD0dami: on neak-ci :at-niyum
publish, print two times in newspaper nationalist
nvu rmaq/
(particle) story
'twice printed the story in.the Nationalist newspaper'.
Once established in modern literary usage, however, /nvu/ began
to be used when there was no need for it because the. object
followed the verb immediately:
/troam-tro: ro:q nvu ' ompY:jioom-jii: /
endure undergo (particle) matter pillage
'put up with the pillaging'.
1+


/ney/. In the colloquial, or informal, language, two nouns
occurring in close junction may represent an object and its
possessor, e.g. /phteeh vi:a/ 'their house'. A more elaborate
way, especially useful if attributes of the first noun follow it
before the occurrence of the possessor-noun, is found both in
colloquial speech and in the literary language: /phteeh thorn
reboh vi:a/ (house big possession they) 'their big house'. In
a literary modern sentence, however, the role of /raboh/ is
played by the particle /ney/ 'of', e.g.
/lo:k ?akkeate:sa?phiba:I ney prete:h ?o:stra:li:
governor general of country Australia
'the Governor-General of Australiav
It may be pointed out that, in the Chronicles, many similarly
long instances of possession and possessor are expressed without
/ney/ by simple juxtaposition, and that /ney/, like /nvu/, came
to be used between short familiar words in modern writing too.
(iii) Among the seemingly unnecessary words which help to
give the verbose impression to modern literature are many
sentence-final nuclei which corroborate a word, often a particle,
which has occurred earlier in the sentence. Examples are in this
case given (see next page) without full sentences, for the most
part, since the point seems clear from the meanings given and
they are divided according to the grammatical functions of the
two elements.
The practice illustrated in these examples, which
occurs freely in the colloquial language as well as the written,
seems to have become a characteristic feature of journalese,
where it has a clarifying role, not so much because the meaning
is reaffirmed as because the sentence-final nucleus, or sometimes
several of them together, confirms that the end of a clause has
been reached.
1+


(a) Meaning of pre-verbal panticle (pre-v.p.) corroborated by sentence final element:
rv>
pre-v.p.
nvu-tae
(still
pum-toan
(not yet
prakaek
argue
raluat
extinguished
pre-V.p.
sot-tae thvv:
(all without do
exception
ti:ahi:an
soldier
srap-tae
(suddenly
dol
arrive
final particle
nvu-laay
still)
nvu-laay
still)
adverbial particle
teaq-?oh
all)
phli:am
immediately)
Verb
pre-v.p.
kuaq prateeh nuiq sombok vi:a pum kha:n
(certain meet with nest they not miss)
translation
fis still arguing about (it)f
fis not yet extinguished'
'all without exception became
soldiers'
'suddenly they were there'
'shall be sure to come across
one of their nests'
(b) Meaning of general particle (gen.p.) corroborated by a final particle (f.p.)
gen.p. f.p.
tae bu:an ponnoh
only four just so) 'there were only four left'
nvu s ol
(left over
extra
'still'
'still'
'all'
'immediately'
'without fail'
'just so'


ro
(c) Meaning of pre-nominal particle (pre-n.p.) corroborated by adverbial particle (adv.p.).
prem.p.
7 oh pu: k
(all group
krup ru:p
(complete person
adv.p.
all)
all)
translation
'the whole group'
'all of them'
(d) Meaning of post-nominal particle (postm.p.) corroborated by adverbial particle:
post-n.p. adv.p.
pum tatu:ol domnvq ?vy bontec-bontu:ec laay
(not receive news any slightest at all.)'did not receive any news
at all'
(e) Meaning of main verb corroborated by final particle or verb:
verb f.p. or verb
kaen
(increase
tvu muk tiot
on further)
'will increase'
(f) Meaning of attributive verb corroborated by an adverb:
yo:k khaet .....bam krup tE9q-?oh mo:k veji
(take province manage complete all come back) 'managed to recover all the
provinces 1
(g) Plurality of a reduplicative compound corroborated by sentence final nucleus:
'many large objects'
reboh thom-thom
(thing big and numerous
ci:a craon
in numbers)
extra
'all'
'all'
'the slightest'
'further'
'all'
'in numbers'


2. A feature which suggests conscious or unconscious imitation
of Thai
In colloquial speech, the only numeral coefficients
which have to he used are those of which the meaning is a term
of measurement, e.g. /ki:lo:/ in the phrase /sko: pi: ki:lo/
'two kilos of sugar'. In a formal or literary context, however,
there is a tendency to use unnecessary numeral coefficients.
Thus, for example, the numeral coefficient for human beings
/neek/ (e.g. /menuh pi: neok/ 'two men'), which has been in
evidence since the seventh century, is used more frequently in
literature; the general word for 'item', /preka:/ is found
(e.g. /haet pi: praka:/ 'two reasons'), and various coefficients
indicating shapes (e.g. /dom/ 'lump'; /sesay/ 'strand'; /daam/
'long thin thing') occur. In addition to the increased use of
numeral coefficients in connection with counting, however, there
is also a different construction which occurs in modern literature,
involving the use of numeral coefficients when counting is not
taking place, e.g.
/ceqcian vuaq nih/ (ring circle this) 'this ring'
/preeh ri:aci:a ?oij nih/ (revered king body this) 'this king'.
In such constructions, reminiscent of Thai, the numeral
coefficients, here /vuaq/ and /?oq/, are behaving as true
classifiers.
3. Features which suggest conscious or unconscious imitation
of French
Slight changes in grammatical usage which seem to be
due to French influence are illustrated below under four heads:
(i) the use of abstract nouns in preference to verbs; (ii) the
attempt to express some nuances contained in the meanings of
various verbal forms in French; (iii) changes of Cambodian
word-order in imitation of the French. Under (iv), the practice
of word-for-word translation of French turns of phrase is
illustrated.
(i) The use of abstract nouns in preference to verbs
The Khmer natural idiom uses a verb rather than a
noun whenever possible. "When an abstract noun is essential Khmer
has its own ways, based on verbs, of supplying the need. Nouns
formed by infixation of the verb are still in use, e.g. /komhoh/
'a wrong' < /khoh/ 'to be wrong'. Sometimes an abstract noun is
produced by juxtaposing two verbs of opposite meaning, e.g.
/tetu:el khoh-tro:v/ (lit. 'to accept wrong-right') 'to take
responsibility for'. Another means is to form a noun by placing
the word /ka:/ 'action' or /?ompv:/ 'activity' or /seckdvy/
1+


'matter1 before a verb or attributive verb, e.g. /ka: pueqriik/
(lit. 'matter open up') 'development'. However, many abstract
nouns were nevertheless borrowed from French and have now been
carefully replaced^ and supplemented by new Indian borrowings.
Also of importance is the position of an abstract noun in a
natural Khmer sentence. It seems that when abstract nouns are
used in Khmer, they do not usually occur, except in philosophical
discourse, as the subject of the sentence. Sentences in modern
writing in which new nouns occur often seem very un-Cambodian
and, especially if nouns do occur as sentence subjects, suggest
French influence to me, e.g.
/kumnut nuiq sakamophiiep nih pum mi:on presvthiphi:op
(idea and action these not have effectiveness
mu:oy ro:y phi:ok ro:y)
one hundred part hundred)
'These ideas and activities were not one hundred percent
effective'.
/sa:rophi:op srok yv: 13 ci:o sa:rophi:op srok vi:ol tumni:op/
(truth country we is truth country plain low)
'The essential nature of our country is that of a low
plain'/ 'Our country is essentially a low-lying plain'.
(ii) The attempt to express some 'nuances' contained in the
meanings of various verbal forms in French
(a) /daoy/ This word may occur as a clause-marker
(conjunction),, and, if so, the unexpressed subject of the clause
is always the same as that of the main clause. In everyday usage
it occurs with the meaning 'through the fact (that); because (of)'.
In novels, /daoy/ has been observed in occurrences such as these
but also in contexts where it seems to indicate that the action
of the verb in the clause it introduces takes place simultaneously
with the action of the main verb and where no cause is given, e.g.
/kraok chb: khvt ceji chga:y, daoy somdaeq
(get-up stand move off afar, (through) show
pre oh-ka:y-vika: s'op-khpv:m
royal-gesture despise)
'He stood up and moved away some distance, making a gesture
of contempt.'
A native Khmer way of expressing simultaneous action which might
have been used is to place the word /doinnae/ after each of the
two verbs. In the example, however, the Khmer sentence seems to
imitate the ability of French to bring the present participle
into play ('faisant' or 'en faisant').
1+


(b) /kampuq/ occurs as a pre-verbal particle meaning 'in the
middle of verb-ing' and precedes a main verb of action in normal
everyday usage. In novels it has been found in occurrences where
it seems to be added in as an extra word, e.g.
/pretsah phnsrk nraq sat-chlu:h mu:ay kampuq cho: si: smau/
(meet eyes with mouse-deer in-the-middle-of standing eat
grass)
'Their eyes lit upon a mouse-deer standing grazing'.
Here /kompu 13/ occurs unusually before a verb used attributively
in close junction with a noun, where the same phrase without
/kompuQ/ would be normal in Khmer and where French might have a
present participle or a relative clause with imperfect tense.
(c) Khmer sentences, spoken or written, tend to depend as
far as possible on context and meaning to indicate relative time,
and can proceed with a minimum use of words such as /ba:n/ (past),
/nrnq/ (future) and /hasy/ (completion)- Time relationships
between two clauses may not need any clause marker, such as
/kraoy-dael/ 'after', if any of the other indicators is present.
It seems probable that it was familiarity with the precision of
the various tenses in French which caused a much fuller use of
indications of relative time in recent written Khmer, e.g.
/kraoy-dael ba:n sdap seckdvy thlaeq cop haey, preah-soq ko:
bam somdaeq seckdvy ri:k-ri:ay/
(after have listened^to matter express through-to-end
already monks then h^ve shown matter joyful)
'After they had listened through to the end of the address,
the monks expressed their delight...'
Since the whole context of this sentence was already known to be
past, /barn/, which occurs in both clauses, was not needed at all.
In speech /haay/ at the end of the first clause would be enough
to indicate the relative time of the actions of the two clauses.
The French 'apres avoir ecoute...' seems possible as a model for
this.
(iii) Changes of Cambodian word-order in imitation of the
French
In spoken, or simply written Khmer, adverbs, adverbial
phrases, attributive verbs modifying a main verb, and post-verbs
completing the meaning of a main verb all characteristically
follow the main verb and its object and occur in clause or
sentence final position. In newspaper style, however, changes
of word-order such as the following may take place made
possible by use of the particle /nvu/ discussed in section 1 (ii).
1+


/cap-phdaam ci:a yu: chnam mo:k haay nru neayo:ba:y trak/
(establish being long years by now (particle) policy water)
'put the water policy into operation many years ago'
This seems to me to reflect French 'depuis de longues annees'
placed between verb and object.
The following example comes from an article in
Kambujasuriya (1966):
/sonisvt baep nih phdol ?aoy nvu lathaphol do: somkhan/
'A conference of this kind produces important results'
(lit. 'produces for CparticleU result').
Normally /?aoy/ 'to give; for' would come after the object; it
suggests 'for (us, people, the participants, one)'. Perhaps the
aim of the un-Khmer-like position of /?aoy/ is to avoid a
construction which is so unlike French?
(iv) Word-for-word translation of French turns of phrase
(a) The Khmer language has its own metaphoric vocabulary,
some of which uses the same imagery as French or English, for
example, in associating the ideas of heat or fire with anger. In
the following examples, from newspapers, however, the metaphors
of French idiom are used rather than those which the Khmer
language offers:
/rumcu:al hariirtey ya:q cri:al-crYu/ (agitated, royal-feelings
manner deep) 'deeply disturbed'-. Here Khmer idiom would
have /khlaq/ 'strong' instead of 'deep'.
/khoq bpriya:ka:h ri:k-ri:ay/ (in atmosphere joyful) 'in a
happy atmosphere' This use of the new Khmer loanword
/boriya:ka:h/ metaphorically is entirely due to French.
(b) Many instances may be found in newspapers of direct
translation of the complete French phrase:
/km: tha:/ (that-is to-say) 'c'est a dire', /kin:/ alone
rather than /km: tha:/ would seem more natural judging from
older Khmer prose.
/mi:an ka: pisaot/ (have matter experiment) 'avoir
1'experience', Use of the simple verb /pisaot/ rather than
the expression would seem more idiomatic in Khmer.
/baa ?aoka:h hoc ?aoy/ (if opportunity pass-across for)
'si l'occasion se presente'. /hoc/ normally has an object
following it, even in a metaphorical use such as in the
Khmer phrase /hoc khlu:an ?aoy praa/ (pass-across self
so-that use-services-of) 'offer one's services'.
1+


/nipoan daoy/ (compose through) 'ecrit) par'. This phrase,
seeming to he a translation from French, now appears on
the title page of hooks instead of the older Cambodian
format: Title of work: Author's name: /riap-riaq/ (prepared).
/kompuq sthvt nvu/ (in-the-mi.ddle-of be-stationed remain) 'se
trouve actuellement'. /kompuq/, which occurs characteris-
tically in Khmer before verbs of action rather than, as
here, before a stative verb, seems unnecessary in any case
in connotation with both /sthvt/ and /nvu/, either of
which means 'be situated'. I suspect /kompuq/ is there
because of familiarity with French 'actuellement'.
/mi:an karanvyakec nmq bompeji/ (have duty will fulfil) 'avoir
un devoir a remplir'. The Khmer form would be /tro:v
bompep karanvyakec/ 'must fulfil a duty'.
4. Features which seem to reflect the desire to embellish, using
the traditional devices of Khmer literature
These features are discussed under three heads: (i)
elevation of style; (ii) reduplication and repetition of ideas;
and (iii) expansion and variation.
(i) Elevation of style
Several features which have been discussed in preceding
sections help to elevate the style; for example, the Thai-
inspired use of numeral coefficients mentioned in section 2. Here,
however, two specific means are considered: the choice of vocabulary
and the use of particles preceding attributive verbs.
(a) Four waves of loanwords from Sanskrit and/or Pali have
come into the language to fill gaps: legal, religious, and polit-
ical terms in the pre-Angkor and Angkor periods; Pali Buddhist
terms, chiefly from the eleventh century onwards; the bulk of the
royal vocabulary from sometime after the end of the ..Angkor period
onwards; and the new vocabulary during the last few decades. The
'high' language of poetry has been drawn from the first three of
these categories. Terms from all four kinds of borrowing are
characteristically to be found in modern prose, causing the style
to differ profoundly from that of the spoken language.7
(b) /do:/ In Khmer an attributive verb follows immediately
the noun it modifies, in the normal structure of colloquial and
informal language, e.g. /mi:on phteeh thorn/ 'had a big house'.
In the literary language, however, /do:/ has long had the role of
preceding an attributive verb and attaching it to the noun. It
is still used and has the effect of highlighting the attribute:
e.g.
/a:riyaprete:h mu:ay do tracah-tracoq/ (civilised country
one particle shining) 'a magnificent civilization'
1+


In newspapers, however, its use is extended so as to attach other
categories of word to the nouns:
/thlaeq ?omno: kun do: kray le:q/ (express joy (at) good-deed
particle very very) 'express (my) great gratitude'
/lathaphol da: ci:e ti: koap cvt/ (result particle being
focus suit feelings) 'a satisfactory outcome'
In the fist of these examples, an adverb /kray le:q/, and in the
second, a phrase /cilB ti: koap cvt/, are treated in the same way
as attributive verbs.
Another extension of the use of /do:/ in journalese is
to allow it to attach two attributes to the noun:
/?oh ka:l do: yu nraq do: lumba:k/ (whole time particle long
and particle difficult) 'during this long and difficult
period'
/do:/ is not really necessary as a clarifier of the
construction even in the long phrases of the literary language;
it is an embellishment used particularly in descriptive passages
in novels or flowery parts of public speeches.
/ya:q/ This word, meaning 'way', is in competition
with /do:/, now as a means to attach attribute to noun: e.g.
/tatu:al para:c.ey ya: q ?a:mah/ 'suffered a humiliating defeat'
Unlike /do:/, however, it occurs also when an attributive verb
occurs post-verbally (usually to be translated into English by an
adverb), e.g.
/ka: ta:n-tvq ba:n kaet mi:an laaq ya:q khlaq-kla:/ (matter
tense has risen is-there rising way strong) 'tension has
arisen to an extreme degree'
Khmer grammar does not require the presence of any
particle before such an attributive verb and it would not be
present normally in informal and colloquial speech (cf. /knit
seckdvy tro:v/ 'think the matter out correctly'). The particles
/nvu/ and /ney/, discussed in section 1 as having a clarifying
role, also, like /do:/ and /yaq/, contribute by their mere
presence to the literary flavour of a sentence.
(ii) Reduplication and repetition of ideas
As is well known, Khmer is a language in which
reduplication occurs structurally at the levels of phonology,9
morphology,10 and syntax.^ Repetition of ideas occurs in the
formation of compounds.Both exact reduplication using
1+


phonaesthetic words, sometimes specially created, and repetition
of ideas using poetic vocabulary,13 occur as literary devices and
have done so since the beginning of Khmer poetic composition.1
The following examples, taken from modern prose, merely illustrate
the use of the same devices in modern writing.
(a) Reduplication
/rmt-tae chra: khlaq laoq khlaq la913/ (increasingly ill strong
up strong up) 'became more and more seriously ill'
/khlaq laoq/ could have occurred without reduplication.
/ci:k kuol ci:k rmh sum/ (dig tree-stump dig roots ask)
'make a thorough investigation'
/ci:k kuol rwh/ could have occurred without the interpolated
reduplication.
/ci:o nec cim ka:l/ 'always1
This use of interpolated reduplication, based on a compound loan-
word (Pali: niccakalaq) produces an understandable phrase because
both components of the compound are familiar to Khmer readers as
separate words. Word-play exercised on Indian loanwords is used
as a poetic embellishment.
(b) Repetition of ideas
The examples under this heading are from both novels
and newspapers:
/ci:9 ?ula:rmk ?athmk-?athijak/ (being grand splendid) 'in a
very grand manner'
/cim/ usually precedes one attributive verb in this much-
used pattern for forming clause final adverbial phrases
with the meaning 'in a manner'
/nvu-tae prokaek mien-toon totuml prb:m nvu-laoy/ still
refused not-yet accept agree still) 'still staunchly
refused'
The whole meaning is in effect said twice in different ways,
very much in the poetic tradition, particularly of the nineteenth
century.
A certain means of expression, which I have noticed partic-
ularly in the written language, and of which I have found
instances in Middle Khmer poetry, consists of following up the
main verb with a short clause which merely repeats the intention
expressed by the verb in another way: e.g.


/prap ke: ... 'aoy dvq/ 'tell someone ... that he may know'
/so:... 'aoy khY:ji/ 'indicate. .. for (one) to see'
/hqjiceji (e.g. tmk cvt kla:ha:n). .'aoy ke: khv:Ji/ 'show (e.g.
courage)...for them'
/bqpceok... 'aoy ke: dvq/ 'make clear... so that people will
know'
/hoqha :ji. . 'aoy yv:q khv:ji chbah/ 'demonstrate... so that we
shall clearly see'
/ba:n seckdvy comraen... do:c pra:thna:/ 'achieve success...
as was their wish'
/tvu: dol... do:c po:q/ 'arrive... as they desired'
This kind of clause seems to occur, in journalese especially,
rather more often than the contexts warrant. However, it may, in
some cases, seem more necessary to Khmers than it does in the
translation because the clause fills out the meaning of the
preceding verb, which, on its own in Khmer, has a limited
meaning, being able to convey neither the sense of finiteness nor
the ideas of completion or result which one can convey in English
or French by use of the past tense.
(iii) Expansion and variation
The simple forms of everyday speech or writing are
constantly expanded and varied in the formal, literary style.
Some examples are given (see next page).
Traditional poetic forms which parallel these examples
are to be found, e.g. /soh sa:/ 'utterly' (pro/soh/ ); /nuh ney/
'that, those' (pro/nuh/), and /hak bvy ru:/ 'as though'
(pro/do:c-ci:a).)
1+


Simple form
/mu:ey tiet/ (one further)
/mya:q tiet/ (one way further)
/miin tae ponnoh/ (not only like that)
/hontec/ (a little, soon)
/knoq pe:l yu: konloiq mo:k nuh/
(in time long past coming along that)
/teeq/ (also)
/kha:q/ (side)
/dol/ (as far as)
/tae/ (only)
/ba:n ci:e/(get to be)...pi:prueh
Expanded or varied form
/mu:ey tiet saot/ ( + moreover)
/mya:q veji tiet/( + again, now)
/mvin. tae ponnoh prb:m teeq...tiet/
( + join also... further)
/bontoep pi: pe:l^nuh bontec mb:k/
(next from time that a little
coming-along)
/knoq ?advtka:l konlo:i3 mb:k yu:
?ogve:q haey nuh/ (in past-time
coming-along long long already)
/prb:m teeq/ (join also)
/thaem teeq/ (add also)
/phnaek kha:q/ (section side)
/dera:p mb:k dol/ (all the way come
as-far-as)
/trvm tae/ (up-to-the point only)
/haet-dael. .kin: daoy/ (cause which.,
was through...)


Simple form
/haet nih haay "barn cira/
(cause this completed get to he)
/cog/ (wish)
/nah/ (very)
00
j
/soh laay/ (particles)
/rnuin kharn/ (not miss)
Expanded or varied form
'This was the reason why'
'wish'
'very'
'at all'
'without fail'
/daoy ?arsray haet nih haey tvrp
mi ran/ (through depend-on cause
this completed so have)
/mi ran bomnorg cog/ (have desire
wish)
/pe.rk.nah/ (too-much very)
/kray lerg/ (very let-go)
/kray perk/ (very too-much)
/yarg kray lerg/ (way very let-go)
/?aoy toal-tae soh/ (give so-far-as
at all)
/daoy kharn mmn "bam/ (through miss
not possible)


A further kind of expansion takes place in the use of
some idioms which have already been formed on a literary pattern
and which take part in a larger grammatical construction. Two
examples are:
/ya:q saen kho: khvu/ (way 100,000 aggressive) fin an
extremely aggressive manner1
/taq-pi: roap satavoat/ (since count centuries) 'for some,
centuries'
These are both based on idiomatic phrases which occur in
literature, /saen/ '100,000' functions in literature as a particle
'very' modifying a following attributive verb, here /kho:-khvu/.
(Khmer particles with this function would occur after the
attributive verb; /saen/ as a numeral would normally occur
preceding a numeral coefficient, not a verb.) In the phrase
/ya:q saen kho: khvu/ the expansion consists in taking this
instance of an established literary idiom and using it in place
of the attributive verb which usually, as was shown in 4(i),
follows /ya:q/. In the second phrase, /rbop/ 'to count' occurs
idiomatically with several words in a slightly formal language
style, e.g. /rbep ro:y/ 'in hundreds'. Here this idiomatic
adverbial form is used after a pre-nominal particle instead of
the noun, which is expected in that position (e.g. /taq-pi: chnam
nuh/ 'since that year'.)
Some features described in the foregoing sections,
particularly the use of particles, of time indicators, and of
corroborative nuclei when there is no need for them, seem to
confirm that there is a tendency to wordiness in Modern Khmer
written style. With regard to the influence of French, examples
have shown that, ironically, although loanwords from French have,
in principle, been discarded, some aspects of French grammar
have been incorporated! In presenting together all the points
which I have observed, I may have given the impression that all
modern writing of news, articles, speeches and novels is packed
with these features. In fact, however, many writers of articles
have used a restrained style and many novelists have used a very
simple style even in descriptive passages; one for example, is
Dik Keam in his novel Broil tae kromaen (1967). In any case, much
of a novel is concerned with narrative and conversation and, for
the presentation of conversation, a good tradition of writing
both naturally and interestingly seems to have been built up by
Khmer novelists, stemming perhaps from the lively little
conversation passages which are found in the verse novels (/satra:
lbaeq/). In fact, for a thorough assessment of a style one would
have to separate the genres of writing which have been treated
together here.
Although it has been interesting for me to attempt to analyse
recent changes in style, I personally prefer the simple, restrained
1+


one. However, perhaps foreigners should confine themselves to
factual description and not presume to judge style qualitatively.
From a Khmer point of view, as section k shows, the long-estab-
lished literary tradition, which is the tradition of poetry, has
encouraged embellishment by the use of long loanwords and of
reduplicative and repetitive patterns, all of which produce more
words.
NOTES
1. The writing of prose in Khmer was traditionally intended
for practical reasons only. Literature for artistic purposes
was always composed as poetry; at first, from the seventh to
the thirteenth centuries, in Sanskrit.
2. My source for this has been a photographed manuscript of the
Chronicles in the library of the School of Oriental and
African Studies. Some historical writing is easily available
in print, however, in Huffman (1977: 38-78).
3. Reproduced in print in the early volumes of the Pragum rioeh
breh khmaer.
b. The transcription used here was based on Professor Henderson's
pioneering analysis of Khmer pronunciation (Henderson 1952).
5. Terms for grammatical categories are as given in Jacob (1968).
6. This is not to say that French loans for which new vocabulary
was invented went out of use completely. They are still
heard.
7. Many Indian loans are so completely naturalized that their
presence in the colloquial language passes unnoticed, e.g.
/bon/ 'festival' < punya 'good work'.
8. /nvu/, /ney/ and /do:/ were briefly treated in Jacob (1978)
together with /ri:/ as literary particles. /ri:/ does not
seem to occur so much in recent writing.
9. That is, in words having reduplicated initial consonants
which are no longer, or have never been, analysable as
consisting of base and reduplicating prefix, e.g. /bobo:/
'soup', /totu:3l/ 'receive'.
10. Reduplicative prefixes consist chiefly of single consonants
(e.g. the frequentative /k(e)ka:y/ from /ka:y/ 'to dig into
the earth with hands, paws') but include some examples of
rhotized consonants (e.g. /tr(e)tuon/ which occurs in
1+


/tr(e)te:h-tr(e)tuan/ 'graceful' from /tuan/ 'soft, supple')
and of consonants followed "by a nasal consonant (e.g.
/s(a)nsaam/ 'dew' from /saam/ 'wet')* Reduplicative compounds
are formed "by the juxtaposition of alliterative, rhyming,
chiming, or repeated word-forms (e.g. /cat-caeq/ 'to organize'
/ri:ay-mi:ey/ 'untidily scattered'); /kme : q-kma: q/ 'children
(in quantity)'; /proh-proh/ 'the menfolk, the hoys'.
11. Reduplication in syntax intensifies meaning, e.g. /craan
laaq craan laaq/ 'increasing in quantity'is more emphatic
than /craan laaq/. Reduplication with interpolation also
occurs, especially with components of compounds; e.g. /prap
ke: prap 'aeq/ 'tell everybody', pro /prap ke:-'aeq/ id.
The reduplicated form is more stylish.
12. Frequently a compound with specialized meaning or with
abstract meaning has been formed of components, each of which
has, when used alone, approximately the same meaning, e.g.
/ciah-vi:aq/ 'to avoid' from /ciah/ 'to dodge', and /viaq/
'to follow a winding course, go round (an object)'.
13. The following are poetic examples /thla: thlaeq co:-ca:
prap/ (explain converse tell); /kru:eq kra:p 'aphiviiet
voanti:/ (shrink prostrate-oneself salute salute) 'greet
respectfully'.
14. By the nineteenth century this poetic device began to be
overdone. See Jacob (1979).
REFERENCES
Dik Keam
1967. Broh tae kamnaen ('Because of
the conscription'). Phnom Penh: Seng
Nguon Huot.
Henderson, E.J.A.
Huffman, F.E. &
Im Proum
1952. The main features of Cambodian
pronunciation. Bull. Sch. Orient. Afr.
Stud. 14, 149-74.
1977. Cambodian literary reader and
glossary (Yale JLing. ser. ). New Haven:
Yale Univ. Press.
Jacob, J.M.
1968. Introduction to Cambodian.
London: Oxford Univ. Press.
1978. Some observations on Khmer
verbal usages. Mon-Khmer Stud. VII,
95-109.
1+


Jacob, J.M.
1979. Observations on the uses of
reduplication as a poetic device in
Khmer. In Studies in Tai and Mon-
Khmer phonetics and phonology: in
honour of Eugenie J.A. Henderson.
(eds.) Theraphan L. Thongkum et al.
Bangkok: Chulalongkorn Univ. Press,
111-30.
Kambujasuriya
Prajum rioeh breh khmaer
Suttantaprija In
1927-- Phnom Penh: Inst. Bouddhique.
1959-74. Receuil des contes et
legendes cambodgiens. krum jamnum
damniem damlap' khmaer. Commission
des moeurs et coutumes du Cambodge.
Phnom Penh: Inst. Bouddhique, .8 vols.
1936. Gatilok. Phnom Penh: Inst.
Bouddhi que, vols. 1-10.
1+




KHASI KINSHIP TERMINOLOGY1
^Lili Rabel-Heymann
Irawati Karve, the acknowledged authority on Indian
kinship, who admits to unfamiliarity with the Khasi language,
lists twelve Khasi words in her study of kinship (Karve 1965),
but none of these words are included in U.R. Ehrenfels (1953)
paper, nor are they recorded in the four Khasi dictionaries
(Nissor Singh 1906, 1920; Lemuel 1965 (= Diengdoh); and
Kharkhongngor 1968) known tc me. Since these same twelve words
were also never mentioned by any of my Khasi informants, they
should probably be considered as non-existent in the Khasi
language; Karvefs erroneous listings may be understandable,
however, as her only sources were Roberts (l89l), Grierson (19(A) 9
and Gurdon (l9lb). U.R. Ehrenfels' article (1953) is, therefore,
the only modern treatment of Khasi kinship terminology; it had
apparently not come to Karve's attention.
Ehrenfels includes an almost complete list of kinship
terms in four dialects Plateau Khasi, that is, the standard
language; War Khasi; Pnar (or Jaintia); and War Jaintia.
indexed under ^3 English glosses, the purpose of which was one
of comparison. Although the author collected the vocabulary, in
an actual field situation, aided by native interpreters, his
spellings are not always reliable and the literal meaning of most
terms remains hidden from the reader unfamiliar with the language.
I, therefore, see my task as that of amplifying Ehrenfels' work,
of correcting some errors, especially in the orthography, and of
explaining the underlying morphemes occurring in the incredibly
complicated Khasi system.2
Khasi kinship terminology is based on three principles:
(1) a differentiation between blood relatives and relatives
by marriage;
(2) classification by relative age of each member with
regard to the person they are related to; and
(3) distinction between terms of address and terms of
reference.
Khasi society is generally considered to be matrilineal and
matrilocal; the kinship system could be called 'bifurcate
merging' since mother is equated with mother's sister, father
with father's brother, while mother's brother and father's sister
are denoted by distinct terms. Ancestry is traced through the
^3


3
mother's clan, ka kur, often used as an 'imitative' ki kur ki
kmie (lit. 'the clans the mothers'). Gender number^ morphemes
(called 'prefixes' by the Khasis) must precede each noun; ka
indicates feminine, u masculine singular; ki_ is used for plural
number and i_ for respect, endearment, and smallness, the latter
two not distinguishing gender.
Upon marriage, the husband remains in his mother's clan
while his children belong to his wife's clan, kur is used in
three phrases: iadei kur 'to have a relationship on the mother's
side, to be related within the same clan'; in the compound para
kur 'children of mother's siblings', and in the verb tait kur
tait jaid,5 an imitative 'to be banished from the clan' (lit.
'reject clan, reject kind, caste').
The most respected member in the clan is the mother's
elder brother, u kfii or kfii rangbah (rang-bah translates as
'adult male, an elder' and is composed of rang, the combining
form of shynrang 'man' and bah 'older brother', probably based
on the verb bah 'to be big'). The kfii is addressed as mama or
ma, a word which seems to be of Indo-European (IE) origin since
Hindi, Bengali, and Assamese use it for 'maternal uncle' as well.
The kfii is consulted on all important decisions and acts as the
ultimate arbiter in disputes. Mother's other brothers are
referred to as kfii pdeng 'middle brother' and kfii khadduh^
'mother's youngest brother, the very last' respectively. The
latter two uncles are addressed as ma-deng7 and ma-khadduh or
ma-duh; duh 'the last, the youngest' is used for blood relatives
only, never for those related by marriage.
The word for 'mother' has two forms which are probably
not related morphologically (see discussion below), kmie is
used for reference and mei for address; mother's sisters are
also addressed as mei plus the appropriate modifier for age-
ranking .
Father is referred to as i kpa and addressed as papa
or pa, a term strangely familiar to speakers of Indo-European
languages. His brothers, as well as the husbands of his sisters
and of his mother's sisters are all referred to as 'fathers'
(pa- used in compounds is explained below). Father's sisters
are all addressed and referred to as kha. kha is based on a
verb meaning 'to give birth', thus, according to Ehrenfels (1953:
Uo8), apparently recognizing the father's biological function in
procreation, kha also functions as the second constitutent in
compounds designating father's mother, kmie-kha, and cousins on
the father's side, shi para kha-shi is the numeral 'one' used
for units and measurements (as opposed to wei 'one'); para
designates brothers and sisters of one's own generation.
Parallel cousins and cross-cousins are distinguished
both in terminology and in marriage practices; marriage between
hh


parallel cousins is sang 'taboo1, while marriage between cross-
cousins is permitted though not common. Mother's brothers'
children are also referred to as ba-kha, since by definition his
children will belong to their own mother's clan.
All brothers and sisters within the immediate family
are designated by terms that specify whether they are older or
younger than the speaker; there are also terms to indicate a
'middle' brother or sister and terms for the youngest brother
and sister (Table l).
The oldest sister is kong or kong ieit, 'sister-
beloved' the oldest brother is bah bah 'brother big', or bah
rangbah 'brother grown-up man', terms that show a position of
respect occupied by the elder siblings. There are actually two
homophonous morphemes bah, one meaning 'brother', the other 'to
be big'; and since modifiers follow the noun in Khasi, kong ieit
would have to be translated as 'sister who is beloved' and bah
rangbah as 'brother who is grown-up.'
Unfortunately, Table 1 shows some gaps. Also, no two
of the young Khasi speakers who recently supplied me with
information agree on all terms; they have all been living abroad
for a long time and have become accustomed to our simplified
Western terminology using aunt, uncle, cousin,-etc. One speaker
also suggested that address by name is coming into vogue among
the younger generation.
khynnah 'child' is used for the youngest brother, bah
khynnah 'kid brother'; i rit (lit. 'little one') and i duh (lit.
'the last one') are best rendered by 'kid sister'. One word
glosses for hep and hynmen are difficult to suggest; older
people, even non-relatives, can address young people as hep; it
is a term of endearment and is roughly equivalent to the American
usage of 'sonny' or 'kid' when used by men for little boys, or
'dear' when used by old ladies for younger women. The morpheme
hyn- occurs in several other Khasi words that relate to 'time
past, ago': folk-tales always begin with hyndai-hynthai...
'once upon a time...' ; hynne means 'a short time ago', hynnin
'yesterday', so that hynmen could perhaps be rendered by 'born
before, born some time ago',. The morpheme -men occurs in tymmen
'old man or woman', so that it may mean 'old human being'.


Elder S/B*
Middle S/B
Younger S/B
The Youngest
Sister
!kong ieit'**
kong, hynmen kynthei
kong-deng
'hep'
hep
i rit, i duh
S's husband
'hynmen kynsi'
kong heh, hynmen kynsi
(by female)
kyn-um (by male)
kong-deng
'para kynsi'
'hep kynsi'
(by female)
Brother
p-
o\
'bah, bah bah, bah heh'
'bah rangbah'
'bah khynnah'
hep (by female)
'bah duh'
B's wife
hynmen shynrang
'kong kynsi'
* Abbreviations used here and in following tables are: B = Brother, F = Father, M = Mother, S = Sister
** Quotation marks indicate terms of address
Table 1:
Khasi sibling terms of address and of reference


kynsi and kyn-um are "best translated "by 'in-law' ;
the minor syllable kyn.occurs in so many words, nouns and verbs
alike, that it is impossible to assign a definite lexical meaning
to it.9 Ego's, i.e. the speaker's,maternal and paternal aunts
and uncles are, as Table 2 shows, classified according to the
same system as siblings; older or younger than parent referred
to, in-between the older and the younger aunt or uncle, and a
designation for the youngest aunt or uncle. Mother's and
father's in-laws take their age-ranking appellations from their
respective spouses rather than from their actual age. Terms of
address and terms of reference are usually identical in the
second and third ascending generation (grandparents and great-
grandparents) and for the first descending generation (children,
nephews, and nieces): terms are, however, as Tables 1 and 2
demonstrate, differentiated for ego's siblings and for ego's
parents and their siblings. The term of address is often a
shortened or reduced variant of the full term which is used for
reference; the shortened variant, also used in compounds, is
derived by loss of the initial consonant (usually k-), or loss
of the initial syllable in bisyllabic words. Examples are:
pa_vs. kpa 'father', rad vs. kynrad 'lord, master', mei vs.
kmie 'mother', -rang vs. shynrang 'adult male'
Most family members, it will be noticed, are not
addressed or referred to by name but by their kin classification;
even husband and wife have traditionally referred to each other
as 'the mother (of) Coldest child's nameU' e.g. i kmie u Dan and
i kpa u Dan 'the father (of) Dan1. One woman informant told me
that a woman can also address her husband's sister's husband,
i.e. her brother-in-law, as the kpa of the first-born child.
Husband and wife address each other by phi, the polite second
person pronoun 'you'. Younger Khasis state that nowadays
husband and wife may use names for addressing each other. Another
Khasi friend gave me the terms of ka lok for 'wife' and u lok for
'husband', but a young man said: 'lok is a harsh word, don't use
it. '
The reference terms for parents-in-law, kiaw for
'mother-in-law' and kthaw for 'father-in-law', do not distinguish
between maternal and paternal ancestry. The terminology for
grandparents is structured parallel to that of one's own parents;
however, distinctions between terms of address and of reference
are not as varied. 'Mother's mother' is mei-rad, 'mother's
father' is pa-rad; their respective siblings add -heh for the
older sister, -deng for the middle sister, but the youngest is
simply another mei-rad. Great-grandparents on the mother's side
are mei-buh and pa-buh;^ father's mother and father are kmie-kha
and pa-kha respectively. Referential terms for parents in all
generations take the respectful 'prefix/article'
hi


Mother*
Mfs Sisters
Mfs Ss! Husbands
M! s Brothers
M!s Bs' Wives
Father
F's Brothers
F's Bs1 Wives
F's Sisters
Ffs Ssf Husbands
'mei'
i kmie
'pa, papa'
i kpa
Older than M/F
mei-san
pa-san
'mama rangbah'
kfri rangbah
fiia
pa san
nah
kha rangbah
pa-kha? mama?
Middle S/B
mei-deng
pa khynnah
'ma-deng'
kfii pdeng
fria
pa-deng
nah
kha-deng
Younger than M
mei khynnah
pa khynnah
'ma, mama'
kfti
nia
pa khynnah
nah
mama?
The Youngest
nah rit
pa khynnah
'ma khadduh'
khi khadduh
fiia
pa-duh
nah
kha-duh
mama?
* For the unfilled slots definitive information is lacking
Quotation marks indicate terms of address; terms of reference have no quotes. When only one term
is listed, address and reference are the same
Khasi usage of hyphens is not systematic; my own usage is to hyphenate 'reduced' morphemes
Table 2: Khasi terms for maternal and paternal aunts and uncles


Parents refer to and address their children's spouses
as pyrsa, the term also used for maternal nephews and nieces,
pyrsa kurim (kurim 'wife'); uncles and great-uncles on the
mother's side refer to a child as pyrsa ksiew while the child
addresses his great-uncle as "bah, ksiew, or khun ksiew, are
the terms for 'grandchild', and ksiew tun for 'great-grandchild'.
Great-grandchildren are referred to as khun miaw (lit. 'cat
children'), and great-great-grandchildren as khun khnai (lit.
'mouse children'). Step-children are referred to as khun ruid
/khuon ruj/ and''stepfather' is u kpa nah according to Nissor
Singh (1906:38,^5). Unfortunately, no sources of information,
informants or dictionaries, could provide a literal meaning for
tun or ruid.
Kinship terms are generally assumed to "be of native
stock, along with "body parts and numerals. However, Khasi has
"borrowed extensively from the geographically surrounding
languages with many loans so well integrated into the native
sound structure that their detection is difficult, if not
impossible. Three words may be of Indo-Aryan derivation:
(i) pa (see above), but a prefixed k either indicates a very
early borrowing or would be counter-indieative;
(ii) para (see above) is listed by Karve (1965) under Hindi,
Sindhi, and Punjabi as referring to blood-related maternal/
paternal grandfathers, while in Khasi this word is used for
younger relatives on the mother's side; and
(iii) kurim 'wife', has an unchecked long vowel in the first
syllable which, according to my earlier findings, points to
Indo-Aryan origin; kurim 'wife' and kur 'clan' are definitely
not related.
GLOSSARY
The following is a complete inventory of all kinship
terms with which my informants supplied me, many o£ which are
listed in Nissor Singh's famous dictionary (1906).12 The
alphabetical order is that commonly used for European languages;
it deviates from the established Khasi alphabet in three
respects: the aspirated stops kh, ph, and th are treated as
separate phonemes and are, therefore, not arranged within the
k, p_, and t_ listings; k and kh are listed after letter and
not after letter b as in the Khasi alphabet; ng_ /q/ follows n_
instead of taking the place of
bah to be big; u bah 'big brother'; a polite form of address
for any man older than speaker /ba?/
bah bah, bah heh, bah rangbah 'older brother' /ba?he?/,
ba?-raQba?/
bah duh, bah khynnah 'youngest brother' /ba^dtf?/, /ba?
khnna9/
1+9


ba-kha children of mother's brothers; relationship between
mother's children and maternal uncle's children (Nissor Singh
(1906:6), also includes 'paternal aunt' (qv. also Kharkhong-
ngor 1968:10b).
bih form of address for a young girl (bi?/
bu affectionate form of address for a young boy (not common)
deng from pdeng 'in the middle, between', used in compounds
/pdeq/
kni pdeng referring to mother's middle brother
kha-deng addressing father's middle sister
ma-deng addressing mother's middle brother
mei-deng addressing mother's middle sister
pa-deng addressing father's middle brother
duh from khadduh 'the last one, the youngest' /khat-du?/
flia khadduh 'mother's youngest brother's wife' /ffa-khat-
du?/
i duh 'the youngest sister, baby sister'
heh 'big'
hep a polite way of addressing and referring to anybody younger
than the speaker; used for younger siblings; woman addressing
and referring to brother-in-law if married to younger sister;
same as kong hep kynsitwoman referring to younger sister's
husband /knsi/
hynmen i_ referring to elder sister /hnmen/
u referring to elder brother
hynmen hynbew (imit.), elder brother or sister /hnbew/
hynmen kynsi younger sister addressing older sister's
husband
hynmen kynthei referring to elder sister
hynmen shynrang referring to elder brother /hnmen snraq/
ieit 'to love, to be loved' /?iet/
mei ieit addressing mother's mother
pa ieit addressing mother's father
kong ieit addressing elder sister
ing, iing, ying 'house'
ka iing ka sem household, family (ka sem 'stable, shed,
shelter') /ka yieq ka sem/ (imit.)
kiaw mother-in-law
1+


kmie 'mother* when referred to (reduced form i mei) /kmi/
kmie hep referring to father's mother
ki kmie ki kpa 'parents' (imit.)
kmie kha referring to father's mother
kmie-nah referring to mother's younger sister; stepmother
also: i mei-nah khadduh /?i mey-na? khat-du?/
kmie-ra'd referring to mother's mother /kmi-raat/
kmie-san referring to mother's elder sister
kfli u, i_ mother's eldest brother (=kfii rangbah. p.44 above),
the most respected person in the clan; he is addressed as ma
or mama
ki kfii ki kpa relatives on mother's side (imit.)
kffia ka, mother's brother's wife; in compounds Hia /kha/
fiia-kha addressing and referring to father's sisters
Hia-kha rangbah father's oldest sister
flia-khadduh father's youngest sister /na-khat-du?/
ffia-pdeng father's middle sister
kong ka, addressing and referring to older sister; man
addressing wife's female relatives; polite form of address for
any woman older than speaker; in compounds also used for males
kong-deng husband of middle sister, brother-in-law
kong heh woman addressing and referring to brother-in-law
kong kynsi addressing older sibling's spouse
kpa u, i_ referring to 'father', pa when addressed and in most
compounds
kpa-nah referring to step-father, nah from khynnah
kpa-rad referring to father's father; /raat/ from knraat/
'lord'
pa-buh i great-grandfather on mother's side
pa-deng father's middle brother
pa-ieit addressing mother's father /pa-?iet/
pa-kha referring to father's father (cf. Nissor Singh 1906:
147; ''pakha, u, n. a male relative (father's side)')
pa-khynnah addressing father's youngest brother and mother's
youngest sister's husband
pa-rad i addressing father's father, mother's father
/pa-raat/
pa-san addressing father's older brother, also mother's
older sister's husband
ksjew ka, u, i_ grandchild /ksiw/
ksiew tun great-grandchild
khun ksiew i_ referring to grandchild; is addressed by name
para ksiew mother's mother's sister's son (grand-nephew)
1+


pyrsa ksiew man referring to sibling's grandchildren; a
grand-niece or grand-nephew on mother's side
kthaw addressing and referring to father-in-law /kthaaw/
kur ka clan; ka kur ka jaid* (imit.), a relative on mother's
side /ka jaj/
ki kur ki karo, ki kur ki kmie (imit.), considered
obsolete
iadei kur to be related within the same clan (lit. 'come
together (in) clan') /yadey/
para kur member of the same clan, children of mother's
siblings
tait kur tait kmie (imit.), obsolete, and
tait kur tait jaid*(imit.), to be banished, excommunicated
from the clan; /taj/ from /kntaj/ 'to reject, set aside'
kurim ka wife, spouse; to have intercourse (probably of Indo-
European origin)
kiaw kurim ka mother-in-law (Singh 1906:*+0)
shong kurim to marry JfLit. 'lie wit£ wife' )
shong kurim shong kup'ai (imit.) kupai\ = ? from Hindi ?)
jingshong kurim marriage (lit. 'lie with wife') /jipsoq/
kynsi addressing and referring to brother- or sister-in-law
hep kynsi = para kynsi woman referring to younger sister's
husband
hynmen kynsi woman referring to elder sister's husband
kong kynsi woman addressing elder sibling's spouse
kyn-um u, i_ man addressing and referring to sister's husband
/kn?um/
kha to give birth, to bring forth
kha deng father's middle sister
kha-duh father's youngest sister
kha-rangbah father's older sister
kmie-kha referring to father's mother; she is addressed
as mei-kha
para kha (shi) father's brother's children; cousins /si/
'one' (numeral)
khadduh to be last /khat-du?/ see duh
kni khadduh referring to mother's youngest brother
ma-khadduh, ma-duh addressing mother's youngest brother
pa-duh addressing father's youngest brother
* From Hindi, see n.5. (Ed.)
1+


kfrfo ka, u, i_ child, girl, hoy, baby /khuon/
khun ksiew grandchild (maternal uncle's child)
khun khnai great-great-grandchild (lit. 'mouse child')
/khuon khnaay/
khun miaw great-grandchild (lit. 'cat child')
khun ruid step-child /khuon ruj/
khynnah ka, u, i_ to be young; girl, boy /khnna?/
pa-khynnah addressing and referring to father's younger
brother, also to mother's younger sister's husband (cf.
P-i+5
khynraw u referring to a young man
ka referring to a young wpman
u khynraw u samla (imit.) (arch*)
lok friend (arch,)
ka lok wife
u lok husband
para lok friend
lud to be young (arch.) = khynr£w /luot/
mama, ma addressing mother's older brother who is referred to
as u kfii
ma-deng addressing mother's middle brother
ma-khadduh, ma-duh addressing mother's youngest brother
ma-Rangbah (or: mama) addressing mother's oldest brother
mei i_ mother, alternate form for kmie, used in all compounds
mei-buh great-grandmother on mother's side
mei-deng addressing mother's middle sister
mei-hep respectful appellation of an older woman (Nissor
Sing 1906:131)
mei-ieit addressing mother's mother
mei-kha addressing father's mother
mei-khynnah referring to mother's younger sister
mei-nah addressing mother's younger sister and father's
younger brother's wife
mei-rad referring to mother's mother
mei-san addressing mother's oldest sister
myngkew ka, i_ used by wife when referring to husband's older
sister (arch-:;)
nah reduced form of khynnah to be small, be the youngest (used
in compounds)
mei-nah _ addressing mother's younger sister; also
father's younger brother's wife; addressing stepmother
nah rit mother's youngest sister
53


Sia reduced form of kMa, referring to mother's brother's wife
/kfia/
fiia kha, ka addressing father's sister
flia kha rangbah referring to father's older sister
hia kha khadduh referring to father's younger sister
fiia kha pdeng referring to father's middle sister
pa father, reduced form of kpa, used in compounds
para ka, u, children, brothers and sisters of same
generation
para ar kmie mother's sister's children (lit. 'children
(of) two mothers') /para ?aar kmi/
para briew u having no relationship either by blood or by
marriage; fellow human being
para ksiew mother's mother's sister's son, i.e. grand-
nephew on mother's side
para kur member of the same clan, mother's sister's
children
para kynsi referring to wife's younger sister and her
husband
para lok friend
para mynshong u wife's sister's husband
para shong-kha husband's sister's husband, two men marrying
into the same family
para trai ka, u brothers and sisters of the same parents
para kha shi children of father's sisters and brothers,
first cousins and half-siblings on father's side
pyrsa referring to one's own children and to sister's children;
and nephews and nieces on mother's side
pyrsa ksiew mother's mother's brother referring to speaker;
i.e. great-uncle on mother's side referring to speaker
pyrsa kurim referring to child's spouse
Phi 2nd person pronoun (polite), used by husband/wife in
addressing each other
rangbah to be grown-up; an older respected person, an elder
bah rangbah addressing older brother
mama rangbah addressing mother's oldest brother (even if
younger than mother)
kha rangbah addressing father's oldest sister
rad reduced alternate of kynrad 'master, lord' used in compounds
kmie-rad, mei-rad mother's mother
kpa-rad, pa-rad mother's father
5U


rit small
i rit = i duh the youngest sister
nah rit mother's youngest sister
samla marriageable young person
san to grow up, grown-up
mei-san mother's older sister
pa-san father's older brother, mother's older sister's
husband
(sang taboo)
tnga ka wife, when referred to (arch., 'a harsh word')
u husband, when referred to
trai u lord, master /?u traay/
para trai ka, u, i_ brothers and sisters of the same
parents
NOTES
1. This article represents a thoroughly revised version of a
paper delivered at the American Oriental Society Annual
General Meeting in Toronto, Canada, 11 April 1978.
2. Bowing to Khasi preference, I have decided to abandon my
usual practice of using phonemic transcriptions for Khasi
sounds; I use instead established Khasi spelling and only
add transcriptions in those cases where the standard spelling
system fails to indicate vowel length or vowel quality.
Sometimes it will be necessary to separate morphemes by
means of hyphens, although some of these hyphens are not
used by the Khasis themselves. Certain morphological
features, necessary for the reader's better understanding,
will be explained at the appropriate places.
3. 'Imitatives' which I called 'redundants' in Rabel (1968),
consist of two nouns whose combined meaning is equal to that
of the first constituent alone; the meaning of the second
constituent is often unknown to the native speaker. Some-
times the second word is a borrowing from Indo-European.
4. The Khasis call the four gender/number indicators 'prefixes';
I called them 'articles' in Rabel (1961). Neither term is
entirely adequate since these words are free morphemes which
also function as 3rd person pronouns.
55


5.
6.
7.
8.
9.
10.
11.
12.
jaid^/jaj /, with a short vowel, is of Hindi origin.
khadduh=/khat-du?/, is a compound of /khat/ 'to dole out' +
/du?/ 'to he last.'
deng is the reduced form of pdeng; see n.10 "below for loss
of initial consonant.
'minor' syllables have no full vowel nucleus, the second
consonant a liquid or nasal functioning as the vocalic
element; they are always unstressed.
Referring to n.10, it may be derived from (k + stem) + infix
-n-.
A theory first proposed by Pater Wilhelm Schmidt (190U) and
elaborated by Henderson (1976) states that initial k should
be considered a fossilized prefix denoting living things
(kinship terms, body parts, animals, plants). Since
simplification of initial clusters is otherwise never
accompanied by vowel change, it seems unlikely that mei
'mother' is related to kmie.
Dictionaries do not list buh; one speaker pronounced this
word /bu/, two others said /bu7/.
Khasi kinship terminology is, as we have seen, an
interesting topic and remains, as this paper demonstrates,
a subject of varying interpretation and discussion. Not
all of the terms mentioned in this article are included
in its glossary, or vice versa, nor do they correspond
precisely when they are; nor, for example, do they always
agree with the dictionary definitions (when these occur) of
U Nissor Singh (1906) or E. Bars (1973), among others.
This is, however, the most recent study of the question and
makes it most intriguing and informative. (Ed.)
REFERENCES
Bars, E.
1973. Khasi-English dictionary.
Shillong: Don Bosco Press.
Blah, V. Edingson
1971. (comp.) Chapala fs Anglo-Khasi
dictionary. Shillong: Chapala Book Stall.
Diengdoh, A.K.
1965. (comp.) Leemuel Ts Anglo-Khasi
pocket dictionary. Shillong: L. Harrison.
56


Ehrenfels, U.R.
Grierson, G.A.
Gurdon, P.R.T.
Henderson, Eugenie J.A.
Karve, Irawati
Kharkhongngor, U.
Iarington
Rabel, Lili
Roberts, H.
Schmidt, W.
Singh, U Nissor
1953. Khasi kinship terms in four
dialects. Anthropos 48, 396-1+12.
190k. Linguistic survey of India.
Vol.11. Calcutta: Govt. Printing
Office.
191^. The Khasis. __ London: Macmillan,
2nd ed.
1976. Vestiges of morphology in modern
standard Khasi. In Austroasiatic Studies
I (Oceanic linguistics spec. pubs. 13)
(ed.) P.N. Jenner et al. Honolulu:
Univ. Hawaii Press, 1+77-522.
1965. Kinship organisation in India.
Bombay: Asia Publ. Ho., 2nd ed.
1968. Ka Dienshonhi (a Khasi-Khasi
dictionary). Shillong: Ri Khasi Press.
1961. Khasi, a language of Assam
(Louisiana State Univ. Studies, Humanities
ser. 10). Baton Rouge: Louisiana State
Univ. Press.
1968. Redundant expressions in Khasi.
In Studies in Indian linguistics (ed.)
B. Krishnamurti. Annamalainagar: Centre,
Adv. Study Ling., Deccan College, Poona
& Annamalai Univ., 257-86.
1891. A grammar of Khassi language.
London: Kegan Paul, Trench, Trubner.
1901+. Grundzug,e einer Lautlehre der
Khasi~Sprache. Munchen: Verlag der K.
Akademie.
1906. Khasi-English dictionary.
Shillong: East. Bengal & Assam Secretariat
Press.
1920. English-Khasi dictionary (eds.)
R.S.D. Ropmay & U H.K. Singh. Shillong:
Assam Secretariat Printing Off.
57




ON PROSODIC RELATIONS BETWEEN FIJIAN BASES
AND VERBAL SUFFIXES1
G.B. Milner
From the earliest days in the study of the Fijian (Fl)
language the origins, functions and the degree of predictability
of its verbal suffixes have bewildered all those who have tried
to understand Fijian grammar, and they continue to do so. The
problem in question is a suggestive example of the interaction
between synchronic and diachronic factors in language and of its
consequence for linguistic analysis.
It will be remembered that a Fijian word-stem or base
(the latter term being widely employed in the description of
Austronesian languages to distinguish 'content words' from
'functors') subsumes both verbal and nominal word classes.^ It
may be disyllabic (CVCV), by far the most common statistically,
or trisyllabic (CVCVCV). The vowel slots are always filled, but
in disyllabic bases the initial and/or the medial consonant is
optional. In trisyllabic bases the medial and/or the final
consonant is optional.1 There is also a small number of bases
of more than three syllables.
When a base is a verb it may occur in any one of these
three standard forms. It is then said to be stative or
intransitive, according to certain syntactic criteria. When a
verb is followed by a monosyllabic suffix (C)V or by a disyllabic
suffix (C)VCV it is said to be transitive. Recent studies,
however, have questioned the applicability of terms such as
'transitivity' to this feature of Fijian grammar5 (Hockett 1976:
192; Naylor 1978: ^05; Schutz 1981: 197-203).
One of the most interesting problems in the comparative
study of Austronesian languages is that on the one hand in
Fijian, as in other Oceanic languages:
1. The occurrence or non-occurrence of transitive suffixes is
subject to certain semantic and syntactic criteria which are not
yet fully understood; and
2. The consonant of a monosyllabic suffix: (C)V,^ and the first
consonant of a disyllabic suffix: (C)VCV, is selected from a
limited series within the total inventory of consonants (cf.
Pawley 1978: 113-Ao).
In many other members of the Austronesian family, on
the other hand, and especially in Indonesian languages, cognate
59


verbs may occur which, in a fairly large number of cases, show
regular sound correspondences between their (non-significant)
stem-final consonants and the consonants of the verbal suffixes
of Fijian and other Oceanic languages.
Thus, Proto-Indonesian *taqit1 Tcryf corresponds to
Fijian tagica? /taqi5:a/ 'cry for (i.e. so as to obtain) something'.
Proto-Indonesian *davat 'reach; obtain' corresponds to
Fijian rawata 'get, obtain'.
Since the stem-final consonants of verbs like Proto-
Indonesian *taqit' and *davat are not known to have had a grammat-
ical function and the corresponding stem-final consonants of
verbs in modern Indonesian languages do not have such a function,
while corresponding consonants in Fijian and other Oceanic
languages occur in suffixes entering into regular grammatical
relations, intriguing questions arise regarding the origins,
nature and the precise functions of these features. In particular,
three immediate questions which arise are:
1. How can one account for the fact that these sound correspond-
ences can be attested in a significant number, but by no means
in a majority of cases?
2. Are the stem-final consonants of modern Indonesian verbs
vestigial in the sense that they might be the extant reflexes of
'archaic' grammatical suffixes which have now disappeared but
continue to function in Oceanic languages such as Fijian? (cf.
Dahl 1973:11). This is a question which should be asked even if
it cannot be answered in the present state of our knowledge.
3. Should the verbal suffixes of Fijian be regarded as an
integral part of the bases to which they may or may not be
attached? That is to say, is the choice of consonant determined:
(a) By the base and suffix considered as an artic-
ulated (and of course separable) but integral lexical
entity or:
(b) By semantic and syntactic factors, that is to say,
by the independently variable relations which can
obtain between a verb and its potential objects or
complements?
In the earliest days of the study of the Fijian language,
Hazlewood (l872: 32-3), in his work originally published in 1850,
after listing 'The Definite-Transitive Terminations' in two
classes, states that:
60


1. Those which consists of one syllable. These are,
-a, -ca, -ga, -ka, -ma, -na, -ra, -ta, -va, -wa and
-ya ...
Later he adds that:
2. There appears to be no certain rule to determine
which termination a verb will take. This must be
learned from the natives, or from the Dictionary.
3. But notwithstanding that there is no invariable
rule, yet we are persuaded that they (sc. the termina-
tions) are not always used arbitrarily ... (l.) It
seems to amount to a rule, that verbs formed from nouns
without prefixing vaka-, shall take na for their
termination ... (2.) It appears also to be a rule,
that verbs of motion will take va for their termination;
as lakova, ciciva, kadava, drodrova, ... Va_ here
means to. It is also true that many other verbs besides
those of motion take va, but for these perhaps there is
no rule. (3.) When verbs reject a termination of the
first or monosyllabic class, and take one of the second,
or disyllabic, they frequently have either a more
intensive sense, or take a different object.
Nearly a century later, Churchward in A new Fijian
grammar (l9*+l: 17-8; 71-2) speaks of: 'definite-transitive verbs',
and he states that:
different verbs take different suffixes and there
seems to be no rule for determining which suffix any
particular verb will take.
This is also the view taken by the present writer in
his Fijian grammar:
There is no known rule to indicate which suffix is
appropriate to what base.. It is advisable therefore
to learn each new base together with its correct
suffix or suffixes. (Milner 1972: 27-8)9
These words, written nearly thirty years ago, must now
be qualified, not only in the context of the result of subsequent
study by the present writer and his colleagues which have become
available in the meantime,10 but also in the light of recent
attention given to the same problem in connection with the
preparation of a new Fijian dictionary.11
It is necessary first to refer to Dempwolff's (l93*+-9)
Vergleichende Lautlehre, which has for over forty years been an
indispenable text in comparative Austronesian linguistics. It
will be remembered that in his first monograph (Dempwolff 193*+:
61


27-8) he distinguishes five categories of word stems (Wortstamme).
The first, which makes up 70% of his field of 1000 items,
consists of those which conform to the pattern CVCVC (e.g.
*laqit). Next in frequency comes word-stems of the same pattern
with the addition of an optional nasal 'connector' (Basal-
verbindung), hence of the pattern CV(C)CVC (*buqsoq, *guntiq).
They make up another 20% of the total. Another 5% consists of
reduplicated items, followed by 3% made up of word-stems of more
than two syllables. The remainder, approximately 1%, consists
of monosyllabic word-stems.
In his second monograph (Dempwolff 1937: 125-66) he
compares two Melanesian languages with his reconstructed Proto-
Austronesian (PAN) word-stems, one of the two being Fijian, the
vocabulary of which is examined in detail in order to arrive at
regular correspondences {ibid., 126-U6). He is struck by the
number of irregular, as well as regular, reflexes of his proposed
reconstructions in Fijian. Of particular relevance to the
problem under discussion here are the following passages:
Phonetic disagreements (lit. non-agreements of sound:
Lautunstimmigkeiten) (occur) especially frequently
with the final consonants of Fijian before a supporting
suffix...
From these data we shall draw the conclusion here that
these phonetic disagreements of Fijian must be
interpreted as 'false' analogy... {ibid., 133-b9
para. 127(a) 6).12
It is interesting that perhaps in order not to give
hostages to fortune, Dempwolff used inverted commas for 'false'
in 'false analogy'. At the time when he was assembling his data,
knowledge of the vocabulary of Fijian was much less advanced than
it is now, half a century later. With hindsight, therefore, and
the advantage of greater knowledge of Fijian grammar than Dempwolff
had either the possibility or the opportunity of acquiring, it
was useful for me to check his data where they bear directly on
the correspondences between Fijian verbal suffixes and the
reconstructed final consonants of PAN verbs.
Looking again at his PAN glossary in detail (Dempwolff
1938) with this particular end in view, I find 1^3 items which are
suitable for comparison. Of these, 6l (i.e. two more than he was
prepared to accept) show 'correct' (i.e. regular) correspondences,
assuming, that is, that one accepts his own criteria for what is
(and what is not) 'regular'.
67 are 'incorrect'. This total subsumes not only cases
where the proposed correspondence is 'irregular' according to
Dempwolff himself, but cases where there is another reason for
rejection. Some of the non-admissible comparisons arise from an
62


incorrect interpretation of the Fijian data (in some instances
because the information available to him was misleading or
inadequate). Other pairs proposed for comparison seem to be
semantically altogether too far-fetched to be acceptable.
15 correspondences are uncertain, in the sense that
they are insufficiently supported, but there is no reason why,
given additional evidence, they could not be confirmed and
accepted; for example, PAN *palu 'beat, strike' and Fijian valu
'fight'; PAN tinCddav 'consider accurately' and Fijian tirova
'look at one's reflection in water'.
Turning now to the regular correspondences, it is of
interest to note that the following occur most frequently between
stem-final consonants:
PAN FI PAN FI
*0 0 (ll instances) *h 0 (7 instances)
*t t ( 9 instances) *k k (7 instances)
*tf c ( 8 instances) *P v (5 instances)
Hit is important to note that in the above table, *t/
in PAN represents a reconstructed palatal: c_ and v in Fijian
represent two fricatives, a voiced interdental and a voiced
bilabial respectively. II
The next important contribution to a better under-
standing of the problems under consideration appeared a decade
after the publication of Dempwolff's third monograph. In A study
in the phonetics of Fijian, Scott (1948: 737-52) presented the
first detailed analysis of Fijian phonology by a modern profess-
ional linguist.pn particular, he was the first to draw
attention to the structural and incidentally remarkably
symmetrical relationship between the classes of consonants.
Though he was not primarily concerned with orthography, his
analysis fully, if only implicitly, vindicates the consistent
and economical alphabet devised by the pioneer missionaries
Cargill and Cross.^
The table (see next page) reproduced from the article
in question (Scott 1948: 743, Table 3: Correlations between
consonantal phonemes and alphabetic script in Fijian) illustrates
the quasi-complete one-to-one relationship between Fijian
consonant phonemes and the letters used in that alphabet.
63


Bilabial Dental Alveolar Velar
Nasalised plosive b d dr a
Non-nasalized plosive V t r k
Nasal m n g
Lateral
Fricative c s
Semi-nowel (w) w
Scott makes an important contribution to the problems
under discussion here, pointing out that not only are the non-
nasalized sounds (v, t_, r_, and k) 'articulated at corresponding
points', (sc. to b, d, dr, and 'and except for v, in a
corresponding manner' but 'v is linked with the t_, r_, k set
functionally; for like them, it enters into "transitive suffixes"
which b, d, dr, ^ never do.' In a footnote he adds that: 'It
does not seem that anything in Fijian indicates that the consonant
does not belong to the suffix, though the large number of forms
serving apparently the same purpose suggests a problem' {ibid.,
742, n.4).
We come now to the most comprehensive contribution so
far to the understanding of this problem, a doctoral dissertation
by the Rev. David Arms (1975)9 a New Zealand missionary who had
already spent several years in close contact with Fijian-speaking
communities. Arms analyses the phonotactic constraints which
govern the occurrence of the verbal suffixes.He shows {ibid.,
130-47) that,with very few exceptions, the place of articulation
of any consonant in Fijian verb rules out the occurrence of a
verbal suffix with a consonant (or first consonant in the case
of disyllabic suffixes )16 with the same place of articulation. 1
His data are significant, both from a diachronic and a
synchronic point of view. It is likely, for instance, to suggest
an explanation for at least some of the cases of non-correspondence
betwen Dempwolff's PAN verbs and Fijian verbal suffixes.18
As Arms points out (1975:l4o), the general constraint
operating on consonants in suffixes, also helps to account for
the fact that the nasalized stops (b, <1, dr, and phonetically
Lmbl, CndU, Cndrd, and CggD respectively), do not occur in verbal
suffixes; if they did, they would be unacceptable after verb
bases which have a nasalized stop in the initial or the medial
consonant position, and these are very numerous.
I have also made a detailed analysis of the synchronic
system of these dissociations ^ n order to discover to what
extent it conforms with Scott's (1948:743) table of Fijian
consonants reproduced above. This shows that except in one or
two cases, it is also possible to classify places of articulation
64


if one treats dissociation as a sole criterion. In the table
below, consonants which regularly dissociate from one another
have been placed in the same column.
Table 1: Consonantal dissociation in Fijian
b d dr q
v t_ r_ k
m EE
1
c_ s
(w) w
The following observations can be made regarding the table above:
1. Consonants which occur in verbal suffixes are those under-
lined. 1_ and s_ are entered in for the sake of completion, but
since neither occurs in monosyllabic suffixes and only occurs
in the disyllabic suffix -laka (which is 'intensive' in its effect
and apparently not subject to any phonotactic constraints), it is
difficult to decide in which column to enter it.
2. In a monosyllabic suffix, zero consonant (0) (i.e. the absence
of a consonant) occurs very frequently. A suffix is then reduced
to -a after a front vowel and -ya after the open vowel or a back
vowel.
3. -ta regularly dissociates from d or t in the base, apparently
with the sole exception of (vaka)dinata 'bear out, confirm'.
4. r_ and n regularly dissociate from each other, apparently with
the sole exception of karona 'take great care of, value greatly.
5. It is necessary to give a separate column from r_, not only
because the suffix -raka can occur after 1_ (and conversely -laka
after r_; cf. Arms 1975: l4l, n.4) but because 1_ and n associate
freely: e.g. lomana, lawana, etc. (cf. ibid., 139).
6. Scott had regarded the interdental place of articulation of
c_ C5D as relatively less important from the point of view of
classification, and entered it in the same column as the two
dental consonants t_ and d_. For the same reason he had regarded
n, which is, in fact, alveolar, as being intermediate between the
dental and alveolar places of articulation. In both cases, his
decision was probably influenced by considerations of structural
symmetry. It is worth noting, however, that the dissociation
principle firmly confirms c_ USD as being distinct from d and t_.
Likewise n (alveolar) is distinct from d and t_. This is consistent
with the articulatory data, even though it entails a sacrifice of
symmetry or 'elegance'.
65


Having recorded our debt to Arms, it is now necessary
to register strong reservations about the remainder of his
analysis, which I now summarize.
Because of his extensive knowledge of spoken Fijian,
based on regular practice in the course of his activities, both
pastoral and informal, his views merit to be treated with special
attention, although still open to rigorous examination. He
implies21 that, granted a reasonable competence in the language,
if a speaker is given any one base, together with its approximate
semantic reference, he should be able to predict with a fair
degree of accuracy what the consonant of its transitive suffix
(or suffixes) is likely to be. He suggests that, subject to the
phonotactic constraints which have already been examined, the
majority of bases which can be followed by the same suffix have
common semantic characteristics.
As noted earlier, this is a view which Hazlewood had
already hinted at and which, in the case at least of verbs of
motion (without defining motion more precisely for the moment),
is relatively easy to substantiate. What Arms posits, however,
is that each of the consonants which occur in transitive suffixes
is associated with one or more semantic notions or connotations.
Thus, for instance:
-c_- is associated with 'pliancy, gentle contact, bodily
experience' (Arms 1975: 1(A)
-k- with 'hardness, force, opening out'.
-m- with 'insertion, going inside', ... 'the idea of
one thing going inside another, whether it be in
order to stay there or to draw it out' {ibid. 107).
-t_- is associated with the use of a limb or instrument,
moderate force, performative' [ibid., 110-12)
-v- has to do with 'motion to, motion for, motion over'.
Difficulties arise, however, when the consonant is zero
(0). There is a large number of bases in this category and at
first Arms considered them to represent a 'spill-over category'.
Later on, he declares, he was able to identify a 'common
denominator': 'mild force, miscellaneous': 'Thus the 0 ending
is very common with verbs of rubbing, tapping, folding, plucking,
taking off, separating'. It also embraces verbs for 'finding and
buying' (ibid., 113).
There are also complications with and -n-. The
former, in particular, (ibid., 105-6) has 'no convincing semantic
correlation', but appears to have exclusively grammatical
functions, like -n-, which often has the function of forming
verbs from nouns (ibid., 107-8), a point already made by Hazlewood
(1872: 32-3).
66


Now, in most cases, Arms has no great difficulty in
providing plausible, if not invariably convincing, lists of
examples in support of his view. In each case he only gives ten
examples, and the more the examples that are produced, the greater
the difficulty of finding a common semantic denominator. This
weakens his argument and, at least arguably, it weakens it
unne c es s arily.
There are, in fact, two main difficulties. The first
is that Arms seems to be under the impression that covert semantic
connotations attach to the actual choice of certain sounds
consonants in this case much as they do in most languages,
including English.22 Yet cases like the suffixes with -g- and
-n-, which point to grammatical rather than semantic functions,
as well as the large number of bases with -0- consonant which do
not have either a clear or an obvious common semantic denominator,
should have alerted him to the possibility that the genuine
semantic burden of verbal suffixes rests, not on their phonetic
character, conferring on the preceding verb the membership badge,
as it were, of a covert semantic category, but on a complex of
grammatical relations which remain to be investigated.
The second difficulty is this: the phonotactic
constraints which Arms discusses militate in many cases against
the occurrence of a particular suffix when semantic considerations
would seem to require it. Although he does consider such cases
(for instance, Arms 1975: 151-*+ 5 esp. note to p. 152), it does not
seem that he has attempted to make a systematic study of what I
shall call replacement suffixes, i.e. those which, for phonotactic
reasons, are substituted for the suffixes which can normally be
expected to occur, and of the effect of those substitutions on
the synchronic system as a whole.
One could even argue that Arms seems to hedge his bets.23
The phonotactic constraints which he has clearly set out are
incontrovertible, but failing a more extensive investigation of
their effect, it is very difficult to accept his thesis as to the
correlation between individual suffixes and specific semantic
notions. He might have chosen to sacrifice the latter to the
former but, in actual fact, he appears to have spoilt his case
by emphasizing the wrong argument.
Stated briefly, one could present the dilemma as
follows: On the one hand, (a) the pattern of verbal suffixes in
modern Fijian could be the result of interaction between
diachronic phonology and synchronic syntactic and/or semantic
constraints. On the other hand, (b) it could represent the
effect of diachronic semantic factors which are inhibited by
synchronic phonotactic constraints.
I should, therefore, like to propose a different
approach to these problems. In view of their complexity, however,
67


one cannot hope to do more than to suggest lines of inquiry which
seem to be more promising than others and to try to adumbrate a
possible solution.
Let me then proceed from known and generally-accepted
facts and examine the general distribution of monosyllabic verbal
suffixes. I shall attempt to establish, first, what grammatical
functions can be determined for a given suffix, and secondly,
what effect phonotactic constraints have on the occurrence of that
suffix, both when the constraints are present and absent.
On Anns' evidence (19T5126) -t_- and -0- are statist-
ically by far the most commonly occurring suffixes. Together
they account for 569 recorded endings from his total field of
I68O... Not only does it seem unlikely that any two particular
'meanings' (i.e. semantic associations or connotations) would so
greatly predominate over the rest, but those are evidently also
the two suffixes to which Arms was hardest put to attach any
particular 'meaning' (cf. ibid. 110-12 for -t-; H3-k for -0_-).
Any attempt, it would appear, to find a common semantic
denominator between all the verbs that take a verbal suffix in
-t_-, or between all those that take a verbal suffix in -0-, is
likely to end inconclusively. If we are looking for a common
'meaning', it will not be a property of the suffix alone, but of
the interplay of syntactic variables within the verb phrase, in
which suffixes play a vital but not an exclusive role. We must,
therefore, look elsewhere and we are given valuable guidance by
two widely-accepted observations of Hazlewood (1872: 33), namely,
that:
1. -n- is a 'denominal' suffix, i.e. it has the function of
providing a method of forming verbs derived from nouns.
2. -v- is associated with verbs of motion; without defining this
class more precisely for the moment.
I have argued elsewhere (Milner I98O: 1-k) that the
slow development of Austronesian studies during the last 100
years is to some extent due to the geographical fragmentation of
the work and also to the intellectual isolation of the scholars
concerned,which can be ascribed to relative lack of communication
and in general to relative ignorance of one another's problems
and progress. There has also been a noticeable lack of comprehen-
sive studies of individual languages as well as too great a
concentration of effort on comparative studies, particularly on
topics such as subgrouping and putative chronology at the
expense, if not the exclusion of detailed description. For the
greater part of the twentieth century, students of Austronesian
languages, while paying lip service to their common origin and
striving to make sense of an extensive common stock of words,
have neglected comparative grammar. It Is only in the last
68


decade, with the organization of international conferences on
Austronesian linguistics, that the syntactic features of
languages as diverse as those of Taiwan, the Philippines, and
Madagascar, have begun to throw light on the solution of problems
that have long baffled students of Oceanic and Indonesian
languages (Dahl 1978; Naylor 1978).
As a case in point, it would appear that the focus and
topic approach to the understanding of Fijian syntax (Naylor
1978) is likely to help us make significant progress. Let us
then examine, if only provisionally and in order to discover if
one can establish prima facie evidence, the hypothesis that Fijian
too has a system of focus marked by verbal affixes.
I propose to use the term 'focus1 in the sense that is
widely, though by no means unanimously, accepted in Philippine
linguistics, i.e. 'the syntactic relationship between the verb
and the surface subject, signalled by the verb's focus affix in
conjunction with the subject form of noun phrases and pronouns.
For example, a sentence is in instrumental focus if the surface
subject is in the role of instrument and the verb has an
instrumental affix; the verb "focuses" on the subject as
instrument' (Naylor 1975: 12-3).
On this hypothesis, by reason of their frequency of
occurrence alone, the two suffixes -t_- and -0- should be examined
afresh in order to establish whether they represent the Fijian
equivalent of what has been identified elsewhere, particularly
in Taiwan and Philippine languages, as goal focus affixes.
A few years ago Dahl (1978) suggested that four types
of focus were perhaps Pan-Austronesian in their distribution,
namely: actor focus, goal focus, referent focus (the person in
whose interest the action is carried out or the place where the
action is performed) and instrument focus which he characterized
as follows: 'The fourth focus, generally called instrument focus
(IF) got its name because it focuses something for performing
the action, for instance an instrument' (Dahl 1978: 384).
In elaborating his interpretation, Dahl (ibid., 385-6)
goes on to explain that one of the separate functions of the
fourth focus has to do with the displacement of a moving object,
either away from the actor (as in Minahasan languages) or in any
direction (within the actor, towards him, or away from him (as
in Malagasy).
In her contribution to the same volume, Naylor makes a
similar point with reference to Tagalog':
What appears to be at play here is not a contrast
between transitive and intransitive, rather it is
whether the action is viewed as centrifugal or
69


centripetal. Like aspect, however, the contrast
between centrifugal and centripetal is situational as
well as a matter of perspective. When the action is
viewed as going outward from the actor and ends outside
of him, then it is centrifugal;... When the action
itself is viewed as beginning and ending with the actor
himself then it is centripetal. (Naylor 1978: 1+05)
One of her pairs of examples is suggestive from the
point of view of Fijian. She mentions two Tagalog verbs in actor
focus but with different affixes: magbili 'sell1 (centrifugal) as
opposed to bumili 'buy' (centripetal). Both are formed on the
base -bili. There is a similar situation in Fijian where a
similar pair is formed from the cognate base voli, namely, volia
'buy' and volitaka 'sell' (cf. veivoli 'buy and sell, market'
(perhaps also 'exchange, barter' in a pre-contact economy).
Earlier in the same article, Naylor (1978: 1+00-01)
identifies, in the case of Tagalog, four types of focus (actor,
goal, locative and instrumental) {ibid. 396) and six kinds of
role (actor, goal, locative, comitative, benefactive and
instrumental). 25
It will be evident from the views quoted from Naylor
and Dahl that there is, as yet, no consensus among the- scholars
interested in this approach, not only as to the exact nature of
the syntactic relations subsumed by focus and topic but also as
to the number to be distinguished and identified and the
technical terms to be used to describe them. Nevertheless, there
is abundant evidence that a rich and promising area of research
lies before us in Austronesian studies (e.g. Dahl 1981; Ferrell
and Stanley 1980-; Lopez 1978; Naylor 1980).
It seems, therefore, that a good case can be made for
a new approach to the problem of verbal suffixes in Fijian. Thus
what Dahl calls the 'moving object focus' clearly has an equiv-
alent marked by disyllabic suffixes such as -vaka and -taka (as
in cicivaka 'run with something', or viritaka 'throw something
(at a target)') but this moving object focus (which might be
termed 'locomotive') will have to be defined rigorously with
special reference to what has also been called 'comitative,
benefactive and instrumental'. Likewise the -ra suffix shows
evidence of being associated with a locative focus. Within the
limited scope of the present article, however, one can hardly do
more than point to the complexity of the problems and to the
direction in which progress is likely to be made.
Let me first make a point of theory and consider for a
moment the phonotactic constraints which restrict the occurrence
of any one verbal suffix with any one verbal base. A thorough-
going attempt to establish beyond doubt that Fijian does indeed
have a topic and focus system will have to distinguish carefully
70


what might "be called the 'canonical' suffixes from 'adventitious'
or 'intrusive' suffixes, that is, those which are imposed by
phonotactic constraints. For instance, it was Hazlewood who
first stated that while verbs of motion take the suffix -va, so
do many others. It follows that, before we can establish a firm
correlation between any one suffix and any category of verbs,
two factors must be taken into consideration:
(a) Assuming that the suffix -va can, under certain circumstances,
mark a type of focus which we might call 'displacive' or
'locomotive', it cannot appear whenever a verb of motion includes
a bilabial consonant. The suffix, -va will then be replaced by
another as the following examples show:
1. -ta instead of -va
cabeta 'go up to'
kabata 'climb up to'
kevuta 'climb down along'
sobuta 'go down along'
volita 'go round sth.'
cumuta 'butt with the head
against'
ribata 'strike against (in
springing back)'
lavota 'score a hit (with
small object), cast
into'.
livata (of lightning)
'flash on sth.'
robota 'extend over,
stretch over'
2. -ca instead of -va
kuvuca 'blow (smoke) against'
vukaca 'fly towards'
yamoca 'grope for sth.'
vuloca 'roll (sennit) over
thigh'
3. -ka instead of -va
virika 'throw(sth.) at'
tebeka (of stone etc.) 'skim
on surface of water,
ricochet'
vidika 'flip (finger etc.)
against sth.'
dromuca 'sink below, go
under sth.'
lomoca 'dip(sth.) into'
mumuca 'swarm towards'
luvuca 'flood over; plunge
under'
dumuka 'raise, lift up (on
end of stick)'
vodoka 'embark on, go
aboard'
butuka 'step on, tread on'
(b) Conversely, where a consonant (other than -v-) is constrained
from occurring as a verbal suffix because a homorganic consonant
occurs in the base, -v- may be adventitious, that is to say, it
71


may "be substituted for a suffix that would otherwise have been
used. Note, for instance, the following cases where -ta or -a_
might have been expected to occur if, that is, we assume that
either of them can signal a goal focus affix:
-va instead of -ta or -a:
talova 'ladle, scoop
(yaqona etc.)'
taqava 'use (two or more
layers etc.)'
todrava '(of sun)burn,
scorch'
dolava 'open'
setiva
dikeva 'study, scrutinize'
tarava
nitiva
'follow in
succession, be
next to'
'slice off (crown
of taro corm)'
though, interestingly, -a does, in fact, occur after tara, but in
a different sense: 'touch etc.'.
At this point, it is worth examining in some detail
what Arms (1975: 106-7) has suggested with regard to the suffix
-ka. He states that 'it is associated with verbs where the
action is by nature a forceful one; verbs of "breaking, squeezing
hard, striking violently" (sometimes involving a missile) are
typical members of this class.'
When Hockett (197*0 reviewed the general problem of
these verbal suffixes in a paper presented at the First
Conference on Comparative Austronesian linguistics in Honolulu,
the occurrence of -ka was one of only two instances where he
concluded that Arms' thesis could be upheld.27 Indeed, if we
look at the semantic distribution of verbs followed by this
suffix, it is difficult at first flush to see how one can arrive
at any other conclusion. A more recent article, however, (Milner
and Nawadra 1981: 186-9*+) shows that of 8l verbs having to do
with 'breaking, splitting, cutting and grating', only 18, i.e.
22.23%, have a suffix in -ka. Of these 18 verbs, 3 also have an
alternant suffix in -a.
The other difficulty is that, in a large number of
instances, -ka occurs in bases that seem to have little to do
with force, violence or disruption. In addition to butuka and
vodoka, the following three verbs represent instances where -ta
or -g might have been expected to occur (if, that is, we assume
that either of them can signal a goal focus affix):
tomika 'pick up'
tevuka 'unfold, open up'
dodoka 'lift up, stretch
out (hand)'
In actual fact, of course, all three bases include a t or d which
72


rules out -ta as a suffix. On the other hand t or d also occur
in bases which do have a violent or disruptive connotation, which
again suggests that phonotactic rather than semantic factors are
relevant:
vidaka 'split, cleave' tunaka Tgut, disembowel'
teveka 'circumcise' muduka 'cut off'
Other verbal suffixes occur much less frequently than those which
have been mentioned so far. They include -ra, -ma, -na, and -ga.
There is a clear association, it would seem, between
the suffix -ra and what may be a locative focus, as the following
examples show:
ciqira 'stick into, slip
into a narrow
place'
tagara 'place, lay (on
top of)'
davora 'lie on'
t.ubura grow on'
tubera 'carry, hold (in the
hand)'
qisora 'poke (with stick
etc.)'
Not infrequently -ra occurs when the base is preceded by the
prefix vaka-:
vakasobura 'put sth. down' vakamocera 'put someone to
sleep'
vakayacora 'carry out, vakadabera 'make someone sit
perform' down'
If r_, or one of the other two consonants subject to the
same phonotactic constraint (i.e. dr or n) occurs in the base,
another consonant must be substituted for an (assumed) r_ in the
suffix. It may be one of the following:
m as in: darama 'slip into'
tanuma 'dip into'
t_ as in: ravita 'lean on'
suruta 'sneeze on'
c_ as in: miraca 'fall gently £ as in: ravoga 'warm (cold
on'
rubeca 'hang sth.
on'
food)
raraga 'heat (banana
leaves) on'
k as in: tonoka 'dab on'
ra.ma.ka 'cast (light) on'
73


The remaining monosyllabic suffixes, namely, -ma, -na,
and -ga, do not occur very frequently. Arms (1975:106)2is
not able to correlate -ga with any special connotation and he
considers it to be similar to -na. In his opinion, the latter
has a grammatical function, that is to say (as Hazlewood had
already suggested), it serves to form verbs from nominal bases.
There is much evidence to support this view as the following
examples show:
baca 'bait' bacana 'bait, entice'
kato 'box' katona 'put into a box'
buka 'fuel, fire' bukana 'add fuel to'
taga 'bag' tagana 'put into a bag'
duva 'plant used as fish duvana 'poison (fish) with
poison' duva'
siga 'day; sun' sigana 'sun; dry in the sun
This is not to say, however, that all bases which can
be followed by the suffix -na are formed from nouns. A relatively
small number of them appear to be verbal bases !in their own
right'. It can hardly be a coincidence that, for most of them,29
an expected suffix in -ta (assuming again that this is the
normal or 'canonical' form of the goal focus affix unless
phonotactic constraint rules it out) does not occur because t_ or
d occurs in the base:
dabana
domona
dagina
tavuna
tokona
'do up in parcels' tukuna
'love, desire' tawana
'bathe (eyes)' tomana
'roast on embers' tuvana
'prop up; stay'
'relate, tell announce'
'occupy, populate'
'accompany; help'
'arrange in order, set
m rows'
Up to this point this article has dealt with the
phonotactic constraints of consonants, and only with those
associated with monosyllabic suffixes. One phenomenon remains
to be mentioned briefly. There is evidence that not only
consonants but vowel quality also is a factor relevant to the
occurrence of the verbal suffixes we have examined.
Of the ten verbs mentioned by Arms (1975:107) which
are followed by a suffix with m for instance, eight end with a
back vowel (u or o_) and two with the open vowel (a.). He also
(ibid., 2Q3-5) gives a fuller list of 35 bases followed by -m,
of which 29 end with a back vowel, 1 with the open vowel and
only 5 with a front vowel. Of the latter, only 2 (silima 'dive
for' and siqema 'suddenly realize') are attested beyond all doubt,
doubt.31
Likewise, a very high proportion of verb bases is
followed by -0a (zero consonant: Arms (ibid., 25^-68) lists 30.
Of those, the vast majority, 261, ends with a front vowel (i_ or
e_) ; only h3 (i.e. approximately ), end with a back vowel,
7*+


and many of those either need to he confirmed as correct, or
admit the possibility of another suffix as an alternant.
In conclusion, I hope to have shown, if only by
implication, first, that comparative studies in Austronesian
languages will rest on a surer foundation (and therefore advance
more rapidly), not only when the quality of the data for compar-
ison, as well as the quantity, is more satisfactory, but also when
more comprehensive studies of carefully selected languages are
available. At present, much of our data can scarcely be said to
be more abundant or more reliable than it was in Dempwolff's
time. Recent studies (Geraghty 1983; Geraghty & Pawley 198l)
show the wealth of hitherto unpublished and hitherto unknown
evidence from a relatively well-known language area like Fiji.
Secondly, studies of individual languages will
increasingly be assessed by the criterion of the extent to which
the author shows that he is, if not familiar at first hand, at
least aware of the whole field of Austronesian grammar and of
recent progress made in areas other than his own and, moreover,
that he has considered its relevance to his own work. In
particular, serious consideration should now be given to the
question of establishing whether the topic and focus approach
to Austronesian syntax is relevant to the understanding of
Oceanic languages.
Two other contributions which are especially relevant to
the present article have been made since it was written (Schutz
1985; Milner 1986).
NOTES
1. A contribution to the solution of the old and refractory
problem of the Fijian verbal suffixes from a phonological
and prosodic point of view seems to be appropriate in this
collection of articles. I should also like to express my
thanks to a number of colleagues who have generously
commented on.,., and suggested improvements to, the original
draft of this article; especially to Professor Bruce G.
Biggs of the University of Auckland, Professor Otto C. Dahl
of the University of Oslo, Professor Charles F. Hockett of
Cornell University, Professor Albert J. Schutz of the
University of Hawaii, and Dr Paul Geraghty of the Fijian
Dictionary Project in Suva, Fiji. My thanks are also due
to the Rev. Dr David G. Arms of the Columban Fathers,
Professor Robert Blust of the University of Hawaii,
Professor Jack Carnochan of the University of London,
Professor Viktor Krupa of the University of Bratislava, and
75


Professor Paz B. Naylor of the University of Michigan.
2. For a detailed analysis of the relationship "between 'word',
'morpheme' and 'base' in Fijian, see Schutz 1975-
3. As, for example,, in uca 'rain' or ba (CW) 'fence'. (I am
aware that in recent years the one-vowel-per-syllable
analysis of Fijian phonology has been criticized, notably
by Schutz and Biggs, though for reasons that will not be
discussed here I remain unconvinced.) There are also rare
instances of bases where both consonants are zero, as in
ia 'proceed, take place' and ua-(ca) 'beat with a stick
(laundry, et c.)'.
4. Only one instance is known to me of a trisyllabic base where
the initial consonant is zero: Paul Geraghty has written to
me that uea exists as a verb 'to fish trap'. This must be
the base Hazlewood gives as wea 'a fish trap' and it could
take its name from the island of Uvea (Wallis), a Polynesian-
speaking community under French administration. The only
other case I know of a trisyllabic base where both the
medial and the final consonants are zero is biau 'wave'
which is almost certainly a Polynesian loan (peau). Other
instances, however, cannot be ruled out.
5. This statement must be qualified by adding that these
suffixes are also found in combination with other affixes,
in which case they may not be 'transitive'. This is not
strictly relevant to the problem considered here.
6. The consonant in parentheses can either be zero, or
following the open vowel a_ or the back vowels o_ and u
the palatal approximant written y. In disyllabic suffixes,
the second consonant is always -k- (cf. n.l6 below).
7. Andy Pawley (1978:120 (also nn.17,179); 135, 136-9, esp.137)
has put forward the view, which others have accepted, that
from a comparative point of view the verbal suffixes in -Ca
and -Caka of standard Fijian are 'irregular' or at least
untypical of Fijian dialects in general, and that they
represent a conflation of two vowels (i.e. *Ci-a to -Ca and
*Caki-a to -Caka). David Arms (1975:28, 31-4) seems to have
come to the same conclusion at about the same time and
independently of Pawley. There is much evidence to support
this view but, in order to avoid confusion with the so-
called 'passive' suffixes in -Ci, I prefer not to quote
examples in the 'canonical' *Ci form as Pawley and others
do, but to give them with an 'active' -Ca suffix. Regarding
Fijian spelling, see n.l4.
8. This view is often associated with the phrase 'thematic
consonant'. Charles Hockett, in a letter about his article
76


(1976, 1977) adds the following comments:
The behaviour of the Fijian thematic consonants was
one of the real facts about languages that led me
slowly but surely to abandon what I now refer to as
the 'atomic morpheme theory'', the theory of grammatico-
lexical structure I helped develop in the 19U0's and
to which I clung for a long time. That theory
proposes that every phonemically relevant piece in any
utterance must be a part of one or another morpheme
(or of the phonemic representation of one or another
morpheme), and that morphemes are minimum meaningful
elements in much the same sense in which we all assumed
phonemes were minimum meaningless but differentiating
elements. By that theory there would be only three
possibilities for Fijian: (l) rai-ca, as with
Churchward, so that the suffix has ten different
alternants; (2) raic-a, as proposed by Bloomfield
for Samoan, so that the stem has two different
alternants (as do most verb stems); (3) rai-c-a, the
thematic consonant being a separate morpheme.
9. Cf..ibid., 67-8; 89-90; 105-6. This work was originally
published in 1956.
10. See, especially, the references to Arms, Geraghty, Hockett,
Naylor, Pawley, and Schutz.
11. Thanks to the sponsorship, first of Mr Raymond Burr, a well-
known American television actor, and later of the Australian
Government Cultural Fund as well as the support of the
Government of Fiji, a monolingual dictionary of the Fijian
language is being compiled and is now approaching completion.
The Director of this Project was, until early 1986, Mr T.R.
Nawadra; Dr Paul Geraghty currently holds the post.
12. 'Besonders haufig sind Lautunstimmigkeiten bei Auslauten
des Fidji vor stutzendem Suffix...'
'Aus diesen Tatsachen wird hier gefolgert, dass diese
Lautunstimmigkeiten des Fidji als "irrige" Analogie zu
deuten sind;...'
13. Earlier studies include a monograph by Kern (.1886). Scott's
data on Fijian were mainly derived from his study of the
pronunciation of Josua Bogidrau, a Fijian civil servant who
had been seconded to the School of Oriental and African
Studies, London University (l9*+6-H8), and who also helped
me to learn his language at first hand.
Ik. The orthography chosen for Fijian by the first two
missionaries, David Cargill and William Cross, in the 1830s,
77


was remarkably advanced for its period and, in particular,
almost anticipated phonemic theory, at least by implication,
by about 75 years. Thus, with one exception (dr), each
consonant phoneme is always represented by one, and only
one, letter, despite the fact that Roman conventions
(supplemented by the conventions of English orthography)
require that digraphs should be used. So, there are three
voiced stops, each preceded, at least in non-initial
position, by a non-phonemic homorganic nasal: /mb/ (written
b) /nd/ (written <), and /qg/ (written g). The voiced
interdental fricative /5/ is written c_ (instead of th) and
the velar nasal is written (instead of ng). A fourth
nasalized voiced 'stop1 is written dr_ (actually /ndr/) .
Apart from using diacritics in a somewhat arbitrary
and unpredictable manner, however, the orthography of Fijian
does not take vowel length into account. Since the latter
is phonemic, it is a serious defect in an otherwise elegant
system. (Native speakers seldom use or require diacritics
since to them the exact pronunciation is usually clear from
the context. )
I am indebted to Professor Otto C. Dahl for drawing my
attention to his article on the origins of Malagasy
spelling (Dahl 1966). This shows that, although Fijian
orthography was much in advance of its time in its economy
and its disregard for non-significant sounds, it was neither
entirely original, nor an isolated attempt to devise an
alphabet based (l) on a one-for-one equivalence between
letters and phonetic values and (2) on internal consistency
without obligatory regard to the spelling conventions of
English, or for that matter, of French orthography.
In actual fact, in the early l820s, the same principles
had been consciously observed by the three Welsh-speaking
pioneers of the London Missionary Society in Madagascar:
David Jones, Thomas Bevan, and David Griffiths, who devised
the first system of Roman orthography for Malagasy. Their
training in England, at a theological academy in Gosport
presided over by Dr David Bogue, included a linguistic
component which owed much to the well-known grammarian
Lindley Murray (17^5-1926). The latter had stated in his
English grammar that:
a perfect alphabet... would contain a number of
letters, precisely equal to the number of simple
articulate sounds belonging to the language. Every
simple sound would have its distinct character; and
that character be the representation of no other
sound. (Murray 1813:15)
78


Murray in his turn was directly indebted to Samuel
Johnson's short 'Grammar of the English tongue' which
precedes his Dictionary of the English language. In
that short essay, Dr Johnson ends his remarks on ortho-
graphy and pronunciation with references to various
attempts made in the past 'to accommodate orthography
better to the pronunciation' and he notes that some
reformers
have endeavoured to proportion the number of letters
to that of sounds, that every sound may have its own
character, and every character a single sound. Such
would be the orthography of a new language to be
formed by a synod of grammarians upon principles of
science. (Johnson 1828:33)
It is interesting to note also that, just as Fijian
spelling (in accordance with the ideal system for a
previously unwritten language recommended by Johnson and
Murray) uses c_, £ and c^, for example, without regard to the
conventions of English spelling, the three Welsh-speaking
pioneers in Madagascar proposed initially to use c_ for an
affricate /ts/, for the velar nasal, and most interestingly
of all, w for the close back vowel as in Welsh, instead of
oo as in English or ou as in French. Regrettably, however,
this imaginative proposal was abandoned in the face of
opposition from other Europeans Whose first language was
English or French and not Welsh.
This resistance in Madagascar has parallels in the
Pacific. Thus, because it is an unconventional alphabet
from a purely Western point of view, Fijian orthography
has long been the target of well-meaning but uninformed
criticism (often aggravated by patronizing ridicule) in
English-speaking circles. At one point during the Colonial
period, in the late 1930s, the desire to 'reform' Fijian
spelling even led to a debate in the Legislative Council of
Fiji (Schutz 1972: lh ff., esp. 20-2). There is not much
evidence, however, that the Fijian people have ever wanted
to introduce spelling changes, though, undoubtedly, some
Fijians are irritated when they hear the names of people
and places mispronounced by ignorant expatriates or over-
seas news-readers.
Aims (19-75:130-1) acknowledges an article (Krupa 1966)
which had appeared eight years previously and mentions
associative and dissociative tendencies between groups of
consonants in Oceanic languages, including Fijian, according
to their place, or their mode, of articulation. He states,
however, that Krupa was concerned with 'groupings of
consonants according to their place of articulation or
after their mode of articulation, not to the associative
79


and dissociative tendencies of individual consonants the
item of particular interest here' {ibid., 131). Albert
Schutz informs me in a letter that Arms' dissertation was
the first full analysis of these phenomena to be published,
but Bruce Biggs was already discussing consonant restrictions
in the early 1960s although he did not publish his findings.
Moreover, Paul Geraghty's (1973) unpublished term paper on
this subject had been heard by 1973, while Peter Lincoln
had also studied the same problem. I am now indebted to
Paul Geraghty for sending me a copy of the term paper in
question. He reminds me that the problem is also discussed
in his doctoral dissertation (now published as Geraghty
1983. See esp. 260-70).
16. The second consonant of a disyllabic suffix is always -k-:
-caka, -kaka, -laka, -rnaka, -naka, -raka, -taka, -vaka,
-yaka. Unlike the first consonant, it is not subject to
any constraints of occurrence. (A comprehensive analysis
of the distribution and function of disyllabic suffixes has
not been possible within the scope of the present article.)
17. In addition to constraints governed by the place of
articulation of consonants, Arms also points to one or two
cases where the constraint seems to be linked with the mode
of articulation. Thus, the occurrence of a velar nasal g_
in the initial or medial consonant of the base, rules out
the-nasal suffix -ma: e.g. gunuva 'drink', whereas the
reconstructed PAN form *inum (as well as Polynesian
reflexes such as Samoan inumia, not to mention Fijian
dialectal variants) would have made one expect *gunuma
(Arms 1975 : 153); see also *ceguma. The occurrence of a
close back vowel before a suffix in -m- is also considered
in the discussion of vowel quality on p.74 above. The
converse is also true: there is no recorded instance of a
base with a bilabial nasal taking the velar nasal suffix
-ga. Curiously, however, the alveolar nasal suffix -na
can occur without restriction following a bilabial or a
velar nasal in the base. The same observations apply to
the labio-velar approximant w-, the occurrence of which in
a base rules out the bilabial (with some exceptions. See
liwava mentioned in n.26 below and the velar nasal suffixes
-ma and -ga, but not the alveolar nasal suffix -na (cf.
Arms 1975).
18. Thus, for instance, one can see at first glance that at
least six of Dempwolff's reconstructed word-stems: *da^ev_,
*giliq, *leqgak, *psdam, *peg'et9 and *taquk have final
consonants which are deemed to be articulated in the same
place as a medial or an initial consonant. This alone
rules out the possibility of finding direct one-to-one
reflexes in the suffixes of the Fijian verb bases which,
in other respects, show regular correspondences with
80


reconstructed PAN forms: rogoca /roqo5a/, qilia /qgilia/,
loqata /loqgata/, moce(ra) /moSera/, vocota /poSota/,
digova /ndiqopa/ respectively (cf. Arms 1975:156).
19. The terms 'associative1, and 'dissociative', with
reference to consonants and vowels that may, or may not,
respectively, occur within the same base, with or without
a suffix, are used by both Krupa and Arms. In a letter,
Professor Biggs points out that prenasalisation did not
occur finally in PAN and that in his view this fact alone,
even without dissociation, is enough to account for the
absence of prenasalisation in the suffixes.
20. In a letter, Paul Geraghty informs me that karona is
probably a modern form of karauna -(cf. qarauna with
similar meaning) and that the restrictions may not be so
strict at a distance of two vowels.
21. This statement is based on a conversation I had with David
Arms some years ago. He was presumably thinking of an
expatriate learner of Fijian like himself who had already
acquired some knowledge of its covert categories. Yet,
when he tried the experiment of making up imaginary bases
(nonce words) and then of asking native speakers to -suggest
appropriate suffixes and 'meanings' for them, he got
replies which sharply contradicted his expectations. (Arms
1973; 1975:l*+7-8).
22. See especially Arms 1975:128-9). He mentions, for instance,
the connotations linking words in English beginning with
si- as in: slick, slip, slime, slide, slouch, slut, etc.
23. See esp. Arms 1975:118, 150, 156-7.
2k. As, for example, when he argues (Arms 1975:122 ff.) that,
for each 'passive' ('spontaneous') prefix, there is a
corresponding, specific and identifiable semantic content,
and what is more, that there is a one-to-one relationship
of identity between the occurrence of some of those
consonants in spontaneous prefixes and their occurrence in
suffixes. Thus: 'The meaning of -c_ was given as "pliancy,
gentle contact, bodily experience". The meaning for ca-
could be regarded as a semantic specialisation: a shift
from bodily experience in general to the particular bodily
experience of sound' {ibid., 122). This view would seem
to require much more supporting evidence than Arms provides
if it is to be accepted.
25. This distinction has to do with the fact that, in Tagalog,
the locative focus has to be analysed in relation to a
number of subcategories: 'locative goal focus, locative
proper, and locative beneficiary (directional or dative)'.
81


The instrumental role is, likewise, subclassified into:
'instrumental goal (portative or displacive), instrumental
proper, and instrumental benefactive'.
26. I am also reminded by A1 Schutz that one aspect of the
complexity of this problem is that most speakers of standard
Fijian use it as a second language and that for them there
is often an element of doubt as to which verbal Suffix is
'correct' or 'appropriate'. Thus in 1953 even a very
senior chief and distinguished Colonial civil servant from
the Lau group of islands (in the south-eastern part of Fiji),
the late Ratu Sir Lala Sukuna, asked his wife Lady Maraia
in my hearing if liwa '(of the wind) blow' was followed by
the suffix -ca (liwaca) or the suffix -va (liwava). She
spoke Standard Fijian as a first language (unlike her
husband) and she immediately replied with assurance:
'liwava'.
27. This paper appeared later as Hockett (1976,1977). In a
recent letter, he adds:
If thematic consonants are separate morphemes they
ought to have determinable meanings. One of Pawley's
students made an assessment ... of semantic
associations of the thematic consonants. I did the
same thing independently, and came out with this.
Of a random set of 500 stems: Of 51 with thematic
consonant k, 28 or 55% denoted breaking, splitting,
or other such forceful operations. For this sort
of meaning no other thematic consonant scored so high.
Of 74 stems in v, 24 denoted motions or positions;
again stems with this meaning but with other thematic
consonants scored much lower. Also,'"I found some
dissimilative tendencies after certain first and
second consonants, certain thematic consonants are
disfavored (this thing having to do with sound, of
course, not sense).
28. He (loc.cit.) identifies a subclass of verbs having to do
with 'opening out, unfolding, extending', which regularly
take the -k- ending. He claims, nevertheless, to see a
semantic connection with the rest on the ground that '"smash-
ing, breaking, cleaving" all involve disintegration
of some entity' I do not accept this view and consider
that it is just as likely that the process of prosodic
constraint is involved here too. Thus, -t_- is ruled out
in tevuka and dodoka.
29. Some verbs with the suffix -n- do not seem to be formed on
nouns and yet the occurrence of this suffix cannot be
accounted for on the ground of prosodic constraint: cuqena
'support', kumuna 'gather, collect', soqona 'assemble'.
82


30. Arms, following Krupa, uses the term 'dissociative';
Hockett speaks of 'dissimitative tendencies', while I have
spoken in this article of 'phonotactic constraints' and of
'prosodic constraint'. The preferential association of
certain suffixes with certain vowels in the "base which is
discussed here is the opposite of a constraint. Hence the
title of the article which attempts to subsume both
phenomena under the term 'prosodic' in the sense first
used by J.R. Firth and the London school of linguists with
which he is associated (cf. Firth 19^-8).
31. In a letter, A1 Schutz writes as follows:
See my article on borrowings (Fiji Museum Publication),
1978 especially on 'natural syllables' ... The point
was that there are phonetic (articulatory, that is)
reasons for certain C + V associations. I was looking
at it from the point of view of the C being fixed and
the V open to choice; your observation ... approaches
the matter from the opposite direction, (cf. Schutz
1978).
REFERENCES
Arms, D.G. 1973. Whence the Fijian transitive
endings? Ocean. Ling. 12, 503-58.
................1975. Transitivity in Standard Fijian
(Ph.D.. Thesis, Univ. Michigan,
Linguistics, 197M Ann Arbor: Univ.
Microfilms Intern.
Churchward, C.M. 19^1. A new Fijian grammar. Sydney:
Australas. Medic. Publishing.
Dahl, O.C. 1966. Les debuts de 1'orthographe
malgache (Avhandl. Norsk. Videnskaps-
Akad. Oslo. II. Hist.-Filos. Klas.,
N.S. 9). Oslo: Universitetsforlaget.
................1973. Proto-Austronesian (Scand. Inst.
Asian Stud. Monogr. Ser. 15). Lund:
Studentlitteratur.
................1978. The fourth focus. In Second
International Conference on Austronesian
Linguistics: Proceedings (eds.) S.A. Wurm
& Lois Carrington, (Pac. Ling. Ser. C,
6l). Canberra: Austral. Nat. Univ., 1,
383-93.
83


Dahl, O.C.
Dempwolff, 0.
1981. Early phonetic and phonemic
changes in Austronesian (Inst. Comp.
Res. Hum. Cult. Oslo Ser. B, 63).
Oslo: Universitetsforlaget.
1934, 1947 & 1938. Vergleichende
Lautlehre des austronesischen
Wortschatzes: 1. Induktiver Aufbau
einer indonesischen Ursprache.
2. Deduktive Anwendung des
Urindonesischen auf austronesische
Einzelsprachen. 3. Austronesisches
Wdrterverzeichnis (Beihefte z. Zeitschr.
f. Eingeb.-Spr., 15, 17 & 19). Berlin:
Dietrich Reimer.
Ferrell, R. &
Stanley, P.
Firth, J.R.
Geraghty, P.
Geraghty, P. &
Pawley, A.
Hazlewood, D.
1980. Austronesian versus Indo-
European: the case against case. In
Austronesian studies: Papers from the
second Eastern Conference on Austronesian
languages (ed.) Paz Buenaventura Naylor
(Mich. Pap. on South & SEast Asia 15,
1979). Ann Arbor: Univ. Michigan, 19-31.
1948. Sounds and prosodies. Trans.
Philol. Soc127-52.
1973. Some aspects of case-marking in
Fijian (Unpubl. term paper, Dept. Ling.,
Univ. Hawaii).
1978. Topics in Fijian language
history (Ph.D. Thesis, Univ. Hawaii,
Linguistics).
1983. The history of the Fijian
languages (Ocean. Ling. Spec. Publ.
19). Honolulu: Univ. Hawaii Press.
1981. The relative chronology of some
innovations in the Fijian languages.
In Studies in Pacific languages and
cultures: in honour of Bruce Biggs
(eds.) Jim Hollyman & Andrew Pawley.
Auckland: Ling. Soc. New Zealand,
159-78.
1872. A Fijian & English and an English
and Fijian dictionary ... and a grammar
of the language ... (ed.) J. Calvert.
London: Sampson Low, Marston.
84


Hockett, C.F.
191k. The reconstruction of Proto-
Fijian-Polynesian. (Paper delivered at
the first international conference on
comparative Austronesian linguistics,
Honolulu.)
1976. The reconstruction of Proto
Central Pacific. Anthrop. Ling. 18,
187-235.
Johnson, S.
Kern, J.C.H.
1977. Proto Central Pacific: Addenda.
Anthrop. Ling. 19, 2k2-k.
1828. A dictionary of the English
language ... (to which are prefixed) A
history of the language and An English
grammar. London: Robinson (New ed. from
1773).
1886. De Fidjitaal vergeleken met hare
verwanten in Indonesie en Polynesie
(Verh. Akad. Wet. Amst. Afd. Letterk.
16), 1-2U2.
Krupa, V.
Lopez, C.
Milner, G.B.
1966. The phonemic structure of bi-
vocalic morphemic forms in Oceanic
languages. J.Polyn. Soc. 75,
^58-97.
1978. A handbook in comparative
Austronesian (ed.) Ernesto Constantino
(Arch. Philip. Langs & Dial, and Philip.
Ling. Circ.: Spec. Monogr. Issue 5).
Quezon City: Univ. Philippines.
1972. Fijian grammar.
Govt. Press.
Suva, Fiji:
1980. On the centrality of Austronesian
syntax. In Austronesian studies: Papers
from the second Eastern Conference on
Austronesian languages (ed.) Paz
Buenaventura Naylor (Mich. Pap. on South
and SEast Asia 15, 1979). Ann Arbor:
Univ. Michigan, 1-17.
85


Milner, G.B. 1986. A focal approach to problems of
verbal syntax in Fijian. In FOCAL I:
papers from the Fourth International
Conference on Austronesian Linguistics
(eds.) Lois Carrington & S.A. Wurm
(Pac. Ling. Ser. C, 93). Canberra:
Austral. Nat. Univ., 1-20.
Milner, G.B. &
Nawadra, T.R.
Murray, Lindley
1981. Cutting words in Fijian. In
Studies in Pacific languages and
cultures: in honour of Bruce Biggs
(eds.) Jim Hollyman & Andrew Pawley.
Auckland: Ling. Soc. New Zealand, 179-95.
1813. English grammar.
& Sons (2Uth ed.)
York: Wilson
Naylor, P.B.
Pawley, A.
Schutz, A.J.
1975. Topic, focus, and emphasis in
the Tagalog verbal clause. Ocean.
Ling. 14, 12-79.
1978. Toward focus in Austronesian.
In Second International Conference on
Austronesian linguistics: Proceedings
(eds.) S.A. Wurm & Lois Carrington
(Pac. Ling. Ser. C, 6l). Canberra:
Austral. Nat. Univ., 1, 395-^2.
1980. Linking, relation-marking, and
Tagalog syntax. In Austronesian
studies: Papers from the second Eastern
Conference on Austronesian Languages
(ed.) Paz Buenaventura Naylor (Mich.
Pap. on South and SEast Asia 15,
1979). Ann Arbor: Univ. Michigan,
33-50.
1978. Some problems in Proto-Oceanic
grammar. Ocean. Ling. 12, 103-88.
1972. The languages of Fiji. Oxford:
Clarendon.
1975. At a loss for words: the problem
of word classes in Fijian. Ocean.
Ling. 14, 100-18.
1978. English loanwords in Fijian.
In Fijian language studies: borrowing
and pidginization (ed.) A.J. Schutz
{Bull. FijiMus. 4), 1-50.
86


Schutz, A.J. 1981. Specification as a grammatical
category in Fijian. In Studies in
Pacific languages and cultures: in
honour of Bruce Biggs (eds.) Jim
Hollyman & Andrew Pawley. Auckland:
Ling. Soc. New Zealand, 197-207.
................1985. The Fijian language. Honolulu:
Univ. Hawaii Press.
Scott, N.C. 19^8. A study in the phonetics of
Fijian. Bull. Sch. Orient. Afric.
Stud. 12, 737-52.
87




Full Text

PAGE 1

Essays in honour of Eugenie J.A. Henderson Edited by J.H.C.S. Davidson [Bill] School of Oriental and African Studies 1989

PAGE 3

Collected Papers in Oriental and African Studies SOUTH-EAST ASIAN LINGUISTICS: Essays in honour of Eugenie J.A. Henderson Edited by Jeremy H.C.S. Davidson Lecturer in Vietnamese School of Oriental and African Studies SCHOOL OF ORIENTAL AND AFRICAN STUDIES UNIVERSITY OF LONDON 1989

PAGE 4

Published by the School of Oriental and African Studies (University of London), Thornhaugh Street, Russell Square, London WClH OXG of Oriental and African Studies, 1989 British Library Cataloguing in Publication Data South-east Asian linguistics: essays in honour of Eugenie J.A. Henderson. -(Collected papers in Oriental and African Studies) 1. Asian languages South-east I. Davidson, Jeremy H.C.S. (Jeremy Hugh Chauncy Shane) II. Henderson, Eugenie J.A. (Eugenie Jane Andrina, 1914-) 495 Printed in the United Kingdom by Hobbs the Printers of Southampton (2262/89)

PAGE 5

CONTENTS CONTRIBUTORS PREFACE EDITOR Is NOTE EUGENIE J.A. HENDERSON: A BIOGRAPHICAL NOTE R.H. Robins PUBLICATIONS OF EUGENIE J.A. HENDERSON Page vii X 1 Helen Cordell 5 FROM "FISHES' EYES" TO "ANKLE BONES" --A VIETNAMESE CALQUE? -Jeremy H.C.S. Davidson 11 SOME FEATURES OF MODERN KHMER LITERARY STYLE -Judith M. Jacob 23 KHASI KINSHIP TERMINOLOGY tLili Rabel-Heyman ON PROSODIC RELATIONS BETWEEN FIJIAN BASES AND VERBAL SUFFIXES -G. B. Milner A SIAMESE LETTER DATED 7 DECEMBER 1776 Egerod TAI NAMES FOR THE OX William J. Gedney FIRST AND LAST IN THAI, OR THE ORDER OF OPPOSITIONS 43 59 89 111 tMary R. Haas 129 LA TONOLOGIE DU LI DE HAINAN -A.G. Haudricourt 133 PROTO-TAI *kh and -Li Fang-Kuei 143 UNCLES AND AUNTS: BURMESE KINSHIP AND GENDER David Bradley THE BULGING MONOSYLLABLE, OR THE MORA THE MERRIER: ECHO-VOWEL ADVERBIALIZATION IN LAHU 147 James A. Matisoff 163

PAGE 6

THE YAW DIALECT OF BURMESE -John Okell ORAL VOWELS AND NASALIZED VOWELS IN LEPCHA (RONG) AS THE KEY TO A PUZZLING VARIATION IN SPELLING Page 199 R.K. Sprigg 219 Frontispiece: Eugenie J.A. Henderson

PAGE 7

CONTRIBUTORS R.H. Robins, Professor Emeritus of General Linguistics, School of Oriental and African Studies, University of London Helen Cordell, Sub-Librarian, South East Asia and Pacific, School of Oriental and African Studies Jeremy H.C.S. Davidson, Lecturer in Vietnamese, School of Oriental and African Studies, University of London Judith M. Jacob (retired), formerly Senior Lecturer in Cambodian, School of Oriental and African Studies, University of London tLili Rabel-Heymann, one-time Associate Professor of Linguistics, University of Calgary, Alberta, G.B. Milner, Professor Emeritus of Austronesian Studies, School of Oriental and African Studies, University of London Egerod, Professor and Director, Scandinavian Institute of Asian Studies, Copenhagen, Denmark William J. Gedney, Professor Emeritus of Linguistics, University of Michigan, Ann Arbor, USA t Mary R. Haas, Professor Emeritus of Linguistics, University of California, Berkeley, USA A.G.,Haudricourt, Directeur de Recherches honoraire au CNRS, (Centre National de la Recherche Scientifique) Paris Li Fang-Kuei, Emeritus Professor of Asian Languages, University of Hawaii at Manoa, Honolulu, Hawaii, USA David Bradley, Lecturer in Linguistics, La Trobe University, Victoria, Australia James A. Matisoff, Professor of Linguistics, University of California, Berkeley, USA John Okell, Lecturer in Burmese, School of Oriental and African Studies, University of London R.K. Sprigg (retired), formerly Reader in Phonetics, School of Oriental and African Studies, University of London vii

PAGE 9

PREFACE Eugenie Henderson is, as we all know, a most impressive phonetician and linguist. Not only that, but, as is obvious from what follows in this collection of papers, she is also a most informative and charming woman. Several years ago, I had the idea of offering this token of awareness to her, and the response was, as can readily be seen, high-powered. I have, therefore, divided the contributions alphabetically into language families (from Austro-Asiatic through Austronesian and Tai to Tibeto-Burman) and within these sections, once again alphabetically, by the surnames of the contributors, in an attempt to express the range and eq_uality of interest that Professor Henderson has in South-East Asian languages. There is no need for me as editor to discuss the topics presented; their authors need no introduction and will, without doubt, attract readers; the papers speak for themselves. With that comment, I can only say that I, among many others, look forward eagerly to Eugenie1s work on Bwe Karen and all else. ix Jeremy H.C.S. DAVIDSON S.O.A.S. 1989

PAGE 10

EDITOR'S NOTE Certain technical details found in this book may clarification. Romanizations: Languages which are not normally written in roman alphabets are romanized. The systems used are as follows: Burmese is a transcription of the spoken language, not a transliteration of the script, and (where possible) follows the system advanced by John Okell, A guide to the romanization of Burmese (London: Luzac 1971, 66-7); Chinese is romanized into p'iny'in, unless indicated otherwise for the sake of clarity in particular articles; while Khmer is written in accordance with the transliteration used by Saveros Lewitz, 1969, BuU. Ea. fr. Extr.-Orient 55, 163-9. Thai follows the phonet.ic systems described in the next section. Variations may occur; naturally, names of places and people may differ. Transcriptions: Thai specialists who have contributed papers may have evolved their own phonetic transcriptions (e.g. Egerod, Gedney, Haas, Li) which vary somewhat from one to another. Since all Thai specialists will be familiar with these representations, I have left them as they are rather than choose a 'standard' form. Bold face was not available for a wide range of IPA symbols for type-setting in this volume. These are set in roman and, for ease of reading, are underlined when they occur in the running text. Translations: Unless otherwise indicated, all transla tions from languages present.ed are by the authors of the papers in which they appear. Translations of entries do not always follow their originals directly but may follow and 'reproduce' a group, for the sake of linguistic connection. All Chinese or nom characters are wri t.ten by Jeremy Davidson. * * * * I should like to thank all those who have helped to make the task of editing the fifteen contributions easier. I also wish to thank the Publications Committee of S.O.A.S. (especially its three consecutive Chairmen, Professors J.C. Wright, R.H. Robins, and C. Shackle, and it.s Secretary Mr Martin Daly) for their help and support. Finally, I owe a particular debt of gratitude to the Editorial Secretary Miss Diana Matias and to the X

PAGE 11

s.taff of the S.O.A.S. Support Section, especially Mrs Joyce Hutchinson, and to the photographer Mr Paul Fox for the production of the present volume. Jeremy H.C.S. DAVIDSON The photograph on the cover is by Hans Hinz and is reproduced by permission from Thai Painting published by Office du Livre, Fribourg. xi

PAGE 13

EUGENIE J.A. HENDERSON: A BIOGRAPHICAL NOTE R.H. Robins This introduction is not the place for a biography or even a full-length appreciation of Eugenie Henderson's scholarly career. Although she formally retired on 3 September 1982, from her post in the School Oriental and African Studies and in the University of London as Professor of Phonetics, no one of her many friends and colleagues, nor she herself, surely, regards her retirement from teaching as the final conclusion of her active participation in linguistic studies and research. But the contributions from the international world of scholarship to this volume of studies, like those in the volume published in 1979 by Chulalongkorn University (Thongkum 1979), testify to the esteem and affection in which she is held by the now large community of her colleagues, former students, and friends. Eugenie Henderson's involvement in the School of Oriental and African Studies began in 1942, when she was appointed to a lectureship in phonetics. This appointment, like those of a number of her contemporaries, was part of the immediate response by the armed services to the requirements created by the Japanese entry into the Second World War. The Department of Phonetics and Linguistics, under the headship of Professor J.R. Firth, was almost wholly given over between 1942 and 1945 to Japanese language courses for Royal Air Force and Royal Navy personnel, including some highly specialized courses in Japanese phonetics for specific intelligence purposes. Eugenie Henderson's part in the development, administration and teaching of these courses was probably second only to that of Firth himself.l Some of its continuing framework of teaching in the Department had its origin in the forced response of its staff to wartime conditions (cf. Firth 1945), and the impetus behind its expansion was maintained after the war by the University of London's implementation of the Scarbrough Report (1947), which recommended a strong development of Oriental and African studies --linguistic, cultural and historical --a development which, naturally, was concentrated in the School dedicated to such studies. The post-war expansion of linguistic work in the Department of Phonetics and Linguistics covered languages from several different areas the Far East, the Middle East, and Africa. While still a postgraduate student under Professor Daniel Jones at University College London in 1937, Eugenie Henderson had begun the study of Thai, which laid the foundation l

PAGE 14

of her lifelong connection with the languages, peoples and cultures of South East-Asia. Her studies in Thai were reinforced after 1945 with studies, mainly through informants in London, in other South East Asian languages, including Vietnamese, Cambodian, and Mon. In 1946 the Department of South-East Asia and the Islands was re-established, and significantly enlarged soon afterwards, with lectureships in Thai, Cambodian, Vietnamese, and Mon; during their first years of appointment, the lecturers appointed to these posts were attached to the Department of Phonetics and Linguistics and came directly under the supervision of Eugenie Henderson. The 'pro-seminar' that she organized as a forum for the presentation and discussion of their research became a model for the later postgraduate seminar of the Department. In 1953 she was appointed Reader and, in 1964, Professor of Phonetics in the University of London. Eugenie Henderson1s involvement in 5outh-East Asian languages was further strengthened by her service, after the retirement of Professor D.G.E. Hall, as Acting Head of the Department of South East Asia and the Islands, from 1960 until 1966, during a crucial period in the consolidation of that Department. At the same time she continued to play a full part as a senior member of the Department of Phonetics and Linguistics and, from 1966 until 1970, after the retirement of Professor N.C. Scott, she was Head of this Department. Throughout the years 1945 to 1981 Eugenie Henderson involved herself in the subject of general linguistics as a whole. Since 1954 she has been a member of the Council of the Philological Society, serving as the Society's Treasurer from 1965 to 1974, and from 1977 to 1980 she was Chairman of the Linguistics Association of Great Britain. Though her studies of the languages of South-East Asia were mainly conducted in this country, she has made two fairly extensive visits to that part of the world. In 1954 she visited Burma as a Visiting Professor of Rangoon University and engaged in fieldwork on Bwe Karen and Chin, this latter research resulting in the publication of her book Tiddim Chin (Henderson 1965a), while in 1975 she taught in Thailand, in a summer institute organized by the Thailand Research Project in Bangkok. Eugenie Henderson1s publications have been mainly concerned with the phonetics and phonology of South-East Asian languages, in some cases breaking entirely new ground and also applying the theory of prosodic phonology, developed by Firth during the 1940s and early 1950s, to different language material. Her 1948 'Prosodies in Siamese: a study in synthesis' (see Henderson l970a) has for long been recognized as one of the most 2

PAGE 15

thorough and revealing studies in prosodic analysis that we have, and is regularly recommended to students embarking on this aspect of phonological theory, especially since its republication in 1970 in Palmer's Prosodic anaZysis. A glance at her bibliography, however, shows that South-East Asian languages by no means exhaust her fields of interest. She has contributed papers on typological and historical topics at international congresses (e.g. Henderson l965b, Indo-Paaifia Zinguistia studies Pt.II), and has published phonetic studies of Ossetic (e.g. Henderson 1949) and of other Caucasian languages. In this latter area, her article 'Acoustic features of certain consonants and consonant clusters in Kabardian1 (Henderson l970b), though quite brief, has become something of a classic in the literature of experimental phonetics. British phoneticians look back with pride on the great nineteenth-century phonetician and Anglist Henry Sweet, and Firth himself often claimed that his prosodic theory was, in part, a development of ideas latent in Sweet's own work. Sweet's writings were voluminous, and in 1971 Eugenie Henderson performed a valuable service to anyone concerned with the history of phonetics and phonology by selecting and making available the most important passages from his books and articles in a single book, The indispensabZe foundation: a seZeation from the writings of Henry Sweet (Henderson 1971). It was also the happiest of circumstances that in the month in which she retired Professor Henderson was honoured by the world community of linguists in speaking as an invited rapporteur, at the Plenary Session on Phonetics and Phonology, on the present state and the prospects of this branch of linguistics, at the Thirteenth International Congress of Linguists in Tokyo in September 1982. In this year she was also elected a Fellow of University College London and, in 1984, President of the Philological Society. NOTE l. The exigencies of wartime language teaching left an indelible and ultimately treasured impression on those of us who were involved in it. I have a vivid personal memory from 1944 of Eugenie Henderson and myself shouting each other down while endeavouring to conduct simultaneous classes in Japanese pronunciation and Japanese grammar as we sheltered with our students (all service personnel) in the corridor of our departmental territory, to avoid the worst effects of VergeZtungswaffen I, affectionately known as 'doodlebugs'. 3

PAGE 16

Firth, J.R. Henderson, E.J.A. Palmer, F.R. REFERENCES 1945. Wartime experiences in linguistic training. Mod. lang. 26, 38-46. 1949. A phonetic study of Western Ossetic (Digoron). Bull. Sch. Orient. Afr. Stud. 13, 36-79. 1965a. Tiddim Chin: a descriptive study of two texts (London Orient. Ser. 15). London: Oxford Univ. Press. 1965b. The topography of certain phonetic and morphological characteristics of South East Asian languages. In Indo Pacific linguistic studies Pt.II: Descriptive linguistics (eds.) G.B. Milner & E.J.A. Henderson. Amsterdam: NorthHolland, 400-34. 1970a. Prosodies in Siamese: a study in synthesis. In Prosodic analysis (ed.) F.R. Palmer. London: Oxford Univ. Press, 27-53. 1970b. Acoustic features of certain consonants and consonant clusters in Kabardian. Bull. Sch. Orient. Afr. Stud. 33, 92-106. 1971. (ed.) The indispensable foundation: a selection from the writings of Henry SWeet (Lang. & lang. learning 28). London: Oxford Univ. Press. 1970. (ed.) Prosodic analysis. London: Oxford Univ. Press. Scarbrough Report 1947. Report of the interdepartmental commission of enquiry on Oriental Slavonic., East European and African Studies. London: HMSO. Thongkum, Theraphan L. 1979. (et al. ed.) Studies in Tai and Mon-Khmer phonetics and phonology: in honour of Eugenie J.A. Henderson. Bangkok: Chulalongkorn Univ. Press. 4

PAGE 17

.. PUBLICATIONS OF EUGENIE J.A. HENDERSON Helen Cordell 1940 Specimen of Thai. Le mavtre phonetique 69, 11-12. 1943 Specimen of Annamese (Tonkinese dialect), with notes on pronunciation. Le phonetique 79, 6-8. 1948 Notes on the syllable structure of Lushai. Bull. Soh. Orient. Afr. Stud. 12, 713-25. 1949 A phonetic study of Western Ossetic (Digoron). Bull. Soh. Orient. Afr. Stud. 13, 36-79. Prosodies in Siamese: a study in synthesis. Asia Major n.s.1, 189-215. (Repr. with later observations and ideas in Prosodic analysis, (ed. ), F.R. Palmer. London: O.U.P., 1970, 27-53.) 1950 & H.W. Bailey. Digoron word-list. Bull. Soh. Orient. Afr. Stud. 13, 381-88. 1951 The phonology of loanwords in some South-East Asian languages. Trans. Philol. Soc. 131-58. (Repr. in Prosodic analysis, (ed.), F.R. Palmer. London: O.U.P., 1970' 54-81. 1952 The main features of Cambodian pronunciation. Bull. Soh. Orient. Afr. Stud. 14, 149-74. 1957 Colloquial Chin as a pronominalized language. Bull. Soh. Orient. Afr. Stud. 20, 323-27. 1959 Karen Languages. Encyclopaedia Britannica. Chicago, London. The tones of the Tai dialect of Songkhla. Bull. Inst. Hist. Philol. Academia Sinica 30, 233-35. 1961 Tonal exponents of pronominal concord in southern Vietnamese. Indian linguistics 22, 86-97. Tone and intonation in Western Bwe Karen. J. Burma Res. Soc. Fiftieth Anniversary publication 1, Rangoon, 59-69. 1963 Notes on Teizang, a northern Chin dialect. Bull. Sch. Orient. Afr. Stud. 26, 551-58. Introduction to Linguistic comparison in South-East Asia and the Pacific, (ed.) H.L. Shorto. London: Sch. Orient. Afr. Stud. (Collected papers in Oriental and African studies; 4), 1-6. 5

PAGE 18

1964 Marginalia to Siamese phonetic studies. In In honour of Daniel Jones: papers contributed on the occasion of his eightieth birthday ... (eds.), D. Abercrombie et al. London: Longmans, 415-24. 1965 Indo-Pacific linguistic (eds.), G.B. Milner & E.J.A. Henderson. Amsterdam: North-Holland. Published simultaneously as Lingua 14 & 15. Part l: Historical linguistics; Part 2: Descriptive linguistics. Papers submitted to the Conference on linguistic problems of the Indo-Pacific area, London, 5-8 Jan. 1965. The domain of phonetics: an inaugural lecture delivered on 5 1965. London: Sch. Orient. Afr. Stud. The articulation of final 1-nh' and 1-ch' in Vietnamese. Proceedings of the Fifth International Congress of Phonetic Sciences (eds.), E. Zwirner & W. Bethge. Easel: Karger, 348-52. Final -kin Khasi: a secondary phonological pattern. In Indo-Pacific linguistic 1, (eds.), G.B. Milner & E.J.A. Henderson. Amsterdam: North-Holland, 459-66. Tiddim Chin: a descriptive analysis of two texts. London: O.U.P. (London oriental series; 15). The topography of certain phonetic and morphological characteristics of South East Asian languages. In Indo Pacific linguistic 2. (eds.), G.B. Milner & E.J.A. Henderson. Amsterdam: North-Holland, 400-34. 1966 Khasi and the l-clusters in seventeenth century Tonkinese: a preliminary glance. In Essays offered to G.H. Luce in honour of his seventy-fifth birthday, (eds.), Ba Shin et al. Ascona: Artibus Asiae. 1, 139-50. Towards a prosodic statement of Vietnamese syllable structure. In In memory of J.R. (eds.), C.F. Bazell et al. London: Longmans, 163-97. 1967 L'etat actuel des etudes ethnolinguistiques anglaises sur l'Asie du Sud-Est. Rev. Ecole Nat. Langues Orient. 4, 85-104. Grammar and tone in South-East Asian languages. Wiss. Zeitschrift Karl-Marx Univ. Leipzig, 16, 171-78. Vowel length and vowel quality in Khasi. Bull. Sch. Orient. Afr. Stud. 30, 564-88. 1970 Acoustic features of certain consonants and consonant clusters in Kabardian. Bull. Sch. Orient. Afr. Stud. 33, 92-106. 1971 The indispensable foundation: a selection from the writings of Henry Sweet, (ed.), E.J.A. Henderson. London: O.U.P. (Language and language learning; 28). 6

PAGE 19

1971 Structural organization of language: I--phonology. In Linguistics at Zarge, N. Minnis (ed.), London: Gollancz, 35-53. 1975 Phonetic description and phonological function: some reflections upon back unrounded vowels in Thai, Khmer and Vietnamese. In Studies in Tai linguistics in honor of William J. Gedney, (eds.), J.G. Harris and J.R. Chamberlain. Bangkok: Central Institute of English Language, Office of State Universities, 259-70. 1976 Khasi initial clusters. In Austroasiatic studies I, (eds.), P.N. Jenner et al. Honolulu: Univ. Press of Hawaii, 523-38. Thai phonetics sixty years ago: gleanings from the unpublished notes of Daniel Jones. In Tai linguistics in honor of Fang-Kuei (eds.), T.W. Gething et al. Bangkok: Chulalongkorn Univ. Press, 162-70. Vestiges of morphology in modern standard Khasi. In Austroasiatic (eds.), P.N. Jenner et al. Honolulu: Univ. Press of Hawaii, 477-522. Vestiges of morphology in some Tibeto-Burman In South-East Asian linguistic studies 2, (ed.), Nguten Canberra: Australian Nat. Univ. (Pacific lfnguistics. Series C; 42), 1-17. 1977 The larynx and language: a missing dimension? In The larynx and (eds.), G. Fant and C. Scully. Phonetica 34, 256-63. 1978 Notes on Yes-or-No questions and allied matters in Karen and Chin. In Spectrum: essays presented to Sutan Takdir Alisjahbana on his seventieth (ed.), S. Udin. Jakarta: Dian Rakyat, 452-68. 1979 Bwe Karen as a two-tone language? an enquiry into the interrelations of pitch, tone and initial consonant. # In South-East Asian linguistic studies, 3, (ed.), Nguy@n Li@m. Canberra: Australian Nat. Univ. (Pacific Linguistics. Series C; 45), 301-26. 1980 How can genetic-comparative classifications typological considerations? (Discussion). and genetics of (eds.), T. Thrane Cercle Lingu. Copenhague 20, 145-53. be based on In Typology et al. Trav. 1981 Towards a history of (eds.), R.E. Asher & E.J.A. Henderson. Edinburgh: Univ. Press. Introduction to A comparative word-list of Old Chinese and by G.H. Luce. London: Sch. Orient. Afr. Stud., i-iv. 1982 Tonogenesis: some recent speculations and the development of tone. Trans. Philol. 1-24. 7

PAGE 20

1983 Phonetics and phonology in the eighties: prospects and problems. Proceedings of the XIIIth International Congress of Aug. 29 Sep. 1982, Tokyo. Tokyo, 209-19. 1985 Feature shuffling in Southeast Asian languages. In Southeast Asian linguistic studies presented to Andre-G. Haudricourt, (eds.), Suriya Ratanakul, David Thomas & Suwilai Premsrirat. Bangkok: Mahidol Univ.,l-22. Greenberg's 1universals1 again: the case of Karen. In Linguistics of the Sino-Tibetan area: the state of the papers presented to Paul K. Benedict for his ?1st birthday, (eds.), G. Thurgood, J.A. Matisoff & D. Bradley. Canberra: Australian Nat. Univ. (Pacific lingu. Series c; 87) 138-40. Patterns of baby language in Bwe Karen. Linguistics of the Tibeto-Burman 8 (2), 55-59. 1986 Some hitherto unpublished material on Northern (Megyaw) Hpun. In Contributions to Sino-Tibetan (eds.), J. McCoy & T. Bright. Leiden: Brill, (Cornell linguistic contributions; 5) 101-34. 1987 J.R. Firth in retrospect: a view from the eighties. In, Language topics: essays in honour ofMichael Halliday, (eds.), R. Steele and T. Threadgold. Amsterdam: John Benjamins. 67-78. A phonetic oddity in Thai. In, suThai: essays in honour of E.H.S. Simmonds, (ed.), J.H.C.S. Davidson. London: Sch. Orient. Afr. Stud. (Collected papers in Orient. Afr. Stud.), 52-62. Forthcoming J.R. Firth's 'First impressions' of Tibetan pronunciation. In, Prosodic analysis and Asian linguistics: to honour Richard Keith Sprigg, (eds.), David Bradley, Martine Mazaudon, and Eugenie Henderson. Canberra: Australian Nat. Univ. (Pacific ling.) & R.H. Robins. John Rupert Firth and the London School. In, Oxford International Encyclopaedia of linguistics, (ed.) ,, W.O. Bright. New York: Oxford Univ. Pr._J Khasi clusters and Greenberg1s universals. Papers from the symposium on Austroasiatic linguistics and literature, (ed.), J.O. Svantesson. Copenhagen: SIAS; London: Curzon Pr. One word or many? A problem for the lexicographer of pre-literate languages. In, Proceedings of the 18th International Conference of Sino-Tibetan Languages and Linguistics. Bangkok. 8

PAGE 21

Forthcoming, continued Problems and pitfalls in the phonetic interpretation of Khasi orthography. In, Austro-Asiatic Linguistics: essays in honour of H.L. Shorto, (ed.), J.H.C.S. Davidson, London: Sch. Orient. Afr. Stud. (Collected papers in Orient. Afr. Stud.) 9

PAGE 23

, ea: FROM "FISHES' EYES" TO "ANKLEBONES": A VIETNAMESE CALQUE? Jeremy H.C.S. Davidson 'mtt ea d.i gi1 (nem= :te ::ff "what are fishes 1 eyes?", 1 asks Vietnames-e" -(VN). The earliest record so far discovered in answer is Alexandre de Rhodes' Dietionarium entry (1651:456): 1mtt artelho do talus,i.', in which the Latin tal us means 'ankle, anklebone 1 Le grand de la Liraye (l874:150b) translates it as 1cheville du pied', Taberd (1877: 27la) as 'Occulus piscis; malleoli', while Ravier (l880:706a) notes 'malleolus #4.mKt d. (ndi chan tay) 1 (=place of the feet/ legsJii and ck (l897:86b) describes it as 1 fleslied bone swelling out on both sides of the foot at the point where it joins the leg'; (l898:q.3, 3a4, p.5l = p.216, 1.19) uses it cryptically to translate the Chinese character 1 ankle 1 ( > 1 anklebone 1 ) huAi in Modern Standard Chinese (=putong hua/MSC). Again, Bonet (l899:407a) records it as 1cheville du pied; litt. oeil de poisson', Hu@ (l937:555a, 2) repeating it as 1cheville du pied, cor, chevillon', while (l960:299b) defines it as 'heads of bones jutting out at the neck of legs (=ankle) (qv. Tru l970:II, 894a) and the H8i Khai Tr{ ( 1968: 337b) tells us that it is 1 the head of a bone budding out near the neck of the leg' (=i.e. the ankle), hence 'the anklebone 2 This seems to be highly figurative language,3 a descriptive pun (qqv. Emeneau 1947; 1955; Davidson l978:37ff., l986a: 354, n. 37) for to find a fish 1 s eye, probably rather dead in appearance, up at you from the anklebone of a person's leg, bulging out dull, then whitish as it stretches from where it peaks to the darker skin surrounding it on legs that have travailled in fields of paddy, sea-salt, or whatever else, in a Vietnamese-style climate, will c.ertainly provide yet another superb and humorous expression of the impression intended, since the Vietnames.e have long associated their vital environment with all the other natural symbolism that they envisage (e.g. Davidson 1978). Such visual association (qv. n.3),5 is frequently complemented by word-play afforded by the opportunities of homophony. An encounter is the Modern Mon (MM) 'anklebone' (lit. 'eye of the elephant'; Shorto l962:170b mbt ref. p.96b coigl 'elephant'), while in inscriptional language one finds cig2 / n. 1 elephant 1 ( Shorto 1971:98) and 'lower limb' (ibid. ,125), where a mat 2/mot/ n. 'eye' (ibid.,285) plus jun would have provided us with the expected 11

PAGE 24

Mon-Khmer (MK) gnd Malayo-Polynesian (MP) 'eye of the leg' for an 'anklebone'! This 'confusion' may well have stemmed from, or led to, folk-etymology and folk taxonomy --an investigation lying in wait --but it further emphasizes the point already made of the desire for the use of figurative language and especially for descriptive punning. Parts of the body are, however, often not identified with any precision in basic and vernacular Vietnamese language, that is, language which excludes or may not include the use of the specific, pointed (HV) vocabulary found in the literary language of the elite, and which is also 'premodern 1 that is, prior to the introduction of western, scientific and other specialist vocabulary. I A general area of the body may be referred to, but not a specific part of it, by a single, special word. So, chin ( var. chci'n, ch1fn) stands broadly for both 'leg 1 and 1 foot 1 :tN for 1 arm' and 'hand 1 while together ch!n tay means 1 limbs 1 8 When an an-atomical feature does captivate attention and is accorded a specific term, structural or functional perception of it is a common reason for the formation of the word or words describing it. So, many South-East Asian languages have captured in their vocabularies' imagination the concept of the or wrist as the 'neck' of the leg or arm and the as the "'eye" of the leg/foot' even though in actuality it is not a part of the ankle itself but a swelling out on both sides of the lower end of the two leg bones (tibia and fibula) at the point at which they articulate with the tarsal bones l98(:86b). Naturally, words for 'hands 1 and 1 arms 1 1 feet 1 and 'legs 1 may well have changed in the language families that inhabit South-East Asia, China, South Asia, Oceania and other related areas as their languages developed from the early through to the modern forms, but the phonetic similarity is even now often noticeable. The same is also evident for the words for the 1eye19 (e.g. Shorto 1971:285, i.a.) since one finds links between the Sino-Tibetan, Sino-Tai, Austroasiatic and Austronesian forms in_ both their phonetics and the semantic functionslO performed. So, as in Vietnamese we have an 'arm' tay, its 'neck 1 c'8 immediately recognizable as the ,/wrist 1 ci tay, -;;:s does the leg 1 s 'neck 1 become an 'ankle 1 eo (e.g. Gouin l957:18lb). other examples are numerous n. but, when we find the 'board, table' (b'An cf. van helping to structure a 'hand' ta why do we not 'eye' the 'leg' (*mit ) make an 1 anklebone 1? After all, we have already met the partially successful Modern Mon attempt of coi9 and the Malay mata kaki. Bahnar provides us with 'cheville du pie4' (Dourisbourne 1899:203; cf. Katu maniit, identified as in the Vietnamese bu{ translated as 'ankle' in the English (Costello 19(1:2(, 30) and probably from ma 'eye', cf. katam (Zoo. ait. ) ) while Sedang gives us ma 1 ankle 1 (lit. 1 eye 12

PAGE 25

of the leg/foot 1 ) and ma k6"ng 'wrist 1 (Smith 1962:15-1.6) and we can note en passant the Chrau 'leg, foot' (Thomas 1966:24),14 unfortunately with no further relevant data. Going south in present-day Vietnamese territory, we discover more support from the Cham mdta takai (cf. takan takay, Moussay l91l:319b 1cheville du pied'; cf. ibid.,224 mUta le 1cheville1 ) with the other common uses of the 'eye/face' word as 'node, joint', etc. The same occurs in Nicobarese (Roepsforff 1884:78, 122, 149, --okloaka-la; --olmat, etc.; cf. Man 1898-9:6, ll8a). This Vietnamese 'neck' ci of the 'arm' or 1leg1 them to 1hands1 or 1feet1-,-finds many other Mon-Khmer language equivalents, two directly apparent examples being Modern Khmer ko: day 'wrist', ko: cYi9 1ankle1 (Jacob 1911:1) and Modern Mon 'ankle', ko? toa 'wrist' (Shorto l962:53a).l5 Of immediate interest, therefore, and suggestively cognate is the Lao khS: 'ankle' (lit. joint of the foot) (Kerr l912:129b; cf. ibid.,ll8b which is supported by the Mbdern Standard Thai (MST) 'ankle; lit. joint of the foot' (not 'leg', (n.b. Li 1911:213, .6.29 among other entries, e.g. .23, 14.11.4, etc.) --with a secondary vernacular taa ttiD 'eye of the node (of the foot)' >'anklebone'. Thus, the possibility of 1link1 cognates --joint-neck-tibia-shin --keeps springing to mind. Indeed, there is an impression of 'joining, uniting, linking' and of 'reaching, attaining, going to' (see GSR 6i5a-d ; e. [=MSC he'J) encouraged by the character for which one also finds as and in the MSC compound kl'fl6u N (= khtl cf. MSC 1ibid.1 =HV cffi; each). Additionally, there is the Vietnamese word for 'joint; physical articulation' (>'bone-joint1 = rv xddhg), which strongly suggests a relation between it and the Vietnamese 'unite, etc.', joining things together. Reverting in this context to a possibly pre-sinitic word (or a very early Chinese loan?) preserved in present-day Chinese territory, one finds MSC j iao ( GSR 1166p *k 'og/ k'au/k1iao =?HV giao); its variant character used for 1joint1 and presented in Ningpo dialect as gao (Morrison l876:20b), in Swatow (Gibson l886:l01c; cf. VN 'join again' =?Tie-chiu/Hokkien) and its parallel MSC jlng e.g. jfnggti 'shinbone' (>'tibia'; (cf. GSR 83lk. *g'ieng/yieng-/hing leg, shank (Lunyu)'), as well as the two pronunciations for MSC (GSR 83ln. kAang:/king, and* I 'ian k'ing ... Tso)), all of which are related, at least phonetica ly. Hence the pronunciations for the word for 'neck 1 in Vietnamese c'CJ, the Thai Lao khg: a large number of Mon-Khmer languages (qv. Shorto and Archaic Chinese *k1og (GSR ll66p), thrust cognacy before us.16 The image of an articulated (=joint), tube-like connection of varying lengths between the torso and its ending appendages does, of course, bring the possible linked, root-origins of such words together, while the closeness in pronunciation of the 'leg' word l3

PAGE 26

in a range of South Eastern Chinese dialects (qv. Bauer 1987), e.g. keUk (=HV cu&c, also read khu6c, giving rise to gfc 'base, foot (of a tree)' cnem:f-aJ (TrlJ. 1960:46a)) reconstructed as Archaic Chinese *kiak (GSR 776g-h), forcefully implies a presinitic origin of that ends up as the Chinese character /fitrl(MSC I j ob But here the Vietnamese mat ea 'anklebone' remains distinctly in the forefront of one 1 s mind. If the base word 'leg, foot' which qualified other associated ideas giving us words like 'ankle', 'foot', 'toe', 'toenail', 'heel', etc., (qv. n.8 below) fulfils its purpose, then why does an 'anklebone' suddenly turn from the '"eye" of the leg' into a 'fish's eye'? Naturally, in a Vietnamese language situation such as this, one looks to the Chinese reservoir of vocabulary for help in solving the problem. In Archaic and in Ancient Chinese one does not find the 'anklebone' separately listed, one finds the 'ankle' (GSR 35lj. *g'lwar/ywa:/hua ankle (Li) = MSC hua:'i), the Vi@t reading for which (Thi@'u but which is.intriguingly identified as an 'anklebone' (ibid.), and clearly recorded as such in \,he t' tnjc ( 1898: q. 3, 3a4=p. 51 1J_ =p.216, 1.19 as : 1khoa mat ea') while the MSC gives us 'ankle' and for 'anklebone' (cf. i.a. Maclay 1929: 1090 Other HV 'pedal 1 terms flourish too. Meanwhile, the MS(?Il e 1 occurs as JfGp /g}8t_ t (var. while the Ningpo kyitb. tsang Bttl 'ankle' (cf. gao relates directly to a kiok and the Hakka I ankle I kiok muk (cf. kiok ng afk_, Maciver 1926: 288b) connects weil1with the Shanghai kyak mok. Yet, focusing on the 1eye1.as the centre of attention in this 'joint', this 'link', strikingly important is the awareness of the '"eye" of the "leg", foot', in Southern, and in particular in South Eastern Chinese dialects. There is Cantonese ketk (kwut) 1ankle(bone)1 (qv. MSC, above; cf. Maciver 1926:344b, kwut 'the anklebone 1), Lungtu kha' muk, FU:zhou k8: ngU (='the [jointJ's ox's eye') (Maclay 1929:1021),18 the Swatow kha-mak 1 anklebone 1 (and variants, Lechler l:883: 9) and the Amoy 'ankle' k1a bak (liter. ki::>k b::>k, ki:,? bak; cf. C. Douglas 1899: lOa, 257b, 1ankle1 R. kiok,col. kioh). The thought came to my mind of a word of pre-sinitic origin preserved in what are termed South Eastern Chinese dialects and then spurred into a calque by the creative imagination of the Vietnamese because it rhymed descriptively with the phonetic and thence the visual awareness of a 'fish's "eye"' (mKt elf <*kh' And, almost instinctively one senses that the Hokkien dialects offer the greatest suggestions. Among them, I think that Tie-chiu [=ChozhouJ enlivens us. In it we find mak 'eye (Goddard 1883:103a)l9 and k1 { 'foot' (ibid., 63a), the tone, pitch and 7.ontour of whose pronunciation conform well with the Vietnamese strengthening the argument since it was the speakers of Tie-chiu who made up the largest Chinese 14

PAGE 27

population contributing to the 'early' Vietnamese vocabulary,20 just as it appears to be Swatow speakers who influenced Thai (Egerod 1959) For here, it seems, we do have a Chinese-Vietnamese calque created by the Vietnamese who, hearing this South Eastern 'Chinese' word ( 1 anklebone1)23 into their own language, reordering the word as mAk to fit their syntax, to create so that the 'anklebones' became 1 fishes 1 eyes 1 ea l. 2. 3. 4. NOTES Immediatelyr, we are alerted to an unusual etymology. Normally, nmt / 1 are classified as animate by ..2!!. so that con mat ea would mean 'the eye of fishes' even though the syntactic rhythm feels heavily clumsy. But here, it is classified as inanimate by cfi, suggesting a variant meaning. What is the reason? ChUll l966:660a Sino-VietnameseJ hoa, (or khda). l. ca.'i mat c. ch'lin 1the Of the legS I; Cf Anh l957:510b '"fishes' eyes" on the two sides of the ankle', Listings of dictionary entries could continue, but they are not being included as they are meaningfully repetitive, if not identical; e.g. Hung l955:414a; Thanh l967:885a: 'balls of bone jutting out on both sides of the ankle'; Hoa l967:2'70a, etc. So, too, is the Palaung jung 'ankle' (lit. 'joint of the leg') (Milne 1931:18), with 'anklebone' (loc.cit.) a ka-ang i-ar, that is, 'bone of the fowl, hen' (?cockspur. [bone]; cf. 'VN .V:. 1 cockspur' ; xu'ong -'astragalus' ) Note, interest.ingly, that in Chinese we have the word 1 eye' MSC reconstructed as GSR 416l. *ng<.n/ng'fn: /yen eye (Yi) which is also a 'loan for *ngan/ngan:/en protrude as a knob (Chouli)' (loc.cit.}. The HV is (qqv. de Rhodes 1651:548, cf. 456; Tu)in?c 1898: p.52 =p.218, 1.5 Davidson i975:59'7, no.487). In passing, 'faces' and 'eyes' do, understandably, go together (e.g. Shorto T?f-d?c 1898: q.3, la5, p.50,j} =p.215, 1.9 I diE; ml}t I) 15

PAGE 28

5. 6. For instance, there is tr,ilig 1t jey 'eggs of fishes (qv. Bonet 1899: 34la tr'ling ea 'oeuf de poisson'), whence comes 'spawn' and then, because of the visual similarity, 'blackhead, comedo' (H6a l967:516a; l937:1088b; cf. Gouin 1957: l4lb). This is possibly,a play on on :symptom of an illness' (nflm/HV :Jfi: ; MSC zheng) plus 'f1shes whence 'small white face pimples'. Another example is hon dS:i tJJ (lit. 'balls/stones of the genitals'>) 'testicles' (HCla l967:84a; l937:19lb) paralleled by tr&hg d{i (lit. 'eggs of the genitals'>) 'syerm' (Bonet l899:34la; cf. Gouin l957:148lb, i.a.). Note 'to urinate' sometimes written Simllaay, ::me finds gb ma literally mound/knoll of the cheek' so 'cheekbone' ( Bohet l899:230b;,'jo1Je, pommette'; l937:3l4a; Gouin l957:482b), but why ea for 'anklebone'? d h 't l The play on words coul also ave g1ven us mo co1y eye the elephant' and coiij2 'eye of the ridgepole of the house' (Shorto l962:96b). Compare the Archaic Chinese use (MSC jiao 'tibia') for 'the tapering end of the a wheel (Chouli)', (GSR ll66p.). of 7. Nowadays, the scientific and medical vocabulary of Vietnamese is as detailed in its anatomical and other descriptions and recognitions (e.g. Thanh Nght l967:1532a; Hba l967:565a-b) as its main new source language, French; in fact, it is probably made richer by ready recourse to Chinese wherever that proved necessary. Thus Vietnamese has a very rich vocabulary and usually treats technical-type terms in three ways: (i) identifying them by direct translation into VN -e.g. 'astragalus'= xudng 'cockspur bone'; (ii) by use of Chinese terms in the HV pronunciation of their characters, although a degree of Vietnamese wordordering may be introduced --e.g. 'haemophilia' = huyit "' / v / hu"u
PAGE 29

9. 10. ll. 12. 13. 14. 15. khu.tu/cti.J. tay 'elbow'; b'itp chlin 'calf (of the leg) 1 ; 'thigh'; g6t cll'an 'heel' (cf. @t 'tendon, Achilles heel ... '); tay 'fist'. And, although there are 'finger' .words, e.g. ng6n tay 'thumb'' ngtn tay nt 'little finger', one can easily manufacture parallels like 'big toe', 'little toe', etc., by substituting for tay. And so on. The versatility is immense, as is readily shown by the Tlf' -Bt!ic. See, e.g. Bauer 1987 on the 1 leg' in South Eastern Chinese dialects and Tibeto-Burmese; and, in general, Davidson 1975: I, 296ff. & nn.; l986b:63, & references). The use of the 'eye' as a node of the bamboo (e.g. Bonet Zoe.eit., tre 1noeud de bambou'; l937:555a2, ibid.; Gouin l957:80lb. ff.) and sometimes as the joint of a part of the body and so on, is also well observed throughout South-East Asian languages (see i.a., Shorto l962:170a ff.; l971:284ff.; Moussay 1971:224). qv. de Rhodes (1651:126) tay: collo da mao: collum manui', [qv. also p.715J; collo dope: collum pedis'. l937:10la; cf. Gouin l957:18lb). In Malay, one apparently circles the leg to form an ankle (Wilkinson l932:I, 338a, cf. 290a; 496a kaki) but the 'anklebone' is still the mata kaki 'eye of the leg' (Zoe. & II, ll4a, mata). The thought that the Vietnamese m'iit c. might embabn an MP variant of this is farfetched (even given the *proto-language forms being proposed at present) because of the distinctly Mon-Khmer word for 1leg1 in Vietnamese. In Bahnar, I understand that ;*kpa:ij/ is a possible reconstruction. Might this imply a MK 1body1 prefix which is now lost in modern Vietnamese? Pacoh deals where relevant with bones but does not seem to have a special term for the 'anklebone'. (Watson 1979:382; cf. parreat (ati).). Note that /h-/ and kh-/X/ variants are common in Vietnamese, e.g. n.l above. 16. Relevant, too, are the semantic similarities and, to a certain extent, the various reconstructed readings MSC jia, ji (9GSR 630a be on both sides (Shi); support (Shu); loan for 6301. all around (Shu) [.: Jrli jia,xi(l and especially ? I tsiep/tsie [.:MSC xi6J embrace (Shl); all around (Chouli).); n.b. GSR 630k, too. 17

PAGE 30

l7. 18. 19. But lba in Anh (l957:5l0b; Tru l970:278a4) which implies an original *kl-initial cluster and a shift from it and kh-variants. cf. n.l5 above. n.b. gi ci-tl (lit. 'foot-eye-ball', Adam 1891/1905: ll; ci.PMaclay 1929:1222) which brings to mind the Burmese khie-myak-ci, transcribed as chi-mye?-s{ 'footeye-ball' ='eyeball' (note Adam 1905:106 mgk-cit-ci). From the Chinese in which there is m';\ 1 eye 1 reconstructed *miok/ituk/mu (GSR l036a-c), muc (de Rh;aes l5l:488; muc, vide mouc; 483 mouc, con mat: -/. . olhos : Tlf' -fJ12c 189 8: q. 3, la4, p. 50 I3J JtS =p. 215 l. 7 mli:'t). En passa:nt, we may note that final -!. and -.. /k/ are often interchanged, especially in southern Vietnamese dialects. 20. A large Tie-chiu refugee migration --one which preceded the late nineteenth century influx of Cantonese who then also provided a vast source of vernacular, material culture vocabulary --is known to have taken place during the Mlng dynasty, and Tie-chiu loans to, and pronunciations of Chinese characters in, Vietnamese --as distinct from the H&n-Vi@t forms one normally encounters --are numerous. Many of such words are readily found in (1897: e.g. 374a thao; 380a "' 1\ / --thau; 402a tho; 433b 1daddy1 [cf. Egerod 1959: no. l59J; 45la l899:323bJ) and in Tru (1970: e.g. II, l7a etc.) but not all dictionaries are as detailedly precise. En passant, no Tie-chiu words have, it seems been recorded in de Rhodes' Dictionarium, a compilation principally of the northern dialect. 21. As seen in the more risque or vulgar puns like for (Paul) Dourner and a{ t-cua I crab Is arse I for the much despised 1discours1 22. And its irony. Note the idiom ( : ngu? muc h'n chiiu *-1 fishes 1 eyes may be taken for 1 meaning -t:nat true and false are confused. Its main application in both China and was to people who used fake materials to make counterfeit, 'real1-looking, goods, thus deceiving others. This was usually jewelry for women and girls. 23. Perhaps this also refers obliquely to notice of a lot of Chinese migrant labour? 18

PAGE 31

Adam, T.B. Anh, duy, Bauer, R.S. Bonet, J. Carstairs Douglas Costello, Nancy A. P. Davidson, J.H.C.S. REFERENCES l905.(from 1891). An English-Chinese dictionary of the Foochow dialect. Shanghai & Foochow: Lacy & Wilson, Methodist Publ. Ho. 9 R ; v" ,, +.A? ? x ,.., l 57. an1.-et tu-a1.-en g1.-an-yeu 1987. 'Leg' in Southeastern Chinese dialects and Tibeto-Burman root 'leg'. Ling. Tibeto-Burman Area 10(1), 169-74. 1899. Dictionnaire annamite-frangais (langue officielle et langue vulgaire). Paris: Imp. nat., E. Leroux., 2 pts. 1899. Chinese-English dictionary of the vernacular or spoken language of Amoy. London: Publ. Office Presbyt. Church of England (new ed.). 1971. Ng)}J-vJng Katu: Katu vocabulary vol.5). sl.: I 1897. qu8c-&m t-v1. Saigon: Imp. Rey, Curiol, & Cie. 1975. A new version of the ChineseVietnamese vocabulary of the Ming dynasty. I: Bull. Sch. Orient. Afr. Stud. 38(2), 296-315; II: op.cit. 38(3), 586-608. 1978. Images of ecstasy: a Vietnamese response to Nature. In: Natural symbols in South East Asia (ed.) G.B. Milner. London: Sch. Orient. Afr. Stud., 29-54. l986a. Collaborateur versus Abstentioniste versus Tr1): a political polemic in poetic dialogue during the French acquisition of southern Bull. Sch. Orient. Afr. Stud., 49(2), 321-63. 19

PAGE 32

Davidson, J.H.C.S. Dourisbourne, P.X. Egerod, S. Emeneau, M.B. Gibson, J.C. Goddard, J. Gouin, E. GSR Ho'a., N ctinh, Eldi Khai Tr( G. Hung, v8.n, Jacob, Judith l986b. Another source for information on Ayutthaya Thai. In: Lat SWThai: essays in honour of E.H.S. Simmonds (ed.) J.H.C.S. Davidson. London: Sch. Orient. Afr. Stud., 63-72. l889. Dictionnaire bahnar-frangais. Hong Kong: Imp. Soc. Miss. Etr. l959. Swatow loan words in Siamese. Acta Orient. 23, l37-56. l94 7. Homonyms and puns in Annamese. Language 23, 239-44. l886. A Swatow index to the syllabic dictionary of Chinese by S. Wells Williams3 Litt.D. & to the Dictionary of the vernacular of Amoy by Carstairs Douglas. Swatow: Eng. Presbyt. Miss. Press. l883. A Chinese and English vocabulary in the Tiechiu dialect. Shanghai: Amer. Presbyt. Miss. Press (SOAS M5224). l957. Dictionnaire vietnamien-chinois jrangais. Saigon: Imp. Ext.-Orient. l957. B. Karlgren, Grammata serica recensa (= GSR). Bull. Mus. Far East. Antiq. 29. l955. Double puns in Vietnamese: a case of 'linguistics play'. Word 11, 237-44. l967. V,Cetnamese-English dictionary. Rutland, Vermont: Tuttle. l968. V-Wt-Nam tu' 7 ( drafted by Ban hoc HBi Khai Tr{ khJi thlo). 1\ Saigon: l9 37. Dictionnaire annami te-chinois sl. : Imp. Trung Hoa. l955. rietnamese-English dictionary. Paris: Eds. Europe-Asie. l974. A concise Cambodian-English dictionary. London: Oxford Univ. Press. 20

PAGE 33

Kerr, A.D. Lechler, R. Legrand de la Liraye, Le R.P. Li Fang-Kuei Maciver, D. Maclay, R.S. & C.C. Baldwin Man, E.H. Milne, Leslie Morrison, W.T. Moussay, G. Ravier, M.H. Rhodes, A. de, Roepsforff, F.A. de, 1972. Lao-English dictionary. Washington: Consortium Press, 2 vols. 1883. English-Chinese vocabulary of the vernacular or spoken language of Swatow. Swatow: Eng. Presbyt. Miss. Press. 1874. Dictionnaire elementaire annamite frangais. Paris: Challamel. 1977. A handbook of comparative Tai. Honolulu: Univ. Press Hawaii. 1926. (M.C. Mackenzie, rev.) A ChineseEnglish dictionary: Hakka dialect as spoken in KWang-tung province. Shanghai: Presbyt. Miss. Press. 1929. (S.H. Leger, rev. & enlarged) Dictionary of the Foochow dialect. Shanghai: Presbyt. Miss. Press. 1888-89. A dictionary of the central Nicobarese language. New Delhi: Sanskaran Prakashak, rep. 1975. 1931. A dictionary of English-Palaung and Palaung-English. Rangoon: Supdt., Govt. Printing & Stationery. 1876. An Anglo-Chinese vocabulary of the Ningpo dialect. Shanghai: Amer. Presbyt. Miss. Press. 1971. Dictionnaire francais: Phan "' /1. Rang: Centre culturel cam: Trung-tam v!l'.n-hcfa Ch'am. 1880. Dictionarium latino-annamiticum. Ninh Ph1f: ex typis missionis tunquini occidentalis. 1651. Dictionarium annamiticum3 lusitanum3 et latinum ope Sacrae congregationis de propaganda fide in lucem editum. Roma. 1884. A dictionary of the Nancoury dialect of the Nicobarese language -in two parts: Nicobarese English and English Nicobarese. Calcutta: Home Dept. Press. 21

PAGE 34

Shorto, H.L. Smith, K.D. Taberd, Rev. Thanh D. & Th-<3 Sa'ng Watson, R., Watson, Sandra & Cubuat Wilkinson, R.J. l962. A dictionary of modern spoken Mon. London : Oxford Univ. Press. l97l. A dictionary of the Man inscriptions from the sixth to the sixteenth centuries (London Orient. Ser. 24). London: Oxford Univ. Press. l962. Sedang: Sedang_ vocabulary ( n/cm-ngt{' dful-t9c Nam, vol. 2) Saigon : N (\ J.:'J I' ngd-hoc trung-tam Kontum, Bo/ . l877. (rev. J.S. Theurel) Dictionarium annamitico-latinum. Ni nh ex typis missionis tunquini occidentalis. l967. minh hoa . / . Saigon : Khai Tr1. / -+-"'7 1?.:7 l966. Han-Viet tu-aien Saigon: Hung Long: (tJ l966 Chrau: Chrau vocabulary: (TJ-sich ngBn-ng\1'> vol. 1). B9 Gi(o-dvc: l962. Saigon: 2nd ed. l970. (et al.) tu' Saigon: ..., . Khai Tr1, 2 vols. l97l. ( from l898). Ching-ho ed., e. tc l t:J:Lh9c gictinghux-ca .. t M -j_ 51:? Hong Kong: Ch1nese Un1v. Hbng ..t l'v l979. Noh Pacoh-Yoan-Anh: Pacoh dictionary: Pacoh Vietnamese-Efglish vol. 25, pt .l). Huntington Beach, Calif.: Summer Inst. Ling. l932. A Malay-English dictionary (romanised). Mytilene: Salavopolous & Kinderlis, 2 vols. 22

PAGE 35

SOME FEATURES OF MODERN KHMER LITERARY STYLE Judith M. Jacob The literary or formal style under review here is that of modern prose, particularly that to be found in journals, speeches and novels. Interest in this aspect of the Khmer language was stimulated by the realization that certain assumptions are made about it without any attempt at justification. For example, one feels, especially when translating, that it is an unnecessarily verbose style, containing far more words than can be represented in the translation. Is this really so, or is it a false impression due to language differences? One also tends to say, without any specific parallels in mind, that modern Khmer journalese is clearly imitative of its French equivalent. This paper attempts to summarize the various linguistic differences between the consciously formal modern style and the spoken, or informal. Material has been collected for the purpose from articles (post-1930) on literature and religion, particularly those in the journal Kambudasuriya, from newspapers (1950-72) and from novels (1938-71). I am very much aware that there may be many constructions, associable with style, which have been missed and I present this analysis as a precursor of others. The increase in the output of prose writing as compared with poetry in the last few decades, during which, for example, both the novel and the printed newspaper in Khmer have developed from scratch, has caused prose writing to be held in higher esteem than it traditionally was in Cambodia.l Authors of novels and writers of articles were trying consciously to raise the standard of prose style in the 1950s and 1960s in response to the, general wave of nationalistic feeling. Interesting changes in style which thus took place in post-war Cambodia, especially since independence in 1953, may be observed by comparison with the factual and simple narratives of the ChroniaZes, composed in the nineteenth century,2 or of the first written versions of folktales3,or of the esteemed early twentieth century writer Suttantaprija In. The analysis of these developments might be presented in many different ways. I have decided to arrange the various points in four sections which suggest my own idea of their possible origin. These are; 1. Features which seem to be present in order to clarify long, involved sentences; 2. A feature which suggests conscious or unconscious imitation of Thai; 3. Features which suggest conscious or unconscious imitation of French; 23

PAGE 36

4. Features which seem to reflect the desire to embellish, using the traditional devices of Khmer literature. 1. Features which seem to be present in order to clarify long, involved sentences In the 1950s and 1960s, when upwards of 3000 new technical terms based on Sanskrit and Pali were invented so as to avoid using French loanwords any longer, the introduction of this unfamiliar vocabulary added to the need to explain some old loans to a general reading public and had some effects on the style of written Khmer: (i) In the early years of the use of the new vocabulary it was common practice to place two wohds, one Khmer, one Indian, side by side. Thus lsmo:m yl:acokl (Khm. + Skt.) 'beggar', ls:tksa: so:t-rleni(Skt. + Khm.) 'education', lprete:h-d:at m1:etophu:ml (old,+ new borrowing) 1one1s native country', all occur in the newspapers of the 1950s. (ii) The unwieldy sentences produced by use of the new words in complex clauses led to the excessive use of the literary particles I nul and neyl which are used only with restraint in earlier prose writing such as that of the ChronicLes, the folktales, or the writing of Suttantaprija In. ln'Yul (spelt nuv). In Khmer, the object, when it is expressed, normally follows the verb immediately, e.g. lm:1 konl 'see a film'. In an informal or colloquial style, no particle connects the object to the verb. The use of ln'Yul as an indicator that the object of the verb follows is a literary device which is useful when, as in journalese especially, several phrases occur between the verb and its object and it is helpful to know that the next word, or words, will be the object, e.g. lcoh phsa:y p1: l'Y:k publish, print two times in nYu (particle) story sa:p;)edem:en nE:ek-d:et-n1yi:lm newspaper nationalist 'twice printed the story in. the Nationalist newspaper'. Once established in modern literary usage, however, ln'Yul began to be used when there was no need for it because the. object followed the verb immediately: ltr;)em-tr;): endure undergo n'Yu (particle) 'put up with the pillaging'. 24 I omp'Y :Jlnam-Jll: I matter pillage

PAGE 37

/ney/. In the colloquial, or informal, language, two nouns occurring in close junction may represent an object and its possessor, e.g. /phtgah v1:a/ 'their house'. A more elaborate way, especially useful if attributes of the first noun follow it before the occurrence of the possessor-noun, is found both in colloquial speech and in the literary language: /phtgah thom v1:a/ (house big possession they) 'their big house'. In a literary modern sentence, however, the role of /raboh/ is played by the particle /ney/ 'of', e.g. /lo:k ?akkEate:sa?ph1ba:J governor general ney prate:h ?o:stra:l1: of country Australia 'the Governor-General of Australia' It may be pointed out that, in the Chronicles, many similarly long instances of possession and possessor are expressed without /ney/ by simple juxtaposition, and that /ney/, like /nu/, came to be used between short familiar words in modern writing too. (iii) Among the seemingly unnecessary words which help to give the verbose impression to modern literature are many sentence-final nuclei which corroborate a word, often a particle, which has occurred earlier in the sentence. Examples are in this case given (see next page) without full sentences, for the most part, since the point seems clear from the meanings given and they are divided according to the grammatical functions of the two elements. The practice illustrated in these examples, which occurs freely in the colloquial language as well as the written, seems to have become a characteristic feature of journalese, where it has a clarifying role, not so much because the meaning is reaffirmed as because the sentence-final nucleus, or sometimes several of them together, confirms that the end of a clause has been reached. 25

PAGE 38

1\) 0\ (a) Meaning of pre-verba l particle5 (pre-v.p.} corroborated by sentence final element: pre-v.p. final particle translation nu-tae (still pli.m-t'Oan (not yet pre-v.p. prakaek argue raluat extinguished sot-tae thv: t1:ah:t:an soldier (all without do exception srap-tae (suddenly pre-v.p. d::>l arrive kUalJ pratEah nmlJ s ::mJ.b.ok v:t : a with nest they (certain meet nu-laay still) nu-laay still) adverbial particle tEalJ-?::>h all) phl:t: am immediately) verb pli.m kha:n not miss) 'is still arguing about (it)' 'is not yet extinguished' 'all without exception became soldiers' 'suddenly they were there' 'shall b e sure to come across one of their nests' (b) Meaning of general particle (gen.p.} corroborated by a final particle (f.p.} nu s::>l (left over tae bu:an only four ponn::>h just so) 'there w ere only four left' extra 'still' 'still' 1all1 'immediately' 'without fail' 1 just so'

PAGE 39

(c) Meaning of pre-nominat partiate (pre-n.p.) corroborated by adverbial particle (adv.p.): pre-n.p. adv.p. translation extra ?oh pu:al:: (all group krup ru: p (complete person 'the whole group' 'all of them' (d) Meaning of post-nominal partiate (post-n.p.) corroborated by adverbial particle: post-n.p. adv.p. bontec-bontu:ac laay pil.m tatu:al domnYJ;J ?vy (not receive new:s .e;ny slightest at all)'did not receive any news at all' (e) Meaning of main verb corroborated by fina t partic 'le or verb: verb f.p. or verb kaan (increase tvu muk tl:at on further) 'will increase' (f) Meaning of attributive verb corroborated by an adverb: y::,:k khae1; ...... ba:n krup t'EaJ;J-?::,h m::,:k veJl (take province manage complete all come back) 'managed to recover all the provinces' (g) Plurality of a reduplicati v e compound corroborated by sentence final nua'leus: raboh thom-thom (thing big and numerous cl::a craan in numbers) 'many large objects' 'all' 'all' 'the slightest' 'further' 'all' 1 in numbers

PAGE 40

2. A feature which suggests conscious or unconscious imitation of Thai In colloquial speech, the only numeral coefficients which have to be used are those of which the meaning is a term of measurement, e.g. in the phrase /sko: 'two kilos of sugar'. In a formal or literary context, however, there is a tendency to use unnecessary numeral coefficients. Thus, for example, the numeral coefficient for human beings /nEek/ (e.g. /menlih nEek/ 'two men'), which has been in evidence since the seventh century, is used more frequently in literature; the general word for 1 item 1 /preka: / is found (e.g. /haet preka:/ 'two reasons'), and various coefficients indicating shapes (e.g. /dom/ 'lump'; /sesa,y/ 'strand'; /daem/ 'long thin thing') occur. In addition to the increased use of numeral coefficients in connection with counting, however, there is also a different construction which occurs in modern literature, involving the use of numeral coefficients when counting is not taking place, e.g. (ring circle this) 'this ring' /prEeh (revered king body this) 'this king'. In such constructions, reminiscent of Thai, the numeral coefficients, here and are behaving as true classifiers. 3. Features which suggest conscious or unconscious imitation of French Slight changes in grammatical usage which seem to be due to French influence are illustrated below under four heads: (i) the use of' abstract nouns in preference to verbs; (ii) the attempt to express some nuances contained in the meanings of various verbal forms in French; (iii) changes of Cambodian word-order in imitation of the French. Under (iv), the practice of word-for-word translation of French turns of phrase is illustrated. (i) The use of cibstract nouns in preference to verbs The Khmer natural idiom uses a verb rather than a noun whenever possible. When an abstract noun is essential Khmer has its own ways, based on verbs, of supplying the need. Nouns formed by infixation of the verb are still in use, e.g. /komhoh/ 'a wrong' < /khoh/ 'to be wrong'. Sometimes an abstract noun is produced by juxtaposing two verbs of' opposite meaning, e.g. /tetu:el khoh-tro:v/ (lit. 'to accept wrong-right') 'to take responsibility for'. Another means is to form a noun by placing the word /ka:/ 'action' or /?omp:/ 'activity' or /seckdyY/ 28

PAGE 41

'matter' before a verb or attributive verb, e.g. /ka: pUe!Jrl::k/ (lit. 'matter open up') 'development'. However, many abstract nouns were nevertheless borrowed from French and have now been carefully replaced6 and supplemented by new Indian borrowings. Also of importance is the position of an abstract noun in a natural Khmer sentence. It seems that when abstract nouns are used in Khmer, they do not usually occur, except in philosophical discourse, as the subject of the sentence. Sentences in modern writing in which new nouns occur often seem very un-Cambodian and, especially if nouns do occur as sentence subjects, suggest French influence to me, e.g. /kilmnut nm!J sakamephl::ep nl:h pilm ml::en presthiphl::ep (idea and action these not have effectiveness mu:ey ro:y phl::ek ro:y) one hundred part hundred) 'These ideas and activities were not one hundred percent effective'. /sa:rephl::ep srok yY:!J d:e sa:rephl::ep srok vl::el tilmnl::ep/ (truth country we is truth country plain low) 'The essential nature of our country is that of a low plain'/ 'Our country is essentially a low-lying plain'. (ii) The attempt to express some 'nuances' contained in the meanings of various verbaL forms in French (a) /daoy/ This word may occur as a clause-marker (conjunction) .. and, if so, the unexpressed subject of the clause is always the same as that of the main clause. In everyday usage it occurs with the meaning 'through the fact (that); because (of)'. In novels, /daoy/ has been observed in occurrences such as these but also in contexts where it seems to indicate that the action of the verb in the clause it introduces takes place simultaneously with the action of the main verb and where no cause is given, e.g. /kraok cho: khrl (get-up stand move prgeh-ka:y-vl:ka: royal-gesture ce.J:! ch!Ja:y .. daoy off afar, (through) s1op-khpif:m despise) somdaelJ show 'He stood up and moved away some distance, making a gesture of contempt. 1 A native Khmer way of expressing simultaneous action which might have been used is to place the word /domnae/ after each of the two verbs. In the example, however, the Khmer sentence seems to imitate the abi.lity of French to bring the present participle into play ('faisant' or 'en faisant'). 29

PAGE 42

(b) /kompuiJ/ occurs as a pre-verbal particle meaning 1 in the middle of verb-ing' and precedes a main verb of action in normal everyday usage. In novels it has been found in occurrences where it seems to be added in as an extra word, e.g. /pratah phn:k nili!J sat-chlu:h mu:ay kompu!J cho: s:t: smau/ (meet eyes with mouse-deer in-the-middle-of standing eat grass) 'Their eyes lit upon a mouse-deer standing grazing'. Here /kompu!J/ occurs unusually before a verb used attributively in close junction with a noun, where the same phrase without /kompu!J/ would be normal in Khmer and where French might have a present participle or a relative clause with imperfect tense. (c) Khmer sentences, spoken or written, tend to depend as far as possible on context and meaning to indicate relative time, and can proceed with a minimum use of words such as /ba:n/ (past), /nili!J/ (future) and /haay/ (completion). Time relationships between two clauses may not need any clause marker, such as /kraoy-dael/ 'after', if any of the other indicators is present. It seems probable that it was familiarity with the precision of the various tenses in French which caused a much fuller use of indications of relative time in recent written Khmer, e.g. /kraoy-dael ba:n sdap seckdyY thlae!J cop haay, prEah-so!J ko: ba:n somdae!J seckdYy r1:k-r1:ay/ (after have listened-to matter express through-to-end already monks then have shown matter joyful) 'After they had listened through to the end of the address, the monks expressed their delight ... Since the whole context of this sentence was already known to be past, /ba:n/, which occurs in both clauses, was not needed at all. In speech /haay/ at the end of the first clause would be enough to indicate the relative time of the actions of the two clauses. The French 1apres avoir ecoute ... seems possible as a model for this. (iii) Changes of Cambodian word-order in imitation of the Freneh In spoken, or simply written Khmer, adverbs, adverbial phrases, attributive verbs modifying a main verb, and post-verbs completing the meaning of a main verb all characteristically follow the main verb and its object and occur in clause or sentence final position. In newspaper style, however, changes of word-order such as the following may take place made possible by use of the particle /n'Yu/ discussed in section l (ii). 30

PAGE 43

/cap-phdaam c1:a yU: chnam mo:k haay nu nEayo:ba:y tWk/ (establish being long years by now (particle) policy water) 'put the water policy into operation many years ago' This seems to me to reflect French 1depuis de longues annees1 placed between verb and object. The following example comes from an article in Kambujasuriya ( 1966) : /son1st baep n1h phdol ?aoy nu lathaphol do: 'A conference of this kind produces important results' (lit. 'produces for [particle] result'). Normally f?aoy/ 'to give; for' would come after the object; it suggests 'for (us, people, the participants, one)'. Perhaps the aim of the un-Khmer-like position of /?aoy/ is to avoid a construction which is so unlike French? (iv) Word-for-word translation of Iir'ench turns of phrase (a) The Khmer language has its own metaphoric vocabulary, some of which uses the same imagery as French or English, for example, in associating the ideas of heat or fire with anger. In the following examples, from newspapers, however, the metaphors of French idiom are used rather than those which the Khmer language offers: /rtimcu:al harmtey cr1:al-cru/ (agitated royal-feelings manner deep) 'deeply disturbed'. Here Khmer idiom would have 1 strong' instead of 'deep 1 b;n-:l:ya:ka:h r1:k-r1:ay/ (in atmosphere joyful) 'in a happy atmosphere'. This use of the new Khmer loanword /bor1ya:ka:h/ metaphorically is entirely due to French. (b) Many instances may be found in newspapers of direct translation of the complete French phrase: /km: tha:/ (that-is to-say) 1c1est a dire'. /km:/ alone rather than /km: tha:/ would seem more natural judging from older Khmer prose. /m1:an ka: p1saot/ (have matter experiment) 'avoir l 1 experience 1 Use of the simple verb /pl:saot/ rather than the expression would seem more idiomatic in Khmer. /baa ?aoka:h hoc ?aoy/ (if opportunity pass-across for) 1 si l 1 occasion se presente 1 /hoc/ normally has an obj.ect following it, even in a metaphorical use such as in the Khmer phrase /hoc khlu:an ?aoy praa/ (pass-across self so-that use-services-of) 'offer one's services'. 31

PAGE 44

/nl:poan daoy/ (compose through) 1ecrit) par'. This phrase, seeming to be a translation from French, now appears on the title page of books instead of the older Cambodian format: Title of work: Author's name: (prepared). sthyt nYu/ (in-the-middle-Of be-stationed remain) I Se trouve actuellement'. which occurs characteristically in Khmer before verbs of action rather than, as here, before a stative verb, seems unnecessary in any case in connotation with both /sthvt/ and /nvu/, either of which means 1be situated'. I suspect is there because of familiarity with French 1actuellement'. /ml::an karanyYakec (have duty will fulfil) 'avoir un devoir a remplir'. The Khmer form would be /tro:v karanYyakec/ 'must fulfil a duty'. 4. Features which seem to reflect the desire to embellish, using the traditional devices of Khmer literature These features are discussed under three heads: (i) elevation of style; (ii) reduplication and repetition of ideas; and (iii) expansion and variation. (i) Elevation of style Several features which have been discussed in preceding sections help to elevate the style; for example, the Thai-inspired use of numeral coefficients mentioned in section 2. Here, however, two specific means are considered: the choice of vocabulary and the use of particles preceding attributive verbs. (a) Four waves of loanwords from Sanskrit and/or Pali have come into the language to fill gaps: legal, religious, and political terms in the pre-Angkor and Angkor periods; Pali Buddhist terms, chiefly from the eleventh century onwards; the bulk of the royal vocabulary from sometime after the end of the .Angk.or period onwards; and the new vocabulary during the last few decades. The 'high' language of poetry has been drawn from the first three of these categories. Terms from all four kinds of borrowing are characteristically to be found in modern prose,causing the style to differ profoundly from that of the spoken language.7 (b) /do:/ In Khmer an attributive verb follows immediately the noun it modifies, in the normal structure of colloquial and informal language, e.g. /ml::an phtEah thom/ 'had a big house'. In the literary language, however, /do:/ has long had the role of preceding an attributive verb and attaching it to the noun. It is still used and has the effect of highlighting the attribute: e.g. /a:rl:yaprate:h mu:ay do (civilised country one particle shining) 'a magnificent civilization' 32

PAGE 45

In newspapers, however, its use is extended so as to attach other categories of word to the nouns: /thlaeiJ ?om.no: kiln do: kray lE: 1J/ (express joy (at) good-deed particle very very) 'express (my) great gratitude' /lathaphol cl:a t1: koap cYt/ (result particle being focus suit feelings) 1 a satisfactory outcome' In the fist of these examples, an adverb /kray lE:!J/, and in the second, a phrase /d ;,a t1: koap cYt/, are treated in the same way as attributive verbs. Another extension of the use of /do:/ in journalese is to allow it to attach two attributes to the noun: /?oh ka:l do: yu nm!J do: l1lmba:k/ (whole time particle long and particle difficult) 'during this long and difficult period' /do:/ is not really necessary as a clarifier of the construction even in the long phrases of the literary language; it is an embellishment used particularly in descriptive passages in novels or flowery parts of public speeches. /ya:IJ/ This word, meaning 'way', is in competition with /do:/, now as a means to attach attribute to noun: e.g. /tatu:al para:cey ?a:mah/ 'suffered a humiliating defeat' Unlike /do:/, however, it occurs also when an attributive verb occurs post-verbally (usually to be translated into English by an adverb ) e g. /ka: ta:n-tYIJ ba:n kaat ml:an laaiJ ya:IJ khlaJJ-kia:/ (matter tense has risen is-there rising way strong) 'tension has arisen to an extreme degree' Khrner grammar does not require the presence of any particle before such an attributive verb and it would not be present normally in informal and colloquial speech (cf. /kmt seckdYy trco:v/ 'think the matter out coY'rectly'). The particles /nu/ and /ney/, discussed in section l as having a clarifying role, also, like /do:/ and /ya!J/, contribute bB their mere presence to the literary flavour of a sentence. (ii) Reduplication and repetition of ideas As is well known, Khrner is a language in which reduplication occurs structurally at the levels of phonology,9 morphology,lO and syntax.ll Repetition of ideas occurs in the formation of compounds.l2 Both exact reduplication using 33

PAGE 46

phonaesthetic words, sometimes specially created, and repetition of ideas using poetic vocabulary,l3 occur as literary devices have done so since the beginning of Khmer poetic composition.l The following examples, taken from modern prose, merely illustrate the use of the same devices in modern writing. (a) RedupZication /rillt-tae chill: (increasingly ill strong up strong up) 'became more and more seriously ill' could have occurred without reduplication. /ci:k kugJ ci:k rillh su:e/ (dig tree-stump dig roots ask) 'make a thorough investigation' /ci:k kuel rmh/ could have occurred without interpolated reduplication. /cl:e nee Cl:e ka:l/ 'always' This use of interpolated reduplication, based on a compound loanword (Pali: produces an understandable phrase because both components of the compound are familiar to Khmer readers as separate words. Word-play exercised on Indian loanwords is used as a poetic embellishment. (b) Repetition of ideas The examples under this heading are from both novels and newspapers: /ci:e ?Ula:rillk ?athillk-?athillk/ (being grand splendid) 'in a very grand manner' /ci:e/ usually precedes one attributive verb in this muchused pattern for forming clause final adverbial phrases with the meaning 'in a --manner' /nu-tae prekaek tetu:el nu-laey/ still refused not-yet accept agree still) 'still staunchly refused' The whole meaning is in effect said twice in different ways, very much in the poetic tradition, particularly of the nineteenth century. A certain means of expression, which I have noticed particularly in the written language, and of which I have found instances in Middle Khmer poetry, consists of following up the main verb with a short clause which merely repeats the intention expressed by the verb in another way: e.g. 34

PAGE 47

/prap ke: ... 'aoy 'tell someone ... that he may know' /s-;J: ... 'aoy khY:J1/ 1indicate ... for (one) to see' /bqpceJ1 (e.g. tUik cYt kla:ha:n) ... 'aoy ke: khY::J1/ 'show (e.g. courage) ... for them' /bqpcek ... 'aoy ke: 'make clear ... so that people will know' =J1 . 1 aoy y'Y: kh :J1 chbah/ 1 demonstrate ... so that we shall clearly see' /ba:n seckdyY C-;Jmraen ... do:c pra:thna:/ 'achieve success ... as was their wish' /t'Yu: d-;Jl. .. do:c I arrlve ... as they desired' This kind of clause seems to occur., in journalese especially, rather more often than the contexts warrant. However, it may, in some cases, seem more necessary to Khmers than it does in the translation because the clause fills out the meaning of the preceding verb, which, on its own in Khmer, has a limited meaning, being able to convey neither the sense of finiteness nor the ideas of completion or result which one can convey in English or French by use of the past tense. (iii) Expansion and variation The simple forms of everyday speech or writing are constantly expanded and varied in the formal, literary style. Some examples are given (see next page). Traditional poetic forms which parallel these examples are to be found, e. g. /s-;Jh sa: I 'utterly' (pro/s-;Jh/ l; /n1ih ney/ 'that, those' (pro/nUll/), and /hak byY ru:/ 'as though' (pro/do :e-el: :e).) 35

PAGE 48

Simple form /mu:ey tiet/ (one further) tiet/ (one way further) /mllin tae (not only like that) /bontec/ (a little, soon) pe:l yti: nuh/ (in time long past coming along that) (also) (side) (as far as) /tae/ (only) /ba:n ci:e/,(get to be) . pi:prueh 'Moreover, ... 'And another thing: ... 1 'Furthermore ... 1 'Soon ... 1 1 In the distant past ... 1 'and also' 'in the direction of' 'until' 'only ... 'The reason why ... was because' Expanded or varied form /mu:ey tiet saot/ +moreover) tiet/( +again, now) /mmn tae ... tiet/ (+join also ... further) pi: pe:l 1nuh (next from time that a little coming-along) ?adtka:l yti: haey nuh/ (in past-time coming-along long long already) (join also) /thaem (add also) /phnaek (section side) /dera:p (all the way come as-far-as) / trYm tae/ (up-to-the point only) /haet-dael ... kill: daoy/ (cause which ... was through ... )

PAGE 49

Simple form /haet n1h haay ba:n c1:a/ (cause this completed get to be) /c-:>rJ/ (wish) /nah/ (very) /s-:>h laay/ (particles) /mmn kha:n/ (not miss) 'This was the reason why' 'wish' 'very' 'at all' 'without fail 1 Expanded or VCJX'ied form /daoy ?a:sray haet n1h haay t:p m1:an/ (through depend-on cause this completed so have) /m1:an b-:>mno:rJ corJ/ (have desire wish) /pe:knah/ (too-much very) /kray le:rJ/ (very let-go) /kray pe:k/ (very too-much) /ya:rJ kray le:rJ/ (way very let-go) /?aoy toal-tae s-:>h/ (give so-far-as at all) /daoy kha:n mmn ba:n/ (through miss not possible)

PAGE 50

A further kind of expansion takes place in the use of some idioms which have already been formed on a literary pattern and.which take part in a larger grammatical construction. Two examples are: saen kho: khu/ (way 100,000 aggressive) 'in an extremely aggressive manner' r2>ep satav2>et/ (since count centuries) 'for some. centuries' These are both based on idiomatic phrases which occur in literature. /saen/ 1100,0001 functions in literature as a particle 'very' modifying a following attributive verb, here /kho:-khYU/. (Khmer particles with this function would occur after the attributive verb; /saen/ as a numeral would normally occur preceding a numeral coefficient, not a verb.) In the phrase saen kho: khu/ the expansion consists in taking this instance of an established literary idiom and using it in place of the attributive verb which usually, as was shown in 4(i), follows In the second phrase, /raep/ 'to count' occurs idiomatically with several words in a slightly formal language style, e.g. /raep ra:y/ 1in hundreds'. Here this idiomatic adverbial form is used after a pre-nominal particle instead of the noun, which is expected in that position (e.g. chnam nUb/ 'since that year'.) Some features described in the foregoing sections, particularly the use of particles, of time indicators, and of corroborative nuclei when there is no need for them, seem to confirm that there is a tendency to wordiness in Modern Khmer written style. With regard to the influence of French, examples have shown that, ironically, although loanwords from French have, in principle, been discarded, some aspects of French grammar have been incorporated! In presenting together all the points which I have observed, I may have given the impression that all modern writing of news, articles, speeches and novels is packed with these features. In fact, however, many writers of articles have used a restrained style and many novelists have used a very simple style even in descriptive passages; one for example, is Dik Keam in his novel tae (1967). In any case, much of a novel is concerned with narrative and conversation and, for the presentation of conversation, a good tradition of writing both naturally and interestingly seems to have been built up by Khmer novelists, stemming perhaps from the lively little conversation passages which are found in the verse novels (/satra: In fact, for a thorough assessment ofa style one would have to separate the genres of writing which have been treated together here. Although it has been interesting for me to attempt to analyse recent changes in style, I personally prefer the simple, restrained 38

PAGE 51

one. However, perhaps foreigners should confine themselves to factual description and not presume to judge style qualitatively. From a Khmer point of view, as section 4 shows, the long-established literary tradition, which is the tradition of poetry, has encouraged embellishment by the use of long loanwords and of reduplicative and repetitive patterns, all of which produce more words. NOTES 1. The writing of prose in Khmer was traditionally intended for practical reasons only. Literature for artistic purposes was always composed as poetry; at first, from the seventh to the thirteenth centuries, in Sanskrit. 2. My source for this has been a photographed manuscript of the Chronicles in the library of the School of Oriental and African Studies. Some historical writing is easily available in print, however, in Huffman (1977: 38-78). 3. Reproduced in print in the early volumes of the Frajum breh khmaer. 4. The transcription used here was based on Professor Henderson's pioneering analysis of Khmer pronunciation (Henderson 1952). 5. Terms for grammatical categories are as given in Jacob (1968). 6. This is not to say that French loans for which new vocabulary was invented went out of use completely. They are still heard. 7. Many Indian loans are so completely naturalized that their presence in the colloquial language passes unnoticed, e.g. /bon/ 'festival' < 'good work'. 8. /nu/, /ney/ and /da:/ were briefly treated in Jacob (1978) together with /rl:/ as literary particles. /rl:/ does not seem to occur so much in recent writing. 9. That is, in words having re.dupli.cated initial consonants which are no longer, or have never been, analysable as consisting of base and reduplicating prefix, e.g. /beba:/ 'soup', /tetu:el/ 'receive'. 10. Reduplicative prefixes consist chiefly of single consonants (e.g. the frequentative /k(e)ka:y/ from /ka:y/ 'to dig into the earth with hands, paws') but include some examples of rhotized consonants (e.g. /tr(e)tuen/ which occurs in 39

PAGE 52

/tr(a)te:h-tr(a)tuan/ 'graceful' from /tuan/ 'soft, supple') and of consonants followed by a nasal consonant (e.g. /s(a)nsaam/ 'dew' from /saam/ 'wet'). Reduplicative compounds are formed by the juxtaposition of alliterative, rhyming, chiming, or repeated word-forms (e.g. 'to organize' /ri:ay-mi:ey/ 'untidily scattered'); 'children (in quantity)'; /proh-proh/ 'the menfolk, the boys'. ll. Reduplication in syntax int,.ensifies meaning, e. g. /craan laalJ craan laalJ/ 1 increasing in quantity 1 is more emphatic than /craan laalJ/. Reduplication with interpolation also occurs, especially with components of compounds; e.g. /prap ke: prap 'aelJ/ 'tell everybody', pro /prap ke:-'aelJ/ id. The reduplicated form is more stylish. l2. Frequently a compound with specialized meaning or with abstract meaning has been formed of components, each of which has, when used alone, approximately the same meaning, e.g. / ciah-vi: alJ/ 1 to avoid 1 from / ciah/ 1 to dodge 1 and /vial;)/ 'to follow a winding course, go round (an object)'. l3. The following are poetic examples /thla: thlaelJ co:-ca: prap/ (explain converse tell); /kru:alJ kra:p 'aphivi:at (shrink prostrate-oneself salute salute) 'greet respectfully'. l4. By the nineteenth century this poetic device began to be overdone. See Jacob (l979). Dik Keam Henderson, E.J.A. Huffman, F.E. & Im Proum Jacob, J.M. REFERENCES l967. Broh tae kamnaen ( 1 Because of the conscription'): Phnom Penh: Seng Nguon Huot. l952. The main features of Cambodian pronunciation. Bull. Sch. Orient. Afr. Stud. 14, l49-74. l977. Cambodian literary reader and glossary (Yale ser.). New Haven: Yale Univ. Press. l968. Introduction to Cambodian. London: Oxford Univ. Press. l978. Some observations on Khmer verbal usages. Mon-Khmer Stud. VII, 95-l09. 40

PAGE 53

Jacob, J.M. Kambujasuriya 1979. Observations on the uses of reduplication as a poetic device in Khmer. In Studies in Tai and MonKhmer phonetics and phonoZogy: in honour of Eugenie J.A. Henderson. (eds.) Theraphan L. Thongkum et aZ. Bangkok: Chulalongkorn Univ. Press, 111-30. 1927-. Phnom Penh: Inst. Bouddhique. breh khmaer 1959-74. ReceuiZ des contes et Zegendes cambodgiens. krum j 8.J!lll11J!l damlap' khmaer. Commission des moeurs et coutumes du Cambodge. Phnom Penh: Inst. Bouddhique, 8 vols. Suttantaprija In 1936. GatiZok. Phnom Penh: Inst. Bouddhique, vols. 1-10. 41

PAGE 55

KHASI KINSHIP TERMINOLOGY1 tLili Rabel-Heymann Irawati Karve, the acknowledged authority on Indian kinship, who admits to unfamiliarity with the Khasi language, lists twelve Khasi words in her study of kinship (Karve 1965), but none of these words are included in U.R. Ehrenfels (1953) paper, nor are they recorded in the four Khasi dictionaries {Nissor Singh 1906, 1920; Leemuel 1965 (= Diengdoh); and Kharkhongngor 1968) known tc me. Since these same twelve words were also never mentioned by any of my Khasi informants, they should probably be considered as non-existent in the Khasi language; Karve's erroneous listings may be understandable, however, as her only sources were Roberts (1891), Grierson (1904), and Gurdon (1914). U .R. Ehrenfels' article (1953) is, therefore, the only modern treatment of Khasi kinship terminology; it had apparently not come to Karve's attention. Ehrenfels includes an almost complete list of kinship terms in four dialects --Plateau Khasi, that is, the standard language; War Khasi; Pnar (or Jaintia); and War Jaintia -indexed under 43 English glosses, the purpose of which was one of comparison. Although the author collected the vocabulary. in an actual field situation, aided by native interpreters, his spellings are not always reliable and the literal meaning of most terms remains hidden from the reader unfamiliar with the language. I, therefore, see my task as that of amplifying Ehrenfels' work, of correcting some errors, especially in the orthography, and of explaining the underlying morphemes occurring in the incredibly complicated Khasi system.2 Khasi kinship terminology is based on three principles: (1) a differentiation between blood relatives and relatives by marriage; (2) classification by relative age of each member with regard to the person they are related to; and {3) distinction between terms of address and terms of reference. Khasi society is generally considered to be matrilineal and matrilocal; the kinship system could be called 'bifurcate merging' since mother is equated with mother's sister, father with father's brother, while mother's brother and father's sister are denoted by distinct terms. Ancestry is traced through the 43

PAGE 56

mother's clan, ka kur, often used as an 1imitative'3 ki kur ki kmie (lit. 'the clans the mothers'). Gender number4 morphemes (called 'prefixes' by the Khasis) must precede each noun; ka indicates feminine, u masculine singular; ki is used for plural number and i for resp;ct, endearment, and smallness, the latter two not distinguishing gender. Upon marriage, the husband remains in his mother's clan while his children belong to his wife's clan. kur is used in three phrases: iadei kur 'to have a relationship on the mother's side, to be related within the same clan'; in the compound para kur 'children of mother's siblings', and in the verb tait kur Wt jaid,5 an imitative 'to be banished from the clan 1 (lit. 'reject clan, reject kind, caste'). The most respected member in the clan is the mother's elder brother, u kfii or kfii rangbah (rang-bah translates as 'adult male, an elder' and is composed of rang, the combining form of shynrang 'man' and bah 'older brother', probably based on the verb bah 'to be big,-)-.-The kffi is addressed as mama or ma, a word which seems to be of Indo-European (IE) origin since Hindi, Bengali, and Assamese use it for 'maternal uncle' as well. The is consulted on all important decisions and acts as the ultimate arbiter in disputes. Mother 1 s other brothers are referred to as kfli pdeng 'middle brother 1 and krli khadduh6 'mother's youngest brother; the very last' respectively. The latter two uncles are addressed as ma-dengl and ma-khadduh or ma-duh; duh 'the last, the youngest' is used for blood relatives only, never for those related by marriage. The word for 'mother' has two forms which are probably not related morphologically (see discussion below). kmie is used for reference and mei for address; mother's sisters are also addressed as mei plus the appropriate modifier for ageranking. Father is referred to as i kpa and addressed as or a term strangely familiar to speakers of Indo-European languages. His brothers, as well as the husbands of his sisters and of his mother's sisters are all referred to as 'fathers' (pa-used in compounds is explained below). Father's sisters are all addressed and referred to as kha. kha is based on a verb meaning 'to give birth', thus, according to Ehrenfels (1953: 408), apparently recognizing the father's biological function in procreation. kha also functions as the second constitutent in compounds designating father's mother, kmie-kha, and cousins on the father's side, shi para kha-shi_ is the numeral 'one' used for units and measurements (as opposed to wei 'one'); para designates brothers and sisters of one's own generation. Parallel cousins and cross-cousins are distinguished both in terminology and in marriage practices; marriage between 44

PAGE 57

parallel cousins is sang 'taboo', while marriage between crosscousins is permitted though not common. Mother's brothers' children are also referred to as ba-kha, since by definition his children will belong to their own mother's clan. All brothers and sisters within the immediate family are designated by terms that specify whether they are older or younger than the speaker; there are also terms to indicate a 'middle' brother or sister and terms for the youngest brother and sister (Table 1). The oldest sister is k0ng or kong ieit, 'sisterbeloved', the oldest brother is bah bah 'brother big', or bah rangbah 'brother grown-up man', terms that show a position of respect occupied by the elder siblings. There are actually two homophonous morphemes bah, one meaning 'brother', the other 'to be big'; and since modifiers follow the noun in Khasi, kong ieit would have to be translated as 'sister who is beloved' and bah rangbah as 'brother who is grown-up.' Unfortunately, Table 1 shows some gaps. Also, no two of the young Khasi speakers who recently supplied me with information agree on all terms; they have all been living abroad for a long time and have become accustomed to our simplified Western terminology using aunt, uncle, One speaker also suggested that address by name is coming into vogue among the younger generation. khynnah 'child' is used for the youngest brother, bah khynnah 'kid brother'; i rit (lit. 'little one') and i duh (lit. 'the last one') are best rendered by 'kid sister'. One word glosses for hep and hynmen are difficult to suggest; older people, even non-relatives, can address young people as hep; it is a term of endearment and is roughly equivalent to the American usage of 'sonny' or 1ki"d1 when used by men for little boys, or 1dear1 when used by old ladies for younger women. The morpheme occurs in several other Khasi words that relate to 'time past, ago': folk-tales always begin with hyndai-hynthai ... 'once upon a time ... '; hynne means 'a short time ago', hynnin 'yesterday', so that Rynmen could perhaps be rendered by 'born before, born some time ago'. The morpheme -men occurs in tymmen 1 old man or woman 1 so that it may mean 'old human being'

PAGE 58

Sister S's husband Brother B' s wife El-der S/B* 1kong ieit'** kong, hynmen kynthei 1hynmen kynsi' kong heh, hynmen kynsi (by female) kyn-um (by male) 'bah, bah bah, bah heh1 1bah rangbah' hynmen shynrang 1kong kynsi1 Middl-e S/B kong-deng kong-deng 'bah khynnah1 Younger S/B 1hep1 hep 'para kynsi' 1hep kynsi' (by female) hep (by female) The Youngest i rit, i duh 'bah duh' Abbreviations used here and in following tables are: B ** Quotation marks indicate terms of address Brother, F Father, M Mother, S Sister Table 1: Khasi sibl-ing of address and of reference

PAGE 59

kynsi and kyn-um are best translated by 'in-law'; the minor syllableS kyn-_occurs in so many words, nouns and verbs alike, that it is impossible to assign a definite lexical meaning to it.9 Ego's, i.e. the speaker1s,maternal and paternal aunts and uncles are, as Table 2 shows, classified according to the same system as siblings; older or younger than parent referred to, in-between the older and the younger aunt or uncle, and a designation for the youngest aunt or uncle. Mother's and father's in-laws take their age-ranking appellations from their respective spouses rather than from their actual age. Terms of address and terms of reference are usually identical in the second and third ascending generation (grandparents and greatgrandparents), and for the first descending generation (children, nephews, and nieces): terms are, however, as Tables 1 and 2 demonstrate, differentiated for ego's siblings and for ego's parents and their siblings. The term of address is often a shortened or reduced variant of the full term which is used for reference; the shortened variant, also used in compounds, is derived by loss of the initial consonant (usually k-), or loss of the initial syllable in bisyllabic words. Examples are: 12.:_ VS. kpa 1 father 1 rad VS. kynrad 1 lord, master 1 mei VS. kmie 'mother 1 -rang vs. shynrang 'adult male 1 10 Most family members, it will be noticed, are not addressed or referred to by name but by their kin classification; even husband and wife have traditionally referred to each other as 'the mother (of) [oldest child's nameJ' e.g. i kmie u Dan, and i kpa u Dan 'the father (of) Dan'. One woman informant told me that a woman can also address her husband's sister's husband, i.e. her brother-in-law, as the kpa of the first-born child. Husband and wife address each other by phi, the polite second person pronoun 'you'. Younger Khasis state that nowadays husband and wife may use names for addressing each other. Another Khasi friend gave me the terms of ka lok for 'wife' and u lok for 'husband', but a young man said: 'lok is a harsh word, don't use it. I The reference terms for parents-in-law, kiaw for fmother -in-law' and kthaw for 'father-in-law', do not distinguish between maternal and paternal ancestry. The terminology for grandparents is structured parallel to that of one's own parents; however, distinctions between terms of address and of reference are not as varied. 'Mother's mother' is mei-rad, 'mother's father' is pa-rad; their respective siblings add -heh for the older sister, -deng for the middle sister, but the youngest is simply another mei-rad. Great-grandparents on the mother's side are mei-buh and pa-buh;ll father's mother and father are kmie-kha and pa-kha respectively. Referential terms for parents in all generations take the respectful 'prefix/article 1 .i 47

PAGE 60

.!="" ()) Mother* 'mei' i kmie Older than M/F Middle S/B Younger than M The Youngest M's Sisters mei-san mei-deng mei khynnah nah rit M's Ss' Husbands pa-s an pa khynnah pa khynnah pa khynnah M' s Brothers 'mama rangbah' 'ma-deng' 'ma, mama' 'ma khadduh' kfli rangbah ld'!i pdeng kfii kfii khadduh M's Bs' Wives f'lia f'lia i'iia flia Father 'pa, papa' i kpa F's Brothers pa san pa-deng pa khynnah pa-duh F's Bs' Wives nah nah nah nah F's Sisters kha rangbah kha-deng kha-duh F's Ss' Husbands pa-kha? mama? mama? mama? For the unfilled slots definitive information is lacking Quotation marks indicate terms of address; terms of reference have no quotes. When only one term is listed, address and reference are the same Khasi usage of hyphens is not systematic; my own usage is to hyphenate 'reduced' morphemes Table 2: Khasi terms for maternal and paternal aunts and uncles

PAGE 61

Parents refer to and address their children's spouses as the term also used for maternal nephews and nieces, pyrsa kurim (kurim 'wife'); uncles and great-uncles on the mother's side refer to a child as pyrsa ksiew while the child addresses his great-uncle as bah. ksiew, or khlin ksiew, are the terms for 'grandchild', and for 'great-grandchild'. Great-grandchildren are referred to as khun miaw (lit. 'cat children'), and great-great-grandchildren as khlin khnai (lit. 'mouse children'). Step-children are referred to as khlin ruid /khuon ruj/ and'' stepfather' is u kpa nah according to Nissor Singh (1906: 38, 45). Unfortunately, no sources of information, informants or dictionaries, could provide a literal meaning for tun or ruid. Kinship terms are generally assumed to be of native stock, along with body parts and numerals. However, Khasi has borrowed extensively from the geographically surrounding languages with many loans so well integrated into the native sound structure that their detection is dificult, if not impossible. Three words may be of Indo-Aryan derivation: (i) above), but a either indicates a very early borrowing or would be counter-indicative; (ii) para (see above) is listed by Karve (1965) under Hindi, Sindhi, and Punjabi as referring to blood-related maternal/ paternal grandfathers, while in Khasi this word is used for younger relatives on the mother's side; and (iii) kurim 'wife', has an unchecked long vowel in the first syllable which, according to my earlier findings, points to Indo-Aryan origin; kurim 'wife' and kur 'clan' are definitely not related. GLOSSARY The following is a complete inventory of all kinship terms with which my informants supplied me, many of which are listed in Nissor Singh's famous dictionary (1906).12 The alphabetical order is that commonly used for European languages; it deviates from the established Khasi alphabet in three respects: the aspirated stops kh, and th are treated as separate phonemes and are, therefore, not arranged within the listings; kh are listed after and not after as in the Khasi alphabet; instead of taking the place of bah to be big; u bah 'big brother'; a polite form of address for any man older than speaker /ba?/ bah bah, bah heh, bah rangbah 'older brother' /ba?he?/, bah duh, bah khynnah 'youngest brother' /ba?du?/, /ba? khnna?/

PAGE 62

ba-kha children of mother's brothers; relationship between mother's children and maternal uncle's children (Nissor Singh (1906:6), also includes 'paternal aunt' (qv. also Kharkhongngor 1968: lOb) bih form of address for a young girl (bi?/ bu affectionate form of address for a young boy (not common) deng 'in the middle, between', used in compounds /pderJ/ kni pdeng referring to mother's middle brother kha-deng addressing father's middle sister ma-deng addressing mother's middle brother mei-deng addressing mother's middle sister pa-deng addressing father's middle brother duh from khadduh 'the last one, the youngest' /khat-du?/ khadduh 'mother's youngest brother's wife' /na-khatdu?/ i duh 'the youngest sister, baby sister' heh 'big' hep a polite way of addressing and referring to anybody younger than the speaker; used for younger siblings; woman addressing and referring to brother-in-law if married to younger sister; same as kong hep kynsi,woman referring to younger sister's husband /knsi/ hynmen i referring to elder sister /hnmen/ u referring to elder brother hynmen hynbew (imit.), elder brother or sister /hnbew/ hynmen kynsi younger sister addressing older sister's husband hynmen kynthei referring to elder sister hynmen shynrang referring to elder brother /hnmen snraTJ/ ieit 'to love, to be loved' /?ret/ mei ieit addressing mother's mother pa ieit addressing mother's father kong ieit addressing elder sister ing, iing, 'house' ka iing ka sem household, family (ka sem 'stable, shed, shelter') /ka yrerJ ka sem/ (imit.) kiaw mother-in-law 50

PAGE 63

kmie 'mother' when referred to (reduced form i mei) /kmi/ kmie hep referring to father's mother ki kmie ki kpa 'parents' (imit.) kmie kha referring to father's mother kmie-nah referring to mother's younger sister; stepmother also: i mei-nah khadduh /?i mey-na? khat-du?/ kmie-rad referring to mother's mother /kmi-raat/ kmie-san referring to mother's elder sister kni i mother's eldest brother (=kfii rangbah, p.44 above), the most respected person in the clan; he is addressed as ma or mama ki ki kpa relatives on mother's side (imit.) knia ka, i mother's brother's wife; in compounds i'l.ia /kiia/ nia-kha addressing and referring to father's sisters nia-kha rangbah father's oldest sister nia-khadduh father's youngest sister /na-khat-du?/ nia-pdeng father's middle sister kong ka, i addressing and referring to older sister; man addressing wife's female relatives; polite form of address for any woman older than speaker; in compounds also used for males kong-deng husband of middle sister, brother-in-law kong heh woman addressing and referring to brother-in-law kong kynsi addressing older sibling's spouse kpa i referring to 'father', addressed and in most compounds kpa-nah referring to step-father, nah from khynnah kpa-rad referring to father's father; /raat/ from knraat/ 'lord' pa-buh i great-grandfather on mother's side pa-deng father's middle brother pa-ieit addressing mother's father /pa-?Iet/ pa-kha referring to father's father (cf. Nissor Singh 1906: 147; ''pakha, male relative (father's side)') pa-khynnah addressing father's youngest brother and mother's youngest sister's husband pa-rad i addressing father's father, mother's father /pa-raat/ pa-san addressing father's older brother, also mother's older sister's husband ksiew ka, i grandchild /ksiw/ ksiew tun great-grandchild khun ksiew i referring to grandchild; is addressed by name para ksiew mother's sister's son (grand-nephew) 51

PAGE 64

pyrsa ksiew man referring to sibling's grandchildren; a grand-niece or grand-nephew on mother's side kthaw addressing and referring to father-in-law /kthaaw/ kur ka clan; ka kur ka jaid* (imit. ), a relative on mother's --sid-;-/ka j aj I ki kur ki karo, ki kur ki kmie (imit. ), considered obsolete iadei kur to be related within the same clan (lit. 'come together (in) clan') /yadey/ para kur member of the same clan, children of mother's siblings tait kur tait kmie (imit.), obsolete, and tait kur tait jaid*(imit.), to be banished, excommunicated from the clan; /taj/ from /kntaj/ 'to reject, set aside' kurim ka wife, spouse; to have intercourse (probably of Indo-European origin) kiaw kurim ka mother-in-law (Singh 1906:40) shong kurim to marry 'lie wi tJ} wife') shong kurim shong kupai ( imi t. ) I, = ? from Hindi ? ) jingshong kurim marriage (lit. 'lie with wife') addressing and referring to brother-or sister-in-law hep kynsi = para kynsi woman referring to younger sister's husband hynmen kynsi woman referring to elder sister's husband kong kynsi woman addressing elder sibling's spouse kyn-um i man addressing and referring to sister's husband /kn?um/ kha to give birth, to bring forth kha deng father's middle sister kha-duh father's youngest sister kha-rangbah father's older sister kmie-kha referring to father's mother; she is addressed as mei-kha para kha (shi) father's brother's children; cousins /si/ 'one 1 (numeral) khadduh to be last /khat-du?/ see duh kni khadduh referring to mother's youngest brother ma-khadduh, ma-duh addressing mother's youngest brother pa-duh addressing father's youngest brother From Hindi, see n.5. (Ed.) 52

PAGE 65

khlin ka, i child, girl, boy, baby /khuon/ khun ksiew grandchild (maternal uncle's child) khlin khnai great-great-grandchild (lit. 'mouse child') /khuon khnaay/ khlin miaw great-grandchild (lit. 1 cat child 1 ) khlin ruid step-child /khuon ruj/ khynnah ka, i to be young; girl, boy /khnna?/ pa-khynnah addressing and referring to father's younger brother, also to mother's younger sister's husband (cf. p.45 khynraw referring to a young man ka referringto a young W()man u-khynraw u samla ( t. ) (arch. ) lok friend (arch,) ka lok wife u lok husband para lok friend lud to be young (arch.) = khynraw /luot/ mama, ma addressing mother's older brother who is referred to as u kfli ma-deng addressing mother's middle brother ma-khadduh, ma-duh addressing mother's youngest brother ma-Rangbah (or: addressing mother's oldest brother mei i mother, alternate form for kmie, used in all compounds mei-buh great-grandmother on mother's side mei-deng addressing mother's middle sister mei-hep respectful appellation of an older woman (Nissor Sing 1906:131) mei-ieit addressing mother's mother mei-kha addressing father's mother mei-khynnah referring to mother's younger sister mei-nah addressing mother's younger sister and father's younger brother's wife mei-rad referring to mother's mother mei-san addressing mother's oldest sister myngkew ka, i used by wife when referring to husband's older sister nah reduced form of khynnah to be small, be the youngest (used in compounds) mei-nah i addressing mother's younger sister; also father's younger brother's wife; addressing stepmother nah rit mother's youngest sister 53

PAGE 66

nia reduced form of kffia, referring to mother's brother's wife /kna/ nia kha, ka addressing father's sister nia kha referring to father's older sister nia kha khadduh referring to father's younger sister nia kha pdeng referring to father's middle sister father, reduced form of kpa, used in compounds para ka, ki children, brothers and sisters of same generation para ar kmie mother's sister's children (lit. 'children (of) two mothers') /para ?aar kmi/ para briew no relationship either by blood or by marriage; fellow human being para ksiew mother's mother's sister's son, i.e. grandnephew on mother's side para kur member of the same clan, mother's sister's children para kynsi referring to wife's younger sister and her husband para lok friend para mynshong sister's husband para shong-kha husband's sister's husband, two men marrying into the same family para trai ka, and sisters of the same parents para kha shi children of father's sisters and brothers, first cousins and half-siblings on father's side referring to one's own children and to sister's children; and nephews and nieces on mother's side pyrsa ksiew mother's mother's brother referring to speaker; i.e. great-uncle on mother's side ref-erring to speaker pyrsa kurim referring to child's spouse phi 2nd person pronoun (polite), used by husband/wife in addressing each other rangbah to be grown-up; an older respected person, an elder bah rangbah addressing older brother mama rangbah addressing mother's oldest brother (even if younger than mother) kha rangbah addressing father's oldest sister rad reduced alternate of kynrad 'master, lord' used in compounds kmie-rad, mei-rad mother's mother kpa-rad, pa-rad mother's father

PAGE 67

rit small i rit i duh the youngest sister nah rit mother's youngest sister samla marriageable young person san to grow up, grown-up mei-san mother's older sister pa-san father's older brother, mother's older sister's husband (sang taboo) tnga ka wife, when referred to (arch., 'a harsh word') u husband, when referred to trai u lord, master /?u traay/ para trai ka, i brothers and sisters of the same parents NOTES 1. This article represents a thoroughly revised version of a paper delivered at the American Oriental Society Annual General Meeting in Toronto, Canada, 11 April 1978. 2. Bowing to Khasi preference, I have decided to abandon my usual practice of using phonemic transcriptions for Khasi sounds; I use instead established Khasi spelling and only add transcriptions in those cases where the standard spelling system fails to indicate vowel length or vowel quality. Sometimes it will be necessary to separate morphemes by means of hyphens, although some of these hyphens are not used by the Khasis themselves. Certain morphological features, necessary for the reader's better understanding, will be explained at the appropriate places. 3. 'Imitatives', which I called 'redundants' in Rabel (1968), consist of two nouns whose combined meaning is equal to that of the first constituent alone; the meaning of the second constituent is often unknown to the native speaker. Some times the second word is a borrowing from Indo-European. 4. The Khasis call the four gender/number indicators 'prefixes'; I called them 'articles' in Rabel (1961). Neither term is entirely adequate since these words are free morphemes which also function as 3rd person pronouns. 55

PAGE 68

5. with a short vowel, is of Hindi origin. 6. khadduh=/khat-dU?/, is a compound of /khat/ 'to dole out' + /dU?/ 'to be last. 1 7. deng is the reduced form see n.lO below for loss of initial consonant. 8. 'minor' syllables have no full vowel nucleus, the second consonant --a liquid or nasal --functioning as the vocalic element; they are always unstressed. 9. Referring to n.lO, it may be derived from + stem) + infix -n-. 10. A theory first proposed by Pater Wilhelm Schmidt (1904) and elaborated by Henderson (1976) states that initial be considered a fossilized prefix denoting living things (kinship terms, body parts, animals, plants). Since simplification of initial clusters is otherwise never accompanied by vowel change, it seems unlikely that mei 'mother' is related to kmie. ll. Dictionaries do not list buh; one speaker pronounced this word /bu/, two others said /bu?/. 12. Khasi kinship terminology is, as we have seen, an interesting topic and remains, as this paper demonstrates, a subject of varying interpretation and discussion. Not all of the terms mentioned in this article are included in its glossary, or vice versa,nor do they correspond precisely when they are; nor, for example, do they always agree with the dictionary definitions (when these occur) of U Nissor Singh (1906) or E. Bars (1973), among others. This is, however, the most recent study of the question and makes it most intriguing and informative. (Ed.) Bars, E. Blah, V. Edingson Diengdoh, A.K. REFERENCES 1973. Khasi-Eng.lish dictionary. Shillong: Don Bosco Press. 1971. (comp.) dictionary. Chapa.la's Ang.lo-Khasi Shillong: Chapala Book Stall. 1965. ( comp. ) LeerrrueZ 1 s Ang.Z.o-Khasi pocket dictionary. Shillong: L. Harrison.

PAGE 69

Ehrenfels, U.R. Grierson, G.A. Gurdon, P.R.T. Henderson, Eugenie J.A. Karve, Irawati Kharkhongngor, U. Iarington Rabel, Lili Roberts, H. Schmidt, W. Singh, U Nissor 1953. Khasi kinship terms in four dialects. Anthropos 48, 396-412. 1904. Linguistic survey of India. Vol.II. Calcutta: Govt. Printing Office. 1914. The Khasis. London: Macmillan, 2nd ed. 1976. Vestiges of morphology in modern standard Khasi. In Austroasiatic Studies I (Oceanic linguistics spec. pubs. 13) (ed.) P.N. Jenner et al. Honolulu: Univ. Hawaii Press, 477-522. 1965. Kinship organisation in India. Bombay: Asia Publ. Ho., 2nd ed. 1968. Ka Dienshonhi (a Khasi-Khasi dictionary). Shillong: Ri Khasi Press. 1961. a language of Assam (Louisiana State Univ. Studies, Humanities ser. 10). Baton Rouge: Louisiana State Univ. Press. 1968. Redundant expressions in Khasi. In Studies in Indian linguistics (ed.) B. Krishnamurti. Annamalainagar: Centre. Adv. Study Ling., Deccan College, Poona & Annamalai Univ., 257-86. 1891. A grarrrnv.r of Khassi language. London: Kegan Paul, Trench, Trubner. 1904. GrundzUg,e einer Lautlehre der Khasi-Sprache. Mlinchen: Verlag der K. Akademie. 1906. Khasi-English dictionary. Shillong: East. Bengal & Assam Secretariat Press. 1920. English-Khasi dictionary (eds.) R.S.D. Ropmay & U H.K. Singh. Shillong: Assam Secretariat Printing Off. 57

PAGE 71

ON PROSODIC RELATIONS BETWEEN FIJIAN BASES AND VERBAL SUFFIXESl G.B. Milner From the earliest days in the study of the Fijian (FI) language the origins, functions and the degree of predictability of its verbal suffixes have bewildered all those who have tried to understand Fijian grammar, and they continue to do so. The problem in question is a suggestive example of the interaction between synchronic and diachronic factors in language and of its consequence for linguistic analysis. It will be remembered that a Fijian word-stem or base (the latter term being widely employed in the description of Austronesian languages to distinguish 'content words' from 'functors') subsumes both verbal and nominal word classes.2 It may be disyllabic (CVCV), by far the most common statistically, or trisyllabic (CVCVCV). The vowel slots are always filled, but in disyllabic bases the initial and/or the medial consonant is optional. In trisyllabic bases the medial and/or the final consonant is optional.3 There is also a small number of bases of more than three syllables.4 When a base is a verb it may occur in any one of these three standard forms. It is then said to be stative or intransitive, according to certain syntactic criteria. When a verb is followed by a monosyllabic suffix (C)V or by a disyllabic suffix (C)VCV it is said to be transitive. Recent studies, however, have questioned the applicability of terms such as 'transitivity' to this feature of Fijian grammar5 (Hockett l976: l92; Naylor l978: 405; Schutz l98l: l97-203). One of the most interesting problems in the comparative study of Austronesian languages is that on the one hand in Fijian, as in other Oceanic languages: l. The occurrence or non-occurrence of transitive suffixes is subject to certain semantic and syntactic criteria which are not yet fully understood; and 2. The consonant of a monosyllabic suffix: (c)v,6 and the first consonant of a disyllabic suffix: (C)VCV, is selected from a limited series within the total inventory of consonants (cf. Pawley l978: ll3-40). In many other members of the Austronesian family, on the other hand, and especially in Indonesian languages, cognate 59

PAGE 72

verbs may occur which, in a fairly large number of cases, show regular sound correspondences between their (non-significant) stem-final consonants and the consonants of the verbal suffixes of Fijian and other Oceanic languages. Thus, Proto-Indonesian *taijit' 'cry' corresponds to Fijian tagica7 /taiji5a/ 'cry for (i.e. so as to obtain) something'. Proto-Indonesian *davat 'reach; obtain' corresponds to Fijian rawata 'get, obtain'. Since the stem-final consonants of verbs like ProtoIndonesian *tal))it 1 and *davat are not known to have had a grammatical function and the corresponding stem-final consonants of verbs in modern Indonesian languages do not have such a function, while corresponding consonants in Fijian and other Oceanic languages occur in suffixes entering into regular grammatical relations, intriguing questions arise regarding the origins, nature and the precise functions of these features. In particular, three immediate questions which arise are: l. How can one account for the fact that these sound correspondences can be attested in a significant number,but by no means in a majority of cases? 2. Are the stem-final consonants of modern Indonesian verbs vestigial in the sense that they might be the extant reflexes of 'archaic' grammatical suffixes which have now disappeared but continue to function in Oceanic languages such as Fijian1 (cf. Dahl 1973:11). This is a question which should be asked even if it cannot be answered in the present state of our knowledge. 3. Should the verbal suffixes of Fijian be regarded as an integral part of the bases to which they may or may not be attached? That is to say, is the choice of consonant determined: (a) By the base and suffix considered as an articulated of course separable) but integral lexical entity, or: (b) By semantic and syntactic factors, that is to say, by the independently variable relations which can obtain between a verb and its potential objects or complements? In the earliest days of the study of the Fijian language, Hazlewood (1872: 32-3), in his work originally. published in 1850, after listing 'The Definite-Transitive Terminations' in two classes, states that: 60

PAGE 73

l. Those which consists of one These are, -ea, -ka, -ma, -.!!:., -ra, -ta, -va, -wa and ... Later he adds that: 2. There appears to be no certain rule to determine which termination a verb will take. This must be learned from the natives, or from the Dictionary. 3. But notwithstanding that there is no invariable rule, yet we are persuaded that they (se. the terminations) are not always used arbitrarily .. (1.) It seems to amount to a rule, that verbs formed from nouns without prefixing vaka-, shall take na for their termination ... (2.) It appears also to be a rule, that verbs of motion will take their termination; as lakova, ciciva, kadava, drodrova, ... Va here means to. It is also true that many o.ther verbs besides those o.f motio.n take va, but fo.r these perhaps there is no rule. ( 3. ) When ;;rbs reject a terminatio.n of the first o.r mo.nosyllabic class, and take one o.f the second, o.r disyllabic, they frequently have either a more intensive sense, or take a different object. Nearly a century later, Churchward in A new Fijian grammar (1941: 17-8; 71-2) speaks o.f: 'definite-transitive verbs', and he states that: different verbs take different suffixes and there seems to be no rule for determining which suffix any particular verb will take. This is also the view taken by the present writer in his Fijian grammar: There is no known rule to indicate which suffix is appropriate to what base. It is advisable therefore to learn each new base together with its correct suffix or suffixes. (Milner 1972: 27-8)9 These words, written nearly thirty years ago, must now be qualified, not only in the context of the result of subsequent study by the present writer and his colleagues which have become available in the meantime,lO but also in the light of recent attention given to the same problem in connection with the preparation of a new Fijian dictionary.ll It is necessary first to refer to Dempwolff1s (1934-9) VergZeichende LautZehre, which has for over forty years been an indispenable text in comparative Austronesian linguistics. It will be remembered that in his first monograph (Dempwolff 1934: 61

PAGE 74

27-8) he distinguishes five categories of word stems (Wortstamme). The first, which makes up 70% of his field of 1000 items. consists of those which conform to the pattern CVCVC (e.g. *laQit). Next in frequency comes word-stems of the same pattern with the addition of an optional nasal 'connector' (Nasal verbindung), hence of the pattern CV(C)CVC (*SUQSOQ, *guntiQ). They make up another 20% of the total. Another 5% consists of reduplicated items, followed by 3% made up of word-stems of more than two syllables. The remainder, approximately 1%, consists of -monosyllabic word-stems, In his second monograph (Dempwolff 1937: 125-66) he compares two Melanesian languages with his reconstructed ProtoAustronesian (PAN) word-stems, one of the two being Fijian, the vocabulary of which is examined in detail in order to arrive at regular correspondences (ibid., 126-46). He is struck by the number of irregular, as well as regular, reflexes of his proposed reconstructions in Fijian. Of particular to the problem under discussion here are the following passages: Phonetic disagreements (lit. non-agreements of sound: Lautunstimmigkeiten) (occur) especially frequently with the final consonants of Fijian before a supporting suffix ... From these data we shall draw the conclusion here that these phonetic disagreements of Fijian must be interpreted as 'false' analogy ... (ibid., 133-4, para. 127(a) 6).12 It is interesting that perhaps in order not to give hostages to fortune, Dempwolff used inverted commas for 'false' in 'false analogy'. At the time when he was assembling his data, knowledge of the vocabulary of Fijian was much less advanced than it is now, half a century later. With hindsight, therefore, and the advantage of greater knowledge of Fijian grammar than Dempwolff had either the possibility or the opportunity of acquiring, it was useful for me to check his data where they bear directly on the correspondences between Fijian verbal suffixes and the reconstructed final consonants of PAN verbs. Looking again at his PAN glossary in detail (Dempwolff 1938) with this particular end in view, I find 143 items which are suitable for comparison. Of these, 61 (i.e. two more than he was prepared to accept) show 'correct' (i.e. regular) correspondepces, assuming, that is, that one accepts his own criteria for what is (and what is not) 'regular'. 67 are 'incorrect'. This total subsumes not only cases where the proposed correspondence is 'irregular' according to Dempwolff himself, but cases where there is another reason for rejection. Some of the non-admissible comparisons arise from an 62

PAGE 75

incorrect interpretation of the Fijian data (in some instances because the information available to him was misleading or inadequate). Other pairs proposed for comparison seem to be semantically altogether too far-fetched to be acceptable. 15 correspondences are uncertain, in the sense that they are insufficiently supported, but there is no reason why, given additional evidence, they could not be confirmed and accepted; for example, PAN *palu 'beat, strike' and Fijian valu 'fight'; PAN 'consider accurately' and Fijian 'look at one's reflection in water'. Turning now to the regular correspondences, it is of interest to note that the following occur most frequently between stem-final PAN *\1) *t *t I consonants: FI \1) (11 instances) t ( 9 instances) c ( 8 instances) PAN FI *h \1) (7 instances) *k k (7 instances) *p V (5 instances) Cit is important to note that in the above table, *t' in PAN represents a reconstructed palatal: in Fijian represent two fricatives, a voiced interdental and a voiced bilabial respectively.] The next important contribution to a better understanding of the problems under consideration appeared a decade after the publication of Dempwolff's third monograph. In A study in the phonetics of Scott (1948: 737-52) presented the first detailed analysis of Fijian phonology by a modern professional linguist.l3 In particular, he was the first to draw attention to the structural --and incidentally remarkably symmetrical--relationship between the classes of consonants. Though he was not primarily concerned with orthography, his analysis fully, if only implicitly, vindicates the consistent and economical alphabet devised by the pioneer missionaries Cargill and Cross.l4 The table (see next page) reproduced from the article in question (Scott 1948: 743, Table 3: Correlations between consonantal phonemes and alphabetic script in Fijian) illustrates the quasi-complete one-to-one relationship between Fijian consonant phonemes and the letters used in that alphabet. 63

PAGE 76

Nasal-ized pl-osive pl-osive Nasal-Lateral-Fricative Semi-vowelBilabialb V m (w) Dental-AZveol-a:r' d dr q t r k n _g_ l -------------c s w Scott makes an important contribution to the problems under discussion here,pointing out that not only are the nonnasalized sounds (v, t, r, and k) 'articulated at corresponding points 1 (se. to Q_-:and Si:) 1 and except for y_, in a corresponding manner', but 'v is linked with the t, r, k set functionally; for like them,-it enters into "transitive-suffixes" which Q_, dr, do. 1 In a footnote he adds that: 'It does not seem that anything in Fijian indicates that the consonant does not belong to the suffix, though the large number of forms serving apparently the same purpose suggests a problem' (ibid., 742, n.4). We come now to the most comprehensive contribut.ion so far to the understanding of this problem, a doctoral dissertation by the Rev. David Arms (1975), a New Zealand missionary who had already spent several years in close contact with Fijian-speaking communities. Arms analyses the phonotactic constraints which govern the occurrence of the verbal suffixes.l5 He shows (ibid., 130-4 7) that, with very few exceptions, the place of artic.ulation of any consonant in Fijian verb rules out the occurrence of a verbal suffix with a consonant (or first consonant in the case of disyllabic suffixes)l6 with the same place of articulation.l7 His data are significant, both from a diachronic and a synchronic point of view. It is likely, for instance, to suggest an explanation for at least some of the cases of non-correspondence betwen Dempwolff1s PAN verbs and Fijian verbal suffixes.l8 As Arms points out (1975:140), the general constraint operating on consonants in suffixes, also helps to account for the fact that the nasalized stops (Q_, dr, and [mbJ, [ndJ, [ndrJ, and [ijgJ respectively), do not occur in verbal suffixes; if they did, they would be unacceptable after verb bases which have a nasalized stop in the initial or the medial consonant position, and these are very numerous. I have also made a detailed analysis of the synchronic system of these dissociations l9 in order to discover to what extent it conforms with Scott's (1948:743) table of Fijian consonants reproduced above. This shows that except in one or two cases, it is also possible to classify places of articulation 64

PAGE 77

if one treats dissociation as a sole criterion. In the table below, consonants which regularly dissociate from one another have been placed in the same column. Table 1: Consonantal dissociation in Fijian b d dr V t m n l s (w) q k w The following observations can be made regarding the table above: l. Consonants which occur in verbal suffixes are those under-lined. l and s are entered in for the sake of completion, but since neither occurs in monosyllabic suffixes and l only occurs in the disyllabic suffix -laka (which is 'intensiv;' in its effect and apparently not phonotactic constraints), it is difficult to decide in which column to enter it. 2. In a monosyllabic suffix, zero consonant (i.e. the absence of a consonant) occurs very frequently. A suffix is then reduced to -a after a front vowel and -ya after the open vowel or a back vowel. 3. regularly dissociates from d or t in the base, apparently with the sole exception of (vaka)dinata 1bear out, confirm'. 4. and dissociate from each other, apparently with the sole exception of karona 'take great care of, value greatly.20 5. It is necessary to a separate column from r, not only because the suffix -raka can occur after (and conversely -laka after; cf. Arms 1975: 141, n.4) but because land n associate freely: e.g. lomana, lawana, etc. (cf. ibid., l39). 6. Scott had regarded the interdental place of articulation of [oJ as relatively less important from the point of view of classification, and entered it in the same column as the two dental consonants t and For the same reason he had regarded which is, in fact, alveolar, as being intermediate between the dental and alveolar places of articulation. In both cases, his decision was probably influenced by considerations of structural symmetry. It is worth noting, however, that the dissociation principle firmly confirms c [oJ as being distinct from d and t. Likewise n (alveolar) is distinct from d and t. This consistent with the data, even though-it entails a sacrifice of symmetry or 'elegance'.

PAGE 78

Having recorded our debt to Arms, it is now necessary to register strong reservations about the remainder of his analysis, which I now summarize. Because of his extensive knowledge of spoken Fijian, based on regular practice in the course of his activities, both pastoral and informal, his views merit to be treated with special attention, although still open to rigorous examination. He implies21 that, granted a reasonable competence in the language, if a speaker is given any one base, together with its approximate semantic reference, he should be able to predict with a fair degree of accuracy what the consonant of its transitive suffix (or suffixes) is likely to be. He suggests that, subject to the phonotactic constraints which have already been examined, the majority of bases which can be followed by the same suffix have common semantic characteristics. As noted earlier, this is a view which Hazlewood had already hinted at and which, in the case at least of verbs of motion (without defining motion more precisely for the moment), is relatively easy to substantiate. What Arms posits, however, is that each of the consonants which occur in transitive suffixes is associated with one or more semantic notions or connotations. Thus, for instance: -c-is associated with 'pliancy, gentle contact, bodily experience' (Arms 1975: 104) -k-with 'hardness, force, opening out'. -m-with 'insertion, going inside', ... 'the idea of one thing going inside another, whether it be in order to stay there or to draw it out' (ibid., 107). -t-is associated with the use of a limb or instrument, moderate force, performative' (ibid., 110-12) -v-has to do with 'motion to, motion for, motion over'. Difficulties arise, however, when the consonant is zero There is a large number of bases in this category and at first Arms considered them to represent a 'spill-over category'. Later on, he declares, he was able to identify a 'common denominator': 'mild force, miscellaneous': 'Thus ending is very common with verbs of rubbing, tapping, folding, plucking, taking off, separating'. It also embraces verbs for 'finding and buying' (ibid., 113). There are also complications with -&and The former, in particular, (ibid., 105-6) has 'no convincing semantic correlation', but appears to have exclusively grammatical functions, like which often has the function of forming verbs from nouns (ibid., 107-8), a point already made by Hazlewood (1872: 32-3). 66

PAGE 79

Now, in most cases, Arms has no great difficulty in providing plausible, if not invariably convincing, lists of examples in support of his view. In each case he only gives ten examples, and the more the examples that are produced, the greater the difficulty of finding a common semantic denominator. This weakens his argument and, at least arguably, it weakens it unnece-ssarily. There are, in fact, two main difficulties. The first is that Arms seems to be under the impression that covert semantic connotations attach to the actual choice of certain sounds -consonants in this case --much as they do in most languages, including English.22 Yet cases like the suffixes with and --, which point to grammatical rather than semantic functions, as well as the large number of bases with consonant which do not have either a clear or an obvious common semantic denominator, should have alerted him to the possibility that the genuine semantic burden of verbal suffixes rests, not on their phonetic character, conferring on the preceding verb the membership badge, as it were, of a covert semantic category, but on a complex of grammatical relations which remain to be investigated. The second difficulty is this: the phonotactic constraints which Arms discusses militate in many cases against the occurrence of a particular suffix when semantic considerations would seem to require it. Although he does consider such cases (for instance, Arms 1975: 151-4, esp. note to p.l52), it does not seem that he has attempted to make a systematic study of what I shall call replacement suffixes, i.e. those which, for phonotactic reasons, are substituted for the suffixes which can normally be expected to occur, and of the effect of those substitutions on the synchronic system as a whole. One could even argue that Arms seems to hedge his bets.23 The phonotactic constraints which he has clearly set out are incontrovertible, but failing a more extensive investigation of their effect, it is very difficult to accept his thesis as to the correlation between individual suffixes and specific semantic notions. He might have chosen to sacrifice the latter to the former but, in actual fact, he to have spoilt his case by emphasizing the wrong argument.2 Stated briefly, one could present the dilemma as follows: On the one hand, the pattern of verbal suffixes in modern Fijian could be the result of interaction between diachronic phonology and synchronic syntactic and/or semantic constraints. On the other hand, (Q) it could represent the effect of diachronic semantic factors which are inhibited by synchronic phonotactic constraints. I should, therefore, like to propose a different approach to these problems. In view of their complexity, however, 67

PAGE 80

one cannot hope to do more than to suggest lines of inquiry which seem to be more promising than others and to try to adumbrate a possible solution. Let me then proceed from known and generally-accepted facts and examine the general distribution of monosyllabic verbal suffixes. I shall attempt to establish, first, what grammatical f'unct.ions can be determined for a given suffix, and secondly, what effect phonotactic constraints have on the occurrence of that suffix, both when the constraints are present and absent. On Arms 1 evidence 126) and -C/1are statist ically by far the most commonly occurring suffixes. Together they account for 569 recorded endings from his total field of 1680. Not only does it seem unlikely that any two particular 'meanings' (i.e. semantic associations or connotations) would so greatly predominate over the rest, but those are evidently also the two suffixes to which Arms was hardest put to attach any particular 'meaning' (cf. ibid., 110-12 ll3-4 for-!-). Any attempt, it would appear, to find a common semantic denominator between all the verbs that take a verbal suffix in or between all those that take a verbal suffix in -!-, is likely to end inconclusively. If we are looking for a common 'meaning', it will not be a property of the suffix alone, but of the interplay of syntactic variables within the verb phrase, in which suffixes play a vital but not an exclusive role. We must, therefore, look elsewhere and we are given valuable guidance by two widely-accepted observations of Hazlewood (1872: 33), namely, that: 1. is a 'denominal' suffix, i.e. it has the function of providing a method of forming verbs derived from nouns. 2. -v-is associated with verbs of motion; without defining this class more precisely for the moment. I have argued elsewhere (Milner 1980: 1-4) that the slow development of Austronesian studies during the last lOO years is to some extent due to the geographical fragmentation of the work and also to the intellectual isolation of the scholars concerned,which can be ascribed to relative lack of communication and in general to relative ignorance of one another's problems and progress. There has also been a noticeable lack of comprehensive studies of individual languages as weJLl as too great a concentration of effort on comparative studies, particularly on topics such as subgrouping and putative chronology at the expense, if not the exclusion of detailed description. For the greater part of the twentieth century, students of Austronesian languages, while paying lip service to their common origin and striving to make sense of an extensive common stock of words, have neglected comparative grammar. It is only in the last 68

PAGE 81

decade, with the organization of international conferences on Austronesian linguistics, that the syntactic features of languages as diverse as those of Taiwan, the Philippines, and Madagascar, have to throw light on the solution of problems that have long baffled students of Oceanic and Indonesian languages 1978; Naylor 1978). As a case in point, it would appear that the focus and topic approach to the understanding of Fijian syntax (Naylor 1978) is likely to help us make significant progress. Let us then examine, if only provisionally and in order to discover if one can establish prima faaie evidence, the hypothesis that Fijian too has a system of focus marked by verbal affixes. I propose to use the term 'focus' in the sense that is widely, though by no means unanimously, accepted in Philippine linguistics, i.e. 'the syntactic relationship between the verb and the surface subject, signalled by the verb's focus affix in conjunction with the subject form of noun phrases and pronouns. For example, a sentence is in instrumental foaus if the surface subject is in the role of instrument and the verb has an instrumental affix; the verb on the subject as instrument' (Naylor 1975: 12-3). On this hypothesis, by reason of their frequency of occurrence alone, the two suffixes and should be examined afresh in order to establish whether they represent the Fijian equivalent of what has been identified elsewhere, particularly in Taiwan and Philippine languages, as goal focus affixes. A few years ago Dahl (1978) suggested that four types of focus were perhaps Pan-Austronesian in their distribution, namely: aator foaus, goal foaus, referent foaus (the person in whose interest the action is carried out or the place where the action is performed) and instrument foaus which he characterized as follows: 'The fourth focus, generally called instrument foaus (IF) got its name because it focuses something for performing the action, for instance an instrument' (Dahl 1978: 384). In elaborating his interpretation, Dahl (ibid., 385-6) goes on to explain that one of the separate fUnctions of the fourth focus has to do with the displacement of a moving object, either away from the actor (as in Minahasan languages) or in any direction (within the actor, towards him, or away from him (as in Malagasy) In her contribution to the same volume, Naylor makes a similar point with reference to Tagalog: What appears to be at play here is not a contrast between transitive and intransitive, rather it is whether the action is viewed as centrifugal or

PAGE 82

centripetal. Like aspect, however, the contrast between centrifugal and centripetal is situational as well as a matter of perspective. When the action is viewed as going outward from the actor and ends outside of him, then it is centrifugal; ... When the action itself is viewed as beginning and ending with the actor himself then it is centripetal. (Naylor 1978: 405) One of her pairs of examples is suggestive from the point of view of Fijian. She mentions two Tagalog verbs in actor focus but with different affixes: magbili 'sell' (centrifugal) as opposed to bumili 'buy' (centripetal). Both are formed on the base -b--ili. There is a similar situation in Fijian where a similar pair is formed from the cognate base voli, namely, volia 'buy' and volitaka 1sell1 (cf. veivoli 'buy and sell, market' (perhaps also 'exchange, barter' in a pre-contact economy). Earlier in the same article, Naylor (1978: 400-01) identifies, in the case of Tagalog, four types of focus (actor, goal, locative and instrumental) (ibid., 396) and six kinds of role (actor, goal, locative, comitative, benefactive and instrumental).25 It will be evident from the views quoted from Naylor and Dahl that there is, as yet, no consensus among the scholars interested in this approach, not only as to the exact nature of the syntactic relations subsumed by focus and topic but also as to the number to be distinguished and identified and the technical terms to be used to describe them. Nevertheless, there is abundant evidence that a rich and promising area of research lies before us in Austronesian studies (e.g. Dahl 1981; Ferrell and Stanley 1980; Lopez 1978; Naylor 1980) It seems, therefore, that a good case can be made for a new approach to the problem of verbal suffixes in Fijian. Thus what Dahl calls the 'moving object focus' clearly has an equivalent marked by disyllabic suffixes such as -vaka and -taka (as in cicivaka 'run with something', or viritaka 'throw something (at a target)') but this moving object focus (which might be termed 'locomotive') will have to be defined rigorously with special reference to what has also been called 1comitative, benefactive and instrumental'. Likewise the -ra suffix shows evidence of being associated with a locative focus. Within the limited scope of the present article, however, one can hardly do more than point to the complexity of the problems and to the direction in which progress is likely to be made. Let me first make a point of theory and consider for a moment the phonotactic constraints which restrict the occurrence of any one verbal suffix with any one verbal base. A thoroughgoing attempt to establish beyond doubt that Fijian does indeed have a topic and focus system will have to distinguish carefully 70

PAGE 83

what might be called the 'canonical' suffixes from 'adventitious' or 'intrusive' suffixes, that is, those which are imposed by phonotactic constraints. For instance, it was Hazlewood who first stated that while verbs of motion take the suffix -va, so do many others. It follows that, before we can establish a firm correlation between any one suffix and any category of verbs, two factors must be taken into consideration: (a) Assuming that the suffix -va can, under certain circumstances, mark a type of focus which we might call 1displacive' or 'locomotive', it cannot appear whenever a verb of motion includes a bilabial consonant. The suffix -va will then be replaced by another as the following examples show: l. -ta instead of -va ea beta 1 go up to' kabata 'climb up to' kevuta 'climb down along' so but a 1 go down along 1 volita 1 go round sth. 2. -ea instead of -va kuvuca vukaca yamoca vuloca 'blow (smoke) against' 1 fly towards 1 1 grope for sth. 1 'roll (sennit) over thigh' 3. -ka instead of -va virika tebeka vidika 'throw( sth.) at 1 (of stone etc. ) 'skim on surface of water, ricochet' 1flip(finger etc.) against sth. 1 cumuta ribata lavota livata robota dromuca lomoca mumuca luvuca dumuka vodoka butuka 'butt with the head against' 'strike against (in springing back) 1 'score a hit (with small object), cast into' (of lightning) 1 flash on sth. 1 1 extend over, stretch over' 'sink below, go under sth. 1 1dip(sth.) into' 'swarm towards' 'flood over; plunge under' 'raise, lift up (on end of stick) 1 'embark on, go aboard' 'step on, tread on' (b) Conversely, where a consonant (other than is constrained from occurring as a verbal suffix because a homorganic consonant occurs in the base, -v-may be adventitious, that is to say, it

PAGE 84

may be substituted for a suffix that would otherwise have been used. Note, for instance, the following cases where -ta or might have been expected to occur if, that is, we assume that either of them can signal a goal focus affix: -va instead of -ta or -a: tal ova taq_ava todrava dolava 'ladle, scoop (yaqona etc.) 1 'use (two or more layers etc.) 1 1 (of sun)burn, scorch' 'open' setiva dikeva tarava nitiva 1cover1 'study, scrutinize' 'follow in succession, be next to' 'slice off (crown of taro corm) 1 though, interestingly, -a does, in fact, occur after tara, but in a different sense: 'touch etc. 1 At this point, it is worth examlnlng in some detail what Arms (1975: 106-7) has suggested with regard to the suffix -ka. He states that 'it is associated with verbs where the action is by nature a forceful one; verbs of "breaking, squeezing hard, striking violently" (sometimes involving a missile) are typical members of this class.' When Hockett (1974) reviewed the general problem of these verbal suffixes in a paper presented at the First Conference on Comparative Austronesian linguistics in Honolulu, the occurrence of -ka was one of only two instances where he concluded that Arms-1-thesis could be upheld.27 Indeed, if we look at the semantic distribution of verbs followed by this suffix, it is difficult at first flush to see how one can arrive at any other conclusion. A more recent article, however, (Milner and Nawadra 1981: 186-94) shows that of 81 verbs having to do with 'breaking, splitting, cutting and grating', only 18, i.e. 22.23%, have a suffix in -ka. Of these 18 verbs, 3 also have an alternant suffix in -a. The other difficulty is that, in a large number of instances, -ka occurs in bases that seem to have little to do with force, violence or disruption. In addition to butuka and vodoka, the following three verbs represent instances where -ta or have been expected to occur is,we assume-that either of them can signal a goal focus affix): tomika tevuka 'pick up' 'unfold, open up' dodoka 'lift up, stretch out (hand) 1 In actual fact, of course, all three bases include at or d which 72

PAGE 85

rules out -ta as a suffix. On the other hand t or d also occur in bases which do have a violent or disruptive-connotation, which again suggests that phonotactic rather than semantic factors are relevant: vidaka teveka 'split, cleave' 'circumcise' tunaka muduka 'gut, disembowel' 'cut off' Other verbal suffixes occur much less frequently than those which have been mentioned so far. They include -ra, -ma, -na, and There is a clear association, it would seem, between the suffix -ra and what may be a locative focus, as the following examples show: ciSJ.ira 'stick into, slip tu bur a 'grow on' into a narrow place' tuber a 'carry, hold (in the taSl.ara 'place, lay (on ---hand)' top of)' davora 'lie on' SJ.isora 'poke (with stick etc.)' Not infrequently -ra occurs when the base is preceded by the prefix vaka-: vakasobura 'put sth. down' vakayacora 'carry out, perform' vakamocera 'put someone to sleep' vakadabera 'make someone sit down' If r, or one of the other two consonants subject to the same phonotactic constraint (i.e. dr or n) occurs in the base, another consonant must be substituted an (assumed) r in the suffix. It may be one of the following: m as in: darama 'slip into' t as in: ravita 'lean on' tanuma 'dip into' suruta 'sneeze on' .. as in: miraca 'fall gently .as in: ravoga 'warm (cold on' food) on' rubeca 'hang sth. raraga 'heat (banana on' leaves) on' k as in: tonoka 'dab on' ramaka 'cast (light) on' 73

PAGE 86

The remalnlng monosyllabic suffixes, namely, -ma, -na, and do not occur very frequently. Arms not able to correlate any special connotation and he considers it to be similar to -na. In his opinion, the latter has a grammatical function, to say (as Hazlewood had already suggested), it serves to form verbs from nominal bases. There is much evidence to support this view as the following examples show: baca kato buka tag a duva 'bait' 'box' 1 fuel, fire' 1bag1 'plant used as fish poison' 'day; sun' bacana katona bukana tagana duvana sigana 'bait, entice' 'put into a box' 'add fuel to' 'put into a bag' 'poison (fish) with duva' 'sun; dry in the sun' This is not to say, however, that all bases which can be followed by the suffix -na are formed from nouns. A relatively small number of them be verbal bases 'in their own right'. It can hardly be a coincidence that, for most of them,29 an expected suffix in -ta (assuming again that this is the normal or 'canonical' form of the goal focus affix unless phonotactic constraint rules it out) does not occur because t or d occurs in the base: dab ana 'do up in parcels' tukuna 'relate, tell announce' domona 'love, desire' tawana 'occupy, populate' dagina 'bathe (eyes)' tomana 'accompany; help' tavuna 'roast on embers' tuvana 'arrange in order, set tokona 'prop up; stay' in rows' Up to this point this article has dealt with the phonotactic constraints of consonants, and only with those associated with monosyllabic suffixes. One phenomenon remains to be mentioned briefly. There is evidence that not only consonants but vowel quality also is a factor relevant to the occurrence of the verbal suffixes we have examined.30 Of the ten verbs mentioned by Arms (1975:107) which are followed by a suffix for instance, eight end with a back vowel or Q.) and two with the open vowel He also (ibid., 203-5) gives a fuller list of 35 bases followed by -m, of which 29 end with a back vowel, l with the open vowel andonly 5 with a front vowel. Of the latter, only 2 (silima 'dive for' and sigema 'suddenly realize') are attested beyond all doubt. doubt.3l Likewise,a very high proportion of verb bases is followed by (zero consonant: Arms (ibid., 254-68) lists 30. Of those, the vast majority, 261, ends with a front vowel (i or ; only 43 (i.e. approximately 1!0 ) end with a back vowel:-74

PAGE 87

and many of those either need to be confirmed as correct, or admit the possibility of another suffix as an alternant. In conclusion, I hope to have shown, if only by implication, first, that comparative studies in Austronesian languages will rest on a surer foundation (and therefore advance more rapidly), not only when the quality of the data for comparison, as well as the quantity, is more satisfactory, but also when more comprehensive studies of carefully selected languages are available. At present, much of our data can scarcely be said to be more abundant or more reliable than it was in Dempwolff1s time. Recent studies (Geraghty 1983; Geraghty & Pawley 1981) show the wealth of hitherto unpublished and hitherto unknown evidence from a relatively well-known language area like Fiji. Secondly, studies of individual languages will increasingly be assessed by the criterion of the extent to which the author shows that he is, if not familiar at first hand, at least aware of the whole field of Austronesian grammar and of recent progress made in areas other than his own and, moreover, that he has considered its relevance to his own work. In particular, serious consideration should now be given to the question of establishing whether the topic and focus approach to Austronesian syntax is relevant to the understanding of Oceanic languages. Two other contributions which are especially relevant to the present article have been made since it was written (Schutz 1985; Milner 1986). NOTES l. A contribution to the solution of the old and refractory problem of the Fijian verbal suffixes from a phonological and prosodic point of view seems to be appropriate in this collection of articles. I should also like to express my thanks to a number of colleagues who have generously commented on., and suggested improvements to, the original draft of this article; especially to Professor Bruce G. Biggs of the University of Auckland, Professor Otto C. Dahl of the University of Oslo, Professor Charles F. Hockett of Cornell University, Professor Albert J. Schutz of the University of Hawaii, and Dr Paul Geraghty of the Fijian Dictionary Project in Suva, Fiji. My thanks are also due to the Rev. Dr David G. Arms of the Columban Fathers, Professor Robert Blust of the University of Hawaii, Professor Jack Carnochan of the University of London, Professor Viktor Krupa of the University of Bratislava, and 75

PAGE 88

Professor Paz B. Naylor of the University of Michigan. 2. For a detailed analysis of the relationship between 'word', 'morpheme' and 'base' in Fijian, see Schutz 1975. 3. As, for example,. in uca 'rain' or ba (CVV) 'fence'. (I am aware that in recent years the one-vowel-per-syllable analysis of Fijian phonology has been criticized, notably by Schiitz and Biggs, though for reasons that will not be discussed here I remain unconvinced.) There are also rare instances of bases where both consonants are zero, as in ia 'proceed, take place' and ua-(ca) 'beat with a stick (laundry, etc.)'. 4. Only one instance is known to me of a trisyllabic base where the initial consonant is zero: Paul Geraghty has written to me that uea exists as a verb ''to fish trap'. This must be the base Hazlewood gives as wea 'a fish trap' and it could take its name from the island of Uvea (Wallis), a Polynesianspeaking community under French administration. The only other case I know of a trisyllabic base where both the medial and the final consonants are zero is biau 'wave' which is almost certainly a Polynesian loan (peau). Other instances, however, cannot be ruled out. 5. This statement must be qualified by adding that these suffixes are also found in combination with other affixes, in which case they may not be 'transitive'. This is not strictly relevant to the problem considered here. 6. The consonant in parentheses can either be zero, or following the open vowel or the back vowels and -the palatal approximant written L In disyllabic suffixes, the second consonant is always (cf. n.l6 below). 7. Andy Pawley (1978:120 (also nn.l7,l79); 135, 136-9, esp.l3T) has put forward the view, which others have accepted, that from a comparative point of view the verbal suffixes in -Ca and -Caka of standard Fijian are 'irregular' or at least -untypical of Fijian dialects in general, and that they represent a conflation of two vowels (i.e. *Ci-a to -Ca and *Caki-a to -Caka). David Arms (1975 :28, 3l-4)""";eems to have come to the same conclusion at about the same time and independently of Pawley. There is much evidence to support this view but, in o.rder to avoid confusion with the socalled 'passive' suffixes in -Ci, I prefer not to quote examples in the 'canonical' *Ci form as Pawley and others do, but to give them with an 'active -Ca suffix. Regarding Fijian spelling, see n.l4. 8. This view is often associated with the phrase 'thematic consonant'. Charles Hockett, in a letter about his article 76

PAGE 89

(1976, 1977) adds the following comments: The behaviour of the Fijian thematic consonants was one of the real facts about languages that led me slowly but surely to abandon what I now refer to as the 'atomic morpheme theory':, the theory of grammaticolexical structure I helped develop in the 1940's and to which I clung for a long time. That theory proposes that every phonemically relevant piece in any utterance must be a part of one or another morpheme (or of the phonemic representation of one or another morpheme) and that morphemes are minimum meaningful elements in much the same sense in which we all assumed phonemes were minimum meaningless but differentiating elements. By that theory there would be only three possibilities for Fijian: (1) rai-ca, as with Churchward, so that the suffix has ten different alternants; (2) raic-a, as proposed by Bloomfield for Samoan, so that the stem has two different alternants (as do most verb stems); (3) rai-c-a_, the thematic consonant being a separate morpheme. 9. Cf .. ibid., 67-8; 89-90; 105-6. This work was originally published in 1956. 10. See, especially, the references to Arms, Geraghty, Hockett, Naylor, Pawley, and Schutz. 11. Thanks to the sponsorship, first of Mr Raymond Burr, a wellknown American television actor, and later of the Australian Government Cultural Fund as well as the support of the Government of Fiji, a monolingual dictionary of the Fijian language is being compiled and is now approaching completion. The Director of this Project was, until early 1986, Mr T.R. Nawadra; Dr Paul Geraghty currently holds the post. 12. 'Besonders haufig sind Lautunstimmigkeiten bei Auslauten des Fidji vor stutzendem Suffix ... 'Aus diesen Tatsachen wird hier gefolgert, dass diese Lautunstimmigkeiten des Fidji als "irrige" Analogie zu deuten sind; ... 1 13. Earlier studies include a monograph by Kern (1886). Scott's data on Fijian were mainly derived from his study of the pronunciation of Josua Bogidrau, a Fijian civil servant who had been seconded to the School of Oriental and African Studies, London University (1946-48), and who also helped me to learn his language at first hand. 14. The orthography chosen for Fijian by the first two missionaries, David Cargill and William Cross, in the 1830s,

PAGE 90

was remarkably advanced for its period and, in particular, almost anticipated phonemic theory, at least by implication, by about 75 years. Thus, with one exception (dr), each consonant phoneme is always represented by one, and only one, letter, despite the fact that Roman conventions (supplemented by the conventions of English orthography) require that digraphs should be used. So, there are three voiced stops, each at least in non-initial position, by a non-phonemic homorganic nasal: /mb/ (written E_), /nd/ (written and /!} (written The voiced interdental fricative /o/ is written c (instead of th) and the velar nasal is written g_ (instead-of gg_). A fourth nasalized voiced 'stop' is written dr (actually /ndr/). Apart from using diacritics in a somewhat arbitrary and unpredictable manner, however, the orthography of Fijian does not take vowel length into account. Since the latter is phonemic, it is a serious defect in an otherwise elegant system. (Native speakers seldom use or require diacritics since to them the exact pronunciation is usually clear from the context. ) I am indebted to Professor Otto C. Dahl for drawing my attention to his article on the origins of Malagasy spelling (Dahl 1966). This shows that, although Fijian orthography was much in advance of its time in its economy and its disregard for non-significant sounds, it was neither entirely original, nor an isolated attempt to devise an alphabet based (1) on a one-for-one equivalence between letters and phonetic values and (2) on internal consistency without obligatory regard to the spelling conventions of English, or for that matter, of French orthography. In actual fact, in the early 1820s, the same principles had been consciously observed by the three Welsh-speaking pioneers of the London Missionary Society in Madagascar: David Jones, Thomas Bevan, and David Griffiths, who devised the first system of Roman orthography for Malagasy. Their training in England, at a theological academy in Gosport presided over by Dr David Bogue, included a linguistic component which owed much to the well-known grammarian Lindley Murray (1745-1926). The latter had stated in his English grammar that: a perfect alphabet .. would contain a number of letters, precisely equal to the number of simple articulate sounds belonging to the language. Every simple sound would have its distinct character; and that character be the representation of no other sound. (Murray 1813:15) 78

PAGE 91

Murray in his turn was directly indebted to Samuel Johnson's short 'Grammar of the English tongue' which precedes his Dictionary of the EngZish Zanguage. In that short essay, Dr Johnson ends his remarks on orthography and pronunciation with references to various attempts made in the past 'to accommodate orthography better to the pronunciation' and he notes that some reformers have endeavoured to proportion the number of letters to that of sounds, that every sound may have its own character, and every character a single sound. Such would be the orthography of a new language to be formed by a synod of grammarians upon principles of science. (Johnson 1828:33) It is interesting to note also that, just as Fijian spelling (in accordance with the ideal system for a previously unwritten language recommended by Johnson and Murray) uses for example, without regard to the conventions of English spelling, the three Welsh-speaking pioneers in Madagascar proposed initially to use c for an affricate /ts/, the velar nasal, and most of all, the close back vowel as in Welsh, instead of oo as in English or ou as in French. Regrettably, however, this imaginative proposal was abandoned in the face of opposition from other Europeans whose first language was English or French and not Welsh. This resistance in Madagascar has parallels in the Pacific. Thus, because it is an unconventional alphabet from a purely Western point of view, Fijian orthography has long been the target of well-meaning but uninformed criticism (often aggravated by patronizing ridicule) in English-speaking circles. At one point during the Colonial period, in the late 1930s, the desire to 'reform' Fijian spelling even led to a debate in the Legislative Council of Fiji (Schiitz 1972: 14 ff., esp. 20-2). There is not much evidence, however, that the Fijian people have ever wanted to introduce spelling .changes, though, undoubtedly, some Fijians are irritated when they hear the names of people and places mispronounced by ignorant expatriates or overseas news-readers. l5. Arms (19:130-1) acknowledges an article (Krupa 1966) which had appeared eight years previously and mentions associative and dissociative tendencies between groups of consonants in Oceanic languages, including Fijian, according to their place, or their mode, of articulation. He states, however, that Krupa was concerned with 'groupings of consonants according to their place of articulation or after their mode of articulation, not to the associative 79

PAGE 92

and dissociative tendencies of individual consonants --the item of particular interest here' (ibid., 131). Albert Schutz informs me in a letter that Arms' dissertation was the first full analysis of these phenomena to be published, but Bruce Biggs was already discussing consonant restrictions in the early 1960s although he did not publish his findings. Moreover, Paul Geraghty' s (1973) unpublished term paper on this subject had been heard by 1973, while Peter Lincoln had also studied the same problem. I am now indebted to Paul Geraghty for sending me a copy of the term paper in question. He reminds me that the problem is also discussed in his doctoral dissertation (now published as Geraghty 1983. See esp. 260-70). 16. The second consonant of a disyllabic suffix is always -kaka, -laka, -maka, -naka, -raka, -taka, -vaka, -yaka. Unlike the first consonant, it is not subject to any constraints of occurrence. (A comprehensive analysis of the distribution and function of disyllabic suffixes has not been possible within the scope of the present article.) 17. In addition to constraints governed by the place of articulation of consonants, Arms also points to one or two cases where the constraint seems to be linked with the mode of articulation. Thus, the occurrence of a velar nasal g in the initial or medial consonant of the base, rules out the'nasal suffix -ma: e.g. gunuva 'drink', whereas the reconstructed PAN form *inum (as well as Polynesian reflexes such as Samoan inumia, not to mention Fijian dialectal variants) would have made one expect *gunuma (Arms 1975: 153); see also *ceguma. The occurrence of a close back vowel before a suffix in -m-is also considered in the discussion of vowel quality on-p . 74 above. The converse is also true: there is no recorded instance of a base with a bilabial nasal taking the velar nasal suffix Curiously, however, the alveolar nasal can occur without restriction following a bilabial or a velar nasal in the base. The same observations apply to the labio-velar approximant w-, the occurrence of which in a base rules out the bilabial (with some exceptions. See liwava mentioned in n.26 below and the velar nasal suffixes -ma but not the alveolar nasal suffix -na (cf. Arms 1975). 18. Thus, for instance, one can see at first glance that at least six of Dempwolff's reconstructed word-stems: *dagaY, *giliiJ, *le!Jgak, *padam, *peg' at' and *taiJuk have final consonants which are deemed to be articulated in the same place as a medial or an initial consonant. This alone rules out the possibility of finding direct one-to-one reflexes in the suffixes of the Fijian verb bases which, in other respects, show regular correspondences with 80

PAGE 93

reconstructed PAN forms: rogoca qilia loqata moce(ra) /mooera/, vocota digova respectively (cf. Arms 1975:156). 19. The terms 'associative', and 'dissociative', with reference to consonants and vowels that may, or may not, respectively, occur within the same base, with or without a suffix, are used by both Krupa and Arms. In a letter, Professor Biggs points out that prenasalisation did not occur finally in PAN and that in his view this fact alone, even without dissociation, is enough to account for the absence of prenasalisation in the suffixes. 20. In a letter, Paul Geraghty informs me that karona is probably a modern form of karauna (cf. garauna with similar meaning) and that the restrictions may not be so strict at a distance of two vowels. 21. This statement is based on a conversation I had with David Arms some years ago. He was presumably thinking of an expatriate learner of Fijian like himself who had already acquired some knowledge of its covert categories. Yet, when he tried the experiment of making up imaginary bases (nonce words) and then of asking native speakers to -suggest appropriate suffixes and 'meanings' for them, he got replies which sharply contradicted his expectations. (Arms 1973; 1975:147-8). 22. See especially Arms 1975:128-9). He mentions, for instance, the connotations linking words in English beginning with sl-as in: slick, slip, slime, slide, slouch, slut, etc. 23. See esp. Arms 1975:118, 150, 156-7. 24. As, for example, when he argues (Arms 1975:122 ff.) that, for each 'passive' ('spontaneous') prefix, there is a corresponding, specific and identifiable semantic content, and what is more, that there is a one-to-one relationship of identity between the occurrence of some of those consonants in spontaneous prefixes and their occurrence in suffixes. Thus: 'The meaning of -c was given as "pliancy, gentle contact, bodily experience"-:The meaning for eacould be regarded as a s.emantic specialisation: a shift from bodily experience in general to the particular bodily experience of sound' (ibid., 122). This view would seem to require much more supporting evidence than Arms provides if it is to be accepted. 25. This distinction has to do with the fact that, in Tagalog, the locative focus has to be analysed in relation to a number of subcategories: 'locative goal focus, locati ve proper, and locative beneficiary (directional or dative)'. 81

PAGE 94

The instrumental role is, 'likewise, subclassified into: 'instrumental goal (portative or displacive), instrumental proper, and instrumental benefactive'. 26. I am also reminded by Al Schutz that one aspect of the complexity of this problem is that most speakers of standard Fijian use it as a second language and that for them there is often an element of doubt as to which verbal suffix is 'correct' or 'appropriate'. Thus in 1953 even a very senior chief and distinguished Colonial civil servant from the Lau group of islands (in the south-eastern part of Fiji), the late Ratu Sir Lala Sukuna, asked his wife Lady Maraia in my hearing if liwa '(of the wind) blow' was followed by the suffix -ea or the suffix -va (liwava). She spoke Standard Fijian as a first language (unlike her husband) and she immediately replied with assurance: 'liwava'. 27. This paper appeared later as Hockett (1976,1977). In a recent letter, he adds: If thematic consonants are separate morphemes they ought to have determinable meanings. One of Pawley's students made an assessment ... of semantic associations of the thematic consonants. I did the same thing independently, and came out with this. Of a random set of 500 stems: Of 51 with thematic consonant 28 or 55% denoted breaking, splitting, or other such forceful operations. For this sort of meaning no other thematic consonant scored so high. Of 74 stems 24 denoted motions or positions; again stems with this meaning but with other thematic consonants scored much lower. Also, I found some dissimilative tendencies -ater certain first and second consonants, certain thematic consonants are disfavored (this thing having to do with sound, of course, not sense). 28. He (loc.cit.) identifies a subclass of verbs having to do with 'opening out, unfolding, extending', which regularly take the -k-ending. He claims, nevertheless, to see a semantic connection with the rest on the ground that '"smash ing, breaking, cleaving" all involve disintegration of some entity'. I do not accept this view and consider that it is just as likely that the process of prosodic constraint is involved here too. Thus, -t-is ruled out in tevuka and dodoka. ----29. Some verbs with the suffix -do not seem to be formed on nouns and yet the occurrence of this suffix cannot be accounted for on the ground of prosodic constraint: cuqena 'support', kumuna 'gather, collect', sogona 'assemble'. 82

PAGE 95

30. Arms, following Krupa, uses the term 'dissociative'; Hockett speaks of 'dissimilative tendencies', while I have spoken in this article of 'phonotactic constraints' and of 'prosodic constraint'. The preferential association of certain suffixes with certain vowels in the base which is discussed here is the opposite of a constraint. Hence the title of the article which attempts to subsume both phenomena under the term 'prosodic' in the sense first used by J.R. Firth and the London school of linguists with which he is associated (cf. Firth 1948). 31. In a letter, Al Schutz writes as follows: Arms, D.G. See my article on borrowings (Fiji Museum Publication), 19'78 especially on 'natural syllables' ... The point was that there are phonetic (articulatory, that is) reasons for certain C +V associations. I was looking at it from the point of view of the C being fixed and the V open to choice; your observation .. approaches the matter from the opposite direction. 19'78) REFERENCES 19'73. Whence the Fijian transitive endings? Ocean. Ling. 12, 503-58. 19'75. Transitivity in Standard (Ph.D. Thesis, Univ. Michigan, Linguistics, 19'74). Ann Arbor: Univ. Microfilms Intern. Churchward, C.M. 1941. A new Fijian grammar. Sydney: Dahl, O.C. Australas. Medic. Publishing. 1966. Les debuts de liorthographe malgache (Avhandl. Norsk. VidenskapsAkad. Oslo. II. Hist.-Filos. Klas., N.S. 9). Oslo: Universitetsforlaget. 19'73. Proto-Austronesian (Scand. Inst. Asian Stud. Monogr. Ser. 15). Lund: Studentlitteratur. 19'78. The fourth focus. In Second International Conference on Austronesian Linguistics: Proceedings (eds.) S.A. Wurm & Lois Carrington, (Pac. Ling. Ser. C, 61). Canberra: Austral. Nat. Univ., 1, 383-93. 83

PAGE 96

Dahl, O.C. Dempwolff, 0. Ferrell, R. & Stanley, P. Firth; J .R. Geraghty, P. Geraghty, P. & Pawley, A. Hazlewood, D. 1981. Early phonetic and phonemic changes in Austronesian (Inst. Comp. Res. Hum. Cult. Oslo Ser. B, 63). Oslo: Universitetsforlaget. 1934, 1947 & 1938. Vergleichende Lautlehre des austronesischen Wortschatzes: 1. Induktiver Aufbau einer indonesischen Ursprache. 2. Deduktive Anwendung des Urindonesischen auf austronesische Einzelsprachen. 3. Austronesisches WoT't(3rverzeichnis (Beihefte z. Zeitschr. f. Eingeb.-Spr., 15, 17 & 19). Berlin: Dietrich Reimer. 1980. Austronesian versus IndoEuropean: the case against case. In Austronesian studies: Papers from the second Eastern Conference on Austronesian languages (ed.) Paz Buenaventura Naylor (Mich. Pap. on South & SEast Asia 15, 1979). Ann Arbor: Univ. Michigan, 19-31. 1948. Sounds and prosodics. Trans. Philol. Soc.3 127-52. 1973. Some aspects of case-marking in Fijian (Unpubl. term paper, Dept. Ling., Uni v. Hawaii) 1978. Topics in Fijian language history (Ph.D. Thesis, Univ. Hawaii, Linguistics). 1983. The history of the Fijian languages (Ocean. Ling. Spec. Publ. 19). Honolulu: Univ. Hawaii Press. 1981. The relative chronology of some innovations in the Fijian languages. In Studies in Pacific languages and cultures: in honour of Bruce Biggs (eds.) Jim Hol1yman & Andrew Pawley. Auckland: Ling. Soc. New Zealand, 159-78. 1872. A Fijian & English and an English and Fijian dictionary .. and a grammar of the language ... (ed.) J. Calvert. London: Sampson Low, Marston. 84

PAGE 97

Hockett, C.F. Johnson, S. Kern, J.C.H. Krupa, V. Lopez, C. Milner, G.B. 1974. The reconstruction of ProtoFijian-Polynesian. (Paper delivered at the first international conference on comparative Austronesian linguistics, Honolulu.) 1976. The reconstruction of Proto Central Pacific. Anthrop. Ling. 18, 187-235. 1977. Proto Central Pacific: Addenda. Anthrop. Ling. 19, 242-4. 1828. A dictionary of the English language .. (to which are prefixed) A history of the language and An English grammar. London: Robinson (New ed. from 1773). 1886. De Fidjitaal vergeleken met hare verwanten in Indonesie en Polynesie (Verh. Akad. Wet. Amst. Afd. Letterk. 16)' 1-242. 1966. The phonemic structure of bivocalic morphemic forms in Oceanic languages. J.Polyn. Soc. 75, 458-97. 1978. A handbook in comparative Austronesian (ed.) Ernesto Constantino (Arch. Philip. Langs & Dial. and Philip. Ling. Circ.: Spec. Monogr. Issue 5). Quezon City: Univ. Philippines. 1972. Fijian grammar. Suva, Fiji: Govt. Press. 1980. On the centrality of Austronesian syntax. In Austronesian studies: Papers from the second Eastern Conference on Austronesian languages (ed.) Paz Buenaventura Naylor (Mich. Pap. on South and SEast Asia 15, 1979). Ann Arbor: Univ. Michigan, 1-17.

PAGE 98

Milner, G.B. Milner, G.B. & Nawadra, T.R. Murray, Lindley Naylor, P.B. Pawley, A. Schutz, A.J. 1986. A focal approach to problems of verbal syntax in Fijian. In FOCAL I: papers from the Fourth International Conference on Austronesian Linguistics (eds.) Lois Carrington & S.A. Wurm (Pac. Ling. Ser. C, 93). Canberra: Austral. Nat. Univ., l-20. 1981. Cutting words in Fijian. In Studies in Pacific languages and cultures: in honour of Bruce Biggs (eds.) Jim Hollyman & Andrew Pawley. Auckland: Ling. Soc. New Zealand, 179-95. 1813. English grammar. York: Wilson & Sons (24th ed.) 1975. Topic. focus, and emphasis in the Tagalog verbal clause. Ocean. Ling. 14, 12-79. 1978. Toward focus in Austronesian. In Second International Conference on Austronesian linguistics: Proceedings (eds.) S.A. Wurm & Lois Carrington (Pac. Ling. Ser. C, 61). Canberra: Austral. Nat. Univ., 1, 395-442. 1980. Linking, relation-marking, and Tagalog syntax. In Austronesian studies: Papers from the second Eastern Conference on Austronesian Languages (ed.) Paz Buenaventura Naylor (Mich. Pap. on South and SEast Asia 15, 1979). Ann Arbor: Univ. Michigan, 33-50. 1978. Some problems in Proto-Oceanic grammar. Ocean. Ling. 12, 103-88. 1972. The languages of Fiji. Oxford: Clarendon. 1975. At a loss for words: the problem of word classes in Fijian. Ocean. Ling. 14, 100-18. 1978. English loanwords in Fijian. In Fijian language studies: borrowing and pidginization (ed.) A.J. Schutz (Bull. Fiji Mus. 4), l-50. 86

PAGE 99

Schutz, A.J. Scott, N.C. 1981. Specification as a grammatical category in Fijian. In Studies in Paeifie languages and cultures: in honour of Bruee Biggs (eds.) Jim Hollyman & Andrew Pawley. Auckland: Ling. Soc. New Zealand, 197-207. 1985. The Fijian language. Honolulu: Univ. Hawaii Press. 1948. A study in the phonetics of Fijian. Bull. Seh. Orient. Afrie. Stud. 12, 737-52.

PAGE 101

A SIAMESE LETTER DATED 7 DECEMBER 1776 Sl&ren Egerod In the Danish National Archives a considerable number of documents are preserved from the files of the (Danish) Asiatic Company and from the Government of Tranquebar. The Asiatic Co. file no. 2188b has the title: Documents with reference to Pegu3 Siam3 Kedah3 Johore3 Cambodia3 and Manila3 1682-1776. Only one letter in this package is written in Siamese. It is dated 7 December 1776, and is accompanied by a Portuguese translation-or perhaps, more precisely, a Portuguese version of the letter-under the same seal as the original, that of the Siamese King's 'Primeiro Ministro 1 ( /phra? khlalJ/ in Siamese). The recipient is the Danish governor of Tranquebar, (David) Brown. On an accompanying sheet is added information that the letters have been extracted from a bamboo tube. The letter was registered by Henning Engelhart in 1780 together with other documents concerning Tranquebar and the following resume of its contents was given (here translated from Danish): 'A letter in the Siamese language, from the King of Siam or his Prime Minister, to Governor Brown, containing requisitions of ammunition, inviting to trade with Siam, and promising assistance. 1 In 1973 the National Archives asked Pensak Chagsuchinda, Lecturer in Thai in the University of Copenhagen, and me for a comment on the accuracy of the above resume, and we supplied the Archives with a somewhat more detailed translation. Recently I had occasion to look at the letters again and decided to translate the original Siamese as well as the Portuguese versions and to supply some notes on both.l Transcription of the Siamese text l.l. nalJSYY nawaabthancawphajaaphra?khla1Jphuujajna?kru1Jtheebphra?miimathura? 2. cidsanidsaneehaa maatheelJjoolJbroonseenaacawdiinmagphuupencawmyaJJtraJJkaabaad duaj 3. phra?baads6mdedphra?mahaakrasadtraathfraadcawthiisuuJJjaj miiphra?raadcha?oolJkaanmaan

PAGE 102

Z..4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. na?maarasi:lm silaa?1igmyYnbaag t 'lhaj kapltanleg? ::bgmaacadsyY-na baad phaaraa mahaanakhaanca?khfdraakhaahajkhrobtaamcamnuankhaapyynpra Notes to the text z.. 1 / 'letter' The /a/ is written by means of the diacritical sign which is now reserved for the tonal marker /majthoo/. In this letter no distinction is made between /majhan ?aakaad/ and /majthoo/. This will not be pointed out 90

PAGE 103

91

PAGE 104

in every single case in the following notes. /nawaab/ 'Nabob'. The first syllable is spelt with the letter /noo neen/ as if it meant 'at, in'. The scribe may well have misunderstood the expression, thinking that /waab/ was a noun designating something like 'residence, domicile'. But cf. /nakhoon/ below. /than/ 'title'. This title is spelt with a long /aa/ as in the modern orthography, but without the tonal marker /maj?eeg/. /caw/ 'lord'. In this word the /majthoo/ has its usual force. All examples of this will not be pointed out below. /phajaa/ 'title' is written with a conventional ligature. /phra?/ 'honorific prefix' is written with a ligature in which the /r/ does not take on its usual shape. /jaj/ 'big' is spelt with /majmalaaj/ rather than /majmuan/, and with /majthoo/ instead of /maj?eeg/. /kruiJ/ 'capital city' is in this letter consistently equipped with a /jaadnamkhaaiJ/. /nakhoon/ 'city'. The first syllable is spelt with /noo neen/ rather than with /noo nuu/, and the second has the vowel /oo/ written out instead of the /r/ serving for vowel plus final consonant. /mii/ 'have' is spelt with a vowel symbol consisting of modern /sara? ?y?/ plus modern /maj?eeg/ (a combination of /sara? ?i?/ with both and /fonthooiJ/). This symbol is used in this style for any of the vowels /i/, /ii/ or /yy/. We shall call it the 'general i/y vowel'. /mathura?/ 'sweet'. Both of the short /a/ vowels are indicated with the There is not much consistency in the use or non-use of this symbol in this style. Z.2 /cid/ is spelt /cidtra/ even though the meaning is probably 'mind, heart' rather than 'variegated, wonderful'. The two are not consistently kept apart in modern usage either. The word is spelt with the general i/y vowel. /sanid/ 'amiable' is spelt with final /coo caan/ instead of /thoo thahaan/ or /thoo thoiJ/, the vowel is written with the general i/y. 'friendship' has the vowel /ee/ placed between /s/ and /n/ and is written without a tonal marker. 92

PAGE 105

'reach' is spelt with a combination of preposed /sara? ?ee/ plus the general i/y. This combination could not indicate any other vowel than /ee/. /braan/ can also be read /bruan/. /sara? and are not clearly distinguished in this style. /diinmag/ 'Denmark' reflects Portuguese Dinamarca. is spelt with rya/ (see also the discussion of the Portuguese forms for 'Denmark' in Egerod 1958:111-14). vowel is spelt with the specific symbol for long /ii/. The /n/ Pidgin The /pen/ 'to be'. The vowel shortener /majtajkhuu/ does not appear in this text. /myaQ/ 'city' is spelt with the usual elements, except that the superscript symbol is the general i/y vowel (perhaps with an added unexplained cf. l.5 below). /traQkaabaad/. The final consonant /d/ is spelt with reflecting an original supradental comes out as -r in Portuguese, Danish, etc. The initial /b/ in /baad/ is joined with the vowel /aa/ in a ligature involving the script form of the letter /b/. l.3 /somded/ 'royal prefix'. The /om/ is spelt with both the superposed /jaadnamkhaaQ/ and the /m/, thus presenting a mixture of Siamese (/m/ only) and Cambodian only) conventions. /krasad/ 'king' has an unetymological /-r-/ (Skt. sya/ rather than ryysfi/. The /i/ of /-athfraad/ is the general i/y. /thii/ is spelt in exactly the same way as /thyy/ would now be spelt. /maan/ is spelt as if it were Skt. 'killing'. The word is, however, of Cambodian origin and is the equivalent of Siamese /mii/ 'to have'. The example used in the Thai Royal Academy dictionary (Photchananukrom 1950:695b) we find here /maan phra?banthuun/ which is parallel to the preceding phrase /mii phra?raadcha?ooQkaan/. l.4 /banthuun/ 'command' is spelt with nuu/ in the first syllable and /laa liQ/ in the second, where the present official orthography has neen/ and rya/. Pallegoix (1854:47) has liQ/ finally as in Cambodian. /sura/ is spelt with a long /uu/ by confusion of /slira?/ 'godly' with /suura?/.'her-c'. The /r/ is doubled for no obvious rea,.son. 93

PAGE 106

/sfQ/ in /sfQhanaad/ 'lion's roar' is spelt with /ii/; /naad/ with chadaa/ through confusion of 'roar' thahaan/ with 'dancing' (though this is normally spelt with 'long t' rather than 'long d'). /damrad/ 'royal speech' has final deg/ instead of sya/. /nya/ 'above' has superscript /ii/ rather than /yy/ the first time, the general i/y the second time. /mam/ in /kramam; 'top of head', no tonal marker. /saQ/ 'command', no tonal marker. 'monsoon'. is spelt out. /ra/ is spelt with the rare letter /ryy/, perhaps by false association with words beginning with designating 'death', and also perhaps with /ryduu/ 'season'. l.5. /pii/ 'year' is written with a special ligature combining features of the superscript part of plaa/ and the superscript vowel which serves the same purposes as the general i/y symbol. /?adtha/ '8'. The thaan/ is replaced by the simpler chadaa/. /sag/ 'year' is written with the /fannuu/ or 'rat's teeth' (also so named in Cambodian) to indicate the short, otherwise 'inherent' /o/. The two consonants are joined in a special ligature with a lower than normal position of the upstroke of the saalaa/. 'Captain'. The first syllable is spelt with (identical with modern /maj?eeg/). The second syllable is spelt exactly like the word /pii/ 'year' above. /leg/ 'Personal Name'. Since no vowel shortener is used in this style, we do not know whether /leg/ or /leeg/ is intended. Whichever it is, it is likely to be wrong (see below: Notes to the translation). /syy/ 1buy', no tonal marker. /pyyn/ 'gun' contains the same ligature as is used for /pii/ and above. /myaQ/. The vowel is /sara? ?i?/ plus the element /jaadnamkhaaQ/, used in l.2 above, is missing.

PAGE 107

/khaw/ 'enter' is written with a ligature which blends khaj/ and /sara? ?aa/ into one symbol. l.6. /thawaaj/ 'give' has a superfluous /majhan?aakaad/ /majthoo/ to indicate the short /a/. 'must', no tonal marker. 'desire'. The 'inherent' /o/ is indicated with and not the /fannuu/ as in /sog/, l.5 above. There is no silent khwaaj/ finally. /silaa/ 'flintlock' is nonsensically spelt with initial /sii/ 'prefix of splendour' --the guns could not have been that exceptional! /myYn/ 110,0001 no tonal marker. / / 'then' with general i /y. l.T. 'arrange' has /majthoo/ instead of maj?eeg/. 'ask' is spelled with khuad/, which would be etymologically correct for the homophonous word meaning 'hook'. There is a /fannuu/ over the consonant. l.S. /kg/ 'to', no tonal marker. /majtrii/ 'friendship' is twice spelt with ?y?/. /chuaj/ 'help', no tonal marker. 'tend', spelt with neen/. 'prepare' is spelt with /majthoo/ for no obvious reason. l.9. /pyyn/ has been equipped with an extra /maj?eeg/. cf. ll.5 and 6 above. 'make to be'. The 'inherent' /o/ is indicated by means of as in l.6 above. /ea?/ 'verbal auxiliary'. In this word the is used for the short vowel /a/ instead of the /w{sanchanii/. l.lO. /k3?/ 'then' is spelt with kaj/ and no markers. /kampan/ 'ship', no tonal marker. 95

PAGE 108

'Place Name'. The short /a/ is spelt with cf. /kapitan/, l.5 above. /hog/ 'six' has to indicate the short /o/; cf. /somded/ l.3 above. 'per' has the vowel written out plus a /fannuu/ above the /t/; cf. /kh;o/ in l.7 above. l.ll. is written with a ligature that joins some features of /choo with the /aa/. /nya/ 'meat' is spelt with /sara? ?y?/ instead of /?yy/ and has no tonal marker. /maj/ 'wood' has /majmuan/ instead of /majmalaaj/, and no tonal marker. /sin/ 'goods' is spelled with the general i/y symbol, as is also 'thing' (no tonal marker). /daj/ 'any' has /majmalaaj/ instead of /majmuan/ and is equipped with a meaningless final /joo jag/. l.l2. /juu/ 'be at', no tonal marker, /banthug/ 'load'. The official orthography writes double /roo rya/ to indicate /an/, a convention not followed here. 'send'. The 'inherent' /o/ is indicated with (cf. ll. 3,10) above); there is no tonal marker. /daj/ 'can' is spelt with /majmuan/ rather than with /majma.laaj I. l.l3. /kh!d/ 'calculate' is spelt with /sara? ?ii/ (or sara? ?i?/ with superposed /raakhaa/ 'price' is equipped with a /majthoo/ on the second syllable --not used in the line above --through confusion with /khaa/ 'price', spelt the same way later in this line, with /majthoo/ instead of /maj?eeg/. /khrob/ 'complete' has for /o/ as in several other words above. /camnuan/ 'amount'. The final /-n/ is written /roo rya/. /m1?/ 'not' has the same elements as /mii/ in l.l, but with the closer to the which is, however, not the case in l.l7 below.

PAGE 109

'unfinished'; no tonal marker. /keen/ 'exceed'. The superscript is /sara? ?yy/ and not /sara? ?1?/. /son/ in/khadson/ 'needy' has for /o/ as in /khr6b/, etc., above. l.l4. 'one' is spelt This is often the actual pronunciation in modern Thai; the word has a complicated history with dialect variants contradicting each other as to tone and initial consonant category. /phan/ 'kind, sort' is lacking the final /noo neen/ in the first, but not in the second instance, whereas a /majhan?aakaad/ is used in the first, but not in the second instance. /phaa/ 'cloth', no tonal marker. /phr/ 'silk' has double /roo rya/ for no obvious reason, surrounded by two occurrences of /phan/, correctly spelled with double /r/. l.l5. /khaaj/ 'sell'. The initial plus the vowel are written by means of a ligature combining /kh;o khuad/, not etymologically correct, and /aa/. /na?/ 'at, in' seems equipped with a differently from ll. 1 and 16. /raakhaa/ is spelt as in l.l3; /phan/ as in the first instance in l.l4. /nya/ is written with the general i/y vowel, differently from l.ll. l.l6. /nan/ 'that' is equipped with two superimposed markers identical with modern /majthoo/, the lower one indicating the vowel, the upper one the tone. /kap1tan/ has no marker on the first syllable. l.17. /theed/ 'final particle' is spelt with /thoo thahaan/ rather than and with superscript /sara? ?yy/ rather than /sara ?1?/. /chuaj/, no tonal marker. /mf?/ seems equipped, on top of the general i/y vowel, with a meaningless /majthoo/. 97

PAGE 110

/phanphaa/ is spelt as in Z.l4. 'of' is spelt with khuad/, which happens to be etymologically correct. 'Place Name'. The last syllable has a meaningless /majthoo/. /khaad/ 'lack' begins with the same ligature as /khaaj/ in Z.l5, again not etymologically correct. The last syllable of /?ankhaad/ 'absolutely' seems to be spelt the same way. The words in between /khaad/ and /?ankhaad/ are hard to make out, because of the seal which is superimposed upon them. Z.l8. /thii/ 'instance' is spelt like the modern /thyy/. 'Mars, Tuesday' is spelt with a long vowel /aa/ found in the first syllable as well. /dyan/ 'month' has superscript /sara? ?ii/ instead of /sara? ?yy/. /?aaj/ 'first', no tonal marker. /s1b/ 1101 has /sara? ?ii/ instead of /sara ?1?/. /phan/ 11000' is written with final /r/, differently from Z.5 above. 11001 no tonal marker. The orthography in the present letter, besides being old-fashioned, is also extremely inconsistent. The usage of the superscript vowels, including the general i/y symbol, is completely haphazard. The tonal marker /maj?eeg/ is not used, but the identical indicates (inconsistently) /a(?)/ or /o/. The tonal marker /majthoo/ is identical with /majhan ?aakaad/, while the /fannuu/ occasionally marks the /o/, or is added above the consonant preceding /sara? Several now obsolete ligatures are employed.

PAGE 111

Translation of the Siamese text l.l. This is a letter from the Nabob, the Lord Grand Minister of the Treasury in the mighty city of Krungthep, with kindness 2. and affection to (John!) Brown, Danish officer who is the Governor of Tranquebar, to the effect that, in accordance with 3. H.M. the King and Supreme Ruler's order and 4. command uttered by him; during this monsoon 5. of the Year of the Monkey, the Eighth year of the Decade, Captain Leit went to arrange for the purchase of 1000 flintlock guns in the city of Tranquebar and has reported back 6. to H.M. and made delivery. As the King now wishes further 10,000 guns, so [. I have charged Captain Leit with sailing to Tranquebar and arranging for the purchase there. I ask you, Lord Governor of Tranquebar, 8. for the sake of H.M.'s friendship to be good enough to assist and take care of arranging for Captain Leit 9. to acquire the 10,000 guns and get them safely here. If you, Lord Governor of Tranquebar, should want tin 10. you may arrange for a ship to come and receive the tin through Captain Leit in Phuket, delivering 6 guns per 11. load. If you should want ivory, wood, or other kinds of merchandise from the mighty city of Krungthep, Captain 12. Leit is informed about the price. Please arrange for the ship to load the guns and carry them to the mighty city of Kru:ngthep. 13. We shall calculate the price in full agreement with the total value of the guns so that there is no deficit or inadequate pay. Furthermore, 14. if you, Lord Governor of Tranquebar, will arrange for the ship to carry various sorts of cloth and silk as well as other merchandise 99

PAGE 112

l.l5. to be traded in the mighty city of Krungthep, then as for the price of the cloth (as compared to) the price of ivory, wood, and other merchandise 16. in the mighty city of Krungthep, Captain Leit is already informed about it. Please arrange for the ship to be loaded and carry 17. the merchandise here. We shall assist and see to it that none whatsoever of your merchandise, Lord Governor of Tranquebar, will remain unsold. 18. This letter was written on Tuesday, the First Lunar Month, the Tenth Day of the Waning Moon, in the Year 1138 of the Lesser Era, the Year of the Monkey, in the Eighth Year of the Decade. Notes to the translation l.l. 'Nabob'. The Siamese form is /nawaab/, very close to Urdu nawab*, and Malay nawab*. The usage of this Mogul title does not seem to have been common in Thailand; it is not recorded in the dictionaries. 'Minister of the Treasury'. is the Treasury, but the actual function of this official was, at the time, more like a Minister of Foreign Affairs. The Portuguese version says 'primeiro Ministro'. 'Krungthep'. In 1776, when this letter was written, the capital was at Thonburi, alternatively called T'anaburi and Dhonburi (see Hall 1966). Taksin installed himself as king in Thonburi in 1767; Chakri was crowned in Bangkok, across the river, in 1794). l.2. '(John?) Brown'. The Siamese text says (or bruan)/. The name of the Governor of Tranquebar in 1776 was David Brown, a member of a Scottish family which had settled in Denmark. David Brown lived from 1734 till 1804 and was Governor of Tranquebar for the period 14 February 1775 -17 January 1779. His elder brother John Brown (1723-1808) was a well-known merchant and shipowner. In 1776 the Danish king bestowed upon nawab, from the broken plural of the Arabic naib 'deputy' was treated as a singular by the Persians, and exported as such by them. (See also Yule 1886:467a-9a). (Ed.). 100

PAGE 113

him the honorary title of Commissioner-General of War. John Brown served as a member of the Board of Directors of the (Danish) Asiatic Company in 1770-75 and 1779-85; he carried on private shipping with East India from 1774. It seems that the author of the letter had somehow mixed up the two brothers. Tranquebar was Danish from 1624 till 1845. l.3. 'The King of Siam' was Phaya Taksin (1767-1782). l.5. 'Captain Leit'. The Siamese text has /le(e)g/, but the Portuguese translation says Leit. The final /-g/ may have seemed more natural to the Thai letter writer, since there is no Thai word /led/. Were the Captain a Dane, Leth would be a likely guess, but cf. to l.5 of the Portuguese translation. As regards the calendar terms, see l.l8 below. l.lO. 'Phuket'. The Siamese text has i.e. Ujong Salang or Junk-Ceylon (see Yule 1886:36lb). l.ll. 'Load'. The Siamese text says /phaaraa/ from Skt. bhara (= 20 tula or 'quintals1). 'Wood'. The Siamese /nyamaj/ means the 'flesh' or grain of the wood, the wood stripped of the bark. The Portuguese translation makes it clear that the fragrant 'eagle-wood' is meant. l.15. 'Other merchandise'. The letter says only /sin/ 'goods 1 the equivalent of ;sinkhaa/ in U.ll and 14. l.l8. The 'Lesser Era' (L.E.) or Minor Era /cunlasagkaraad/ began in A.D. 638, so L.E. 1138 = A.D. 1776 or 'Buddhist Era' (B.E.) 2319. 'Year of the Monkey'. /pii w3ag/ is the ninth year in the cycle of twelve. 'Eighth Year'. /?adthasog/ is the eighth year in the cycle of ten. 101

PAGE 114

102

PAGE 115

The Portuguese version Excellentissimo Senhor l.l. Supost esperando a boa receba V.E.a estas minhas regras, com alguCma 2. caricia, dezando em primeiro Lugar Logre V.E.a prospera, e felix saude com muitos 3. augmentos e [aJJges de felicidades de dia em dia majores no seu nobre Estado, e feliz 4. governo. 5. Veio nesta o M.r Leit neste Reino de Siam, e fez som-6. baya a S. Majestade, que elle dito tem comprado em Trangabar mil espin-7. gardas e aprezentou aS. Majestade, e S. Majestade ficou bastantemente conten-8. te comas ditas Espingardas: de prezente S. Real Majestade executou ordem 9. a mim primeiro Ministro de S. Majestade que escrevesse esta Carta a V.E.a pedindo que 10. ao dito Leit em Trangabar, porque vai o dito M.r Leit 11. com ordem expressa de S. Majestade procurar mais dez-mil Espingardas em Trangabar: 12. a V.E.a por boa amizade, que o Leit coma faculdade de V.E.a se ache elle dito 13. Leit as ditas dez-mil Espingardas, e se V.E.a quizer Calaem mande barco 14. de V.E.a a recebar Calaem das de Capitao M.r Leit, elle dito se entre-15. gara por seis Espingardas, hum bar de Calaem, e se quizer marfim, aguilla, e mais fazen-16. das deste Reino de Siam, o M.r Leit entende, e sabe o das ditas fazendas, e 103

PAGE 116

z.n. 18. 19. 20. 21. 22. 23. 24. 25. V.E.a mande trazer espingardas no barco de V.E.a a este Reino de Siam, e chegan do se ajustara as contas puntualmente das ditas espingardas sem difficuldade algua; e se praz V.E.a mandar barco a este Reino de Siam o preco das roupas, e pessas, etc. o Capitao M.r Leit sabe ja o preco dos ditos panos, e pessas, etc. e tambem o dito Capitao M.r Leit sabe o preco de mar-fim, e outras mais fazendas deste Reino de Siam, pode V.E.a mandar barco a comenciar neste Reino, e nos Ministro de S. Majestade de Siam ajudar, e beneficiar as ditas fa-zendas, e satisfazer puntualmente. --Ecomo por horas se offerece mais ficamos apprecando a V.E.a muita vida, saude com majores progresses de bens, e felicidades como lhe dize-ja com devidos respeitos quem sempre he De V.E.a Todo affectuozo Assignamos com nosso sello Siam a os 7: de Dezembro de 1776 Notes to the Portuguese version Z.l. V.E.a, i.e. Vossa Excellencia Lugar Logre. In the transcript, the capitalization of the original has more or less been followed, and these two words are clearly capitalized in the original. There are, however, quite a few cases where it has been impossible to decide whether an initial consonant represents a stylistically enlarged small lette r or a true capital. For the interpretation, this luckily, plays no role. Z.3. Ca?Jges. Because of a fold in the paper, the word is not clearly legible. 104

PAGE 117

l.5. M.r, i.e. Mister. Ms Diana Matias has kindly called to my attention the recent article by Professor E.H.S. Simmonds (Bull. Soh. Or. Afr. Stud. 50 (3), 1987, 529-531) entitled 'A letter in Thai from.Thalang in 1777' which describes a transaction very similar to the one treated in the present article. No doubt our Capitno Leit is identical with Captain Francis Light in the 1777 letter. Iri both letters the Siamese transcription is /leg/ with /h;o nam/. See also Professor Simmonds' article for further bibliographic information on Francis Light. l.l5. l.l8. l.l9. l.20. l.24. hum, i.e. um. algua, i.e. alguma. pessas, i.e. panos, i.e. pannos. apprecando, i.e. muita and majores are written by means of ligatures. The first one could perhaps be read melhora instead of muita. 105

PAGE 118

Translation of the Portuguese version Most excellent Sir L.l. Hoping for your kind acknowledgement, I ask your Excellency to receive these lines of mine with some 2. favour, while at the same time I express the wish that you may enjoy a prosperous and joyous health with much 3. success and plenty of happiness increasing day by day in your noble state and felicitous 4. government. 5. During this present monsoon the Captain Mr. Leit arrived in this Kingdom of Siam and paid 6. obeisance to His Majesty, [reporting] that he, the said captain, had bought one thousand flintlocks in Tranquebar, 7. and he delivered them to His Majesty, and His Majesty was rather satisfied 8. with the said flintlocks: now His Majesty has issued an order 9. to me, His Majesty's Prime Minister, to write this letter to Your Excellency, asking you 10. to favour the said Captain Mr. Leit in Tranquebar, so that he 11. with His Majesty's express order shall procure further 10,000 flintlocks in Tranquebar: 12. I ask Your Excellency, for the sake of good friendship, that Captain Leit through your authority acquires 13. the said 10,000 flintlocks, and if Your Excellency wants tin, 14. send a ship of yours to Junk-Ceylon to receive tin from the hands of Captain Mr. Leit, who 15. in exchange for six flintlocks will deliver one load of tin, and if you wish ivory, eagle-wood and other merchandise 16. from this Kingdom of Siam, Captain Mr. Leit is acquainted with them and knows the price of the said merchandise, and 106

PAGE 119

l.l7. Your Excellency should order the transfer of the flintlocks on your boat to this Kingdom of Siam, and when they arrive 18. the accounts for the said flintlocks will be settled without any difficulty; and if it pleases your Excellency 19. to send a ship to this Kingdom of Siam, [as regards] the price of linens, firearms, etc., Captain Mr. Leit knows 20. already the price of the said cloth, firearms, etc. And as the said Captain Mr. Leit also knows the price of ivory, 21. and other further merchandise of this Kingdom of Siam, Your Excellency can direct a ship to trade 22. in this Kingdom, and we, Minister to His Majesty of Siam, [willJ assist and facilitate the said dealings, 23. and give satisfaction punctually. --And since, for the time being, there are no more items of business at hand, we shall continue 24. to esteem highly a long life for Your Excellency, good health and greatest progress in possessions and happiness, such as 25. with due respect we wish for you that it may always be. Your Excellency's devotedly we endorse with our seal Siam, the seventh of December, 1776 Notes to the translation l.3. 'plenty of'. This translation is based on a reading auges 'apogee', for [a?Jges, suggested to me by Luis de Vasconcelos. l.6. 'obeisance'. Translates sombaya, probably from Malay sembah or sembahyang.

PAGE 120

Z.l2. 'friendship'. From the Portuguese version of the letter one could gain the impression that the Minister and the Governor were well acquainted. The Siamese original says, however, expressly, 'His Majesty's friendship'. Z.l3. 'tin'. The term is calaem, from Arabic. Z.l4. 1 Junk-Ceylon 1 i.e. Phuket from Malay Uj ong Salang. (See also Yule 1886:36l(b).) Z.l5. 'load'. The Portuguese version uses bar, corresponding to the Siamese original /phara/, both from Skt. bhara. 'eagle-wood' translates Portuguese aguilla. (See also Yule 1886:258ab who traces the word to Malayalam. ). Z.l9. 'firearms' translates the Portuguese 'pieces, firearms', here evidently referring to the flintlocks already mentioned. Z.24. 'long' translates muita. If we read melhora instead, the meaning will be 'best possible'. 'devotedly'. The Portuguese says todo affectuozo 'totally devoted 1 NOTE 1. I have benefited from correspondence with Dr Theraphan L. Thongkum and, through her, Professor Preecha Changkhwanyuen, concerning the Siamese text; as well as with Palle Kroman M.A. and, through him, Lecturer Luis de Vasconcelos, concerning the Portuguese text. The final responsibility is, in all cases, my own. 108

PAGE 121

Egerod, S. Hall, D.G.E. Pallegroix, J.B. Photcha:nanukrom Yule, H. & A.C. Burnell REFERENCES 1958. Pidgin Portuguese A.D.l621. T'oung Pao 46, 111-4. 1964. A history of South East Asia. London: Macmillan. 1854. Dictionnarium linguae thai sive siamensis; interpretatione gallica et illustratum. Paris. (Republ. 1912, Farnborough: Gregg Internat. Publ.) 1950. Photchananukrom chabap ratchabanditsathan. Bangkok: Min. Educ. (6th ed.) 1886. Hobson-Jobson: being a glossary of Anglo-Indian colloquial words and and of kindred terms ... London: John Murray. 109

PAGE 123

TAI NAMES FOR THE ox1 William J. Gedney For many of the co=on domestic animals, languages of the Tai family use names which show complete uniformity and agreement throughout the Tai-speaking domain. For the 'pig', for example, forms cognate with Siamese (Si.) muu5 occur everywhere, always showing the phonological developments expected in the particular language or dialect. Such regularity is exhibited also in the termt for the 'water buffalo' (Si. khwaayl), the 'horse' (Si. maa ), the 'dog' maa5), the 'chicken' (Si. kay2), and th-;-rduck' (Si. ) One infers that these animals were known to the speakers of Proto-Tai, the parent language of the family, and that these names go back to that period. But the Tai names for some other domestic animals show irregularity from one language to another. For example, the term for the 'goose' (Si haan2) often exhibits forms which violate the usual phonological correspondences, suggesting late acquisition of the animal and its name in the various branches of the Tai family. The name for the 'cat' (Si. is also sometimes irregular, suggesting onomatopoetic innovation from place to place. For the 'goat', most Tai languages use a name which ought to be in Siamese, but Siamese uses instead the form which cannot be cognate with the usual Tai term; one suspects that the old name was lost --goats are rarely seen in Central Thailand --and replaced by a loanword. The Shan term (Cushing 1914:419) is also aberrant. One of the most interesting examples of disagreement in domestic animal names among the Tai languages is found in terms for the 'ox', terms which show not only phonological irregularity but often completely different words from one area to another; a situation described here by studying the various forms of the names used and determining their geographical distribution; and also suggesting a reason for this disparity. In general, three different names for the 'ox' are found in Tai languages, each one used over a wide area. These three areas coincide closely with the three branches which Li Fang-Kuei (1977) has postulated for the Tai family: South-Western, Central, and Northern. South-Western type: Siamese South-Western Tai languages are spoken in Thailand, Laos, Burma, Assam, southern Yunnan, and north-western 111

PAGE 124

Languages of this group all agree in having in their words for the 'ox' the diphthong -ua (as in Si. gual) everywhere that this diphthong exists, i.e., in Siamese and all other Tai dialects of Thailand, all dialects of Lao, and in Black Tai and Red Tai. The diphthong has changed to -oo in all dialects which have made the regular change of the high diphthongs ia, ia, ua to the mid monophthongs ee, ea, Q respectively. Dialects which have made this change of high diphthongs to lowered monophthongs cover a wide arc extending from White Tai in the east, through Lu and Tai Nua in southern Yunnan, to Shan and Khamti in the west. So far as tone is concerned, all South-Western languages agree in giving this word the tone (whatever it happens to have become phonetically) belonging to A-tone words of the parent language with an originally voiced initial, so that in each modern dialect the word for 'ox' has the same tone as, for example, 'to have' (Si. miil), 1ricefield1 (Si. naal), 'boat' (Si. and so on. Because of this absolute tonal regularity, which implies that the word existed with the A-tone and a voiced initial in South-Western Tai languages before the tonal many hundreds of years ago, tone marks will be omitted from the forms cited. Cognates of Siamese gua in other South-Western Tai languages and dialects agree, then, with Siamese in the vocalic nucleus and in tone. If only these constituents were involved, we would feel confident in reconstructing a Proto-South-WesternTai form *gua. But this turns out to be impossible because the initial shows considerable variation and irregularity. Besides forms with initial other forms occur, sometimes in addition to the form with initial and sometimes as the only recorded form for the particular language or dialect. The other initials that appear are w-(or v-) and h-. Siamese has gua as the usual colloquial form, but also wua which is used in more formal and educated speech. How this situation came about in the Standard Thai dialect of Bangkok is a puzzle which ought to be addressed by students of the linguistic history and dialectology of Thailand. The old Pallegoix dictionary of 1854 (481, 869) gives both qua and wua, as do modern dictionaries. Egerod (1961:80) in his paper on Thai dialects cites wua for Bangkok (he must have had an educated informant, or one given to putting on airs), and also for Loei in the North-East, gua for Yuan (the dialect of Northern Thailand), and hua for Phatthalung, Phuket, and Trang. Initial h-in these dialects in Thailand is the regular reflex of earlier so that these forms with initial h-are really examples of earlier Purnell's glossary (1963:73) of the dialect of Northern Thailand gives Both the Reinhorn (1970:396) and Kerr (1972:266b) dictionaries of Lao give gua for that language (see also p.ll5 below). White Tai has goo {-Di'eu 1970:253, Gedney, fieldnotes); Black Tai has (Gedney, fieldnotes); Red Tai has 112

PAGE 125

IJUa (Robert 1941:130). For Lii (in Yunnan, directly north of Northern Thailand) I have recorded voo and hoo in the dialect of Chiang Rung, but ]2 in the Lii dialect of Moeng Yong across the border in Burma. For Lii, Li (1977:240) also gives variants: goo, voo, hoo. Tai Niia, located still further north in Yunnan, has voo in the dialects which I recorded, but Harris (1975:211) found ]2 For Khiin (the dialect of Chiang Tung in Burma) Egerod (1961:80) gives woo, which he also gives for Shan (ibid.); while Cushing's Shan dictionary (1914:623, cf.l54) both woo and ]2; he also enters a form (see below p.ll5). Harris (1976:126) in Tai Mao and in Tai Khamti; Weidert (1977: 51) also records !JOO for Tai Khamti. For Ahom, both the earlier much-cited dictionary of Borua (1920:298-9) and the more recent lexicon by Barua and Deodhai (1964) give the transliteration hu; the vowel appears in Ahom forms which have cognates in other South-Western Tai languages with the diphthong ua. There seems to be no way within the usual rules of comparative Tai phonology to reconcile the three different initials occurring in the forms gua (or goo), wua (or woo or voo), and hua (or boo). So far as the alternation]._ ...... :!:!:_ is concerned, both Egerod (1961:80) and Li (1977:239-40) have suggested a reconstructed initial cluster *gw-; a possibility mentioned much earlier by Haudricourt (1948:218). It seems to be an ad hoc guess for which there is really no very convincing evidence or argument from other examples, especially since this word occurs only in the South-Western branch of Tai. And besides the ]._-:!:!:_alternation, the initial hof hua and hoo is just as baffling, and makes matters even worse. The initial h-in dialects of Southern Thailand is, as indicated above, a local matter, a regular development of voiced initial But initial h-elsewhere cannot have any such explanation; in these other dialects all other words with initial such as 'snake' (Si.guul), 'sesame' or 'ivory' (both Si.Qaal), have regularly modern initial (An unusual exception is Tai Niia, where I found the initial palatal nasal J1-, and Harris (1975: 211, e.g.) in words which elsewhere have the initial velar nasal In many of these dialects modern initial hwith a tone indicating an earlier voiced initial corresponds to Siamese initial r-; this is true, for example, in Lao, White Tai, Black Tai, Red Tai, Lii, Shan, and so on. It is not true, however, for Ahom, where initial Siamese r-regularly corresponds to Ahom =_, but Ahom has, as we have seen, initial h-in the words for 'ox 1 In any case, initial r-, which has changed to h-in some languages, clearly has nothing to do with our word. The voiceless counterpart of Proto-Tai voiced that is, voiceless preaspirated *hiJ-, regularly changes to h-in many Tai dialects. It seems impossible that this change 113

PAGE 126

can have anything to do with the occurrence of initial h-in the word for 'ox', which everywhere has a tone reflecting an earlier voiced, never voiceless, initial. Voiced yield everywhere, except for the unusual local developments in Southern Thailand and in Tai Nua. Modern h-, with a tone reflecting an original voiced initial, is from earlier *=.in dialects where the change *r-> h-has occurred. Elsewhere, as in Siamese, initial h-with a tone reflecting an original voiced initial is a historical impossibility; the few words of this kind that occur have to be modern innovations. The whole phonological picture involved in these forms with alternating initial w-, and h-is bizarre in the extreme. Whether further detailed research into the forms for 'ox' in every local dialect would throw light on this problem seems doubtful; more information on the geographical distribution of initial w-(or v-), and h-in this word would not ameliorate the basic phonological anomaly. The likeliest explanation is probably that this is a loanword, borrowed by different Tai dialects from different nonTai source languages or dialects, perhaps at different times. The different initials would then be due to differences in the source forms, either in space or time. What looks like the same word is found in the Be language of Hainan, which is geographically remote from the SouthWestern Tai languages and genetically somehow remotely related to the Tai family. The Savina-Haudricourt dictionary (1965:62) gives the Ee :f:orm as ngJu, and Hashimoto's lexicon (1980:20) gives it as Haudricourt identifies this Be form as a Chinese loanword. Central type: moo or maa Languages of Li's Central branch of Tai are spoken in the north-eastern part of where they are known by such names as Thg and Nling, and across the Chinese border in adjacent areas of Guangxi province. There is also a displaced Central Tai dialect farther west in the neighbourhood of Lao Kay, about half-way across the northern border of Viet-Nam, on which I have done fieldwork and which I have designated Western Nung. All Central Tai dialects, whether in Vi%t-Nam or China, exhibit amazing uniformity among themselves in calling the 'ox' by the term moo or maa. These are the same form, phonetically [ma:J. In some dialects there is a distinction between the vowels oo and aa; in such cases our word has the form maa. But in many areas there is no such distinction, and phonetic [a:J is commonly phonemicized as /oo/, giving our word the transcription moo. This vowel [a:J, whether phonemicized corresponds historically to the vowel aa in Siamese, as in ll4

PAGE 127

Siamese 'father', 'pot'. This Central Tai word moo or always has in these dialects (like gua above) the which developed from an earlier A-tone with originally voiced initial, so that the Siamese cognate of this Central Tai word, if it occurred, would be To give a few examples, this form of the word for 'ox' is given in Savina's (1924:6la) dictionary of Nling and also in his (1910) dictionary of Tay. Li (1977:240) also cites it from Diguet's glossary of Tho. It is the word for 'ox' in the displaced Western Nling dialect mentioned above. Across the border in China, Li (1940:26lb; cf.l977:172) gives this form for the dialect of Longzhou, and I have recorded it in the dialects of Lung Ming, Lei Ping, Ning Ming, and Ping Siang. The complete agreement among these Central Tai dialects in the form of this word suggests that it is old in this branch of Tai. Whatever its ultimate origin, whether native or a loanword from elsewhere, its shape suggests an onomatopoetic creation. This word for 'ox' is recorded also as occurring in two languages of the South-Western branch, which otherwise and more commonly have for 'ox' words of the type gua described above. Both the Reinhorn (1970:1672) and Kerr (1972:930a) dictionaries of Lao have brief entries for 'ox', marking it as 'archaic' -one would like to know when and where and how this word was used in Lao --and Cushing's (1914:509) Shan has an entry for glossed as 'a couplet for woo, as ka5 woo ka5 "to traffic in bullocks"', apparently roughly equivalent to Siamese khaa4 gual khaa4 khwaayl ('to trade in cattle', lit. oxen and water buffalo). Both the Lao and the Shan forms agree in tone with the Central Tai word. Can it be that in early times traders from Viet-Nam far to the east brought oxen (and their name for the ox) to sell in Laos and even as far as Burma? ? Li (1977:240) forms from Nling, and Shan under his entry for Siamese gua with a query. There seems to be no doubt that and gua are actually entirely different words. The phonological problems involved in the word gua, discussed above, are bad enough without dragging in The Dioi dictionary (Esquirol 1908:319) of a Tai dialect of the Northern branch gives a form mo1 glossed as 'cri du boeuf'. The example given makes it clear that the word is a verb. The tone is that of an original B-tone word with an original voiced initial, so that the cognate in Siamese, if it existed, would be We may safely assume that this is a local onomatopoeia, no more significant for comparative purposes than the English 'to moo'. Interestingly, the Saek 2 word for 'ox' is which is at least partially similar to our Central Tai type or 115

PAGE 128

but it may well be that the Saek form has nothing to do with the Central Tai word. The tone of Saek boo4 marks it as a non-native word. The Saek 4th tone (phonetically high rising-falling) occurs on words that had the earlier A-tone with an originally voiced initial (like, for example, Saek naa4 'ricefield', cognate with Si.naal; or raan4 'house', cognate with no native Saek words with initial b-occur with this tone. Saek bwords with the original A-tone have now the first tone (phonet: ically mid level with a slight rise), e.g., blianl 'moon', cognate with Siamese It looks as if Saek speakers during their southward migration from the Northern Tai areas of China lost (or neYer had) the typical Northern Tai word for 'ox', and somewhere acquired the Vietnamese name for the animal. Northern type: Yay sia Tai languages of the Northern branch are spoken in a fairly large area of southern China --especially in eastern Yunnan, southern Gu1zhou, and western Guangxi --with a small spill-over across the border into Vi%t-Nam, and there is also the displaced Northern Tai language, Saek, now located far to the south. So far as available data show, all these Northern Tai languages, except Saek, agree in using the same word for the 'ox'. In Yay, spoken in the neighbourhood of Lao Kay the Chinese border in the form of this word is sia (Gedney, fieldnotes), while the old much-cited Dioi dictionary by Esquirol and Williatte (1908:46) gives the form chie2 for this Gu1zhoii dialect. Northern Tai forms cited in Li (1977:169,172) are Po-ai sii, Dioi chie, Hsi-lin sie, and Wu-ming sii, all with Li1s tonal category A2. For Po-ai the index to the book gives a different vowel, sii (Li 1977:339). In the glossary of Li's monograph on the dialect of Wu-ming the form is transcribed, with two tones indicated, high falling or low falling (Li 1956:222b, Buyi, the Chinese publication on forty dialects of Pu-yi in southern Gu1zhou also enters the word for 'ox' (Buyr 1959:223,0219). These forms are: at Points 1, 2, 4-7, 9, 16, 19; at Point 3; at Points 8, 13-15, 17, 30; tsie2 at Points 10, 20; tsm2 at Points 11, 12, 22, 24-26, 28, 29, 36, 40; sm2 at Points 18, 38; tsei2 at Points 21, 39; tsme2 at Point23; tshm2 at Point 2 at Points 31-35; si 2 at Point 37. ----There would seem to be no doubt that all these Northern Tai forms are variants of the same word. Many of the apparent differences among them are due merely to different transcriptional conventions; a example of this is found in 116

PAGE 129

Li's two different (1956 v. 1977) transcriptions of the Wu-ming form. Some of the Pu-yi forms obviously reflect minute allophanic variants rather than genuine phonological differences, while in other cases there seem to be genuine phonological differences; for example, four different vocalic nuclei see:.n to be represented, the monophthongs ii and (high back unrounded) and the diphthongs ia and --For our purposes the interesting question here is whether the genuine differences among the various Northern Tai forms (setting aside mere differences in transcription and nonphonemic phonetic variants) agree with, or violate, regular phonological correspondences among these various Northern Tai dialects. If they agree, then the inference is that this name for the 'ox' is old in this branch of Tai. If they fail to agree, then we must suspect a later innovation borrowed from outside or borrowed from one dialect to another, or both. To test this question we must examine three elements, the tone, the initial, and the vocalic nucleus. So far as tone is concerned, virtually all the forms have tones which are the regular development of the earlier A-tone in words having an originally voiced initial. This is true of the 4th tone in Yay, Li's A2 category, the 2nd tone in Pu-yi, and the raised 2 symbol in Dioi. We find only two apparent exceptions. Pu-yi Point 3 has not 2nd tone but 3rd. The Pu-yi 3rd tone is the one normally developed from the earlier C tone with an originally voiceless initial. This might be an error in transcription, though usually the editors of the Buy'i material mark forms where they suspect error, but make no comment here. Other 2nd tone words in Pu-yi have the expected 2nd tone at Point 3, with a great many examples, but with one other exception: Pu-yi 'expensive' (BUy'i 1959:197,0025, cognate with has 3rd tone at Point 3, just like 'ox'. This inconsistency can hardly be the result of mishearing by the recorder; at Point 3 the 2nd tone is low falling, the 3rd tone high rising (ibid.l6). And for we have the puzzling indication of two tones on the word for 'ox' in Li's monograph (1956:222b), although in his Handbook (1977:240) he seems to disregard the first of the two tones given in the glossary. Of the two tones in the glossary, the latter one, low falling, is the one expected for Li's A2 category. The former, high falling, is the one developed from the earlier C tone in words having an original such as Wu-ming ram 'water' (cognate with Si.naam ); this tone would be historically wrong for our word. It is difficult to decide whether these two apparent tonal deviations, in Pu-yi Point 3 and in Wu-ming, are significant enough to cast doubt on the age of our word in Northern Tai. If it turns out that our study of the initials 117

PAGE 130

and the vocalic nuclei shows phonological regularity, then we would probably be justified in disregarding these two minor tonal deviations and assume that they have some explanation. Turning to the initial, Northern Tai languages all have two contrasting sounds of this general phonetic type (sibilants and similar sounds). In Yay, these two sounds are ._ and respectively. In Li's Po-ai material cited throughout the Handbook (1977), they are sand() (a voiceless lateral In his Wu-ming (1956), they are In Dioi (Esquirol 1908) they are represented respectively by eh (presumably meant, in the French manner, to indicate the sibilant !land... In the first is represented by a great many different phonetic symbols, as shown in the list of forms cited above; the second is represented by... The first of these two Northern Tai sounds (Yay s, Po-ai s, etc.) is the one occurring as the initial of our :;ord for 'ox'. This initial occurs both in words whose tone indicates an original voiced initial and in words whose tone indicates an original voiceless initial. In the former category, the one which concerns us, this initial usually reflects an original voiced .::_j_, appearing in modern Siamese as eh. There are a few other Northern Tai words in whjch this initial has other sources, e.g., Yay saw4 'evening meal', 3cognate with the obsolete Siamese word phrawr:and a few other words believed to have had an original voiced cluster initial (see Li 1977:95). It seems unlikely that these need concern us, since the vast majority of words with this Northern Tai initial go back to .::_j_, but anyone seeking the ultimate origin of the Northern Tai word for 'ox' will have to keep such alternative sources of the initial in mind. In attacking the question of regularity of phonological correspondences in initial in the word for 'ox' among the Northern Tai dialects, one looks, of course, for words in each dialect having the same initial. One is at first overwhelmed by the abundance of available examples; each source has dozens of words with this initial. Therefore, a selection of half a dozen of the more familiar examples is made here to try to render the material manageable. 4 'early' (Si.chaaw4): Yay saw 6 ; Po-ai sau C2 (p.l68); Dioi chaou3 (p.36); Wu-ming sau(p.l70) <;:eu (p.217); Pu-yi ( 8) 4 -6p.219, 01 0 <;:au at Points 1-7, 1 17, 19; sau a oln s 13, 18, 38; tsau4 at Points 22-25, 30, 32-35, 40; tsau3 at Points 31, 36; form missing at Points 9-12, 20, 21, 26-29, 37, 39. 118

PAGE 131

'rope' Yay saak5 ; Po-ai saak D2L (pp.31, 35, 168, 282); Dioi cha1 (p.23); Wu-ming saak (p.l71), 5 QckJ (p.25); Pu-yi (p.219, 0192) Qek 8 at Points 1, 2: Qa6 at Points 3-7, 16, 17; sa6 at 9, 13-15, 37; at Points 10, 22, 24, 40; tsak8 at Points 11, 12, 20, 21, 26-28, 39; sak8 at Points 18, 38; Qa? 8 at Point 19; tsa?7 at Point 23; tsak7 at Point 25; tsa6 at Points 29-31, 36; tsa3 at Point 32; tsa4 at Points 33-35. 'name' Yay soo5 ; Po-ai srr (pp.l69, 265); Dioi cho1 (p.54); Wu-ming so (p.l72), Q.0J (p.221); Pu-yi (p.220, 0196) Qo6 at Points 1-7, 16, 17, 19; so6 at Points 8, 9, 13-15, 18, 38; tso6 at Points 12, 20, 22-24, 26, 28, 32-34, 36, 40; tseu6 at Points 11, 29; tso2 at Point 27; 6 --6 -tsu at Point 29; tsue at Points 30, 31, 35; form missing at Points 21, 25, 37. 'hole' Yay Po-ai B2 (p.l69); Dioi chong (p.60); Wu-ming (p.l71), (p.22l); Pu-yi (p.220, 0202) at Points 1-7, 16, 17; at Points 8, 9, 13-15, 37, 38; tsog6 at Points 10-12, 30, 39; suag6 at Point 18; Quig 6 at Point 19; tsuag6 at Points 20, 21, 23-29, 31-36, 40; tsuag5 at Point 22. 'to soak' (Si.ch3): Yay see5 ; 6 Dioi che"1 (p.38); Pu-yi (p.221, 0204) Qe6 at Points 1-7, 16, 17, 19; se6 at 6 -, Points 8, 9, 13-15, 37; tse at Points 10, 20, 22-26, 32, 36, 40; tsei6 at Point 11; tQe6 at Points 21, 28, 33, 34; tse1 at Point 27; tQi6 at Point 29; Qie6 at Point 30; tQie6 at Point 35; form missing at Points 12, 18, 31, 38, 39. 'female organ' (a Northern Tai word; Central and South Western Tai use another word, hii5 in Si.): Yay siat5 ; no Po-ai form available; Dioi chat1 chieut1 chiet1 (p.37); Wu-ming siat D2L (p.253), QmatA (p.223); Pu-yi (p.222, 0213) *9i:t8 at Points 1-2, 4-7, 9, 13, 19, 37, 38; 9et at Points 8, Buyr 1959: 222, 0213 nusheng Zhlql 'female genitalia'. (Ed.) 119

PAGE 132

14, 39; 33; 16; tset8 at Point 15; tiat8 at Point 21; Point 26; form missing 10; tsi:t8 at Points 11, 12, 20, 23-25, 9it8 at Point 17; 9iat8 at Points 18, tsiat8 at Points 2;-:--40; t9iat8 at at Points 3, 27-32, 34-36. The initials of our word for 'ox' and the six other words cited show exact corresondences, with no exceptions, in four languages: Yay s, Po-ai s, Dioi eh, and Wu-ming sin Li's Handbook in the 1956-glossary. The initials in the Pu-yi material are more complicated. The transcription apparently aims at extreme phonetic refinement, in the manner traditional in dialect geography. The Editors of Buyr (1959) recognized that these initials are all the same phonologically (or perhaps historically) by grouping the fortyodd examples together (ibid., 217-23). To study in detail the allophones and geographical variants exhibited in the initials of our seven words ('ox', plus 'early', etc.) would also divert us from our purposes. What matters to us is that, as in Yay, Dioi, Po-ai, and Wu-ming, the initial of 'ox' corresponds with that of the other words; although Pu-yi shows some slight variation from point to point in one word or another, it is clear that in general these seven words have the same initial, or possibly at most two phonemic initials. The facts concerning the Pu-yi initials in our seven words may, therefore, be summarized as follows, disregarding occasional missing forms at one point or another: there are, in general, two phonetic types, the first or !, and the Initial 9 is found in all seven words at Points 1-7, 16, 17, 19. s is found in all words except 'female organ', which has 9, Points 37, 38. Initials is found in all words except 'ox' and 'female organ', which have 9, at Points 8, 9, 13-15. Initial s is found in all the words except for 9 in 'to soak' and 'femal; organ' at Point 18. Environment the following vowel) seems to account for some alternations --that is, 9 tends to occur before front vowels --but in many cases the fluctuations seem to be due merely to accidental shifts in the recorder's choice of symbols. The second type, ts, is found in all seven words at Points 10-12, 22-25, 36, 39-,-40. Initial ts is replaced by t9 at Points 21, 28, 29, 33-35 in 'to soak', at Point 26 in 'female organ', and at Points 31-35 in 'ox'. ts is replaced by 9 at Point 30 in 'ox' and 'to soak', and at:Point 33 in 'female organ'. Again, there is a tendency for 9 and t9 to occur (or for the recorder to think he hears it) ts when a front vowel follows. tsh rather than ts in 'ox' at Point 27 is surprising, 120

PAGE 133

as is 'female organ' at Point 21. We cannot be sure without looking at the entire inventory of Pu-yi initial consonant.s, but it seems as if a phonemicist attempting a synchronic analysis of these data would probably conclude that there are two initial consonant phonemes involved, each in a different geographical area, one /'if, with [QJ as an allophone occurring especially before front vowels, and the other /ts/, with various allophones, many of them likewise conditioned by a following front vowel. Each of these contrasts with all other Pu-yi initial consonant phonemes, except for one complication. This arises out of the occurrence now. and then in the data of [tQJ. This sound coincides with an en]irely different Pu-yi initial consonant found in such words as tQa 'rice 3 seedling, shoot' (Buyt 1959: 277, 0625) (cognate with Si.klaa ) and 'middle' (Buyt 1959: 278, 0637) (cognate with Si. ), 7 a phonemic problem fortunately irrelevant to our subject. The evidence, then, is for phonologically regular correspondences in the initial of the word for 'ox' in the Northern Tai languages, supporting a view that this word is old in that branch. We turn now to the vocalic nucleus of the word, which as we have seen varies among the two monophthongs ii and :i,:i, and the diphthongs ia and :i,a, with here and there some other vowel. Ideally, we would like to find other Northern Tai words exhibiting forms rhyming exactly in each dialect with the word for 'ox'. Such evidence would argue conclusively that these words are old in this branch, going back at least to the period of Proto-Northern Tai unity. Unfortunately, a search of the available Northern Tai material has failed to turn up any other words rhyming everywhere with 'ox'. But two other wo:r:ds, those for 'ear' and 'snake', agree rather closely, sufficiently so to justify a study of them. 8 The forms for 1ear1 (Si.huu5 ) are: Yay r:i,a 4 Dioi theueu (i.e. o:i,a) (p.506); Po-ai lii A2; Wu-ming ri (Li 1977: 2 233-4). The Pu-yi forms (p.246, 0392) are ome at Points l, 2, 4, 10, 16, 19, 20, 23; om2 at Points 3, 8, 14, 18, 24-29, 39, 40; oie2 at Points 5-7, 9; rm 1 at Point ll; rm 2 at Points 12, 37, 38; oi2 at Points 13, 15, 17; oei2 at Point 21; Ym2 at Point 22; Ji2 at Points 30-35; form missing at Point 36. The forms for 'snake 1 (Si .1Juu 1 ) are: Yay IJ:i,a 4 ; Dioi gueueu (i.e. 9:i,a) (p.l45); Po-ai !Jii A2; Wu-ming !Jii 2 (Li 1977: 204-5); Pu-yi (p.295, 0759) !Jme at Points l, 2, 4-7, 10, 16, 19, 20, 23; at Points 3, 8, 9, 11, 13-15, 17, 18, 121

PAGE 134

22, 24-28, 30, 31, 37-40; gam2 at Point 12; gei2 at Point 21; 2 at Points 29, 32-36. We must now undertake a comparison of the vocalic nuclei in our three words in all these Northern Tai dialects, a situation which turns out to be less complicated than in the case of the initials. Three different patterns of correspondences emerge from the data. In the first pattern, all dialects have the same vocalic nucleus in all three words. In the second type, 'ear' and 'snake' agree, but 'ox' is different. In the third pattern, 1ox1 and 1ear1 agree, but 'snake' is different. FIRST PATTERN (same vocalic nucleus in all three words): in all three words: Po-ai (if is the correct vowel in the Po-ai word for 'ox'), Wu-ming, and Pu-yi Points 3, 11, 18, 22, 24-28, 38, 40. ei in all three words at Pu-yi Point 21; in all three words at Point 23. SECOND PATTERN (same vocalic nucleus in 'ear' and 'snake', but a different one in 1 ox') : ia in 1 ox' but in 'ear' and 'snake' in Yay, Dioi, and Pu-yi Points l, 2, 4, 10, 16, 19, 20. ii in 'ox' but in 'ear' and 'snake' in Po-ai (if ii is the correct vowel in the Po-ai word for 'ox') and Pu-yi Points 8, 14, 37. ei in 'ox' but in 'ear' and 'snake' at Pu-yi Point 39. THIRD PATTERN (same vocalic nucleus in 'ox' and 'ear' but a different one in 'snake'): ia in 'ox' and 'ear' but in 'snake' at Pu-yi Points 5-7, in 'snake' at Point 9. ii in 'ox' and 'ear' in 'snake' at Pu-yi Points 13, 15, 17, 30, 31, and e in 'snake' at Points 32-35. in 'ox' and 'ear' but (diphthong ending in a high back unrounded semi-vowel) in 'snake' at Pu-yi Point 12, and e at Point 29 (perhaps also Point 36, for which the form for 'ear' is missing?). It is clear from these data that 'snake' is the most conservative of the three words so far as its vocalic nucleus is concerned. This word has the high back unrounded vowel or the diphthong in almost all dialects (all except ei Point 21 and Points 32-35); the vocalic nucleu;-has undergone fronting in many dialects in the word for 'ox', and in a somewhat smaller number of dialects in the word for 'ear'. There is, of course, no doubt that 'ear' and 'snake' are genuine inherited native Tai words; cognates occur in every Tai language or dialect. Hence, our data permit us to formulate 122

PAGE 135

the following hypothesis: these two words, and also the word for 'ox', all had, in the prehistoric Proto-Northern Tai parent language of this branch, the dipthong *:i:a. In some of the Northern dialects all three words still:Preserve the agreement in vocalic nucleus. In a long list of Northern dialects, however, the vocalic nucleus has been fronted to ii or ia in 'ox' because of the palatal initial, while in a somewhat shorter list of Northern dialects, such vowel fronting has taken place in the word for 'ear'. The word for 'ear' is believed to have had, in Proto-Tai, an initial cluster of obstruent plus liquid; change 9 of the post-initial liquid to -y-then caused the vowel fronting. Since Northern Tai *r:i:a 'ear' is cognate with Siamese huu5, and Northern Tai 'snake' corresponds to Siamese one must conclude that if Siamese had the cognate for the Northern Tai word *s:i:a 'ox' (as, of course, it does not have, since *s:i:a is limited to the Northern Tai group), the shape of the Siamese word would be (mirabile dictu!) *chuul, or perhaps *phruul. South-Western Tai gua and its variants are limited to this branch of the Tai family; the various initials occurring in this word cannot be accounted for by the rules of comparative Tai phonology. The word must, therefore, be an innovation of some sort within this branch, and the various scholars who have reconstructed Proto-Tai forms for this word have surely been mistaken in taking it back to the parent language; they have probably been misled by its occurrence in the more familiar Tai languages. It seems probable that gua and its variants were borrowed at an early period from other non-Tai sources. Central Tai maa appears from the complete regularity of correspondences among the Central dialects to have probably been present in this group of Tai languages from very early times. The sporadic occurrences of this form to the west, in Shan and Lao, admits of various possible explanations. One is that the Central form underwent some slight diffusion to the west, perhaps in connection with trade. Another is.that mao was the word for 'ox' at a more remote period of Central/SouthWestern Tai unity, but was later replaced almost completely by gua and its variants in the South-Western languages, except for vestigial marginal survivals of maa in Lao and Shan. Still another possibility is that the resemblance is due to coincidence; either the Central form, or the marginal Lao-Shan form, or all of them, may havP been independent onomatopoetic creations. Northern Tai s:i:a dates back to the period of Northern Tai unity, which would mean that Saek broke a:way from the group at an earlier time, or that Saek may have had the word at one time, but later replaced it by the loanword baa4. Did the speakers of Proto-Tai have a word for the 1ox1 as they certainly had for the buffalo? In view of our findings, it is difficult to imagine that either moa or s:i:a (gua being 123

PAGE 136

excluded as certainly a later innovation) could have been the Proto-Tai word for 'ox' and then vanished without a trace except in a single branch. It seems much more plausible to assume that the speakers of the Proto-Tai parent language had no knowledge of this animal, and no name for it. Then later, after the separation into the major Tai groups, the animal came to be known, either through domestication or through introduction from outside, and a name was acquired by each of the major Tai groups. We are unable to say much about -dates, but if Proto-Tai is put at about two thousand years ago, as many scholars believe, then this introduction of the various names would have occurred somewhat later, perhaps some fifteen hundred years ago. The close correlation in the geographical distribution of our three words for the 1ox1 with Li Fang-Kuei's three branches of Tai perhaps lends some support to his view of a tripartite genetic branching. On the other hand, the whole matter of the names for the lox1 may be -irlVolved not so,much with genetic branches aswith ethno-linguistic areas. The scope of this paper has been intentionally restricted to seeking out and presenting the Tai linguistic evidence on the subject. Two other areas have been deliberately excluded. For one thing, it would be tempting to rummage around in dictionaries and wordlists of non-Tai languages of South-East Asia and the Far East for possible sources of our three words as loans from foreign languages into Tai. But it is felt wiser to leave this to specialists in those other languages, especially as considerable background in their historical phonology would be necessary to identify with certainty possible outside sources, at some time perhaps a millennium and a half ago, for and its variants *wua and *hua in the South-West, in the Central area, and in the North. For another, we understand that students of South East Asian archaeology and prehistory have some theories and knowledge as to when the ox was domesticated in, or introduced into, South East Asia. It is hoped that our evidence that each branch of the Tai family seems to have acquired a name for the 'ox' as an innovation some time after the break-up of the original parent language may be of interest and use to scholars in those disciplines. NOTES 1. This paper was originally presented at the XIVth International Conference on Sino-Tibetan Languages and Linguistics held at the University of Florida in Gainesville, Florida, U.S.A., on the 30th October to lst November, 1981. 2. Saek is a language belonging to the Northern branch of Tai 124

PAGE 137

(see the next, third type), now displaced far to the south and surviving in a few villages in the province of Nakhon Phanom in north-eastern Thailand and in a few villages on the Lao side of the Mekhong River near Tha Khek. 3. cf. 1959: 217, 0179: vau2 at Points 1-7, 16-17, 19; tyau2 at Points 8-12, 20, 22-4, 39-40; sau2 at Points 13-15, 18; tsau2 at Points 21, 25-8, 30-5; Point 29; tsm2 at Point 36. ( Ed. ) 4. Yay forms are from Gedney fieldnotes; Po-ai from Li 1977; Dioi from Esquirol 1908. For Wu-ming, two forms are given (more transcriptional variants), the first from Li 1977, the second from Li 1969; Pu-yi forms are from 1959. In these word entries they are coded, after Pu-yi, as e.g. (p.219, 0180) instead of (Buyt 1959:219, 0180). 5. This is a typographical error; in the Handbook (Li 1917), long is written elsewhere with a single letter short the letter and a breve. 6. This is not treated in Li (1956 & 1917), so there are no Po-ai or Wu-ming forms available. 7. 8. 9. Abundant examples of this other initial are given in Buyt It is represented by in Yay (caa3, caaql); ki in Dioi (kia, kiang); i in Po-ai (eaa Cl, caaiJ Al); and kl in Wu-ming ( kla, klaiJ) Words with vocalic nuclei similar to those of 'ox', 'ear', and 'snake', but with a final consonant, appear to have undergone completely different changes, and are not helpful here. Some scholar is, indeed, going to have great fun working out the vicissitudes of the word 'ear' in these and other Tai dialects! REFERENCES Barua, B. & 1964. Ahom lexicons (based on original Tai manuscripts). Gauhati: Dept. Hist. Antiquarian Stud. in Assam. N.N. Deodhai Phukan Borua, R.S. Golap Chandra 1920. Ahom-Assamese-English dictionary. Calcutta: Baptist Miss. Press. 125

PAGE 138

Buy'i Cushing, J.N. fliE:lu Ch{nh Nhlm & Jean Donaldson Egerod, Esquirol, Jos. & Gust. Williatte Harris, Jimmy G. Hashimoto, Mantaro J. Haudricourt, A.G. Kerr, A.D. 1959. BUy'i-yu diaocha baogao (ed.) Zhongglio kexueyuan, Shaoshu minzu yliyan yanj1usuo. Beijing: Kexue. 1914. A Shan and English dictionary. Rangoon: Amer. Baptist Miss. Press, 2nd ed. 1970. Pap san khhfim pak Tay-Keo-Eng: Ngfi, vllng Tai-Vietnamese ?/ "'"' English vocafulary (Tu-sach ngon-ngu 4). Saigon: '/ B? Giao-d':-c 1961. Studies in Thai dialectology. Acta Orient. 26, 43-91. 1908. Essai de dictionnaire dioi3 reproduisant la Zangue parlee par ies tribus Thai de la haute Riviere de suivi d'un vocabulaire Hong Kong: Impr. Soc. Miss. etr. 1975. A comparative word list of three Tai NUa dialects. In Studies in Tai linguistics in honor of William J. Gedney (eds.) J.G. Harris & J.R. Chamberlain. Bangkok: Central Inst. Eng. lang., Office State Univ., 202-30. 1976. Notes on Khamti Shan. In Tai linguistics in honor of Fang-Kuei Li (eds.) T.W. Gething (et al.). Bangkok: Chulalongkorn Univ. Press, 113-41. 1980. Be-go go-i shu (The Be language a classified lexicon of its Limkow dialect) (Ajia Afurika kiso go-i shu 11). Tokyo: Ajia Afurika Gengo Bunka Kenkyu-jo (Inst. Study langs. cults. Asia & Africa). 1948. Les phonemes et le vocabulaire du thai commun. J. asiat. 236, 197-238. 1965. Le vocabulaire Be de F.M. Savina (Publ. Ec. fr. Extr.-Orient 57). Paris: Ec. fr. Extr.-Orient. 1972. Lao-English dictionary. Washington: Consortium Press, 2 vols. 126

PAGE 139

Li Fang-Kuei Pallegroix, J.B. Purnell, H.C. Reinhorn, M. Robert, R. Savina, F.M. Weidert, A. 1940. Longzhou tUyu (The Tai dialect of Longzhou: texts, translations, and glossary) (Academia Sinica, Inst. hist. philol. Monogr. ser. A 16). Shanghai: Corn. Press. 1956. tuyu (The Tai dialect of Wu-ming). (Academia Sinica, Inst. hist. philol. Monog. ser. A 19). Taipei: Ac. Sinica. 1977. A handbook of comparative Tai. (Oceanic ling. spec. publ. 15). Honolulu: Univ. Press Hawaii. 1854. Dictionnarium linguae thai sive siamensis; interpretatione latina, gallica et anglica, illustratum. Paris. (republ. 1972, Farnborough: Gregg Internat. Publ.). 1963. A short northern Thai-English dictionary (Tai Yuan). Chiengmai: s.n., mimeo. 1970. Dictionnaire laotien-franaais. Paris: CNRS, 2 vols. 1941. Notes sur les Tay Deng de Lang Chanh (Thanh-hoa --Annam) (Inst. indochin. pour l'Etude de 11Homme, Mem. 1). Hanoi: Impr. Extr.-Orient. 1910. Dictionnaire tay-annamitefranqais precede d'un precis de grammaire tay et suivi d'un voaabulaire franaaistay. Hanoi: Impr. Extr.-Orient. 1924. Dictionnaire etymologique franqais nung-chinois. Hong Kong: Impr. Soc. Miss. etr. 1977. Tai-Khamti phonology and vocabulary (Beitrage zur Slidasien-Forschung B 27). Wiesbaden: Franz Steiner Verlag. 127

PAGE 141

FIRST AND LAST IN THAI, OR THE ORDER OF OPPOSITIONS tMary R. Haas In 1943 the well-known Chinese writer, Lin Yutang, published a book entitled Between tears and laughter. To me this title was vaguely indeed, it seemed quite awkward. Then I realized that the opposition of 'tears' and 'laughter' is normally rendered 'between laughter and tears' in English. In other words, the order is the reverse of that used in Lin's title. Perhaps in Chinese the order is the reverse of English--this I do not know.* In Thai, however, binomials (Malkiel 1968) within this semantic range are regularly the reverse of those in English, and this paper is concerned with binomials containing opposites. Thai is a language in which the head always precedes the attribute in modifying constructions (e.g. man-good = good man) whereas English is a language in which the opposite order usually pertains. Both languages also make considerable use of various kinds of what are often described as coordinate constructions, and in many of these where the languages appear to have equivalent expressions the word order is the same in both languages. On the other hand, in the case of coordinate constructions involving oppositions, it frequently happens that Thai uses the reverse of the English order, something that has even been observed in expressions which may be calques from English, such as 'black and white (photograph)', rendered in Thai as 'white (and) black (photograph)'. In studying examples of the order of coordinates which are opposites, I have included coordinate phrases as well as compounds. In the following examples (Haas 1964) Thai word order is the reverse of the English: 1. thugsug 'sorrow-happiness', Eng. 'joy and sorrow' 2. pr1awwaan 'sour-sweet', Eng. 'sweet and sour' 3. 'silver-gold', Eng. 'gold and silver' 4. 'fly-go fly-come' or 'fly on, fly back' Eng. 'fly back and forth' It is xiao jian 'between tears and laughter', which is also better rhythmically in putonghUa (Mandarin). (Ed.) 129

PAGE 142

5. thli 'some bad, some good'' Eng. 'some are good, some are bad' 6. 'now near, now far', Cf. Eng. 'far and near' Theoretically speaking, it would seem that coordinate items could be listed in any order in both English and Thai, but most, all, languages have a preferred order which in any given expression will remain fixed. Nevertheless, the question regarding the examples above is: why does Thai have the reverse of English (or English the reverse of Thai)? It appears that while English prefers the order 'favourable:unfavourable' (good and bad, light and dark, etc.), Thai prefers 'unfavourable: favourable.' Lin Yutang's use of 'tears and laughter' rather than 'laughter and tears' thus fits the Thai pattern nicely, but goes against the normal English one. There are, however, other examples of coordinate oppositions in which somewhat more flexibility exists. Even in examples which are very close semantically, one may have one order and another its reverse, both in Thai and in English. 7. khuu 'pair groom-bride', Eng. 'bride and groom' S. (elegant term) 'husband-wife', Eng. 'husband and wife' 9. 'both women and men', Eng. 'men and women' 10. father' 'father and mother, parents', (but also 1dad and mom')* Eng. 'mother and So examples involving male and female oppositions in both languages show a little more flexibility between expressions, even though a given expression tends, as already noted, to remain fixed. An example showing such variation within one expression is the following variant for (S): Sa. 'wife-husband', Eng. 'husband and wife' This latter has the advantage of showing inner rhyme (-jaa saa-), a device favoured by some speakers. The examples given here are chosen to show some variation, though it turns out that Thai and English expressions involving male and female opposition tend to place the male term first. But one does find a few expressions placing the female term first, as in English (7) and (10) and Thai (Sa) and (9). Most commonly 'mum and dad' in England. (Ed.) 130

PAGE 143

In undertaking this study I was hoping to find some explanation for those cases where Thai and English have reverse order in coordinates, but beyond the observation already made that English prefers 'favourable:unfavourable' and Thai prefers the reverse, it would appear that the treatment of oppositions is not subject to generalization. Instead, each expression tends to have its own explanation in both languages. One disruptive influence that occurs in Thai is the desire of many speakers to use inner rhyme, as in (Sa), and this may entail the rearrangement of the items in order to bring about the adjacency of the rhyming segments, a device which also accounts for a few cases where two orders are possible in Thai; thus (S) vs. (Sa). Further studies of Thai binomials might, indeed, profit from an areal approach, i.e. a comparative study of the problem in other languages of the area (see Haas l97S). Haas Mary R. Lin Yutang Malkiel, Yakov REFERENCES 1964. Thai-English student's dictionary. Stanford: Stanford Univ. Press. 197S. Areal linguistic characteristics of East Asia. In Chinese language use (ed.) Beverly Hong (Contemp. China papers 13).Canberra: Contemp. China Centre, Austral. Nat. Univ., 1-S. 1945. Between tears and laughter. Garden City, New York: Blue Ribbon Books. 196S. Studies in irreversible binomials. In Essays on linguistic themes. Oxford: Blackwell, 311-55. 131

PAGE 145

LA TONOLOGIE DU LI DE HAINAN A.G. Haudricourt Les enquetes de F. Savina (1931) et de Wang (1951) notaient trois tons dans les syllabes sonores des dialectes* qu'ils decrivaient. C'est seulement en 1963 que Ouyang et Zheng indiquent que si la majorite des dialectes ont trois tons, il en existe a six tons, et ce n1est qu1en 1980 que ces auteurs publient une etude avec un vocabulaire compare d'un dialecte a trois tons (le parler de dialecte Ha3) et d'un autre a six tons (parler de Tongza, dialecte Gei4) (see Table 1, below). La comparaison nous permett.ra de montrer que le systeme a trois tons est le plus ancien et que le systeme a six tons resulte du dedoublement des tons comme dans les langues du continent (chinois, miao-yao, tai-yai et dong-sui).** Ce dedoublement provient toujours de la fusion d'initiales sonores et d1initiales sourdes, mais le dialecte qui n'a pas mute n'a pas forcement conserve les anciennes initiales, et quatre cas se presentent: l. Le dialecte de a conserve les sonores, assourdies dans celui de Tongza (comme thai-commun y devenant ). Exemples: Baad:Zng Tongza van 5 3 fan ll terre vau53 f ll au difficile veu53 feu ll nombril vo:n 53 fo:n ll abattre ve:IJ55 fe:!J 131 breche va55 fma 131 epaule ll fe:!J 14 vetement ve:IJ vi:p55 fi:p 13 vivant, cru Savina described the regional dialect of the Southern day, Wang Ll that of Baisha, hjm:nl dialect (MSC bendl). (Ed.) ** D
PAGE 146

... Baod1-ng Tongza vat 55 fat13 arc VUI:k55 fm: ?13 os va:t55 fa:t13 pauvre contrastant avec fei53 fei 33 feu; marcher fou53 fou 33 pou fun 5 3 fun 33 pluie, pleuvoir fan53 fan 33 graine, dent fo :m53 fo:m 33 tenir dans la bouche fi: IJ53 fiaiJ 33 porter sur le dos fi :u55 fi:u51 crete fo:i55 f 51 o:l oncle uterin cadet fo:n55 fo:n 51 semer fu:n55 fan 5 1 cribler fall fa 55 ciel f ll am fam 55 neuf; cadette du pere fan ll fan 55 poitrine fu:t55 fu:t13 dix; ecume fe :k55 fe?l3 vomir fm:k55 fm: ?13 tisser fat 5 5 fat13 ... inserer dans Nous pouvons done, sur ces exemples opposer, a Tongza, une serie tonale basse: 11, 131, 13 a une serie tonale haute: 33, 51, 55. L'ordre dans lequel nous avons range les tons de Baodtng est celui de l'ancien chinois: le premier ton le plus frequent est egal, le second est montant (il cofncide avec celui des mots a occlusive finale, done il avait a l1origine un glottal stop final) et le troisieme est descendant; done meme origine probable que le chinois et le vietnamien. Nous avons les meme correspondances tonales avec les sifflantes si nous remarquons que la sifflante sourde s se realise en laterale sourde comme dans nombreux dialect;s du continent. 134

PAGE 147

Exemples: V .. Baod-z,ng zai53 zeilJ53 Zi!J53 za55 zaur55 za:u55 zi :m55 zi:!J55 zu:i11 zok55 contrastant avec ia53 iai53 ian53 io:i53 ieur 53 iur:n53 ie,p.53 ia:n55 iun55 ia:u11 "ll rau "ll reJ;J ii:n11 im:!Jll ia:c55 Hp 55 io:k55 iu:t55 iu:k55 Tongza ".11 "ll ren ll ie: lJ H!Jll ial3l "131 raur ia:ul3l "-" 131 rl:m ,_. 131 rlalJ ".14 :>:U:l iok13 ia33 hi33 ian33 ia:i 33 "33 :>:eur iur:n33 ien33 ia:n51 iun5l ia:u55 iau55 ieJ;J 55 ii :n55 im:!J55 ia:t13 ,_. 13 rlp io :?13 iu:t13 iu: ?13 135 oreille givre jaune sangsue serpent changer devetir lecher doigt voir voler, derober poisson Li (peuple, langue) choisir salive; plusieurs gendre s'eveiller bon bras se chaud, tiede coeur deux chapeau de paille langue clair (temps) sang foudroyer profond; sourd entrer cerveau

PAGE 148

2. La confusion des initiales s1est faite au profit des sonores en Tongza (comme thai-commun hw devenant Exemples: Baad'?.ng Tongza za53 11 vieux; remede za zm:m53 11 oeuf zm:m 11 14 grenier za:u za:u 11 14 mari de ainee zau zo soeur ll 14 rapide zmn zmn s1opposant a hja53 33 Imperata, herbe a za paillotte hjau53 33 haricots zau hjo:n53 33 chanter zo:n hja:m55 za:m 51 franchir hjan55 51 zan ver 33 zau reins hja55 za5l jo:m55 zo:m 51 avaler Il s 1 agi t d 1 anciennes semi voyelles : j_; _et nous avons un cas analogue avec les anciens Y, W (note _g_, comme en vietnamien puisque b,d sont des Exemples: .... .. Baodmg Tongza gwa53 11 planter gwa semer, ga:m53 11 demander ga:m ga53 11 (inclusif) gau no us gwa:i53 .11 laver gwa:1 gwa:n53 11 cent gwa:n gei53 .11 de palmi er ge1 esp. geu53 11 mince geu gia53 gia 11 tousser go:i53 .11 fer ga:1 gou53 11 huit gou gom53 11 son de riz go m 53 11 cadets gu:!J gU:!J 136

PAGE 149

.., Baodmg Tongza ga:i55 .131 gourde ga:l gwai55 .131 gwal manquer gou55 131 courir gou 11 14 tete gwou go 11 14 go m gam viande gwiu 11 14 gwiu pus gi:!J 11 gia!J 14 mur .11 .14 huile, gwel gu:l gras gat 55 gatl3 couper (legumes) gip 55 gip 13 chasser ge:k55 ge:? 13 taro gu:k55 gm: ?13 chauve-souris contrastant avec 53 33 bo1 wa:u gwa:u hwa:i53 .33 depasser gwa:l hwan 53 33 ver de terre gwan 53 33 fumee hwo:n go:n hwom53 gom33 rond hwat55 gwat13 dur 3. Mais nous avons un cas ou l1ancienne sonore, conservee en Tongza, est devenue sourde en Baad1..ng Tongza hwan 53 11 jour van hwo:n53 11 offenser -vo:n hwi:!J53 via!J 11 gauche hwi :n53 vi:n ll mauvais augure hwan 55 131 typhon van hwi:!J 55 via!J 131 boucle d1oreille hwe:k55 ve:? 13 banane C1est que ces dialectes n1ont plus de semivoyelles sonores mais seulement: !hi_, hw, ?w. 137

PAGE 150

4. Enfin dans le cas suivant ou nous n'avons plus que des sourdes actuelles, nous ajouterons les mots recueillis par Savina (1931) dans un dialecte du sud actuellement disparu qui pourra nous eclairer. Exerrrp'les: Baod"ing 53 pou 53 pa 53 pa pm:n53 peu53 po:i53 pui53 pou55 .11 pal 11 pom 11 pua:m 11 pan put 55 po:t55 tiu53 tem53 tom53 ta55 teq55 11 ta:u tseu53 kan53 kem 53 Tongza 11 pau 11 pa 11 pa 11 pm:n 11 pa:u .11 pa:l .11 pUl 131 pau .14 pl 14 pam 14 pua:m 14 pan t13 pu 13 po:t tiu11 t 11 em tom11 ta131 t 131 en 14 ta:u 11 tseu kan11 k 11 em pore chien cinq venir revenir stupide ivre annee mere bouche bar be humide fourmi puce rat des sus six riziere droite long tirer*** argent sesame Dia'lecte de Savina mB:u* / ma / ma mli(:Jn m &ut I. mul m6'In "'' men nlU "' nom 7 na I nln nao ngb:en ngW Savina1s transcriptions are often irregular (e.g. 1931: 14Tb 'pore'' 108a m!u; 109a mau 1annee1 148a mu, 18lb so entries here are based on his 1Lexique and 'Differences dialectales ... ', 107-10, not other parts of the article. (Ed. ) ** See, esp. Savina 1931: 104, 148b. (Ed.) *** Savina 1931: 170b tho 'tirer', 154b nhSu 'fleche'. (Ed.) 138

PAGE 151

Baod":ing Tongza Dialecte de Savina kau55 kaul31 couche kall kal4 cheval nga kun 11 kun 14 bois de ngun, chau:ffage ... ngun ka:i11 ka:i14 enrouler ngay ( 1rouler') kut55 kut13 aiguille ngut Aucun dialecte actuel ne presente les attendu mais les nasales que Savina nota vers 1925 devaient etre des nasales fortes longues qui ont donne des seminasales: mb, nd .. engendrant les sonores necessaires a la mutation tonale. -Certain de ces mots ont en thai commun une initiale complexe: hm-(v. 'pore, chien'), hn-(v. 'rat'), 1riziere1 est rna en khmu, et 'cinq'' 'six' ontdu etre l.ma, ?nom (austronesien: lima, ?onom); parmi les etymologies indonesiennes on peut lnga (indonesien: lenga) (Dempwolff l938:95a). Un autre paradoxe est la rarete des correspondances pour les occlusives de la serie tonale haute. Cela ressortait des comparaisons de R. Shafer (1957:389-90): Exemples: iaoding Tongza Savina ku:n 53 ku:n 33 chemin """ cuon ko:i53 ka:i 33 abeille (cti 1miel 1 ll9a) te:1J 53 te:1J 33 peigne ttfng pok55 pok55 ventre (b6ng, ll4b) En realite, ces anciennes occlusives sourdes sont aussi issues de groupes de consonnes, comme le suggere le nom du 'chemin' (ku:n), thai-commun: hron; Dong-Sui: khwen. A la difference des langues du continent, la mutation des initiales qui a produit le dedoublement des tons n'a concerne que les anciennes occlusives longues (issues de groupes) et les spirantes. Les anciennes occlusives simples sont representees par les occlusives aspirees (comme l1a montre R. Shafer, 1957:391-95), qui n'ont pas participees a la mutation, non plus que les occlusives nasales (puisqu'il n'y avait pas de nasales sourdes), formant ainsi une serie moyenne dont les realisations tonales se confondent avec la serie haute. 139

PAGE 152

Tongza 2,.Jl31 3\5 4Jll 6A14 7\5 8J13 9\3 (gei4 ) c Baod"ing l'f53 2155 3Jll 1\3 2155 3111 7155(8Jll) 7155 7\5 (ha3 ) BfdslUi lJll 2 J31 3 11n ( 4 'i51) ( 5\5) 8Jn ( 7\5 l 8Jn 7\5 1 (hjUl:n ) f-' _,.. 0 x?fang 1\3 2\5 3124 1\3 2\5 7\5 7\5 (mo:i1 fau1 ) Table 1: Ma,jor differences between tones o_f the Li dialects: 4 h 3 1 d .l f 1 gel _il;, hJ UI: m an .!!!Q..!..1. (Adapted from Ouyang 1980: 88. Ed. )

PAGE 153

Dempwolff, 0. 5uyang Jueya & Zheng Y1q1ng Savina, F.M. Shafer, R. Wang Ll & Qian xlin BIBLIOGRAPHIE CITEE 1938. Austronesisches Worterverzeichnis (Beihefte z. Zeitschr. f. Eingeb.-Spr., 19). Berlin: Dietrich Reiner (Andrews & Steiner). 1963. L1-yu gaikuang. zhonggUo yuuen 126 (1963/5), 432-41. ft-yu jfanzhr (Zhongguo shaoshu m1nzu yuyan jianzh1 congshu). Beij1ng: M1nzu chubanshe. 1931. Lexique Bull. Ec. fr. Extr.-Orient 31, 103-99. 1957. Quelques equations phonetiques pour les langues li d1Hainan. Rocznik orient. 21, 385-408. ...... ...... ..,. .... Hainan-dao Baisha L1-yu chutan. Lmgnan xuebao 11 (2) (1951/6), 253-300. 141

PAGE 155

PROTO-TAI *kh-AMD *x-Li Fang-Kuei In A handbook of comparative Tai3 published in 1911, I reconstructed two Proto-Tai (PT) consonants *kh-and These two consonants merge into kh-in most of the South-Western and Central Tai languages, such as Siamese and Lungchow, but they are distinguished in the Northern group of Tai languages, such as Po-ai, where PT *kh-becomes (unaspirated) and PT *xbecomes h-. A few examples may be given as follows: *kh-1 Siamese Lunge how Po-ai PT 'to kill' khaa Cl khaa Cl kaa Cl 'horn' khau Al kau Al PT *x-'to enter' khau Cl khau Cl hau Cl 'to step across' khaam Cl khaam Cl haam Cl The reconstructions are obvious enough, but there is some disturbing evidence from the South-Western group of languages which makes me reconsider the whole problem, particularly the problem of aspiration in Proto-Tai. In White Tai and some of the Lu dialects, and also, I believe, in Old Siamese, i.e. of the Sukhothai period or about that time, there is a split of both PT *kh-and *x-. For example: Siamese White Tai Po-ai A. PT *kh-'to kill' khaa Cl xa Cl kaa Cl 'arm' Al Al ceen Al PT *x-'white' khaau Al xau Al haau Al 'to enter' khau Cl xau Cl hau Cl 143

PAGE 156

Siamese White Tai Po-ai B. PT *kh-'ditch, pit' khum Al khum Al kum A2 'to ride on horseback 1 khii Bl khi, khii Bl kiii B2 PT *x1 excrement 1 khii Cl khi Cl hai C2 'rice' khau Cl khau c1 hau C2 'bitter' khom Al khum Al ham A2 Similarly my Lti material, as yet unpublished, from Cheng Tung in Yunnan province, shows the same split as in White Tai. It is evident that khrepresent both PT *kh-and White Tai andLu. This is not a random split, as it has its reflexes in the Northern Tai languages. x-appears in words which show tone series l (indicating original voiceless initials) in all dialects, but kh-appears in words which show tone series 2 (indicating original voiced initials) in the Northern Tai languages. Now, White Tai and Lu are South-Western languages, far removed from the Northern group; the split which agrees with the tone alternation in the Northern group must indicate some.early phonological condition in the Proto-Tai period. It is for this reason that I wish to amend my previous reconstruction. Siamese orthography formerly provides two symbols for the modern consonant kh-; one corresponds to the Indic letter kh-( 'IJ ) and the other is a modified form of the Indic kh ). There must have been a difference in between these two letters, although they merge in modern Siamese. At the present time the modified form of Indic kh-is no longer used. From the Sukhothai inscriptions, some words which show the modified form of kh---such as khaa Cl 'to kill': khEEn 'arm'; khaau Al 'white'; khin Cl 'to ascend'; khaam Cl 'to step across', etc. --all agree with x-in White Tai and Lu. This fact had already been noticed by-Burnay and Coedes (1928:125). It may imply two things. First, we may simply assume in modern Siamese (Modern Standard Thai, i.e. MST), that the two distinct sounds of the Sukhothai period have merged, and, therefore, also assume that modern Siamese is a direct descendant of Sukhothai speech; or rather we may conclude that modern Siamese is not a direct descendant of the Sukhothai dialect, but a closely related dialect. It seems that what has been reconstructed as *kh-and must be given two different forms for each of the PT consonants, i.e. two forms for PT *kh-and two forms for The reconstructed forms must be able to explain the split into 144

PAGE 157

kh-and in White Tai, Lu, and Old Siamese, and also the alternation of tone, from series l to series 2 in the Northern dialects. It has been known that such alternation of tones occurs chiefly in words with aspirated initials, hence it is quite possible that the aspiration may be the cause of such tone alternations.2 I make the assumption that there were two kinds of aspiration in Proto-Tai, namely and an assumption based on the aspiration as reflex in the South-Western and Central dialects. is perhaps a voiced aspiration or murmur. Thus, we have in Proto-Tai both *kh-and *kfi-; the latter type of consonant being quite common among the Wu dialects in China. We may now amend our reconstruction of *kh-and *x-and their development in the following way: Siamese White Tai Lunge how Po-ai PT *kh-kh-X-kh-k-PT *kfi-kh-kh-kh-PNT *g-> k-3 PT *x-kh-X-kh-h-PT *fi-kh-kh-kh-PNT *Y> h-3 It seems reasonable to assume that all the aspirated stops may show two forms; one shows no tone alternation, and the other an alternation of tone series l to series 2. For example, what have been reconstructed as *tfi-and also be given two forms PT *th PT *tfi PT *12!!_ each. Siamese thth-phph-White th-th-ph-ph-Tai Lunge how Po-ai th-t-th-t-(< PNT *d-) ph-p-ph-p-(< PNT *b-) This assumption seems to explain a fairly large number of tone alternations 4 of this type, but by no means explains all alternations of tones.5 NOTES l. In the following examples, letters A or C after the cited forms indicate the tone class, and the numeral l indicates that the initial consonant was originally voiceless. (Li 1977:25).

PAGE 158

2. See appropriate sections on PT aspirated stops and on *xin Li (1977:63-5, 102-4, 192-8, 207-14). 3. PNT (Proto-Northern-Tai) *and *y are assumed because they give the same in the-modern Northern Tai dialects as PT *and *Y 4. There are other types tone alternation, such as B2 -C2, Cl Bl, Cl B2, etc. It is impossible to take them into consideration at this juncture. 5. Professor William Gedney proposed a new series voiced initials in Proto-Tai in a paper at the Sino-Tibetan Conference in Paris in 1979. The paper is not published* but, while his idea agrees in general with mine, the reconstructions are Burnay, J. & G. Coedes Li Fang-Kuei REFERENCES 1928. rrJ et Q'1 et leur onglnes. J. Siam Soa. 21 (2), 119-26. 1977. A handbook of comparative Tai (Oceanic Linguistics Spec. Pubs. 15). Honolulu: Univ. Press Hawaii. Gedney1s paper, 'Evidence another series of voiced initials in Proto-Tai1 has now been published in his Selected papers on comparative Tai studies ed. R.J. Bickner et al. (Michigan papers on South and Southeast Asia 29, 1989, 229-69). (Ed.) 146

PAGE 159

UNCLES AND AUNTS: BURMESE KINSHIP AND GENDER1 David Bradley Introduction: The position of Burmese in Sino-Tibetan Burmese is the Sino-Tibetan (ST) language with the second largest number of speakers after Chinese. It was the fourth to develop an orthography --preceded by Chinese, Tibetan, and extinct XIx1a (Tangut); surviving Burmese inscriptions date from 1112 AD onwards. Its historical linguistic position within Sino-Tibetan is represented in the following language tree: Sino-Tibetan Tibeto-Burman (TB) Sinitic (Chinese) Bodic (Tibetan,etc.) BURMESE (dialects) Baric (Garo,etc.) Karenic Burmic Ugong Naxi-Burmese-Lolo Naxi Burmese-Lolo xrx1a Tosu Burmish Loloish At si Maru Lashi Southern Central (Akha,etc.) (Lahu, Lisu, etc.) Fig.l: Sino-Tihetan Uxnguage tree 147

PAGE 160

This classification is based on patterns of sound correspondence found in non-borrowed vocabulary, and on proportions of shared basic vocabulary.2 Burmese has a number of regional dialects. The 'standard' language, or central dialect, has subdialects: that of Upper Burma centred on Mandalay, and that of Lower Burma centred on Rangoon. The Arakanese dialect, spoken along the north-western coast and into Bangladesh, has the second largest number of speakers, and is archaic in a number of ways; there are also several other dialects. Quite closely related to Burmese are Atsi (Tsaiwa), Maru (Lawngwaw) and Lashi, spoken in north-eastern Burma by smaller groups which are part of the 'Kachin' culture complex. These languages show extensive influence from Jinghpaw ('Kachin'), a Baric Tibeto-Burman language according to Burling (1971), and of particular interest within these languages is the wide range of terms used for uncles and aunts. Terms for Uncles and Aunts A. Burmese The system of kinship terms for parentsf siblings is an area of substantial dialect difference in Burmese, and of extensive changes observable by comparing older and more recent sources on these dialects. Inscriptional data, mostly summarized in Luce (1981), with some data in Ba Shin (1962) and Than Tun (1958), provide early evidence for some forms though the exact referents of the terms are often hard to determine. Judson (1953) provides early nineteenth century data, and Tun Nyein (1906) gives normative early twentieth-century forms. Two anthropological studies have investigated modern Rangoon usage: Brant and Mi Mi Khaing (1951), and Burling (1965). Most recently, Spiro (1977) discussed the kinship system in depth, with 1960s usage for a village near Mandalay in Upper Burma, reporting 'old' Upper Burma forms, and current Rangoon forms. The three last sources disagree extensively and, indeed, my Burmese informants have always had trouble with these terms, which are in a state of flux: Tun Nyein (1906) actually contains a basic error, calling the father's sister terms 'maternal' and the mother's sister terms 'paternal'. Table 1 below sUmmarizes the data: 148

PAGE 161

Period: Early Late c.l850 c.l900 'old' (Spiro) c.l960 (Spiro) Village Source: Inscriptions Judson Tun Nyein Upper Burma Upper Burma Mandalay Rangoon .. .. Uyl X X MeB MB MeB1 .... """ MeB1 uylJl ....... MeB2 weylJl ...... MeB MeB2 PeB2 UJl u MB MB 1-' ruejl. PeB 3 +="" \0 Uml:n MyB rue MyB MyB Pyl3 weyl.le MyB bajl. X FeB FeB FeB FeB FeB PeB 1 baba FeB (address) PeB4 (intimate) babajl. PeB 5 (intimate)

PAGE 162

Period: Early Late c.l850 c.l900 'old' (Spiro) c.l960 (Spiro) Village Source: Inscriptions Judson Tun Nyein Upper Burma Upper Burma Mandalay Rangoon ba thwe X FyB FyB FyB FyB FyB .1' ... X my1 ... X X FS MS FS FyS FS ey1 ....... FeS eylJl FeS ey1le FeS I-' nUjl X MeS FS MeSl \.Jl 0 j1do FS MeS2 MeS MeSl PeSl '-o'\. FS FeS (address) PeS JlJl (intimate) .... PeS 2
PAGE 163

I-' \J1 I-' Period: Early Late c.l850 c.l900 'old' (Spiro c.l960 (Spiro) Village Source; Inscriptions Judson Tun Nyein Uppe r Burma Upper Burma Mandalay Rangoon (a)dole FS MyS 2 MyS P y S 2 dodo F S Table l: Burmese uncle and aunt t erm s* Forms which occur in inscriptions are indicated by X. Abbreviations are; P =Parent; F = Father; M Mother; B =Brother; S =Sister; Sb =Sibling; W =Wife; H =Husband; e =elder; y = younger.

PAGE 164

Apart from the basic form /y1/, all the above are historically analysable compounds. /u/ is 'head'; /j1/ is 'big' or 'elder'; /le/ is 'little' or 'younger'; /m1n/ is 'king', /ba/ is a now archaic form for 'father', and /mf/ similarly for 'mother'; /thwe/ is another word for 'younger'; /do/ is probably derived from the royal honorific suffix /to/; and /e/ is a formative prefix used with stative verbs and bound suffixes to form nouns. In addition to their use as kin terms /u/ and /do/ are now used as honorific prefixes to male and to female names respectively. The radical restructuring of the Rangoon kinship system, with bilateral extensions of all surviving terms, results in extensive confusion among Burmans about the referents of these terms: some of this confusion is reflected in the data of Brant and Mi Mi Khaing (1951). An additional factor is the possibility of kin numeratives (with /j1/ for the first, /la?/ for the second last, and /le/ for the last; or in inscriptions, /j1/ and /thwe/, /o/ 'old', and /thwe/, or /o/, /thwe/ and 'little', in relative order of birth). These numeratives are no longer used in 'standard' Burmese, though Arakanese and other dialects still have them. B. Arakanese Forms from the Arakanese dialect are found in Bernot (1967), for the Marma dialect as spoken by a group who fled to what is now Bangladesh at the time of the Burmese conquest in the 1780s, and for 1Magh1 the Bengali name for the Arakanese, in Levi-Strauss (1952), who does not indicate tones. The least Burmanized Arakanese is probably Marma; the Arakanese spoken along the coast in Bangladesh and northern Arakan shows several innovative terms, while in southern Arakan, Burmanized forms are used, as seen in Table 2 below. The Arakanese /ri/ is regularly cognate with Burmese /y1/, as are /gri/ and /j1/, /thwi/ and /thwe/. The innovative mother's brother term is used for address only in Marma, which also retains conservative forms for the other aunts and uncles. There is a Burmese couplet for husband and wife, /khin bun/, whose first syllable 'husband' may be related to the Arakanese term; this semantic shift may be connected with the Arakanese preference for mother's brother's daughter-father's sister's son marriage. Arakanese /bye/ could be derived from an alternative form of 'father'; /ywe/ is more problematic, though there is a Burmese bound couplet /ywe/ 'to associate familiarly and affectionately' (Judson 1953:864), but if this is the source, the Burmese spelling is etymologically wrong, although it is not unparalleled for the Burmese spelling, when representing a dialect which has merged /r/ into /y/, to 1respell1 words incorrectly. Arakanese further has traces of suffixes for birth-order differences which provide for up to four ordered possibilities, as shown for Rangoon Burmese by Brant and Mi Mi Khaing (1951). 152

PAGE 165

Dialect: Marma Northern Arakanese Southern Arakanese Source Bernot (Bangladesh)* (Akyab)* ( Sandoway) A + MB MB url ekhan MB MB PeB/MB (address) bagri FeB FeB FeB bathwi FyB FyB ebye FyB FyB eri(shan) FS FS FS FS megrima MeS ... MeS megrl egri(shan) MeS MeS methwima MyS methwi MyS eywe MyS MyS Table 2: Arakanese uncle and aunt terms* c. Burmish In other Burmish languages, many of the uncle and aunt terms are loanwords from Jinghpaw ('Kachin'). Burling (1971) demonstrates that the kinship structure of Maru has been rearranged into the Jinghpa! pattern,3 and that Maru has borrowed a number of Jinghpaw terms. Table 3 shows the overall pattern. Sites. (Ed. ) 153

PAGE 166

Language: Maru At si Lashi Jinghpaw Source: Okell Bur ling Benedict Bur ling MeB ya!JYl: ... ts?a yuk-pho ts?a ny1 .. ... ts?a ts?a MyB ny1 yuk-pho .. FeB ph ph6m6 ma!J/phamo pha-mo wa(=F) FyB yagan phakan ma!J/phatha!J pha-thang wa(=F) MeS ya!Jyl: m6 m?i mo mye-mo nu(=M) .. MyS m?J:kan m?i tha!J mye-thang nu(=M) FeS & na m;:, na moi ning-mo moi FyS na thf; moi ning-thang moi na Table 3: BW'mish and Jinghpaw uncle and aunt terms Atsi shows two Jinghpaw loanwords, for the crossuncles and cross-aunts; and Maru dialect reported by Burling has shifted the meanings of some terms and, like Atsi and Jinghpaw, does not distinguish relative age for cross-uncles and aunts. Under the 1Kachin1 system of marriage, there is a strong preference for mother 1 s brother 1 s daughter-father 1 s sister 1 s son marriage, so it is not too surprising that the Lashi term for mother's brother is, in fact, cognate with the usual Burmish term for wife's father. In the 1Kachin1 system,each lineage is in a wife-receiving relationship with one other patriline.,and is in a wife-giving relationship with another patriline. The Atsi are the Burmish group most tightly integrated into this system. As in Burmese, the terms for parallel uncles (father's brothers) and aunts (mother's sisters) are mostly compounds containing the term for father or mother respectively. The Maru mother's brother/parent's elder sibling term /!J-ylf or /n-y1/ may be almost regularly cognate with Burmese /y1/; Maru occasionally shows additional prefixes in other etyma too. Maru /na/ or /na/ and Lashi 'ning' (father's sister) suggest *ni2 which has cognate forms in Loloish, Naxi, Ugong and elsewhere in Tibeto-Burman, but not in Burmese. For Atsi, for which I have more data, it seems that the order of birth suffixes is quite regular and productive: /mo/ 1first1 /l?at/ 'second', /nu/ 'third' and /tha!J/ 1fourth1 154

PAGE 167

D. Burmic For the Loloish languages, quite closely related to Burmese and Burmish, the mother's brother and father's sister etyma found in Burmish are also represented, as well as several other terms;5 Akha Common Lisu Naxi Ugong Lahu Lijian8 Yanggin8 MeB a g1oe I g1ui (pa_ )* I I kudaJJ V vYvY e v MyB a g1oe I kujE:? V FeB a ui I uv(pa_) wii(pha) I kudaJJ V V f.. bo FyB V wU.wu puJJjE:? a zaw V MeS aui (=M) I I mE:? dalJ V V wU.(ma) (=M) MyS a mui meh ?ajE:? V FeS a k1o kuv(ma) I neda]J any a V V ny1nya FyS a k1o nyi a el}i nejE:? V V nyanya Table 4: Other Burmie uncle a:nd aunt terms Where appropriat.e, Central Loloish terms oft.en have male or female suffixes, too. These are shown in parentheses. Various shifts of meaning can be seen, such as the generalization of father's sister to parent's sister in Lijiang Naxi; the of Lisu /wu/ to parallel uncle or aunt, unlike Lahu and Akha; or the Akha generali4ation of /avuivf to parent's elder sibling ( s.ame sex) =father s elder brother/ mother's elder sister; .also, the extension of Ugong /ku/ to refer to father's elder brother, in addition to mother's brother. Analysable forms in Ugong contain /dalJ/ 'big' (elder); the mother's elder sister term is composed of /m?/ 'mother' and /dalJ/ 1big1 (in the Sangkhla dialect it is /brur/ 'mother' plus / dalJ/) / j E:? / patterns like a 1 small 1 (younger) suffix for some of these terms, though not generally in Ugong; its core meaning seems to be mother's younger sister. 2. Proto-Sino-Tibetan and Burmic terms Of the various etyma for uncles and aunts reconstructed for Proto-Tibeto-Burman in Benedict (1972) and for Chinese in 155

PAGE 168

Karlgren (1957), only two have Burmic cognates. Karlgren (1957) GSR 106Tb 'maternal uncle; wife:s Benedict (1972: #255) *kew 'mother's brother/w1fe s father 1s found in Ugong /ku/ and Naxi /gr/ and has shifted its meaning in Burmese to form part of the 'elder brother' term /eko/. Karlgren (1957) GSR 359d *nier/niei; 'mother'/;Benedict (1972:#316) *niy 'father's sister/mother's brother's wife/wife's mother' shows a strange semantic shift in Chinese, and has been replaced by another term there (cf. Karlgren 1957: GSR 49g *ko/kuo/ 'father's sister; mother-in-law'). This etymon has no Burmese cognates, but Maru /na/, /na/, Lashi 'ning', Lisu Lahu nyi a-, Naxi /fti/ or /ni/, and Ugong /ne/ all provide support for it in its reconstructed meaning, within Burmic. Benedict (1942) speculates that the Burmese-Lolo term for mother's brother is derived from the etymon for 'big' (elder), *k-ri2 (L=LoloishJ Bradley 1979b:756, without prefix). Unfortunately for this hypothesis, the Loloish forms suggest *Tone 3, while the Burmish forhlB imply *Tone 2. There are a few other cognates which show this pattern of tonal difference between Burmish and Loloish. Thus, Burmish *ri2 and Loloish *ri3 must be reconstructed, with Burmese, Arakanese, Maru, Akha, Common Lahu and Lisu cognates implying a Proto-Burmese-Lolo origin for this mother's brother etymon (L 196/7), not found in Ugong or Naxi which retain cognates of the Proto-Sino-Tibetan form. To trace the apparent process of development through l the various stages, it seems that Proto-Burmic may have had 'mother's brother' (L 202-2) and *ni1 'father's sister'; that the former shifted its meaning and-;as replaced by *ri2 (L 196/7) at the Burmese-Lolo stage, with a subsequent development to *ri3 in Proto-Loloish; then came various subsequent independent developments, involving analysable forms, for example, 'head' 1 88A *u2 being used for 'mother's brother', and subsequently 'uncle'-in Burmese; and conversely being used for 'father's brother' in Loloish (L 192/3 *u2). Comparison of the Loloish forms further suggests some po;7ible Central/Southern Loloish innovations, such as L 199 *me1 3 'mother's (younger) sister', and so on; and the loss of the sex-specific use of terms in Central Loloish, leading to the addition of productive male or female suffixes, generalizing 'mother's brother' and 'father's brother' terms to 'mother's sibling' and 'father's sibling' in some dialects of Lahu (e.g. Shehleh, Bradley 1979a), and generalizing 'father's brother' to parallel 'aunt' ('mother's sister') in Lisu. Prior to the inception of inscriptions, the further Burmese developments involved replacing the 'father's sister' cognate by generalizing the 'mother's brother' cognate to 'cross-aunt' ('father's sister'i, then eventually eliminating the use of this form as a 'mother's brother', replacing it with 156

PAGE 169

the 'head' form as noted above. The 'mother's brother' term also survives in the term for 'elder brother's wife' or 'wife's elder sister' in Burmese, /mayl/; this is less paradoxical given the survival of the Proto-Sino-Tibetan term for 'mother's brother' in the Burmese term for 'elder brother', /ako/. In Rangoon, /ayl/ is not an 'aunt' term, having been eliminated when the aunt and uncle terms 'went bilateral'; but /mayl/ survives-two steps removed from the original meaning of /yl/. The various compound forms also show an interesting pattern of development. The suffix /jl/ 1big1 (elder) has survived, but the earlier /thwe/ 'younger' has mostly been replaced by /le/. Interestingly, the former form for 'father' /baj has been replaced by a fused form /phe/, from /pha ?e/ in its core meaning, but it survives in some 'father's brother' or, in Rangoon, 'parent's brother1,terms. And just as /u/ 'head' has replaced /yl/inthe meaning 'mother's brother', /jl/ 1big1 seems to be acquiring the additional meaning of 'mother's elder sister' (or, in Rangoon, 'parents' elder sister'). Also, the 'royal' suffix /t-:J/ in a voiced form /d-:J/, seems to have acquired the meaning of 'mother's younger sister', and in Rangoon 'parents' younger sister'. The radical restructuring of kinship terms in Rangoon has resulted in the total elimination of /yl/ as an 'uncle/aunt' term; in the final elimination of the /thwe/ 'younger' suffix, due to the loss of the 'father's younger brother' term which survives in Mandalay; in the generalization of /ujl/ 'mother's elder brother' and /ba(ba)jl/ 'father's elder brother' so that both are used for 'parents' elder brother', while /Ule/ 'mother's younger brother' takes over 'father's younger brother' as well. And, as noted, /jl/ compounds become 'parents' elder sister', while /d-:J/ compounds become 'parents' younger sister' --both generalized from 'mother's sister' terms, eliminating the 'father's sister' terms found elsewhere. Arakanese developments are more conservative in some ways, but more innovative in three new terms: /abye/ 'father's younger brother', /aywe/ 'mother's younger sister', and /akhan/ 'mother's brother', in most dialects. 3. History of Burmese marriage and kinship Based on comparative evidence summarised in Benedict (1942), it seems likely that Proto-Sino-Tibetan society was patrilineal, with a preference for matrilateral cross-cousin marriage. Bradley (1979b) has discussed the Loloish groups, and concludes that Proto-Loloish society was also patrilineal, preferring mother's brother's daughter --father's sister's son marriage. It also appears likely that there was a bride price (payment by the groom and his family to the bride's family) and, in addition, or instead, a requirement for several years of bride 157

PAGE 170

service (the groom living with, and working for, his parents-inlaw). The location of residence after marriage was, thus, at first uxorilocal, but subsequently virilocal, that is, with or near the groom's family. Some modern Loloish societies have changed certain aspects of these patterns; for example, most Lahu groups are bilateral, and regard cross-cousin marriage as incestuous, while most Lisu groups prefer bride price and virilocal residence. However, some Lahu groups, such as the Shehleh (a Black Lahu subgroup) and the Banlan (a Yellow Lahu subgroup) still allow, or even prefer, matrilineal cross-cousin marriage, and many Lisu grooms do bride service, as do nearly all Lahu ones. The Lisu are patrilineal, and do allow cross-cousin marriage, while the Akha, for example, prefer it. Not surprisingly, the 'Kachin1-influenced Maru, Atsi and Lashi show Jinghpaw-like patterns for the non-reciprocal exchange of spouses; one lineage always, and only, provides the grooms to another which provides brides to the first. Hence, a hierarchy of lineages is created, with obligations created by the receipt of wives. This pattern is not characteristic of Burmese society, despite the close historical linguistic connection within Burmish, including Burmese, Maru, Atsi and Lashi. In Arakanese society, as in most Loloish societies, Naxi society and Ugong society, there is a stated preference -frequently carried out --for marriage between mother's brother's daughter and father's sister's s.on. Moreover, the evidence of the Burmese kinship terms suggests the same at an earlier stage for the rest of Burmese society: it is only with this marital pattern (mother's brother's daughter and father's sister's son) that the mother's brother's wife will normatively be the father's sister (and,of course, father's sister's husband= mother's brother) and so the mother 1 s brother t.erm can generalize to mother's brother/father's sister, as /y'i/ does in Burmese. In this marriage pattern, there would be a direct exchange of women for women between two linked patrilines. Spiro (1977) postulated exchange siblings as the earlier Burmese pattern, based on kin term equivalences which equally support reciprocal cross-cousin marriage. It is, of course, possible that the earlier pattern of exchange between two specific lineages, which comparative evidence supports, could have developed into a general possibility of exchange between any two lineages, particularly with the vast expansion of Burmese society entailed by the politico-military success of the Burmans. After this proposed intermediate stage of sibling exchange, the Burmese marriage and inheritance system has again changed to a bilateral one, in which cross-cousin marriage is at least regarded unfavourably, with many, especially in Rangoon, regarding it as incest. However, there is still some feeling that 158

PAGE 171

patrilineal ties are closer, and incest taboos are stronger patrilaterally. It was traditional until fairly recently for Burmans to do about three years of bride service and to pay a substantial bride price. These are exactly the postulated Proto-Burmese-Lolo customs, which are appropriate in a patrilineal system. Other aspects of the system of kin terms also support this conclusion. A final property of the Burmese system, found also in other Burmish and some Loloish societies, is a differentiation of terms based on relative age: elder or younger. There are separate, unrelatedterms for younger sister, younger brother, and elder brother in Burmese; and a Proto-Burmese-Lolo term for elder sister which is not represented in Burmese. There is also widespread use of verb-adjectives (i.e. stative verbs) such as 1big1 and 'small', as suffixes to indicate relative age of the parent and the aunt or uncle. Perhaps this age-grading, which reaches its extreme among the Atsi, is a relatively recent characteristic of Burmese-Lolo societies. The suffixes used differ in different languages and are generally productive; so it would be risky to postulate very early agegrading. In conclusion, Burmese kin terms for uncles and aunts show extensive differences between dialects and considerable variations within some dialects and the comparison of these kin terms with one another, 1vi th the corresponding terms from closely related Burmic languages, and from reconstructed Sino-Tibetan, has permitted the formulation of a hypothesis that pre-Burmese society was patrilineal, with cross-cousin marriage. Similarly, one may also postulate the presence of certain other characteristics in pre-Burmese society, such as bride service and bride price, from an analysis of comparative cultural evidence within the linguistic groups under study. NOTES 1. This paper was originally prepared for the Language, Gender, and Society Panel of the Asian Studies Association of Australia, to be presented at its Fourth National Conference held at Monash University, Melbourne, on May 10-14, 1982. The support of the Australian Academy of Humanities and of the Myer Foundation (1976), the Australian Research Grants Committee ( 1977) and the Social Science Research Council of the American Council of Learned Societies Joint South East Asia Program, sponsored by the Ford Foundation ( 1980) is gratefully acknowledged. 159

PAGE 172

2. Some useful sources on Sino-Tibetan genetic classification include Benedict (1972) on Sino-Tibetan; Bradley (1975) on Naxi-Burmese-Lolo; and Bradley (l979b) on Burmese-Lolo, which which is known as Lolo-Burmese in Burling (1971), qv. 3. Indispensable information on, and discussion of, Jinghpaw kinship terminology is to be found in Leach (1954) and Leach (1977). (Ed.) 4. Further Maru data were provided by John Okell. Benedict (1942) has collected older data on Maru, Atsi, and Lashi, and Bradley (n.d.) has more recent data on Atsi. 5. Data are drawn from Southern Loloish Akha (in manuscript); Central Loloish Lahu (reconstructed Common Loloish Lahu from Bradley (l979a), in manuscript), and Central Loloish Lisu (Bradley and Hope, 1986). Naxi, which is less closely related, is represented by two dialects, those of Lijiang (Bradley 1975) and Yangning (Fu 1979). The data on the language most divergent from Burmese within Burmic, Ugong (Kok Chiang dialect), was collected in Thailand by the author in 1980-81. Ba Shin Benedict, P.K. Bernot, L. Bradley, D. REFERENCES 1962. Lokahteikpan: early Burmese culture in a Pagan temp le. Rangoon: Burma Historical Comm. 1942. Kinship in Southeastern Asia. Unpublished thesis, PhD, Harvard Univ. 1972. Sino-Tibetan: a conspectus (Princeton-Cambridge studies in Chinese Ling. 11). Cambridge: Cambridge Univ. Press. 1967. Les paysans arakanais du Pakistan oriental. The Hague: Mouton, 2v. 1975. Nahsi and Proto-Burmese-Lolo. Ling. Tibeto-Burman Area 2 (1), 93-150. l979a. 23). Lahu Dialects (Orient. Monograph Canberra: Austral. Nat. Univ. Press. l979b. Proto-Loloish (Scand. Inst. Asian Studies. Monogr. ser. 39). London/ Malmo: Curzon Press. 160

PAGE 173

Bradley, D. Bradley, D. & Hope, E.R. Brant, C.S. & Mi Mi Khaing Burling, R. Fu Maoji Judson, A. Karlgren, B. Leach, E.R. Levi-Strauss, C. Luce, F.H. n.d. Field data on Burmese, Arakanese, Tavoyan; Maru, Atsi; Lahu, Akha, Lisu; and Ugong. 1986. (eds) Lisu-EngUsh dictionary. Canberra: Pacific Ling. 1951. Burmese kinship and the life cycle: an outline. Southwest J. Anthrop. 7, 434-54. 1965. Burmese kinship terminology. Amer. Anthropologist 67 (5, Pt.2), 106-17. 1971. The historical place of Jinghpaw in Tibeto-Burman (Occ. papers Wolfenden Soc. on Tibeto-Burman Ling. 2). Urbana: Univ. Illinois. 1979. La famille matriarcale et les termes de parente chez les Naxi de la commune populaire Yongning. (Paper, 12th Internat. Sino-Tibetan Conference Paris). 1953. Judson's Burmese-English dictionary (rev. & enlarged by R.C. Stevenson & Revd R.H. Eveleth). Rangoon: Baptist Board Publ. 1957. Grammata serica recensa (GSR). Bull. Mus. Far East 29. 1954. Political systems of highland Burma: a study of Kachin social structure. London: G. Bell. 1977. Jinghpaw kinship terminology. In E.R. Leach, Rethinking anthropology (London Sch. Econ. Monogr. on soc. anthrop. 22). London: Athlone Press, 28-53. 1952. Kinship systems of three Chittagong hill tribes (Pakistan). Southwest J. Anthrop. B, 40-51. 1981. A comparative word-list of Old Chinese and Tibetan. London; Sch. Orient. Afr. Stud. 161

PAGE 174

Okell, J.W.A. Spiro, M.E. Than Tun Tun Nyein n.d. Maru field data. 1977. Kinship and marriage in Burma: a cultural and psychodynamic analysis. Berkeley/Los Angeles/London: Univ. Calif. Press. 1958. 1287. Social life in Burma, A.D. 1044J. Res. Soc. 41 (1-2), 37-47. 1906. The students English-Burmese dictionary. Rangoon: Burma Secretariat. 162

PAGE 175

THE BULGING MONOSYLLABLE, OR THE MORA THE MERRIER: ECHO-VOWEL ADVERBIALIZATION IN LAHU James A. Matisoff l. Sesguisyllabism in South-East Asian Languages The languages of South-East Asia are overwhelmingly monosyllabic in structure (the notable exception being the Austronesian family) at least in the sense that their morphemes are only one syllable long.l Yet, as all South-East Asianists can ruefully testify, this 1monomorphosyllabism1 is by no means to be equated with phonetic simplicity! The South-East Asian monosyllable often seems to be bulging at the seams with phonetic material: consonantal, vocalic, and supra-segmental. Diachronically, phonemic features frequently bounce back and forth from one segment of a South-East Asian syllable to the other.2 Adjacent vowels and consonants unidirectionally or mutually influence each other's articulation --something I have been known, rather inelegantly, to refer to as 'intersegmental stopover'. More strikingly, decaying consonantal contrasts in syllable-initial or -final position may be 1transphonologized1 into the suprasegmental realm, so that previously redundant tonal features acquire a compensatory phonemic status.3 The intersegmental attraction is by no means confined neatly within the boundaries of individual syllables. A voracious South-East Asian monosyllable may also absorb phonetic material from a neighbouring syllable, incorporating it into its own substance. These adJacent syllables belong originally to other morphemes --these are after all 1monomorphosyllabic1 languages --yet the fusional process respects no lexical boundaries. A. Prefixization of compound constituents Especially vulnerable to trans-syllabic absorption are weakly stressed morphemes that stand in a modifying or subordinate relationship. A modifying syllable in a lexical compound may undergo such radical phonological reduction that its original morphemic identity is obscured. Once this happens, it can become more like a meaningless affix or 'formative' than like a full noun or verb and,although elements in compounds may also sutfer this fate, even in non-monosyllabic languages like English, the process of 1affixization' or 'cliticization' comes into full flower only in truly monomorphosyllabic language families like 163

PAGE 176

Tibeto-Burman (TB): --The now meaningless in Written Burmese (WE) parwak 'ant' (>Mod.Bs. payw?) descends from the free morpheme *buw 'insect, bug' (Benedict 1972:#27), reflected by forms like Written Tibetan (WT) 'bu' and by the independent WE word pui 'bug'. The fully dissyllabic compound for ant is explicitly attested by forms like Lahu (Lh.) pu-g;? 'ant' (<*buw-rwak), where the vowel of the first syllable retains its quality and has not been reduced to shwa.5 --The prefixal element sain WE samak 'son-in-law' (>Mod.Bs. 8ame:?) is a reduction---;f the full morpheme *za **tsa 'child, son' The unprefixed root is reflected in forms like WT mag-pa and Lushai ma.k-pa. The dissyllabic prototype *tsa-mak is directly preserved in the Ch'iang cognates ts'l:l.-me and ts'l:l.-mja. Forms like Dhimal hma-wa and Lahu have taken the fusional process even further than Burmese, preserving only indirect traces of a sibilant feature before the root.6 Sometimes the telescoping of two proto-syllables into one is so complete in TB that the dissyllabic prototype can only be established at the cost of considerable comparativehistorical toil. The TB root for lungs that Benedict reconstructed as *tsywap or *tswap largely on the basis of Lushai tsuap, has since been shown to be a fusion of two separate roots *tsi-wap, with a second element that originally meant 'spongy, porous'. (Matisoff 1978:113-23, esp.ll5). In Jinghpaw (='Kachin'), an important TB language of northern Burma and adjacent areas of China and India, the process of prefixization has been carried very far, to the point where hundreds of nouns and verbs have the phonological shape CpaCi (G) V (Cf)*.7 Most of these unstressed prefixal or 'pre-lnitial' syllables (C a-) seem to be relatively recent accretions to their roots,Pand have no plausible etymologies or definable meanings. Others are reductions of semantically obscure but fully syllabic prefixes like glimor with which they sometimes alternate. One prefix, has a clear-cut grammatical function and has been generalized to all verbal roots with the meaning causativizer/transitivizer.B Still others and these cases are the most interesting in the present context are reductions of fully syllabic root-morphemes that were once the first constituent in compounds. An example of this latter type is the pair lakhra 'right' and lapai 'left', where the la is a reduction of the widespread Proto-Tibeto-Burman (PTB) root Refer to linguistic symbols/abbreviations at the head of the NOTES. (Ed.) 164

PAGE 177

*lak 1hand1 (Benedict 1972:#86) (cf. WB lak-ya 'right hand'). In some TB languages, the vowel of the reduced prefixal syllable is not exactly a centralized shwa, but rather an unstressed front or back vowel, like Angami Naga -e(e.g. mekru 'dove'; 'causative prefix') or Gyarung -.2.-(e-:-g. korok __ 'ant'). In any case,these prefixal syllables are so stressless that they usually cannot bear a full tone --and it is universally true that even if a tone contrast does exist in such syllables, as has been claimed for Jinghpaw (Maran 1971) it is on the rudimentary side (never more than a two-way, high vs. low distinction). Important as these reduced pre-syllables are in TB, they seem to play an even more basic structural role in the languages of the Austro-Asiatic (AA) or Mon-Khmer (MK) family.9 In almost all branches of MK, except Vietnamese, these 'minor syllables' abound.10 Compared to TB, there is generally a wider variety of possible consonants in MK minor syllables, and it is seldom possible to derive such a syllable from any semantically plausible fully syllabic prototype. This makes it look as if the MK minor syllables go back to remote antiquity, and were present in the family ab initio. In any event, it is clear that untold thousands of words in South-East Asian languages are neither monosyllabic nor dissyllabic, but rather what we might call sesquisyllabie: a 'syllable-and-a-half' long.ll B. Suffixization and Tused vocalic nuclei This paper is concerned primarily with bulging at the other end of the syllable, i.e. the vocalic nucleus. In SOV, the postpositional languages,12 grammatical functors like case-and aspect-particles follow the nouns or verbs with which they interact. Since these functors have a high textual frequency and are naturally unstressed by comparison with their preceding root-word, they are prime candidates for sloppy articulation and phonological reduction. In a language like Lahu --a member of the Central Loloish group of the Lolo-Burmese branch of Tibeto-Burman -postpositional particles are apt to lose their initial consonant in rapid speech.l3 Sometimes both variants are used almost interchangeably (e.g. ta a I perfectiVe aspect I ; tha? -a? 1 accusative case 1 ) (Matisoff l973a: 38) ; but the disproportionately large number of modern Lahu particles that begin with a vowel all the time makes one suspect that in some cases an old initial consonant has been lost for good.

PAGE 178

Once its initial consonant has disappeared, a functor is so phonologically slight that it may be helpless to resist the pull of the voracious, fully-stressed noun or verb that precedes it. If circumstances are right the functor's vowel and/or tone may be incorporated bodily into the vowel of the head-syllable, resulting in a complex, fused vocalic that is 'a mora-and-a-half' long. The 'sesquimoral11 syllables of Lahu are synchronically anomalous and marginal, but the strains to which they subject the phonology are of fundamental importance, since they might well presage an eventual radical restructuring of the entire vowel system. In what follows, we shall focus on an elusive type of Lahu sesquimoral syllable, which was only recognized after many years of work on the language: eoho-vowel adverbials. First, however, we should consider the whole phenomenon of vocalic fusion in Lahu in terms of the forces operating to tear down or build up the phonetic substance of the syllable. 2. Diachronic dimensions of the phonological system of Black Lahu:l5 intersegmental influence and the economy of the syllable The Black Lahu dialect of Chiangdao, Chiang Mai Province, Thailand, has a system of nine simple vowels rather similar to that of Siamese:l6 i u e 0 a o. If only syllables with these nuclei are taken into account (and they comprise the vast majority), the Lahu syllable-canon can be formulated as: This starkly simple syllabic structure, maximally comprising an initial consonant, a vowel, and a tone, represents an extreme reduction from the point of view of the complex syllables reconstructed for PTB: [TJ *(P1)(P2)ci(G) V (:)(Cf)(s).17 The final stops */-p -t -k/ of Proto-Lolo Burmese (PLB) have lost their oral occlusion in Lahu, though they have usually left their trace in the form of a post-vocalic glottal stop.l8 Although such syllables are transcribed with the symbol 1?1 166

PAGE 179

written on the line after the vowel (e.g. t3?, kh?>?), I prefer to consider this glottalization to be a suprasegmental, 'tonal' feature from the synchronic point of view (see Matisoff 1973a: 25-6).19 There is a two-way tonal contrast in these Lahu 'checked syllables' (as in similar syllables in most other Loloish languages),20 symbolized by the (highstopped tone) and/'?/ low-stopped tone). Black Lahu is thus a seven-tone language, with 5 open tones and 2 checked ones.21 Although the former three-way positional contrast in PLB final stops has been neutralized in Lahu, leaving only stopped tones as the pale reflection of the proto-occlusion, the position of articulation of the *Cf has differentially affected the quality of the preceding vowel, so that in many cases there has been no loss of contrast with respect to the syllables as a whole (Matisoff 1972): PLB *-ak > Lahu -a? PLB *-at > Lahu -e? PLB > Lahu -o? PLB *wak ) '?22 'pig' (ibid., #168 > Lh. PLB *k-r-wat 'leech' (ibid., #167) > Lh. ve? PLB *k-rap 'needle' (ibid., #191) > Lh. g_Q_?. Similarly, the three *nasal Cf's of PLB have totally departed from the segmental scene, leavlng their traces only in the quality of the preceding vowel: PLB > Lahu -::::> PLB *na!J 1 'you' > Lh. n?> PLB *-an > La.hu -e PLB *wan 1 1dhole' > Lh. ve PLB *'-am > La.hu PLB *;3'am 1 'otter' Lh. -0 > As these examples indicate, La.hu shows assymetry in the degree of decay of original syllable-final consonants according to whether they were *stops or *nasals. While almost all originally *stopped syllables have preserved at least a postvocalic [?J --the exception being 'doubly glottalized' syllables, which lost their occlusion entirely (qv. n.l8 above) --the *nasal C 's have not even caused a nasalization of the preceding vowel; so that the feature of nasality has disappeared completely from the rhyme of such syllables. (But see 2_.A.c,d,_ e, below.) These transphonologizations may be symbolized by and [where cf was /p t k/J [where cf was /m n !J/].

PAGE 180

That is, final *stops have been transphonologized into both suprasegmental (tonal) and vocalic dimensions of contrast, while final *nasals have influenced only the oral quality of the preceding vowel and have not superimposed any coarticulatory feature upon it. A. Secondary glottalization and nasalization in Lahu Counteracting the overall decline in syllable-final occlusion in native inherited lexical material is a variety of disparate phenomena which are 'conspiring' to increase or introduce the suprasegmental features of glottalization and nasalization. (a) Loans :f'rom Tai and Burmese where the donor language had /-p -t -k -?/ are usually borrowed into Lahu under a stopped tone (e.g. ha? 'iove' [< Shan; cf. Thai rakJ), and this same treatment is sometimes accorded to loans where the donor language had a short vowel followed by a nasal (e.g. ta?-n(i? 'police' < Thai tamrU.at), or even a liquid (e.g. motos:i:,kt"? 'motorcycle' < Eng.). (b) Any Lahu action-verb that is under one of the five open tones may be given imperative force by shortening its vowel and pronouncing it with a glottal stop: d?> 'drink' > d?>-? 'Drink some tea!'23 (c) Loans from Tai and Burmese with a nasalized vowel are sometimes pronounced with a nasalized in the careful speech of Lahu who have a fair knowledge of the donor language (e.g. ?>-bo ?>-ban 'merit; advantage' [< Shan (cf. Thai bun), ultimatel0P8.li); a-khwa -a-khwan 'permission' [< Bs .J). (d) Syllables whose initial is Q_or zero, and whose vowel is -.:_ or -E_, are optionally nasalized by the widespread phenomenon I have called rhinog"lottophi"lia: :>(n) 'four', ?>-ha.(n) 'spirit', h3(n)? 'to coil', ha(n)Lha(n)? 'fast', etc.25 (e) A: few verbs acquire v0wel nasalization in vivid aclverbial expressions involving the particle ka? (Matisoff 1973a:4.44): _!E (V) 'spread open' > lj)iin (AE)---..;ide open'; the (V d.) 'be stTaight' > then ka7 'straight as an arrow'. a J Marginal and heterogeneous as such secondarily glottalized and nasalized syllables may be as far as their historical status is concerned, they are of considerable potential importance for the future of Lahu phonology. As a general rule of thumb, once some feature is present phonetically in a South East Asian monosyllable, no matter how redundant or trivial it may appear, it is avai.lable for future exploitation and transphonologization. 168

PAGE 181

3. Complex vocalic nuclei in Lahu (Matisoff 19T3a: 15-20) T Our streamlined syllable canon, (Ci)V, fails to account for any but the nine simple vocalic nuclei mentioned above. In addition, Lahu has a wide assortment of 'complex' nuclei consisting of more than a simple vowel. These may be roughly subdivided into 'intrinsically complex' (i.e. occurrent within a single morpheme) vs. 'fusional' (the result of phonetic telescoping across morpheme boundary). Almost all intrinsic complex nuclei have come into the language through borrowing. A. Intrinsic falling diphthongs The most frequent diphthong of this type is /ay/ found in a great many loanwords from Tai: lay 'several', hay I evil, fierce I vay I fast I thay I to plough I etc. 26 Other, rarer,diphthongs found in loanwords include /aw/ [a2J (g_aw 'tell, narrate'' ma?-paw 'coconut') and /ew/ (ma?-tew 'gambling, card-playing', khe-mew? 'a Meo, a Hmong'). B. Intrinsic rising diphthongs in loanwords Lahu rising diphthongs always begin with a labial (never a palatal) semivowel. We write this phonemically as /w/, though its precise phonetic quality depends on the height of the following vowel, thus: /w/ [!P.J, /we/ [2eJ, /w./ [?e:J, /wa/ A large proportion of these syllables are loans from Burmese or Tai, e.g. pwE 'festival' [< Bs.J, a-khwa(n) 'permission' [< Bs. J, hw-;j'f 1 oyster' [< Shan (cf. Thai h:5,; )J, kwa(n) 'govern' [ < Bs. J, naLwe:-si I candy' [< Shan (for 2nd syll. cf. Thai ?o,j 1sugarcane1 )J, etc. The labial element strikes the ear as more vocalic than consonantal, especially before non-high vowels. It is articulated laxly, without very pronounced puckering of the lips, so that the syllable sounds 1sesqui-moral.1 Syllables which begin with this labial glide (as in 'candy', above) are deemed to have zero initial. That is, the is a feature of the vocalic nucleus rather than of the C .. 2 1 C. Labialized doublets of syllables with back vowels (Matisoff 19T3a:l9, #1.43) Further support for the analysis of /w/ as part of the syllable's rhyme is provided by an extremely interesting and rather productive type of doublet formation, wherein syllables with simple back vowels /u o may also be pronounced with nuclei consisting of /w/ plus the front vowel of the corresponding height /i e e:/: i.e .. u-wi, o-we, Thus, 13a-kuga-kwi 'dried fish'; -lo?:Qo --13?=9we 'stick inserted in bobbin of spinning-wheel to take wound thread off' ; yg-mi-to yg-mi-twg 'a bear'; rice' a-gha-wE (lit. 'rag-weed

PAGE 182

rice') 'ritual rice sent to grave of dead man', etc. This doublet-making is still a living process in Lahu, and is even applied to loanwords: ku-kwi 'bed' [< Shan < Bs. 1bedstead1J; eo-ewe 'era, period of time' [< Shan (cf. Thai chUa.)J; :>--13? :-::r-1:;f? 'terraced field' [< ShanJ; 13 'swim' [< Shan (cf. Thai luj 'wade')J. As far as native lexical material is concerned, the 'basic' variant in these pairs is clearly the one with the simple back vowel. Thus the word for 'cooked rice 1 is always ; ( < PLB except in the single compound a-gha-wE. Yet for certain words the labialized variant occurs more frequently, and may even have displaced the simpler one entirely (e.g. 'barking deer (CervuZus muntjac)', but never Among the loanwords which receive this doublet treatment, some had complex nuclei involving a labial element in the donor language (e.g. era, swim), but others originally had only a simple back vowel (e.g. bed). There are also cases where a Tai syllable with a falling diphthong consisting of a back vowel plus palatal semivowel (e.g. -o(o)j) gets borrowed into Lahu with a prelabialized nucleus with no non-labialized doublet (e.g. oyster, candy)(3.B above). This doublet formation seems originally to have developed internally within Black Lahu, though it clearly has been reinforced and encouraged by attempts to approximate the complex vocalic nuclei in the languages with which these Lahu have recently been in the closest contact: Burmese, Shan, Northern Thai, and Siamese. The question arises whether the labialized variants should be considered 1 intrinsic 1 or 1 fusional' complex nuclei. It will be recalled that the labial element we write abstractly as /w/ is really a semi-syllabic vowel whose phonetic height is the same as the following full vowel: /wi/ = /we/= /we/= [oeJ. These complex nuclei are,therefore, 'fused' in the sense th;t s o > $2 o > This 'something' which usurps the peak of sonority is the front vowel of the corresponding height /i e e/. But this front vowel does not mean anything-it does not belong to any separate morpheme. It is essentially a meaningless extrusion or extension from the original nuclear monophthong. All it does is provide some phonological bulk, a benign bulging of the syllable 1 s substance. 29 As indicated in Section 3 above, we prefer to reserve the term 'fusional' for cases where there has been a phonetic telescoping across morpheme boundary, and these labialized doublets do not quite meet this criterion. 170

PAGE 183

D. Revising the syllable canon In any event, this doublet formation has conspired with the influence of foreign words (3.B above) to reintroduce the slot 1 G 1 into the Lahu syllable. Revising our syllable canon to accommodate these cases, we get: T* T* (c. ) 1 (G) V or, more specifically, (c. ) (w) V 1 where T* may include the feature [?J, and where everything except C. belongs to the rhyme of the syllable. 1 While we are at it, we should also add to our canon the feature of secondary vowel nasalization and the intonational feature of imperative glottal stop (2.A above), as well as the post-vocalic semivowels /-y -w/ that we have encountered in loanwords (3.A above). By now, we are faced with an overall syllable structure of surprising complexity: 30 We shall continue to refine this formula as we go along. E. Fusions of verb-particles to their verbs (a) With o (Pv). When the verb-particle (P ) o (Matisoff 1973a: #4.64), indicating 'change of state' or 'completed action', follows a verb under the same tone as itself/'/, the two syllables are fused into a single sesqui-moral nucleus without affecting the quality of either vowel. We write these sequences with a hyphen: pa-o 'Cit'sJ :finished now'; ga-o 'CWe'veJ arrived already'; ma-o 'Now [IJ see [itJ. I Phonetically 1a-o1 is identical to the intrinsic diphthong found in loanwords like gaw 'narrate' or caw 'lord' (3.A above), but there is no reason to obscure the morphemic structure of verb-plus-Q sequences by such spellings as paw, gaw, or maw. (b) With e (P ). The verb-particle e (Matisoff 1973a: #4.61) indicates 'traXsitive motion', or 'departure from the center of interest; departure into a new state'. It usually maintains its syllabic integrity with respect to its verb: ha e 'go to spend the night'' ph::l away'' s;i, pass t3? e 'go out from'. In a few cases, however, the two into a sesqui-moral unit: pa-e 'fall down. fall over' na-e 'get well, recover, heal'. Although '--these fused syllables in 'a-e' rhyme exactly with intrinsically diphthongal loanwords like lay 'several' or may 'wood' (3.A above), 171

PAGE 184

their bimorphemic status induces us to write them with a hyphen. The most interesting case of all is provided by one of the commonest and most important verbs in the language, 'go'. This word rhymes perfectly both with the monomorphemic loanwords in (3.A) and with the bimorphemic fused nuclei pa-e and na-e. The does not occur. Since there is no evidence that is of non-Lahu origin, I have suggested (Matisoff 1973a:l5-16) that it might well represent an ancient fusion of a now obsolete verb the directional verbparticle *ga-'e.31 Unlike the cases of 1 fall down 1 and na 'be cured, which occur independently in other contexts than before modern Lahu has no simple with the meaning 'go'. For this reason, I write the with the same symbols used for the rhyme of the monomorphemic loanwords in This is a striking instance of a fusion that has occurred so thoroughly that all traces of the 'seam' or 'suture' have disappeared, (somewhat analogous to the loss of previously existing morphemic boundaries among English speakers; qv. n.4). When fusion reaches this point (which we might call superfusion), the nucleus achieves a new wholeness or 'intrinsicality', no less real because it is diachronically secondary. We have thus identified three stages of intimacy between a verb and the following verb-particle e: (1) separate. and equal moras (2 syllables) ha e 'go to spend the night' (2) fused sesqui-moral unit (one-and-a-half syllables) with sense of morpheme boundary preserved na-e 'get better' (3) 'superfused' sesqui-moral unit (one-and-a-half syllables) with sense of morpheme boundary obliterated 'go' By now it will be evident that the whole distinction between 'intrinsic' and 'fusional' complex nuclei is a fuzzy one, since it can be no more precise or stable than the concept of morpheme boundary on which it is based. F. Fusion in lexical compounds (Matisoff 1973a: 18-9, #1. 42g) There are a number of nouns in Lahu which exhibit a peculiar complex vocalism: the syllable peak is a central vowel (esp. !_ or which is then followed by a non-syllabic palatal offglide: /-3:y/C3:iJ, This second element is suffixal, since most of these words have alternate pronunciations with simple vowels /3: a a/. It is impossible to assign any definite meaning to the suffix, however, so we are again faced with an elusive 'morpheme boundary'. It seems in fact that the 172

PAGE 185

/-y/ may represent a merger of more than one originally independent lexical item. Our examples are almost all native Lahu words, but include at least one borrowing from Tai. (a) ?l-ley This nouns means 'something extra, something special, something left over', and is also pronounced ?l-le. It is derived from the verb le 'be left over 1 [ < Shan (-;:;r.-Thai lya J It is safe to assume that the second element was originally the verb-particle e (3.Eb above), since the meanings of le and e are highly compatible ( 1 go on being there to the point of excess'). (b) ?l-mey This word means 1 powder' and occurs without the prefix in such compounds as sa-ma-mey 1cornmeal1 v?-mey 'pollen' ('flowerpowder'), key-mey 'glass fragments' (kew'glass' <:Tai), jeLmey 'dust' ('earth-powder'), etc.32 In this case there is no doubt at all that the -y_ is a reduced and incorporated version of the Pv since there exists an independent verb me 'to powder, reduce to a powder' and the fused form mey ( < me + retains full verbal force, as in mey V:e I to pOWder Salt 2 redUCe rock-Salt tO pOWder I (c) This interesting word, meaning 'mirror, hand-mirror', seems to involve a fusion with quite a different morpheme, namely the diminutive M g.33 pfx The basic form is me?-ga?, comprlSlng the morphemes (< PLB *s-myak (Matisoff 19!2: and gf?. (M fx) 1 someth1ng shiny; shadow' ( < PLB *k-np ( 1-bid., #189) ,, 'something shiny to the eye'. A synonymous with se-condary high-rising tone also occurs 34 The meaning of l:( m8?-ga ranges from (l) 'glass as a material, a glass object' to (2) 'a sheet of glass' to (3) 'a looking-glass; mirror' to (4) 'eyeglasses. Since this is a broad range, it is easy to see how a diminutive coinage -ga-t 'little (sheet of) glass' (big enough for a hand-mirror but too small for a window) came to be specialized both phonetically (fused nucleus gay) and semantically ('hand-mirror'): > mt?-ga > > (d) ?l-@y; ?l-kh; -gay Other words of this type, where the morphemic identity of the fused element is not so clear. include something in colour'; 'a scar'; 'a doubled or forked digit' 'forked finger', 'forked toe' c173

PAGE 186

G. The complex covalent nucleus after palatal initials In a few words beginning with palatal initials, there is a distinctive complex nucleus that is 'almost two moras' in length--i.e. the nucleus takes longer to utter than a sesquimoral one like /ey/ or /we/, but somewhat less time than two full vowels in hiatus. The first segment of is the superhigh buzzing vowel [1], the normal allophone of after palatal initials (Matisoff 1973a: 6). The second segment is a true low front vowel not merely a palatal semivowel. This diphthong is thus neither 'rising' nor 'falling' in the usual sense, but rather what we might call in that each mora receives more or less equal prominence each by itself is somewhat less than a full vowel).3 The principal words with this vocalism are as follows: 'measles'; 'drizzling (of rain)'; te ve 1 to whisper 1 ; and -cg 1 object smaller than the----;}"orm. This last morpheme occurs as a bound constituent in compounds like 'little finger, pinkie-finger'; 'little toe'; 'small species of figtree'; 'small-leaf banyan' (a species smaller than the ordinary banyan, which is called na-qu-c). The meanings suggest that the fused element is our diminutive morpheme g, appearing in only 'semifused' guise --not totally fused a; in m?-gey 'mirror' (3.Fc above). 38 This is of especial interest since it has two variants with simple vowels, and -cg. Collectively the various allomorphs exemplify the process of phonetic fusion at several different stages. The compound 'small sp. of chilli-pepper' has the allomorph with This is presumably the basic root-form of the morpheme, probably related to the important word (Mpfx) 'a section of a long object' [< PLB *?dzik (Matisoff #45). 9 In contradistinction to this unfused form, there has also developed a superfused variant -cg, which is optionally used in the words for the smallest digits: la? -no-cg 'little finger', ('little toe.' Here,the vowel of the superadded diminutive morpheme fgf has actually displaced the original nuclear vowel of the root-syllable entirely.40 In the process the initial consonant becomes a phonetically palatal affricate instead of the dental allophone it had been when the was still there: c:i: [tsh:J > cg However, the Lahu palatal series /c eh j 8 y/ has dental allophones before viz. [ts tsh dz s zJ (qv. Matisoff 1973a:6). 174

PAGE 187

To recapitulate: 1 joint, section' + 'little' UNFUSED > 'less than norm' SEMIFUSED > CE 1id. I SUPERFUSED In order to accommodate these nuclei with vocalism, another complication must be introduced into our syllable canon: T* y (w) V w ? (N) g Constraint: E occurs as second member of a complex nucleus only if the' preceding vowel is In such syllables the Ci is usually a palatal, but may also be a non-palata.l voiceless spirant. 41 4. Fusions in adverbial expressions One of the most important bits of grammatical hardware in Lahu is the subordinating particle which serves to mark a wide variety of structures as attributive to a nominal or verbal head. (see Matisoff 1973a: #1.42d, 1.8, 3.612, 3.617c, 3.618, 3.62, 4.2b, 4.42, 5.424, 6.U4c, 6.493'). In keeping with its high frequency, predictability in well-defined syntactic constructions, and slight phonological shape, is a phonetically unstable morpheme. Though usually under the open low-falling tone /'/, it sometimes acquires a glottal closure in rapid speech, becoming j'g?/.42 There is another particle pronounced which occurs in quantified noun-phrases with a 'minimizing' meaning that is often best translated 'only'. This is also a kind of subordinating function, and I believe this particle to be of the same historical origin as the subordinator Synchronically, however, they should be distinguished, since they eo-occur in attributive constructions involving the 'minimized extentives,143 To complicate matters, our minimizing 'g is also tonally unstable. It,too,is sometimes pronounced unde; the low-stopped tone j'g?j. Furthermore in the minimized extenti ves ( 4. A below) it has developed high-rising or high-falling tone // or making it look like the etymologically distinct 'diminutive' M fx f we have already encountered. p The same factors (phonetic slightness, high frequency) which make these particles tonally unstable also make them prone 175

PAGE 188

to fusion or incorporation into the preceding syllable. The rest of this paper will be devoted to fusional nuclei involving the various constructions in which subordinating and minimizing E: occur. Most of these will be passed over very briefly, since they have already been discussed in detail in Matisoff (l9'73a). Only in the case of which were not discovered until l9'7'7, will we expatiate at greater length. A. Th . d t t 44 e nunlnuze ex en lVes A few Lahu adjectives referring to measurable quanti ties have morphophonemically related forms that occur together with the determiner chi 1this' to yield expressions of 1extentive1 meaning. ma 'be many' 3, 'be big' ;yi 'be long' vi 'be far' I chi ma I chi h:i, I chi s:j, I chi f3, ----'this many, this much, this amount' 'this big, this size' 'this long, as long as this' 'this far, this distance' mu 'be tall, high' lchi mu 'this high, this tall'. To these extentives may be added a further morpheme, to be identified with the 'only', which serves to minimize the degree of the quantifiable characteristic in question. The resultant nuclei are fused, and acquire the high-rising tone (at least in their onset). Two stages in the fusional process are exemplified in these words: in the partially fused items the second element retains the vowel quality 1-gl; in the totally fused words, the second element has been reduced to the palatal semivowel 1-yl. Of our five extentives, two are minimized only with 1-gl, two either with 1-gl or 1-yl, and one only with 1-yl. See Table l. ORDINARY EXTENTIVES MINIMIZED EXTENTIVES Partiall;z: Fused Totall;z: Fused Gloss chi si 'this long' chi ---chi sH45 'only this long' chi mu 'this high' chi mw -'only this high' chi M 'this big' chi hH -chi h'E chi h!i ChilJ 'only this big' chin 'this far' chi f4 -chi fH chi cniJ 'only this far' chi 'this much' chi !!!fuL 'only this much' Table l: The voaalism of extentive morphemes in Lahu

PAGE 189

It is noteworthy that the two stages of fusion we observe in these expressions are exactly paralleled in the non-extentive lexical compounds we have already discussed, e.g. me-gey 'mirror' (total fusion: 3.Fc above) vs. 'smaller than the norm' (partial fusion: 3.G above). ---Furthermore, if we look more closely at the partially fused minimized extenti ves, it becomes clear that further fusional substages can be recognized --and that indeed it is artificial to try to compartmentalize a gradual process into discrete stages. When a native speaker feels there is a tonal difference between the two parts of the nucleus (e.g. chi "less fusion has occurred than if there is no such perceptible difference (e.g. chi hH). 47 B. Types of adverbial/adnominal expressions with subordinating Lahu has several kinds of adverbial structures which can also be subordinated to nouns via the genitive/relative particle ve. We refer to these collectively as 'subordinate expressions' (or SEs). (Matisoff l973a: 278-301, #4.42). They include gha-adverbials, stative adverbials, reduplicated verbs, intensified adjectives, and verbal elaborate expressions. It is characteristic of SEs to include the subordinating particle In the case of gha-adverbials, this is occasionally fused into the vocalic nucleus of the preceding morpheme. In order to lay the groundwork for our discussion of the fusional type of SE pm' exceZZence --echo-vowel adverbials we should first list examples of these other kinds of modifying structures. See Table 2 below. TYPE OF SE minimized extentives gha-adverbials48 stative 4 adverbials 9 ADVERBIAL USE chi + Nd. + + Vh --liD -ext chi ph? 1be so small 1 gha bu.? i ea 1 eat to satiety' n'l 1be red, become red 1 ( lit go redly' )5 1'7'7 ADNOMINAL USE gha bu? i ve j.!_ 1a satiety of liquor' M + + ve + N --h n'l i ve a-po 'a red shirt'

PAGE 190

TYPE OF SE reduplicated verbs 51 intensified adjectives 52 verbal elaborate expressions 53 ADVERBIAL USE V V +E:+V 1 + l h da?-da? E: te 1do :;ery well' V d. + B a J v + + V h qha-t6 1 ph? 1be bitter as gall' V XV X+E:+V 1--2-h or ha-lE:-ha-ga l te 1be happy and relaxed' l phE:? 1be unjust and dishonest' ADNOMINAL USE V + V + + ve + N 1 1 h da?-da? E: ve 'a v;cygood knife' V + Bv + E: + ve + Nh adj gha-t6 l ve a-pht 'a damn bitter cucumber' V -X-V -X + E: + ve + Nh 1 2 or ha-lE:-ha-qa l 81-gwe 'a happy and relaxed meeting' l ve gha?-sg 1 an unjust and dishonest headman' Table 2: Previously recognized types of subordinate expressions in Lahu . C. Echo-vowel adverbialization Still another type of subordinate expression exists in Lahu, though its subtle and elusive phonetic realization prevented me from becoming aware of it until 1977. Syntactically these expressions behave identically to the other kinds of SEs. In their adverbial guise they occur only before 'dummy' abstract verbs like 1 go 1 phE:? 1 be 1 te 1 do 1 As adnominals they are connected to their head-noun via the relative/genitive particle ve. The big difference lies in the way the subordinating particle realized. Instead of always having the vowel quality the particle takes on the same vowel quality as the nucLeus of the previous syllable. For this reason I would like to call them echo-vowel adverbiaLs, or EVAs for short.54 By the end of my 1977 fieldtrip I had collected 22 examples of EVAs. Though this is a relatively large number,55 I have the distinct feeling that this is only the tip of the iceberg, and that many more remain to be identified. By happy chance, the present corpus of EVAs includes examples of the echoing of all nine of the basic vowels of Black Lahu. The largest number have back vowels /u o o/, with /o/ 178

PAGE 191

being especially frequent. The least-often echoed vowels are central jf, e a/. As for the cases with /e./, it becomes an intricate matter to distinguish an echo-vowel formation from the usual kind of stative adverbial with the 'intrinsically' ._-coloured particle l (see below) Nearly half of our echoed syllables are under checked tones '?f; phonetically the glottalization here is perceived as a. slight interruption of the phonation 6somewhere in the middle of the prolonged sesquimoral syllable.5 An echoed vowel sounds like a single long vowel with a double 'tonal pulse. 1 To my ears the tone contour of the syllable usually changes noticeably at a point about halfway through the vowel, ending up almost always as mid-tone (unmarked in the orthography).57 Morphologically, all the EVAs consist of two parts. The first element (the 1head1 ) is a fully meaningful morpheme (usually an adjectival verb) that can occur independently in other contexts. The second element (the part that is fused with its echo-vowel--let us call it the 'tail') is typically a bound, restricted entity that occurs nowhere else than after its particular head. (I have called such hapless, meaningless formatives 1morphans1 i.e. orphan morphs. See Matisoff l973a: 60-l, #3.333.) In this respect,the EVAs most resemble intensified adjectives (Matisoff l973a: #4.424) where the otherwise meaningless intensifier is also selected by one, or at most two specific head-morphemes.58 (We adopt the orthographic convention of separating the head from its tail by an equals sign'='.) long. 59 The total EVA is almost always 'two-and-a-half' syllables An EVAs 1ideal1 shape is: HEAD=TAIL-ECHO. l l l/2 Often, however, variations are possible (as we shall see when the EVAs are listed individually). The echo may optionally be omitted: HEAD=TAIL(-ECHO). Sometimes, the tail may be followed either by an echo-vowel or by the ordinary unfused subordinating particle g: HEAD=TAIL-ECHO -HEAD=TAIL + g And sometimes, the head and tail may reduplicate with each other (either ABAB or AABB), with nothing following at all (neither an echo-vowel nor unfused l): 60 HEAD=TAIL-ECHO -HEAD-TAIL-HEAD-TAIL N HEAD-HEAD-TAIL-TAIL. One thing we never find, however, is an echo-vowel eo-occurring in the same expression with an unfused l: *HEAD=TAIL-ECHO + g. 179

PAGE 192

This proves that the echo-voweZ is nothing more than a superfused reduction of the underZying particZe l itseZf. D. The EVAs listed according to their tail-vowel [IJ with u-u (1) qho?=otu-u 'hollowed, emaciated (as a face)'. / < gho? (V ) 1be concave, cupped, sunken, dented' ; also adJ gho? l (AEstat). E.g. gho?=tu-u ve 'have a sunken face'/ (2) ci=cu-u 'all sour' /< ci (V d.) 1be s,oUJr."''; also ci-cu-ci-cu/ -a J (3) sa?=qu-u 1all rough (as an unplaned board)' / 'This housepost isn't properly smooth yet --it's still all rough. 1 NB: The same head occurs in another EVA with a different tail and a slightly different meaning (see sa?=gg-e), IX.20 below).61/ CIIJ with o-o (4) 'great big; on a large scale, in a big way' /< (V d.) 'be big' and -16 (B ) 'big thing, something big' .62 -a J n E.g. te gQ 1if (you) make it so big' [adverbial useJ; E ve ;-phu? ea Se ve 'He has eaten up the biggest rice-cake' [adnominal useJ. NB: also occurs as the unitary head of (qv. n.59)./ (5) 'all scarred up' /< (M .f ) 'scar' (cf. -C3.Fd aboveJ); go? -p X --is perhaps an allofam of _g_ (V d.) 1be dried up and brittle a J (as fallen leaves) and ( B ) 1 Ciried thing' (as in '2>-g3;-gwe n (N) 'hide, dried animal-skin'). NB: The same head occurs in a synonymous EVA with a different tail (see IV-12 below)./ 180

PAGE 193

(6) cu=ko?-o 'all shrivelled up; dry and puckered' /< cu (V) 'be tightly closed, puckered up; close something tightly' as in t5-cu 1 sew up tight' m"E? cu ve 1 shut one 1 s eyes tight (OV); be blind (N -V). I E.g. a-phe? d spec o-pha? cu=koLo ve 1When there's blight on the chillipeppers the leaves get all shrivelled up.'/ (7) t8?=po?-o 'short and broad (of a person or thing); stumpy, squat' /t8? is not a free morpheme' but it recurs in o-t8?-n' (N) 'a short person, a shorty', and t8?-n (AEstat) 'short', a-p3-t8? (N) 'kind of short, stubby bananaJ; the compound t8?-po? is a true verb, and can be negated (ma t8?-po?). t8?-po?-o zQ 'He's a dumpy little guy. I guy. ; (8) pi?=cho?-o 'tasteless, insipid (of food); barren (of land)' /< p"E? (V) 'dissipate, lose its power; be tasteless, get stale (food); get flat (beer, soda); be infertile (land); be shallow (objects, water)'. E.g. chi p"E?=cho?_o 'This bread has gone stale.'/ CIIIJ with (9) 'all bruised' /< cho (V) 1be bruised (people, fruit)'; the compound cho-no is also a full verb. E.g. a-po chi 'These bananas are all bruised.'/ (lO) 1grayish (like clouds)' (V d.) 1be gray'; also ph;i, (B ) 'something a J n gray', phf (AEstat) 'gray(ly)', pM-pho? (AEstat)/ (ll) 'big and round; big and lumpy' /l"E is not a free noun or verb, but appears in l"E (AE stat) 'cylindrical' and st?-l"E-lo (N) 'hunk of wood'. NB: This EVA apparently occurs only in combination with another one of similar meaning, chu=pe?_e (VIII.l.7 below). lBl

PAGE 194

CIVJ with :i:-3: (12) 'all scarred up' /This EVA is evidently quite synonymous with (qv,) E.g. ;y_ ve mt?-phu face is all scarred up. 1 / [VJ with a-a (13) 'on a grand scale; so that it's big' (14) /The roughly synonymous morphemes (Vd.) 1be big' and -a J -16 (B ) 1 something big' here form a compound head for the -n EVA. This is a more complex variant of [qv. J, where -16 is in the tail. E.g. te .BQ ni sa _j]_ I If it Is made nice and big it'll look very fine.'/ CVIJ with a-a 'too thin, watery (of a liquid which should be thick, as honey, paint, soup)' /There is no independent morpheme che? with a meaning anything like this. E.g. chi che?=ga-a ma ne ve honey isn't watery--it's thick.'/. CVIIJ with i-i (15) 'light green; light blue64 /< (N; B ) 'green, blue; something green or blue'; n cf. also n5 e (AEstat) 'blue' green I I CVIIIJ with e-e (16) ge?=le-e 'all scraped up; scraped and abraded; red and raw (skin) ; mangy (of animal) /< ge? (V) 'get scraped, abraded; irritated, mangy-looking'; also occurs reduplicated as ge?-ge?-le-le/ (17) chu=peLe 'plump, chubby' /< chu (Vadj) 'be fat'; also chu-pe? le? (AEstat);6 5 chu-pe? is a free verb in its own right. E.g. chu=pe?-e te ve 1be chubby' [adverbial use]; chu=pe?-e ve 1 a chubby 182

PAGE 195

person' [adnominal useJ. NB: This EVA occurs as a constituent in a number of more complex and emphatic expressions: chu=pe?-e-dl-qU (N) 'big fat person; butterball'; chu=pe?-e g (AE h + AE t t) 'fat and round'; -ec o s a chu=p?-e (AE h + AE h ) 'rolling with fat; ec o ec o hulking and blubbery.'/ (18) pa=ne?-e 'very thin (as paper); sharp (of a blade)' d.) 1be thin (of people, objects); be sharp (of a J blades)'; also occurs unfused as (AEstat)/ CIXJ with E:-E: When the vowel of the tail-syllable is intrinsically /E:/ anYWay, it is sometimes hard to decide whether we are dealing with an EVA or simply an ordinary stative adverbial. If the echomora is mid-tone, it is safe to assume the expression is an EVA [#l9,20f.J Otherwise we have a problem [it21,22f.J. (19) qa=pE:-E: 'spread out, splayed; swooping (as with spread wings)' /< .9...: (V) 1be forked, branch out from [archaic] 1 (M f) .9.fu:. (B) 'something forked' C3.Fd aboveJ; p x n ult. related (V) 'go' < *.9...: + (Pv) (3.Eb above). The bound morpheme E_ .PE_ recurs in a few other compounds (e.g. sP-ga-pE: I fork in a tree; forked stick'' pE-ll-ka 'armpit'. E.g. pu ga=pE:-E: ve 'have one's thighs spread apart'; te kh:i:: lE: he-b1 qa=pE:-E: CO t3? la ve ;yQ 1All of a sudden a plane came swooping out of the sky over there. 1 / (20) 'rough and raspy (as the voice of an adolescent boy) 1 I< sa? (V adj) 11Je rough I ; this same head occurs in another EVA with a different tail and a slight semantic difference (cf. sa?=gu-u). E.g. ;yE_ ?>-kh3 sa?=gE-E: ve 'His voice is getting rough and raspy.'/ 183

PAGE 196

(21) h;=v6?-6(?) 'warped, twisted out of twistedly, spirally' I< h; (V) 'get bent (esp. of metal) 1 shape (as wood, metal); and v6? (M f ) 'a --p X screw'. E.g. so-ba chi h;=v6L6? 'This sheet of metal is warped'; v:i:-16 te khE qh::> h;=v6LE? to? la Ve 'A big snake came twisting out of the hollow tree.' NB: Since the last element is not under mid-tone, this expression might better be analysed as an ordinary AEstat: h;-v6? g(?). I (22) 'peacefully; quietly' I< (Vadj) 'be cool; be miserable, in trouble'; 6 (AE t t) 'cold; silent, quiet', ta.?-J: .. -s a 'in absolute silence'. E.g. chf; m'E 'Please stay there quietly. 1 NB: This is the most problematic of our EVAs. The last element is not under mid-tone. Furthermore, 16(?) can be an allomorph of the subordinating particle 6 itself, (qv. n.45) as illustrated by the variant pronunciations of E. Secondary vowel length and the Lahu syllable canon The echo-vowel adverbials are no doubt marginal to the Lahu phonological system, but they have nevertheless introduced the feature of contrastive vowel length into the language.61 T* Although our original simple syllable canon, (C.)V is still adequate to characterize the 'core-system' of Lahu l phonology, a much more elaborate schema is required to accommodate all the secondary or marginal features we have been discussing. Adding vowel-length 1:1 to our formula, we arrive at the following monstrous result: i T* y I I w (c. ) I (w) V ?** I l E*** (N) (* where T includes the checked tones '?I; ** this is imperativelintonational glottal stop, distinct from the checked tones; ***E occurs as second element of a complex nucleus only if l:i:l precedes). 184

PAGE 197

5. The bulging monosyllable: decay and rebirth There is no reason to feel sorry for the poor little monosyllables of languages like Lahu. Despite the phonological reduction they have undergone, these syllables teem with the seeds of new life. Among all the marginal features floating around these syllables, some will certainly catch on and eventually penetrate to the core of the system. The monosyllabic languages of East and South-East Asia show an uncanny homreostatic ability to regulate themselves in cyclic swings of expansion and contraction. What is absorbed and incorporated here will be diffused or extruded there .68 The accretional or augmentative tendencies do not of course stand in a simple one-to-one replacement relationship versus the tendencies toward reduction and attrition. Things are more indirect and slow-moving than that. Nonetheless, it is hard not to believe in some kind of overarching regulatory principle which eventually ensures that things will not go too far in any one direction. There is no harm in referring to this by some functional label like the 'economy of the syllable' .69 In a more cosmic vein, these phenomena furnish one more bit of reassuring evidence that the forces of creativity have nothing to fear from the forces of destruction. Symbols and abbreviations not explained in the text are as follows: belongs to the same word family as; as an allofam of AE adverbial expression AEstat stative adverbial morpheme B bound nominal morpheme n B V bound verbal morpheme cf final consonant c. initial consonant l c prefixal consonant p G glide M morpheme M pfx prefixable morpheme

PAGE 198

MSC Modern Standard Chinese n,N nasal(ization) N noun Nh noun-head; head-noun N specifying noun spec OV object-plus-verb P prefix pn noun-particle P verb-particle V PST Proto-Sino-Tibetan s suffixal-s SE subordinating expression SOV subject-object-verb ST Sino-Tibetan SVO subject-verb-object T tone V vowel; verb Vadj adjectival verb; adjective vh verb-head; head-verb VP verb phrase NOTES l. However, in most languages of the region (including Chinese) the pervasive process of compounding has ensured that a large percentage of the words in the lexicon are polysyllabic. A language may thus be simultaneously monomorphosyllabic but polylexosyllabic. 2. 'There is something about the tightly structured nature of the syllable in monosyllabic languages which favors the shift in contrastive function from one phonological feature 186

PAGE 199

of the syllable to another' (Matisoff l973b:78). Henderson (1975) has dubbed this phenomenon 'feature shuffling'. 3. See Haudricourt (l954a, l954b, 1961). This process has been called tonogenesis (Matisoff 1970, 1972, l973b). A good summary of recent work in this area is Mazaudon (1977). 4. Familiar examples include 'bonfire', 'daisy', 'hussy', 'window', 'nostril' (see Partridge 1978, entries). 5. The Lahu independent morpheme for 'bug' is .. The highrising tone in already represents an incipient fusional process since it occurs (albeit sporadically) as a sort of sandhi-tone in several other compounds, e.g. h 'field, swidden', but h-ga? 'wild chicken' ('field-chicken'). 6. The Dhimal voiceless nasal hm-must descend from *s-m-, while the high-rising tone of Lahu ma bespeaks a ProtoLoloish prefix*?-(< *s-), ultimately< *za or *tsa 'child, son'. (See Ben;dict 1972:#59; Matisoff 1970; and 1972:#153.). 7. For example, fully 36 pages of Hanson (1906/1954:242-78) contain words beginning with ke-; there are 25 pages of sewords (ibid., 631-56), 8. This causative prefix appears as before roots beginning with sibilants or aspirates, and as seotherwise. This formation descends from a sibilant causative prefix that must be set up for PTB itself (Benedict 1972:105). This is one of the rare cases where a morphological element with a well-defined meaning can be imputed to the proto-language already in 'reduced' prefixal form. At a still earlier time-depth, however, we may speculate that even this *.. prefix derived from an independent full syllable, maybe the prototype of the Old Chinese causative auxiliary verb 1l. MSC shr, reconstructed as (Karlgren 1957: GSR 9/5n). (See Maran l97l:l5lff., 1976; Matisoff 1916:431). 9. These designations are not quite synonymous. In current usage, 'Mon-Khmer1 comprises all branches of the AustroAsiatic family except for the Munda J..,anguages of India. 10. We owe this convenient and widely accepted term to Henderson (1952), who first applied it to Cambodian. For a brief discussion of minor syllables in Old Mon, see Shorto (1971: xv); for Khmer, Huffman (1972); for Northern Mon-Khmer (Palaung-Wa), Shorto (1960, 1963); for Khasi, Rabel (1961: 17-9); for Senoic, Diffloth (1973). 11. The phrase 'syllable-and-a-half' I first heard from the lips of Gordon Downer (LSA Summer Institute, 1967);

PAGE 200

the Latinized version 'sesquisyllabic' was introduced in Matisoff (l973b:84ff.). 12. This includes all the branches of TB except Karen, which under heavy Mon and/or Tai influence, has evolved into a prepositional SVO language. 13. This also happens with the high-frequency negative adverb ma, very often in colloquial style. Adverbials are a 'prepositional' class in Lahu, preceding the verbs they modify. 14. Cacophonous as this term may be, it is certainly better to be sesqui-moral than utterly immoral, and less equivocal than to be bi-moral. 15. The discussion in this section is based on Matisoff l973a: 10-38. 16. These 9 vowels are, however, compressed into a much higher and narrower range of phonetic space than in Thai. Thus the vowel written /g/ is like that of Eng. bed (not like bad, as in the Siamese vowel often transcribed with the same symbol). The mid-vowels /e a o/ are so high that they often vary with /i u/. 17. P =prefix (up to two 'prefix-slots' are posited even at the PTB stage); Ci = (root-)initial consonant; G =glide /w y r 1/; V = vowel; :' = vowel length; Cf =final consonant /-m -n -p -t -k -r -1 -s/; s = suffixal which could occur after root-final Cf's. It is still controversial what status to impute to tone at the PTB stage, though this language family seems always to have been 'tone-prone'. 18. In certain types of syllable with a glottalized initial, even this postvocalic [-?J has disappeared by 'glottal dissimilation', leaving a compensatory high-rising tone (Matisoff 1970; 1972). 19. One reason out of many is that glottal stop disappears in Lahu singing, as do all other tonal features (pitch, contour). Additional support for this suprasegmentalist approach is provided by the phonetic behaviour of the echovowel adverbials,4.C below). For a clear account of glottalization in the context of 'phonation types' in general, see Egerod (1971). 20. This contrast has been explained in terms of the influence of the voicing or voicelessness of various elements in the syllable-initial (see Matisoff 1972, passim.). 188

PAGE 201

21. We use the same diacritics in our digraphs for the stopped tones as are used for two of the open tones, This is entirely a matter of orthographic parsimony, since there is no historical or synchronic connection whatever between r I and r? I' or between /'I and ('?I. We regard the open and checked tones as constituting quite separate subsystems in Lahu phonology. 22. Note that the functions here as the PLB *Ci, not as a *G. For 1pig1 the proto-rhyme is *::.ak and the Lahu reflex is -a?. Contrasting to this are syllables like *twak 'emerge', where the proto-rhyme is *-wak, i.e. where the is functioning as a feature more closely associated with the nuclear vowel than with the Ci. Here the vowel quality is changed, and the Lahu reflex is -si!_. 23. The after the completion of most of the verb's tonal contour, so that there is usually no question of confusing these imperatives with other verbs having 1 intrinsic 1 checked tones (Matisoff 1973a: 352-3). 24. These are written with orthographic convenience, though we conceive of the nasality as a suprasegmental or coarticulational feature. 25. As these last examples show, this blind phonetic process operates even in syllables under stopped tones, so that the same vowel can be nasalized and glottalized simultaneously. This is another bit of evidence for the suprasegmentality of the feature of glottalization. Similarly, Burmese 'creaky' tone may occur on syllables with nasalized vowels (e.g. ?ekhwinl 'permission'). (See Matisoff 1975.) Most Lahu syllables that begin with nasal Ci 1 s do not show pronounced nasalization of the following vowel, with the notable exception of /mu/, where the vowel is so nasalized that it is almost completely swallowed up by the initial: [mYJ. (Lahu labials /p ph b m/ are affricated before /u/ to [pf phf b v mYJ .J Parallel developments have occurred in other Loloish languages like Akha. Thus, PLB *s-muwl 'mushroom' > WB hmui, Lahu mu cmvl:l.J, Akha hm. ---26. Phonetically, the second element is a semivowel intermediate in height 27. /w/ never occurs before the back vowels /u o or the central vowels e/. 28. PLB initial consonantal *wdeveloped regularly into Lahu I: as in PLB *wak 1pig1 > Lh. va?, PLB *wa2 'bamboo' > Lh. va, etc.

PAGE 202

29. This is not to say that there is no difference whatsoever between the two variants. Sometimes the prelabialized form seems to convey a stylistic nuance of familiarity, a more colloquial or folksy tone thanthe plain variant. Strictly speaking, however, this is not a 'morphemic' difference. It is somewhat similar to the 'dropping of the in the present participles of verbs in certain varieties of American English (e.g. singin1 [Si!JenJ instead of singing [Sl!Jl!JJ). 30. To keep things simple, certain constraints have not been built into the formula: (a) No Lahu syllable has yet been encountered with both a prevocalic glide and a postvocalic -L or (Such syllables do exist in Tai, e.g. Siamese duaj 'together', diaw 'single'); (b) A verb that already is intrinsically under a checked tone undergoes no change in the imperative. 31. Subsequent research has thoroughly borne out this hypothesis. A general TB root *ka now be set up with the meaning 'go' deriving ultimately from the notion of or spreading the legs (cf. Benedict 1972:#469, and below 4.D (19)). 32. It also appears without the suffix, either as or with the variant under very-low tone in the compound na?-me 'gunpowder.' 33. A 'Mpfx' or 'prefixable morpheme' is a root which occurs either in 'general' form with the or in 'specified' form modified by another noun (Matisoff l973a: 3.34). The general prefixed form means 'child, baby'. In specified form it serves as a productive diminutivizer 'puppy', 'small intestine', y:-g 'little house'). 34. For alternations between /A?/ which are 'mechanical' in nature, see Matisoff (l973a:l.63, p.28). It is the open-toned variant that occurs in the compound mE?-ge-lwE 1 firefly. 1 35. This term is used by Lahu in Burma. Lahu resident in Thailand now tend to say mE?-kew for 'eyeglasses'. The same semantic association between 'shadow' and 'mirror' is displayed by the Japanese root kage 'shadow' and its derivative kagami 'mirror'. (I owe this observation to Sus an Matisoff. ) 36. If anything is more like a r1s1ng diphthong, since C1J is such a high vowel that it is quasi-consonantal (almost a semivowel), with less acoustic energy than [gJ, This nucleus is somewhat comparable to the three centralizing diphthongs of Siamese, sometimes written 190

PAGE 203

/ia ua/ and sometimes /ia ua/. As the latter transcription suggests, the second element is less prominent than the first, but it is still a true vowel, not a semi vowel. Rather than calling these 'falling diphthongs', the term 'co-valent' seems appropriate here also. 37. The meaning of the fused element in 'measles' is obscure! For 'drizzle' and 'whisper', it is undoubtedly the adverbializing particle i (4.B below) which has become amalgamated with the preceding syllable in underlying stative adverbial constructions (ibid., esp. n.49) of the form 38. For complex nuclei which show free variation between 'totally fused' /-y/ and 1semifused1 /-E/ vocalism, see the discussion of the minimized extentives, below 4.A. 39. The word la?-no-c3; (lit. 'finger-joint') means 'knuckle, phalanx' (in the sense of a finger-section from joint to joint). The word for 'little finger' (la?-no-c3;la?-no-c) seems thus to be derived from a fuller form *la?-no-c3;[< la?-no 'finger' and c3;'small joint'J, i.e. 1 small-knuckled finger; the finger with the smallest phalanges 1 40. This is quite analogous to a phenomenon at the other end of the syllable that I have called prefix preemption, whereby a prefix comes to drive out the original root-initial consonant. (See, e.g., Matisoff 1979.) 41. For such syllables as h3; and see the discussion of minimized extentives (4.A below). 42. Lahu checked syllables take less time to utter than open ones. A number of Lahu syllables under high-rising or very-low tones allegro variants under highchecked or low-checked respectively. See Matisoff Q973a:#l.63l, p.28), and the discussion of mt?-ge? -mt?-ge (3.Fc above). There is evidence that this particle E derives historically from a syllable with lateral initial /lE/ which still survives as an alternant of E in certailin collocations. As noted above (l.B), sev;ral functors optionally drop their Ci in Black Lahu, including such essential items as ma a (Adv) 'negative', tha? a? (P ) -----n 1 accusative 1 and ta a (P ) 1perfecti ve. 1 V 43. E.g., chi g ve cho 'such a small person'. See below 4.A and 4.B (esp. Table 2) and Matisoff (l973a:l30). 44. See Matisoff l973a:#l.42f, 3.62. These words are called 'diminutive extentives1 throughout that work, but we adopt 191

PAGE 204

the label 'minimized' here to avoid confusion with the distinct morpheme g (M f ) that we are calling 1 diminutive'. -px 45. My most reliable informant (1977) insisted that he felt a slight drop in pitch on the second mora, and suggested it be written with the symbol for high-falling 46. The initial consonant is here affricated, the fused vowel [e;J is nasalized to Cln, and the original nuclear vowel, unrounded to [tLJ in this environment, is deprived of its syllabici ty. 4 7. The high-rising tone acq.uired by the extenti ves in their minimized form is similar to what we find in a number of stative adverbials (4.B below), where a mid-tone base form becomes high-rising tone before the subordinating particle phu 'white' > phu 'whitely', chu 'fat' > chu 'fatly', etc. See Matisoff (l973a:#l.641, p.30). 48. The adverb qha means 1all1 ; Vh =verb-head, Nh =noun-head. Examples of fused nuclei in gha-adverbials include: gha 'completely' < gha pa < 'to finish', qha 'until it is reached, up to the point that' ba E: 1 clearly, brightly' J. Sometimes it may be more nounlike Csi (N) 1gold1 > s{ E: 'yellow, gold-coloured']. In the morpheme-only occurs in the adverbial construction, so it is hard to tell what its intrinsic form-class is. To cover all these contingencies we are using the non-committal symbol 'M' (for 'morpheme'). 50. The verb following a stative adverbial is usually one of a handful of highly abstract i terns like 1 go 1 te 1 do 1 phE:? 1be', la 'come 1 The chief semantic burden of the VP is borne by the adverbial, with the verb merely providing a cushion for its adverbiality to rest upon. The same is true for most of the other types of SEs in their adverbial function. (The exception is gha-adverbials, which occur freely before any semantically appropriate verb.) 192

PAGE 205

51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. Especially reduplicated adjectival verbs (V .). adJ B bound verbal morpheme. We conventionally connect tKis intensifier to its adjective by a hyphen. These are four-syllable constructions at least two of which are verbs, such that the first and third, or the second and fourth elements,are identical. There will probably be little danger of confusing this term with the similar acronym used by astronauts for 1 extravehicular activity'! Larger, for example, than the number of adjectiveintensifiers so far discovered. As we shall see, the EVAs resemble intensified adjectives more than any other type of subordinate expression. The auditory impression is rather similar to that of the Vietnamese nga tone, though the Lahu syllables have less 'creakiness' than the Vietnamese ones. Exceptions are ga=p-, where the root-part of the syllable is already mid-tone; and a few items like ga=lE?-E, h3=vE?-E, where the contour ends as a low-falling tone/'/. With these last two there is some question as to whether they are really EVAs C4.D (21-22) below]. These a.re very similar in status to such lexically specific English intensifiers as jet (black), scot (free), stock (still), luke (warm), etc. In our corpus,the exception is i-lo=me-a 'on a grand scale' (3-and-a-half syllables). But here i-16 is functioning as a tight lexical unit. See 4.D (13) below. Since reduplication by itself is already a widespread morphological technique to achieve adverbialized status in Lahu (reduplicated verbs are a kind of 'SE': see Table 2), the reduplication obviates the necessity for any subordinating particle. We may symbolize the cases where a given head may have multiple tails as: HEAD=TATL1-ECHO HEAD=TAIL2 -ECHO. This again is similar to what goes on with ordinary intensified adjectives, where the same Vaa may occasionally take more than one intensifier, with some differentiation. Tbus.h 'be hard' may be intensified as h-ku 'stiff (as cramped muscles or an erect penis)' or h-taw? 'hard and chewy (as sugarcane)'; 'hard but resilient (as a pig's sternum)'. 193

PAGE 206

62. This morpheme is from Shan (cf. Thai lua!J), but is thoroughly integrated into Lahu, fU..11.ctioning as the antonym of g (M f ) 'something small'. px 63. In this example, the EVA occurs 1 independently', wi t.h no following verb. Since Labu s.dverbials tend to be the semantic centre of interest of their clauses (the following verb is usually an abstract dummy), it frequently happens that the verb is omitted, so that the adverbial becomes more verblike. (See Matisoff 1973a:#4.421I, 4.422(3), 4.424, 6.47.) 64. The last syllable of this EVA was correctly understood to be a fusional variant of the subordinating particle e in MatisoffQ973a:l7,#1.42e). However, the identification of vi with the morpheme vi (Mpfx) 'something sharp' (as in a-tho-vi 1knife-blade I, pg-vi I stinger Of bee I) WaS quite wrong. Contra Matisoff, op.cit., there is no AEstat of the shape vi [vi-iJ meaning 1 sharp 1 and vi-i, in fact, occurs only as the tail of no in this particular EVA. 65. The le? is to be regarded as the fuller (presumably original) form of the subordinator (qv. n.42). 66. I analysed this lg(?) as a variant of the subordinator e which came to as part of the root-morpheme of the adverbial, so that another /e/ may directly follow it' in Matisoff(l973a:565, n.l44.) 67. We should perhaps say 'reintroduced', since vowel length is set up as a feature of PTB (Benedict 1972:70f.), though only in syllables with final consonants. 68. There are even excellent examples of new syllable-final oral stops developing from semivowels. In Maru (Burmish group) PLB *-uw and *-ix_ have become -uk and -it, respectively. (See, e.g. Burling 1967:59-61). This is similar to a development that has been traced for Archaic Chinese millennia ago, where *-and been plausibly derived from and *-z (Benedict 1948). 69. I do not believe that A. Martinet (Economie des cha:ngements phonetiques: traite de phonologie diachronique. Berne; 1955) actually uses the term L'economie de La syllabe, though I doubt he would object to it. 194

PAGE 207

Benedict, P.K. Burling, R. Diffloth, G. Egerod, S. Hanson, 0. Haudricourt, A.G. REFERENCES 1948. Archaic Chinese *. Harvard J. As. Stud. 11 197-206. 1972. Sino-Tibetan: a conspectus (Princeton-Cambridge Studies in Chinese Ling. II). Cambridge: Cambridge Univ. Press. 1967. Proto-Lolo (Indiana Publ. Anthrop. & Ling. 43). Bloomington: Indiana Univ. (=The Hague: Mouton). 1973. Minor syllable vocalism in Senoic languages. In Austroasiatic Studies (Pacific Linguistics Spec. Publ. 13) (eds.) P.N. Jenner et al. Honolulu: Univ. Hawaii Press. 1970. Distinctive features and phonological reconstructions. J. Amer. Orient. Soc. 90(1), 67-73. 1971. Phonation types in Chinese and South East Asian languages. Acta Ling. 13(2), 159-72. 1906/1954. A dictionary of the Kachin language. Rangoon: Baptist Board Publ. (Reprinted 1954). 1954a. De l'origine des tons en viet namien. J. Asiat. 242, 68-82. 1954b. Comment reconstruire le chinois archaique. Word 10(2-3), 351-64. 1961. Bipartition et tripartition des systemes de tons dans quelques langues d'Extreme-Orient. Bull. Soc. Ling. Paris 56, 163-80. Henderson, Eugenie J .A. 1952. The main features of Cambodian pronunciation. Bull. Sch. Orient. Afr. Stud. 14, 149-74. 195

PAGE 208

Henderson, Eugenie J.A. 1985. Feature shuffling in Southeast Asian languages. In Southeast Asian linguistic studies presented to Andre-G. Haudriaourt (eds.) Suriya Ratanakul, Huffinan, F.E. Karlgren, R. Mar an, LaRaw & Clifton, J.M. Matisoff, James A. D. Thomas & Suwilai Premsrirat. Bangkok: Mahidol Univ., 1-22. (Paper presented at Eighth Internat. Conf. on Sino Tibetan langs. & ling., Berkeley, 1975). 1972. The boundary between the monosyllable and the disyllable in Cambodian. Lingua 29(1), 54-66. 1957. Grammata seriaa reaensa (GSR) Bull. Mus. Far East. Antiq. 29. 1971. Burmese and Jingpho: a study of tonal linguistic processes (Occ papers Wolfendon Soc. on Tibeto-Burman Ling. 4). Urbana: Univ. Illinois. 1976. The causative mechanism in Jinghpaw. In The grammar of causative constructions (ed.) M. Shibatani. New York: Academic Press, 443-58. 1970. Glottal dissimilation and the Lahu high-rising tone: a tonogenetic case-study. J. Amer. Orient. Soc. 90(1), 13-44. 1972. The Loloish tonal split revisited (Research monograph 7, Center for South and Southeast Asia Studies). Berkeley: Univ. Calif. Press. 1973a. The grammar of Lahu (Univ. Calif. Publ. in Ling. 75). Berkeley and Los Angeles: Univ. Calif. Press. 1973b. Tonogenesis in Southeast Asia. In Consonant types and tone (Southern Calif. Occ. papers in Ling. 1) (ed. L.M. Hyman. Los Angeles: Univ. Calif. Press, 71-96. 1975. Rhinoglottophilia: the mysterious connection between nasality and glottality. In Nasalfest: Papers from a symposium on nasals and nasalization (Stanford Univ. Universals Project) (eds.) C Ferguson et al. Stanford: Stanford Univ. Press, 265-87. 196

PAGE 209

Matisoff, James A. Mazaudon, Martine Partridge, E. Rabel, Lili Shorto, H.L. 1976. Lahu causative constructions: case hierarchies and the morphology/ syntax cycle in a Tibeto-Burman perspective. In The grammar of causative constructions (ed.) M. Shibatani. New York: Academic Press, 413-42. 1978. Variational semantics in Tibeto Burman: the 'organic' approach to linguistic comparison. Philadelphia: ISHI Publ. 1979. Problems and progress in LoloBurmese: Quo Vadimus? Ling. Tibeto Burman Area 4(2), ll-43. 1977. Tibeto-Burman tonogenetics. Ling. Tibeto-Burman Area 3(2), l-123. 1978. Origins: a short etymological dictionary of Modern English. London: Book Club Assoc. 1961. Khasi: a language of Assam (Louisiana State Univ. Studies, Humanities ser. 10). Baton Rouge: Louisiana State Univ. Press. 1960. Word and syllable patterns in Palaung. BuZZ. Sch. Orient. Afr. Stud. 23(3), 544-57. 1963. The structural patterns of northern Mon-Khmer languages. In Linguistic comparison in South East Asia and the Pacific (ed.) H.L. Shorto. London: Sch. Orient. Afr. Stud., 45-61. 1971. A dictionary of the Man inscriptions from the sixth to the sixteenth centuries. London: Oxford Univ. Press. 197

PAGE 211

THE YAW DIALECT OF BURMESE John Okell 1. Introduction It was no doubt his admitted lack of information which made Forbes (1881:56) think that the Yaw dialect of Burmese was 'certainly unintelligible to any Burman. In fact, as Houghton (1897: 456) and Taylor (1921:91) observe, Yaw had few differences from Standard Burmese. A British Settlement Officer reported that, 'at first, the dialect is difficult to understand, but after a few days one finds oneself speaking it and it presents no difficulty' (Abigail 1932:6). Few and unspectacular though the differences are, they are important, as they place Yaw (YW*) several steps closer than Standard Burmese (SB*) --or any of the recognised dialects to Written Burmese (WB*), and this feature makes it a valuable ingredient in comparative studies. Despite its importance in this respect, Yaw has not been described in any detail till relatively recently (Kya Htlin 1969:; Yabu Shiro 1980). Ono (1969) allows just over a page to Yaw, and earlier studies went no further than noting a few forms in comparative lists (Buchanan 1799; Houghton 1897; Grierson 1928; Taylor 1921). A description in English, and a comparison with Standard Burmese and Written Burmese has been lacking. I was fortunate in being able to record two texts of spoken Yaw in Burma, and to have the assistance of the speakers in transcribing them. These texts, supplemented by my informants' answers to queries, provide the material for the outline description presented here.l 1.1 Features of particular interest For many, the most spectacular feature of Yaw is its rhymes /ak/ and corresponding to Standard Burme9e /E?/ and /in/. (For systems of transliteration and transcription see Okell 1971.) Yaw not only reflects more closely the ak and an of Written Burmese, but also --with occasional lapses --has velar closure for both. It also has velar closure in its YW /auk/, WE ok, SE /au?/, though not, curiously, in its YW /aun/, These are the abbreviations used in tables and examples. (Ed.) 199

PAGE 212

WB on, SB /aun/. For the array of eight final consonants in Written Burmese, none of the other major dialects now has anything more than glottal stop (for Written Burmese obstruents) or nasalised vowel (for Written Burmese nasals). In these circumstances, Yaw's velar closure is a remarkable relic. Yaw has also not yet allowed initial, or medial /w/ to alter the quality of the following vowel. For WB at an wat wan and am wap wam where SB has /a? an wu? wun/ Yaw stoutly preserves /8? 8n WE:? wE:n/ One other point worth mentioning here is the consistency of the Yaw rhyme /E:/ (rarely /e/) for Written Burmese Like the reflexes of this rhyme in other dialects, Yaw's consistency shows how eclectic Standard Burmese has been with its /i/, /e/ and /E:/ realizations. 1.2 Location and numbers of the Yaw My informants said the main town centres of the Yaw are at Yaw itself, and at Ht1-11n and Gan-gaw, which places them at the head of the Myit-tha valley, with the Chin Hills on the west, and the Pon-daung and Pon-nya ranges on the east --a geographical setting that might be expected to isolate them somewhat. This location is confirmed by Kya Htiln ( 1969:142) and by the Swe-zon kyan (1970:2). The Linguistic survey of Burma (LSB, Webb 1917), however, records the majority on the plains side of the watershed, 'between Saw and Seikpyu', with a few outliers further north on the western edge of the plain, reaching as far as Kani on the Ch1n-dw1n river. It is difficult to reconcile this discrepancy without a further survey. Some of the more obvious possibilities are that respondents to the LSB questionnaire in the valley, which the survey did cover, did not fully understand what was being asked; or my informants being valley men, may have been unaware of the numbers of Yaw on the plains side; or there may have been appreciable population movements in the sixty years since the survey itself. There is also mention in the literature of a group of Yaw who fled to the upper Mu valley in the Katha District (Harvey 1925:262) and crf two Yaw villages way up near Myit-ky1-na (Webb 1917:33). The present number of Yaw speakers is unknown. The LSB (Webb 1917:55) recorded over 24,000. This stands in marked contrast to the Census (1933) figures, but these can hardly claim serious consideration anyway, in view of their incredible fluctuations: in the five decades from 1891 to 1931 they give 200

PAGE 213

370, 5, 0, 2 and 877 respectively! The low response to the Census (1933) is presumably due partly to uncertainty over the criteria that qualifY one as a Yaw, and partly to a reluctance to identifY oneself as a Yaw anyway (Scott 1900:569; Hardiman 1912:29). 1.3 Background The antecedents of the Yaw are obscure, and have attracted some divergent speculations: Dr Mason classes the Yaw as a Burmese tribe. In this he is followed by Dr Cushing. Mr Houghton is inclined to doubt the accuracy of this classification. The Shan chronicles of Mong Kawng (Mogaung) seem to claim them as Shans, though perhaps they may be the Nora spoken of as earlier owners of the land. They themselves have a legend that they are descended from a clan of the Palaungs called Parawga or Payawga. This in time was shortened through Yawga to Yaw. There are still to be found Parawga sayas among them, oracles or mages, who make their divinations on the Tai cycle tables, which is significant. The common folk say that the reason why their dialect differs from Burmese is that they drink the water of the mountain streams .... The dialect is a hybrid, nearest to Burmese now; possibly it was at one time nearer to Shan or to some of the Chin dialects (Scott 1900:569) Other writers choose one or other of the alternatives offered here: primitive Burmans (Forbes 1881:56), from the Irrawaddy valley (Houghton 1897:456), captive Shans sent from Mogaung (Owens 1913:16), Burmese-speaking Chins (Taylor 1921:91), or Chin-tainted Burmans (Saw Shwe Boh 1973:18) --all indicative of a notable absence of hard facts. Equally fanciful is the identification with Tavoyan (Symes 1800:2, 235), and the attempt to derive Yaw from Standard Burmese /Yo-naka/ 'Shan' (Saw Shwe Boh 1973:16). The only real evidence one has for the origins of the Yaw is the dialect, and the closeness of this to Standard Burmese and Written Burmese indicates fairly strongly that as Kya Htlin (1969:141) suggests, the Yaw are nothing more exotic than a group of plains Burmans cut off from the mainstream of Standard Burmese development comparatively recently by a degree of geographical isolation. The 1Payawga' derivation sounds like a folk etymology, based on the fact that the Yaw have a reputation for skill in the magical arts, one form of which --by no means a Yaw monopoly --is called in Standard Burmese, from Pali payoga. Among the many varieties of Palaung listed in the LSE, none has a name resembling this word. Again, the idea that the 201

PAGE 214

Yaw once spoke some more distant or unrelated language (Chin, Palaung, Shan) and subsequently adopted Burmese is an implausible explanation of the peculiarities of their dialect. Such differences as there are between Yaw and Standard Burmese in phonology, grammar and lexicon are not of a kind that can be considered vestiges of an unrelated language. Nor are alleged resemblances to Tavoyan persuasive. In better documented times, there are records of a kind of Yaw autonomy. There was at one stage a Yaw-le1-my6-wun, with jurisdiction over the four towns of Pauk, Ht1-11n, Saw and LaU:ng-shei, and each of these towns was governed by a sa:w-bwa (shades of the alleged Shan connection) The Saw-bwa were replaced by after a rebellion in the reign of Naung daw-gy1 (1763-65) (Owens 1913:15f.), and at the turn of the century the Yaw were still described as 'governed by chiefs of their own, but tributary to the Burmans1 (Buchanan 1799:224) perhaps not wholly inconsistent with Symes1 (1800:1, 235) information that 1 the Yoos are subjects of the Birman state, and observe the same religious worship'. Their relative remoteness from central government at this stage is perhaps indicated by the attitude of the clerk who told Symes that they were 1 exceedingly ugly, having protuberant bellies and white teeth 1 My informants were, in fact, quite good-looking. 1.4 Source of material I did not go to the Yaw area, but made some recordings in Rangoon of unprepared speech by some students from Gan-gaw who had arrived in Rangoon for the first time only a few days previously. One recording describes some distinctive features of the Yaw area and its people, and the second, by a different speaker, is a folktale. 202

PAGE 215

2. Outline }2honolog;y: 2.1 Phoneme inventory 2.1.1 Tones: low /+) high plain !+! high creaky high stop /+?/ or /+k/ weak /e/ 2.1.2 Rhymes: open syllables weak: e full: i e a :J 0 u closed syllables nasal: ain ein ED alJ in oun aun stop: ai? ei? ? ak i? ou? auk 2.1.3 Heads: with with med /y/ med /w/ all but g d b j z by w k t p c s py ? (=/w-/ kh th ph eh sh e phy h (=/hw-/) lJ n lllJl my h!J hn hm hJl hmy ? 1 w y r (ly) h hl hw .r (hly) 2.2 Phoneme descri12tion As in Standard Burmese, except for: /En, ?/: vowel lower than in SB /?/ /a!J, ak, auk/: final velar consonants --the stop not released --clearly audible in slow speech, but sometimes realized by nasalization and glottal stop respectively. The vowel in /a!J, ak/ is more open and back than in SB /an, a?/, and close to the SB /a/ of open syllables. Given /auk/ one would expect /au!J/ to match, but it is not attested in my material. /(ly, hly)/: rare variants for /y, J/, used in formal styles. 203

PAGE 216

2.3 Comparison with Written Burmese 2.3.1 Tones: as for Standard Burmese 2.3.2 Rhymes: open WB i ai .2. ui u YW /i e E: a ::> 0 u a/ closed nasal \ WB in im afifi an an am ari ori uiri un wit ;V--L--YB eln in E:n a!J aun ain oun/ closed stop WB V ac at ak ok uik uv. YW /ei? i? E:lrak auk ai? ou?/ 2.3.3 Heads: as for Standard Burmese 3. Notes 3.1 Phonemes 3.1.1 WB afifi corresponds regularly to Yaw /E:/, but not also to /i/ or /e/-;; in SB, e.g. WB YW SB gloss lhafii:i: hlg hlg 'cart' safiii 8E: Oi, di 1this1 khyaiii:i chE: chi 'cotton (thread) 1 mrafiii: myg ... 'taste' myl krafiii. CE 'look' Cl tarit-mrak-cafifi: tabyak-sg tabyE:?_sJ: 'broom' tacafifi: tazg tazl: 'a bundle' krafifi CE: ci 'be clear' nafifi: ng ... 'way' nl caiifi SE: si 1be crowded' prafifi pyE: pyi 'Ni (measure) 1 praiifi. py pye 'be full' 204

PAGE 217

I did encounter a few words with Yaw /e/: :2accai'ii'i: pyi ?_se pyi?_Sl 'thing' man-k;yai'ii'i: meje mej1 'tamarind' cai'ii'i:-we: se-we s1-we 'meet' tuin:-:2rai'in ta1n-pye ta1n-pye 'country' There seems to be no environment feature corresponding to the use of Yaw /e/ rather than fg/; in the list above the heads .E!:_-, .s:_-, and k;y/kh;yoccur with both rhymes. 3.1.2 Initial and medial w do not affect the pronunciation of the rhymes WB an, am and as they do in SB, e.g. WB wam: kam: lwat tat YW wEn kEn SB g"loss .... wun 1 stomach' kan 'bank' lu? 'be free' ta? 'know' 3.1.3 Yaw has only /8/ where Standard Burmese has /8/ and /'0/' e. g. swi.t: sori.: 8oUn.-8ai:in 8oUn.-oaUn. 'thirty thousand' 3.1.4 The continuant series has some examples of aspiration in words not aspirated in Standard Burmese, e.g. m;ya: hmya my a 'be many' 'anarh ehngn enan 'side' ri.ari. h!Jin !Jin 'draw' lwam: hlwEn lUn 'too much' 3.1.5 Loanwords from Standard Burmese occasionally cut across the general pattern, e.g. YW & SB /pyi-ou pyi?-s1 for *YW /pyg-8u pyi?-se 'People's Trade Corporation' 3.2 Mor:2ho:2honemics 3.2.1 Voicing occurs, as in Arakanese, only with plain initials, not with both plain and aspirate as in Standard Burmese, nor with /8/, e.g. 205

PAGE 218

WB YW SB gloss ne-t a-ka ne-da-ga ne-da-ga 'staying' ta-khu-khu takhu-khu takhu-gu 'something' tha:ta;y tha-dE tha-dE 1 (they) place' ;yok;ya : thak yauk-ca-thak yau?-ca-dE:? 'more than men' lTI.-pyui lu-byo lu-byo 'bachelor' twe.Ehu: twe-phu twe-bu 'ever see' lu-kri: lu-j1 lu-j1 'elder' Ero-kh;yan pya-jin 'want to say' hnahnEn-zE? hnanan-za? 'sewn together' 'aEo-churil apa-sholin apa-zolin 'most abundant' hail: si: haJJ-e1 h1n-o1 'vegetable' This feature reveals some aspirates that one would not expect from Written Burmese and Standard Burmese: laiiii-Eaii: lapha!J lE-b1n 'neck' wam:Euik wE:n-phai? wlin-bai? 'stomach' ran-Eat ya!J-pha? yin-ba? 'chest' Possibly Written Burmese spelling has been altered from a more etymologically correct form. Some words in Standard Burmese have voiced initials even when they are not in close juncture. Their counterparts in Yaw often have voiced initials in the same way, but a number have aspirate initials, e.g. Voiced initial in Ya:w: WB YW SB gloss gwE:n ... 'cotton' gun gui: go go 'goal' khwa .. gwa 'fork' gwa dhat dE? da? 'relic' byuiri: bya1n bya1n 'paddy bird, egret' khye: .... ... 1dirt1 Je Jl jhe: ... ze 'market' ze chak-krui: zak-co ZE:?-co 'reins' 206

PAGE 219

Aspirate initial in Yaw: khon: dha: bha, bhay bhon:bhi bhu-ran bu: bhi: khyit khrui khyok jhi:si: khalin galin tha da pha, phs ba, bs phalin-ph1 batin-bi (sic) beyin phu bu phl: bl: chei? jei? cho jo chau? jau? shl:-el: zl:-ol: 'head' 'knife' 'what? which? 1 'trousers' 'king' 'gourd' 'comb' 'hook' 'horn' 'chasm' 'wild plum' There is hardly enough evidence here to sort out the various factors involved. Some of the Yaw voiced initials can be accounted for by being loans, e.g. /go/ from English goal /ds?/ from Pali dhatu others may have been voiced in Standard Burmese by the formative voicing and then borrowed in this form into Yaw (e.g. /gwa/) -assuming, on the evidence of words such as Yaw /chei?, chau?/, that the voicing formative does not operate in Yaw. Others again have probably been voiced in Standard Burmese by close juncture and later lost the first syllable but retained the voicing (cf. e.g., the older forms, Written Burmese: u:khon:, u:khrui). In Yaw the first syllable may have been dropped in the same way, but no VOlClng remains as aspirate initials would not have been voiced in this position. Forms like Yaw /tha, phu, phs, shl:/ are the most interesting. They could well be survivals of a hypothetical period when Burmese had no voiced obstruents. By this theory, voicing has supervened in Standard Burmese while Yaw holds out against this development. Such words also suggest an explanation for the otherwise bizarre spellings of Written Burmese. 3.2.2 Weakening and induced creaky tone occur much as in Standard Burmese. (See the texts for examples.) 3.3 Grammar 3.3.1 Particles seem to match those of Standard Burmese closely. Two with not quite regular reflexes occur in the recordings. 207

PAGE 220

In Written Burmese and Standard Burmese both have different forms for colloquial and formal styles. WB WB YW SB SB gl-oss coll. form. coll. form. nai. nhari. na nE hn1n 'with' la: lo 1?> 1a 1?> ( interrog. ) Yaw /na/ corresponds to Arakanese, Tavoyan, and In-tha /na/; Standard Burmese is the odd man out here. Yaw preserves the older form /lo/ which is not used in colloquial Standard Burmese and even in formal Standard Burmese is being ousted by the colloquial /la/. 3.3.2 Some selectives show different forms from Standard Burmese: WB YW SB gl-oss sanii 9E: oi, di 'this' bhay, bha phE:, ph a bE:, ba 'which?, what? 1 da tha da 'that (thing) 1 sail. 9a!J 9in 'that (year) 1 Of these, the first two show regular correspondences, with Yaw voicelessness for Standard Burmese voicing. The last item is rare in Standard Burmese; it is not used in colloquial, but occurs in older literature, and then only with the noun nhac 1year1 Yaw preserves it in speech. 3.4 Vocabulary There were a few i terns that differed from Standard Burmese. 3.4.1 Forms which appear not to have cognates in Standard Burmese: YW Je-da!J kezauk-kazak SB poun-byin omya-j:t 208 gl-oss 'story' 'much, a lot1

PAGE 221

3.4.2 Forms with different meaning or use in Standard Burmese: YW SB gZoss eln-hmu khan 1 greet, welcome' cf. .. hmu 'visit, do' e:, loun-gwalJ loun-ji 'lon-gyi1 cf. kwl.n 'circle' si-shalJ si-zin 'arrange' cf. shin 'set up, set out' Other vocabulary items are listed in Kya Htlin (1969: 152f.) and Yabu Shiro (1980 :l69f.) 4. Recordings The following two texts are extracts from the recordings mentioned in the introduction, the lower line showing the Standard Burmese equivalents. In the transcription of the Yaw version, the rhymes /auk, ak, alJ/ are written /au?, a?, an/ where this seems closer to the actual pronunciation on the recording. Irregularities such as the occasional voicing of aspirate initials, presumably the result of Standard Burmese influence, are marked with a (B) and shown as they are spoken. False starts, where the speaker corrects himself, are enclosed in square brackets. Sentences are numbered to facilitate reference to the translation. 4.1 Texts 4.1.1 Yaw customs: courting and hospitality (1) (l) (B) Gu py6-chaJJ-da-ga Jl:-d hl-d dele-le phyi?-pa-de: Jl:-d lu-byo hl-d dele-le phyi?-pa-de: ( 2) Lu-byo hl-d del.e-ga phe:-lo-lE: sho-d5 cen::>-do (2) Lu-byo hl-d dele-ga be:-lo-lE: sho-d5 cen::>-do Jl:-d ka-la-ea-de ka-la-ea sho-da-ga Jl:-d ka-18.-oa-de ka-la-oa sho-da-ga [eywe: yauk-t echein 0 -J eywe: yauLt ache in eeak [eywe: yauLt echein 0 -J eywe: yauLt echein aee:? Jl:-d oJ Jl:-d oJ she:-j::> she:-j::> hneshe: pe:LwE:n-jaJJ-ma Jl:-d lu-byo-de-ha hneshe: pa?-wlin-jin-ma Jl:-d lu-byo-de-ha ka-la-ea-de-ha ka-la-oa-de-ha epyo le:-d ele Jl:-ja-de:. epyo le:-d ale Jl:-ja-de:. 209

PAGE 222

(3) Cepyo 1-d e1e Jf-da-ga g -J (3) Cepyo 1-d e1e Jf-da-ga g -J epyo 1-9wa-16 Jf-ya!J, epyo ein tak.-t. epyo 1-owa-16 Jf-yin, epyo ein t?-t. ( 4) epyo ein tak.-16 Jf-ya!J' epyo-na sega py?>-ya-d. (4) apyo ein t?-16 Jf-yin, apyo-n sega py?>-ya-d. (5) apyo-ha la-d 1u-byo-da1n-go, k.a-la-9a-da1n-go, (5) epyo-ha la-d 1u-byo-da1n-go, ka-1a-oa-da1n-go, -hk.n-b1 sega py?l-d, -hk.an-b1 sega py?l-d. (6) g-1o -k.hn-b1 sega py?l-d ek.ha-ma (6) g-1o -k.han-b1 sega py?l-d ekha-hma 1epha?-ye-jgn-d6 ethu-9ephya!J pg-hla, leph?-ye-jan-d6 ethu-oephyfn pE-h1a, Y2>--n-ma pE-jEn-hla, thw?-t pE-jan-h1a, nau?-p1-d6-ma (B) emya-shotin t-d. nau?-p1-d6-hroa amya-zotin t-d. (7) Nau?-p1-d6-ma E-lo -k.hn saga py?l-d akha-ma-1E (7) Nau?-p1-d5-hma E-1o -khan sega py?l-d akha-ma-1E 9u-d6-ha lak-na e1ou?-na py?-t maJf-phu. 9u-d6-ha lLn alouLn pyaLt achein-y-16 meJf-bu. (8) Sega py?l-d ek.ha-ma, ba1n-go h!JaJJ-16-ea-lagatin, (8) Saga py?>-d ak.ha-hma, ba1n-go !Jin-16-oa-legatin, nau? takhu-(B)gu e1ou?-ko 1ou?-p1-ma -khn-d ele J1-d. nau? takhu-gu a1ou?-ko 1ou?-p1-hma -k.han-d e1e J1-d. (9) g-tha-ma 9u-do-ma-1g saga py?l-ya-da 1W-kU-d, (9) E-da-hma 9u-d6-ma-1E saga py?>-ya-da lw-kU-d, (10) Nau?-p1-d5 ba1n h!Ja!J ch-cha-p1-d5-ma -khn-d (10) Nau?-p1-d5 ba1n !Jin chi-cha-p1-d5-ma -khan-d e1e Jf-d. ele Jf-d. (11) E 1apha?-ye-de pha-de 9auk-p1 E-k.hn-d, pha-ywe-d (11) E 1aph?-ye-de ba-de 9ali?-p1 -khan-d, pha-ywe-d 210

PAGE 223

(12) Techa-ng-ga la-dE lu-da1n-go-1E eu-d6-ga ceg ng-ma (12) Techa-ng-ga la-dE lu-da1n-go-1E eu-d6-ga Cdi ng-ma Jl-dE -J tekha-ga, eya!J tekha twe-phu-dE-eta.ln, twe-phu-eelo Jf-dE -J tekha-ga, eyin tekha twe-bu-dE-eta1n, twe-bu-oelo coun-(B)bu-eelo shak-shgn-dg, pho-ywe-dg. coun-bu-oelo shg?_shan-dg, pho-ywe-dg. (13) E-eE ale-ha sho-16 Jf-ya!J (13) E-di ale-ha sho-16 Jf-yin chi?-seya kalin-dE ale tekhu phyi?-tg-16 sho-cha!J-dg. chi?-seya kalin-dE ale tekhu phyi?-tg-16 sho-jin-dg. (14) Nau?-p1 ephe-eme-ha sho-16 Jf-ya!J-lE (14) Nau?-p1 ephe-eme-ha sho-16 Jf-yin-lE E-lo ka-la-ea-de eh la-dE: echein-ma, eu-d6-ga E-lo ka-la-oa-de elg la-dE echein-ma, eu-d6-ga ca-dg: dele ene-na sho-16 Jf-ya!J WEll myauk-ca-dg. ca-dg: dele ene-nE sho-16 Jf-yin WUn myauLca-dg. (15) E-lo ka-la-ea-de eh la-dE eu-d6-ha s1-z1 (15) E-lo ka-la-oa-de el la-dE echein-ma eu-d6-ha s1-z1 ei?-ya wa!J-dg?-tE ale Jf-dg. ei?-ya win-da?-tE ale Jf-dg. (16) Ka-la-ea-de elg la-dE echein-ma (16) Ka-la-oa-de elg la-dE echein-ma ka-la-ea-de-na ka-la-oa-de-nE 0 -J ca-dg sho-16 Jf-ya!J, 0 -J ca-dg sho-16 Jf-yin, eu-d6-go pyo-byo JWalJ-JWalJ ne-ze-cha!J-dg, eu-d6-go pyo-byo JWin-Jwin ne-ze-jin-dg, lwg?-lwg? lg?,...lg? ne-ze-cha!J-dE lu?-lu? ne-ze-jin-dE (17) Hta-jatin-hm6-16 E-eg ephe-eme-de-ha (17) Da-jatin-hm6-16 E-di ephe-eme-de-ha sho-16 Jf-ya!J-lE eu-do-ha-na eu-d6 sho-16 Jf-yin-lE eu-do-ha-nE eu-d6 wa1n-phwE-16-eo-legalin, lu-j1-cha!J wa1n-phwE-16-oo-legalin, lu-j1-j1n wa1n-phwE-16-eo-legalin, tha-hma mehou? s1-z1 wa1n-phwE-16-oo-legalin, da-hma mehou? s1-z1 ei?-ya wa!J-dE ene-na-eo-legalin, ei?-ya win-dE ene-nE-oo-legalin, 211

PAGE 224

eu-d6-ha lwE?-lwE? lE?-lE? eu-d6-ha lu?-lu? la?-la? ceu ea e -J eu eem1-de-go ceu ea e -J eu eem1-de-go pe-tha-dE. pe-tha-dE. (18) -tha sho-16 Ya-nE-y lu-byo hle-d dele (18) -da sho-16 J1-yin Yo-nE-y lu-byo hle-d dele phyi Lpa-dE. phyi Lpa-dE. 4.1.2 The tale of the tiger and the elephant (1) Je-doUn-ga ca-na Jl-dE kw. (1) Je-dOUn-ga ca-n shin-ga Jl-dE kw. (2) -tha eu-d6 twe-ja-d5, tego sho-p1-ma (2) -da eu-d6 twe-ja-d5, tego pyain-ya-aun sho-p1-ma ca-ga ca-ga pyO-pya-dE (3) E ea 1phE-lo pyain-ja-melE' sho-d5 (3) E ea 1bE-lo pyain-ja-melE1 sho-d5 [ho -J [lak -J sheyak-tei [ho -J [lE? -J zeyE?-tei (B) emya-jl na-ne-dE kwa. emya-ji na-ne-dE kwa. (4) -tha-go ho sheyak-tei-go ceu -J (4) -da-go -ho zeyE?-tei-go ceu -J eeEn-ga pyEn-hnain-aun 1M1n eean-ga katin-la m1n-ga pyan-hnain-aun pyEn-hnain-aun lou?-hnain-mela, pyan-hnain-aun lou?-hnain-mela1 sho-b1 lou?-ca-d5, sho-b1 lou?-ca-d5, (5) [hewa -J :::> kwE1 16 (5) [hew a -J 'M1n-ga eyin :::> kwE1 16 -shin-go. (6) -tha ea (6) -da ea shin-ga ( T) Tekha :::>-d5 [ho -J sheyak-te-ga (T) Tekha :::>-d5 [ho -J zeyE?-te-ga en1n-kan shu-ne-ya-ga-ne-p1-d5-ma tei?-ewa-dE. shu-ne-ya-ga-ne-p1-d5-ma tei Lewa-dE 212

PAGE 225

(8) (8) Nau?-tekha o-1ai?-t5 Nau?-tekha o-1ai?-t5 [ho -J tagaun hnekaun [ho -J tegaun hnakaun tha-pyEn-(B)owa-dE. tha-pyan-owa-dE. ( 9) NauLtakha :>-1ai ?_ t5 eoUn.-1e-gaun, 1e-IJa-shE-to-gaun ( 9) Nau Ltekha o-1ai Lt5 eoUn.-le-gaun, 1e-IJa-shE-ko-gaun pyEn-9wa-dE-1e. pyan-owa-dE-1e. (10) g-nau? ea-d5 [ho -J sheyak-te-ga ahmya-jl-ha-go: (10) g-nau? ea-do [ho -] zayELte-ga emya-jl-ha-go: akoun mepyEn-(B)bu-1e. ekoun mepyan-bu-1e. (11) g ea 'Ne-oun kwa: ma!J :>-da eoUn.-kha s1-9wa-bi. (11) ea 1Ne-oUn. kwa: mln o-da eoUn.-ga s1-owa-bi. IJa o-pha1g yau?-pi1 sho-pl-do-ma, ea-ga-ne-pl-do-ma IJa o-b6 ehlg yau?-pi1 sho-pl-do-hma, ea-ga-ne-pl-d6-ma o-pa1ai?-ta. takha-dg-ng o-pe1ai?-ta. (12) CShayak-oun-ma ho -J 1ak-pEn-balJ-ma sheyak-ha (12) [ZayE?-oun-ma ho -J lE?-pan-bin-ma zayE?-ha tegaun-ma meeEn-(B)bu: pyEn-(B)owa-dE. tegaun-ma meean-bu: pyan-owa-dE. (13) g ea-d5 eu-d6 esa-shoUn ked1 tha-da-ga [hewa -J (13) ea-d5 9u-d6 esa-zoUn kad1 tha-da-ga [hawa -J 1Ma1J-ga JOUn-16 Jf-ya!J, malJ-gO !Ja sa-mE; 1Mln-ga JoUn-16 Jl-yin, mln-go IJa sa-mE; IJa-ga JoUn-16 Ji-ya!J, malJ-ga IJa-go sa' sho-pl-ma !Ja-ga JOUn-16 J{-yin, m1n-ga r;)a-go sa1 shO-pl-ma g-tha-myo g-da-myo Chewa -J 1ou?-tha-da-1e. [hewa -J 1ou?-tha-da-1e. (14) [g hawa -J nau? ea-d5 (14) [g ea-d6 hewa -J nau? ea-d5 'PhE-ne-ga-ne-bl 'BE-ne-ga-ne-bl sa-hma-lg' sho-d5 sa-hma-11 sho-d6 'CKhun-nayak hewa -J 'CKhun-nayE? illln-go hawa-ehan-oa -J khun-nayak-ko malJ-gO pe-mE kwE: khun-nayE?-ko illln-go pe-mE kwE: ayak-shaln eyE?-shaln khun-neyak pe-mE' khun-ne-yE? pe-mE' sho-p1-d6-ma sho-pl-do-ma 1ouLea-d6, 213

PAGE 226

(l5) E Cn-ya-maya-na eu-ga (l5) E shin-ga-lE can-ya-maya-nt eu-ga thin-ya-de Jauk-ewa-pl:-ma Jau?_ewa-pl:-ma [khun-nayak s1-kha-lE -J khun-nayak [khun-nay? s1-ga-lE -J khun-nay? s1-mt yak-ko-lE eu-ga hmya-ne-ya-d. s1-mt y?-ko-lE eu-ga hmya-ne-ya-d. (l6) (l6) E-ma, E-ma, 4.2 4.2.l 8e-d5-hma ea-de sa-pl:-kha-ma' sho-pl:-ma, 8e-d5-hmaea-de sa-pl:-ga-ma' sho-pl:-ma, Cca-ga -J Cca-ga -J shin-ga-lE Translations Yaw customs: courting and hospitality (l) What I would like to talk about now is the way courting is done in the Yaw region. (2) Now, these courtship customs are that the young men in our Yaw region --by young men I mean boys who have come of age, boys in their teens --these young men have a custom of visiting the girl. (3) When they visit her they go into her house. (4) and when they're inside they talk to her. (5) The girl greets everyone that comes, and talks to him. (6) While she is receiving him, mostly she will offer plain tea, and always roasted beans --the roasted butter beans that grow in the Yaw region --and jaggery. ('7) Then, when she is rece1v1ng him in this way, there is hardly any time when she lays down her work. (8) While talking, she will be spinning or doing some kind of work-that's the custom when receiving visitors, (9) and it makes it easier for them to make conversation. (lO) They have this custom of spinning and preparing yarn while they receive visitors. (ll) At the same time they drink tea and so on, and are very friendly. (l2) Anyone who comes from outside the region too is treated in this friendly way, just as if they had met and known him before. 2l4

PAGE 227

(13) This Yaw custom strikes me as very attractive. (14) Then, the parents t.oo are pleased when the young men come visiting--it makes them happy, according to Yaw custom, (15) They usually go to bed early when the young men come visiting. (16) The idea is that when the young men come visiting they want them, if they approve of them, to have a good time, to be at ease. (17) That's why the parents either sit by themselves --just the older people together --or go to bed early, allowing their daughters freedom to talk without constraint. (18) That's the way courting is conducted in the Yaw region. 4.2.2 The tale of the tiger and the elephant (l) Long ago there was a tiger and an elephant. (2) When they met the tiger suggested they should have a contest. (3) The contest was like this: there were a lot of mynahs in a silk-cotton tree. (4) They were to see which of the two had the strongest voice, which one could make the mynahs fly away. (5) So the tiger said 'You shout first' to the elephant. (6) So the elephant shouted. (7) The first time he shouted, the mynahs, who had been making a terrible noise, all went quiet. ( 8) The next time he shouted one or two of them flew away. ( 9) And the next time several more flew away. (10) So then there were lots of mynahs, you see, and they hadn't all flown off --(ll) So then the tiger said 'Hold it now: you've had three shouts. Now it's my turn to shout', and he let out one terrific shout. (12) There wasn't a single mynah left in the silk-cotton tree they all flew off. 215

PAGE 228

(13) So then --they'd made this agreement at the very beginning --'If you lose I'll eat you; if I lose you eat me' -that's how they'd fixed it. (14) Then, when the elephant asked when he was to be eaten, the tiger said he would give him seven days --he'd give him seven days' grace. (15) The elephant couldn't think what to do and just wandered aimlessly about waiting for when the seven days up. (16) 'Soon I shall die, when the tigers eat me up', the elephant thought, and he stood where he was and wept. NOTE l. It is a pleasure to record my gratitude to U HtUn Tfn of the Burmese Department of Rangoon University, himself a Yaw man, for his help with my queries and for arranging the recording session; and to Ko Hpon My{n and his friends who generously recorded for me. Abigail, R.P. Ahtet Bama-naing-ngan sa-yel:-hsaya athl:n. Buchanan F Census of India 1931 REFERENCES 1932. Report on the summary settlement operations of the Ganga:w3 TiUn and Sa:w townships 1930-31. Rangoon: Govt. Printing. 1969. Kabya-akya'Ung sa-akyaU.ng. Mandalay: Lu-du u Hla. 1799. A comparative vocabulary of some of the languages spoken in the Burma empire. Asiatick Researches 5, 219-40. 1933. (Vol.xi), Rangoon: Govt. Printing. Forbes, C.J.F. Smith 1881. Comparative grammar of the of Further India. London: Allen. Grierson, G.A. 1903-1928. Linguistic Survey of India. Calcutta: Govt. Printing. Hardiman, J.P. 1912. Gazetteer of the Lower Chindwin District. Rangoon: Govt. Printing. 216

PAGE 229

Harvey, G.E. Houghton, B. Kya Hti:in Okell, John Ono Toru Owens, F.C. Saw Shwe Boh Scott, J.G. & J. P. Hardiman Swe-zon kyan Symes, M. Taylor, L.F. Webb, C. Morgan Yabu Shiro 1925. History of Burma. London: Longmans. 1897. The Arakanese dialect of the Burmese language. J. Royal Asiatic Soc. 453-6l. 1969. Yaw-ne saga athon-ahnon. In Kabya-akyai:mg sa-akyai:mg ( eds. ) Ahtet Bama-naing-ngan sa-ye1-hsaya athJ:n. Mandalay: Lu-du u Hla, 139-60. 1971. A guide to the romanization of Burmese. London: Luzac. 1969. Dei-than-tara-thon Bama saga-mya. In Kabya-akyaUng sa-akyai:mg (eds.) Ahtet Bama-naing-ngan sa-ye1-hsaya athJ:n. Mandalay: Lu-du u Hla, 43-67. 1913. Gazetteer of the Pakokku District. Rangoon: Govt. Printing. 1973. Beyond the Pon-daung Pon-nya. Forward 11-13, 16-19. 1900-0l. Gazetteer of Upper Burma and the Shan States. Rangoon: Govt. Printing. 1970. (Vol.11). Rangoon: Sa-pei Beik-man. 1800. An account of an embassy to the kingdom of Ava .. 1?95. London: Debrett. 1921. The dialects of Burmese. J. Burma Res. Soc. 11, 89-97. 1917. Linguistic Survey of Burma: preparatory stage or linguistic census. Rangoon: Govt. Printing. 1980. Biruma-go Yo hogen no shiryo. ('Linguistic data of the Yaw dialect of the Burmese language' ) Ajia Afurika gengo bunka kenkyu 19, 164-82. 217

PAGE 231

ORAL VOWELS AND NASALIZED VOWELS IN LEPCHA (RONG) : AS THE KEY TO A PUZZLING VARIATION IN SPELLING R.K. Sprigg l. Lepcha and related languages Lepcha has been classified by Shafer (1955:104-7; see also Henderson 1957, 1963) as belonging 'rather precisely' to the same 'section' as the Lushai (cf. Henderson 1948) and the Tiddim and 'Teizang Chin languages (idem, 1957, 1963, 1965), though not to the same 'branch' of that 'section'; he subclassified Lepcha as belonging to the Ao 'unit' of the Northern Naga 'branch' of Kukish, with Tengsa Naga as the language most closely related to it (Shafer 1955:106, 109). Earlier L.A. Waddell (1899:42 ff.)l had the Arleng (or Mikir) language, spoken in the Garo and Khasia hills, as the most closely related language to Lepcha, and since Shafer classified Mikir as forming a 'branch' of Kukish, Waddell1s proposal would still place Lepcha within Shafer's Kukish 'section'; but the list of comparisons of Lepcha with thirteen other languages, including Lushai and Mikir (and four reconstructed languages) by Bodman (1968) shows Lepcha as most closely related to a language, or language group, that Shafer classifies not as Kukish, or even Burmic, but as belonging to the Misingish 'section' of the Bodic 'division', the Adi group of languages, formerly termed Abor-Miri, spoken in the new state of Arunachal Pradesh. According to these three views, Lepcha, spoken in Sikkim and the Darjeeling District of West Bengal, is a western outlier, separated by three or four hundred miles from the languages to the east to which it is most closely related; and Shafer (1955:109-10) asks: Were the Rong left behind when the Northern Naga Branch (and perhaps all the Kukish peoples) migrated from the Himalayas to their present location on the Indo-Burmese border, or are the Rong a remnant left behind from a time when the Northern Naga extended clear across the Valley of Assam? P.K. Benedict (1972:7-8) on the other hand, associates Lepcha with the Magar language, to the west, in west-central Nepal: Dzorgai (western Szuchuan), Lepcha (Sikkim), and Magari (Nepal) all appear to be closer to Tibetan-Kanauri than to any other nucleus. Lepcha (or Rong) ... might equally 219

PAGE 232

well be regarded as a separate nucleus linking TibetanKanauri with Bahing-Vayu and groups on the south. Finally, and especially because of Henderson's research interest in Khasi (1967) it should be mentioned that R.A.D. Forrest (1962:333) attempted to classify Lepcha as partly AustroAsiatic: it will be seen that Rong has in common with Austroasiatic languages as large a proportion of its phonetically identifiable prefixes as those languages have with each other. If there remains any doubt as to the reality of the Austroasiatic provenance of this feature in Rong, the probability of its affinity is corroborated by a plentiful series of lexical correspondences. He supports this claim with a list of 70 Lepcha lexical items and their proposed Austroasiatic cognates, of which 22 are from Khasi; and the most remarkable of which are: (i) 'Water: R. un C'ung in my romanizationJ, Khasi um, Riang om, Palaung om, Hua Miau au', (ii) 'Dog: R. ka-ju Cka-ju in my romanizationJ, Khasi ksew, Stieng sou, Biat cho, Riang sho, etc. I' (iii) 'Dung: R. ft [ in my romanizationJ, Khasi cit, Khmer ac, Bahnar ik, Stieng ech, Biat ac, etc.' Cibid., 333-4J.2 'It is clear that we have in Rong a very mixed form of speech, ... it is much less easy to determine whether the Austroasiatic or the older Tibeto-Burman (or Tibetan?) stratum is the more fundamental. 1 Cibid., 335J. From these four conflicting attempts at classifying Lepcha, it is clear that its precise classification is still something of a mystery, from which my present phonetic, and grammatical observations may possibly derive an interest that the number of speakers of Lepcha would not justify: Siiger (1967:33) gives the number of Lepchas in Sikkim and the Darjeeling District of West Bengal as 25,780 according to the 1931 census, of whom about 13,000 were estimated to be in Sikkim, but it does not follow that all 25,780 spoke Lepcha; and in any case, by now, some two generations later, the number of speakers must have declined under the influence of Nepali.3 2. Variation in spelling I have found it useful to present these observations of mine in the form of an orthographic problem. The late General Mainwaring refers to the pronunciation of the vowel symbol o as follows: (g 0 has the sound of o in no, as: X((-lj Q:mo, mother, K((O abo father, ((J go I &c. 220

PAGE 233

The Lepchas are apt to pronounce this letter as u, and hence when writing, to confound it with u, this error should be avoided, and corrected in the Lepchas (Mainwaring 1876:9). In some instances, this 'error' appears to be due to an attempt to assimilate loanwords from Tibetan; e.g. 'yok 'work' (Tib. g.yog) (Mainwaring 1876:95); cf. 'yuk (Macdonald 1899, in Grierson 1909:244); thop 'receive' (Tib. thob) (Mainwaring 1876:88); cf. thup 'getting' (Macdonald, op.cit.: 242). These variant spellings correspond to differences in pronunciation, e.g. ?jok versus 2_juk, thop versus thup, in which the former phonetic form of each pair is an attempt to imitate a Tibetan pronunciation, while the latter is more in keeping with 4 the vowel distinctions of what one might term 'original' Lepcha. The examples of variation in spelling that I wish to try and account for in this article, however, are not the same as the half-assimilated loanwords such as 'yok/'yuk and thop/thup cited in the preceding paragraph, for, on the one hand, there is, in their case, no variation in pronunciation parallel to the variation in spelling, and, on the other, the variation results from the important distinction in Lepcha between syllables containing nasalization as a vowel feature (and therefore nasality as an initial-consonant feature) and syllables containing an oral vowel (and therefore only oral syllable-initial consonants), e.g. ngo 'fish' (Mainwaring 1876), but ngu (Sitling 1929; Tamsang 1981); 'a-mo 'consonant', 'mother' (Mainwaring 1876; Sitling 1929), but 'a-mU (Sitling 1970; Tamsang 1981); fa-ngo 'five' (Mainwaring 1876; Sitling 1929; Tamsang 1981), but fa-ngu (Sitling 1970), with which can be compared fo 1bird1 cho 'book', cu 1 (snow) mountain 1 prU. 1 Bhutan 1 for which there is no variation in spelling. It is this distinction that I have taken as the subject of this study; and I have further limited it to open syllables. 3. Open syllables and open/closed-syllable lexical items The characteristic qualities of the vowel units that need to be phonologically distinguished are (i) for oral vowels: 1:_:' !!!.4 .:!:!:_:' Q_:' 5 .2. and (ii) for nasalized vowels: 1=, Y.=, !=, !:/a:, .Q:; but (iii) for closed syllables they are: (where alternatives are given, the vowel sounds concerned are complementarily distributed in relation to differences in initial 221

PAGE 234

consonant, especially palatal and palatalized versus the other types of initial consonant, and to differences in final consonant, velar versus labial and dental, and liquid versus nasal and plosive), e.g. ( i) (ii) (iii) i: Zi, H 6 I: nyi, nyt _!_ d'ing, ding e: YE_ e: nye Zem g: [J]J!!_, ff1i2:Y f!1f.!}_ Y: ma, ma lem, lyam m: YE:. a: ma gum e: yg_, ygjJ& u: ngU., ngo e: lam a: aa 3: ny8 mat u: 'u zuk o: tho .2. rok ::>: []_ nong (i) 1speak1 'chew', 'win', 1 descend', 'know', 1 sleep', 'fry', 'put', 1happy1 (ii) 1have1 1 afterwards 1 'call 1hide1 'stew', 'borrow' (iii) ;stand', 'pile up'' 'play', I is I' I fly1 'do', 'make', 'read', 'go' The vowels r, Y, and u are characteristically closed-syllable vowels, though they are shared with the open-syllable type when nasalized. Open/closed-syllable lexical items A number of verb lexical i terns have both open-syllable and closed-syllable forms: (a) the open-syllable forms when 222

PAGE 235

colligated with a particle, apart from the nominalizing particle ('a-), e.g. barn, syo; and (b) a closed-syllable form (i) when colligated with the auxiliary-verb category, e.g. khu, kOn, or the nominalizing particle ('a-), or (ii) when in the negative form, in -n, e.g. a. Zi-bam 1 am speaking'; di-syo 'shall come' b. i. l/in ma-khun 'cannot say'; ryWri kon 'may it turn out well' 'a-zom 'food', 'meal'; 'a-yam 'knowledge', 'knowing' ii. ma-zun 'is not burning' ma-yan 'do not know'. This type of verb includes a number of lexical items that are in very common use; indeed, having a consonant-final form like those shown at (b.i), -n, -m, -t, can almost be considered as a criterion of L;pcha status; but the same cannot be said for those at (b.ii), where the final consonant -n of the negative form is shared with lexical items that may well be loans, e.g. go 'rejoice' (Tib. dga'), ma, ma/ma 1pray1 (Tib. smon). The following is a representative set of examples : a. b.i. b.ii. a. b.i. b.ii bi: bin bin bi, by'i b'in, by 'in b'in, by in 1 give' li: lim lin u, u Um Un 'heavy' di: dit din di, di/d'i d'it, d(y)'it d'in 'come' de: de: m de:( e )n de de m den 'soothe' t<;:m: t<;:Ym t<;:Yn cu cum cun 'small 1 bm: bYD bYn bu bun bun 'carry' djm: djyt djYD dyu dyut dyun 'fight' k= je:m ,je :n ya, ya/ya yam yan 'know' da: da: dan da da dan 1sleep1 rru: rrum rrun ryu ryum ryun 1good1 du: dun dun du dun dun 1dig1 zo: zo:m zun zo zom zon 'eat' bo: boan buan bo b6n bon 'give' l2= l2= goan go go gon 'rejoice' A similar variation applies to certain pronouns: they have (i) a yowel-final fprm, and (ii) a consonant-final form 223

PAGE 236

(objective) in !!!_, e.g. hux: hYin heju: hejum kedo: kedom hu hum hu-yu hu-yU:m ka-do ka-dom 'he' 1him1 'they' 1them1 'myself' 'to oneself' Verb and pronoun lexical items such as these can be classed as a sub-category of the open-syllable lexical item, an alternating sub-category: each has a closed-syllable form in addition to its open-syllable form; for verbs a form in -!!!_, or (b.i) and (b.ii), and for pronouns a form in-!!!_; closed-syllable lexical items, on the other hand, are invariably closed by a consonant, and do not alternate in this way. 4. The 'oral syllable-initial piece', and oral vowels From the list of syllable-final oral vowels given in section 3, it appears that nine phonological vowel units need to be distinguished, thus forming a nine-term system, and that the phonetic exponents of each one of them are comprised in a pure vowel sound: i:, e:, .::, ux:, e:, a:/a:, Q_:, Indeed, Siiger and-Rischel-(1967:23) state: i UI u e 0 a The vowels thus form a symmetric system of 3 x 3. However, not all of these nine vowels can be treated as functionally comparable; they do not all combine with the same preceding consonant sounds and non-syllabic vowel sounds; so that from this point of view, a syntagmatic point of view, some of them have quite different implications trom others as regards the possible set of preceding sounds. A. i: The for example, with closeness, frontness and lip-spreading as its features, does not, in my data from K. P. Tamsant (qv. n. 4) combine with a syllable-initial nonsyllabic front spread vowel or with the cluster ?j-. In Tamsang (1981), however, I do find examples of yi (his ye), the following three: yi-dfxm., yi-dO (mfA.ng} .. yi-she (tsha-thup}; but it is significant that none of them is a verb; and, in fact, all three are loanwords, religious terms, from Tibetan: yi-dam., yi-dwags., and ye-shes respectively. I do not, therefore, consider these counter-examples powerful enough to upset my 224

PAGE 237

syntagmatic generalization that -i: does not regularly combine with (and 2i-) in Lepcha. B. i:, e:, On similarly syntagmatic grounds the vowels i._: , belong to a different type of syllable-initial 'piece' from the remaining six: these last can combine with syllableinitial ts, tsh, and z; but the front vowels i:, e:, do not. Thus, .!_: These six types of vowel occur in the same type of syllable-initial 1piece1 i.e. under the same prosodic conditions; hence, they are syntagmatically comparable, and form a six-term phonological vowel system applicable to that type of 'piece'. The term 'back' can usefully be applied to their type of 'piece', as opposed to the 'front piece', to which the vowels i._:, of section (B) belong (but with the vowel i._: assigned to a separate sub-section of the 'front piece', because of the syntagmatic difference stated in section (A)); and six symbols such as Y, 8, A. U, 0, and W, need to be allotted to the terms of the 'back piece' phonological vowel system, e.g. Y: tju1: thyu 'mix'; VU!: vu 'buzz around' 8: plya 'produce'; fle: fla 'narrate' A: va: va 'swing'; blja: blya 1 smear' U: zu: zu 'burn'; :ru: hrU. 'warm' o0: lo: lo 'dry'; tho: tho 'put' W: t<;l::>: CO 'go'; kh: khyo 1 overcook' 225

PAGE 238

5. The 'nasal syllable-initial piece', and nazalised vowels The next task is to analyse lexical items that have nasalized vowels and, therefore, syllable-initial nasal consonants, by the same syntagmatic method as was used in Section 4 for the oral-vowel lexical items. Within this second prosodic class of lexical items, I find that I need to draw a distinction between (A), those which have only a nasal consonant in the syllable initial (NV:), and (B) those in which the syllableinitial nasal combines, in a cluster, with a lateral or a rolled consonant, or a non-syllabic front spread vowel, or both a lateral and a non-syllabic front spread vowel: (Nl(j_)/!}jV:, Nijv:). A. NV: The set of vowel units, six in number, that need to be distinguished in this type of 'piece' has already been listed, with examples, in Section 3, but to recapitulate, it comprises: i:, e:;E:, Y:, a:/0::, U:/U:, o:. If the same syntagmatic principle is applied to these six as was applied to the oral vowels, it will be found that: (1) i: and ..: ;i combine with only two types of nasal' the labial and the rr), and on that account, can be grouped together in what can be termed the 'front syllable-initial piece' (cf. also (4.c) above); and (2) the remaining 'a:/a:, u:/u:, combine not only with the labial th; but also with the dental and the velar, a total of four !!.. lJ)' e.g. 1. 'front syllable-initial piece' a. mi: b. mi, m nyi, ny 2. 'back syllable-initial piece' a. J:!!: ma3 ma/mCi nya3 nya b. ma: Jla(r::>): m& nya(-ro) c. ?mu: ..n!'!_: rrrU3 mo nyu3 nyo d. ?m\3: ?,n): mo nyo 226 me: me nV: nu na: na nu: nu3 no -8 nye !J!:9 ngu .!;J!: nga ngu3 !.E_: ngo 3 ngo ngo/nga

PAGE 239

l.a. 'fire', 'have'; b. 'that', 'afterwards' 2.a. 'pray', 'stagger', 'suck', 'get threadbare' b. 'hide' 'i:riini stering to' 'go' (imp. ) 'weariness' c. 'plough', 'snot', 'sharpen', 'thirsty' d. 'sore' 'borrow' ,-, 'be time' ('early' Mainwaring 1898). The NV: type of syllable, then,needs a two-term phonological vowel system for its 'front' type of syllableinitial 'piece'; and, for the 'back' type, it needs a four-term system, two of the four members of which have lip-spreadj_ng as a phonetic exponent, : and a:/a:' while the other two, u:/u: and .Q.:, have --B. Nl(j)/r/jV:, NI(})V: I have left this type of nasal-initial syllable until last because it is not clear to me whether it should be classified as belonging to the nasal syllable-initial piece, the oral syllable-initial piece, or, perhaps, to a third type separate from either of those two. The phonetic criteria that have thus far been used for classifying a lexical item as being an example of the nasal syllable-initial piece are: (i) nasalization as a feature of the syllabic vowel in association with nasality as a feature of the syllable-initial consonant; and ( ii) a: twofold or fourfold distinction in syllabic vowel, twofold for the front piece (i:, e: /8: ) fourf'old for the back piece ( y: ..: /0.: u/u: '5: ) aS" in (A)-above. In this second type of syllable,-in ;hfch labial nasality occurs in association with a lateral or a rolled consonant or with a non-syllabic front spread vowel and velar nasality in association with a rolled consonant(]-), I have noted examples in which the nasal resonance extends from the syllable-initial nasal consonant to the syllabic vowel via the intermediate sounds, e.g. mZu3 mZo 'thing', rni}: mlya3 mlya 'efface', but they are comparatively rare. It would seem that the articulatory stretch, or span, of non-nasal sounds is a formidable obstacle; and beside the example mi}v: ('efface') given above, ,I have also noted ml,jY: in which the non-nasal consonant and the vowels, both non-syllabic and syllabic, are purely oral, together with such other examples as Qru: ngru 'groan' and se'mju: sa-myu 'man' (Tamsa:ng l98l). In comparison with the NV: type of syllable analysed in section (A), the number of examples of syllables of this cluster type is very small; in fact, there are none containing the two types of vowel t: and distinguished in the front syllableinitial piece (A.l), but this type of syllable does, however, seem to have the same fourfold distinction as was made for vowels 227

PAGE 240

in the back syllable-initial piece (A.). If, therefore, a Y-A-U-0 vowel system is accepted for this nasal-cluster type of syllable too, examples of these four vowels can be given as follows: mi3: Y: mljY: mZ.ya slant' mZ.ya, mZ.ya 'efface' A: mlja: mZ.ya 'level' rrrya 'versed in' U: miii: mZ.o 'thing' !Jru: ngrU. 'groan' sa'mju: sa-myu 'man' 0: !!!:.1.: rrryu/rrryo rrryo 'course' 6. Oral syllable-initial piece and nasal syllable-initial pi:ece compared The various vowel units can be compared, and grouped in systems as follows, according to the type and sub-type of syllableinitial piece in which they function, and especially oral (4.A-C) versus nasal (5.A.l-2; 5.B): lip-spread Z. ip-rounde d 4. A. i: B. e:. E: c. !!!_: ' !::!_: Q_: :::>: 5. A.l. r=, 2 !: 0: B. !:/:!_:' o:. Thus, in the back syllable-initial piece (4.C., 5.A.2, 5.B.), the oral type (5.C.) has three lip-spread vowel units as against two for the nasal type (5.A.2, 5.B.): m:, a:, a:/a: versus !: !_: and it also has three lip-roU"nded units -as against two for the nasal type: !::!_: Q_: _e.: versus /TJ..: E_: To provide for the former three, there are the three Lepcha symbols romanized as u, a (or a)' and a, one too many for the needs of the corresponding nasal-piece vowel units; and for the latter three there are the three symbols romanized as u, o and 9, also one too many for the corresponding nasal-piece vowel units. It is from this lack of balance between the two sets of vowels, 228

PAGE 241

three versus two, that fluctuation in spelling has arisen between (a) u and a (or a)' on.the one hand, and (b)' more prominently, u and o on the other. a. Lip-spreading For u and a (or a) the fluctuation is especially to be seen in weak-stress syllables, in which the vowel is central and half-close, e.g. ma-rum, ma-rum/mu-rum ma1rYm 'life-span', mu-zu, ma-zu/mu-zu ma'zm: 'body', of which mu-rum and mu-zu are preferable on etymological grounds because the first lexical item in each of these compounds is mu 1body1 b. Lip-rounding For the fluctuation between o and u there are examples in Section (2) above, ngo versus ngU 'fish', 'stew', 1a-mo v. 'a-mU 'consonant', 'mother', etc.l2 There are oral-initial piece lexical items that show a corresponence of Lepcha o with Tibetan a, e.g. i. oral-initial: Lep. fo 1bird1 1a-bo 'father', zo 'eat' Tib. bya pha za so that it is tempting to suppose that the threefold distinction in back rounded vowels might be a comparatively recent development in Lepcha, whereby one of the six terms of the vowel system appropriate to the back syllable-initial piece (4.C.) developed lip-rounding as one of its phonetic exponents instead of lipspreading.l3 While a resulting threefold distinction (u:, o:, _e.:) would present no difficulties in the articulation of-oral vowels, the well-known muffling effect of nasal resonance might have been responsible for making such distinction too fine for the language to bear, whence a reduction from threefold to twofold for nasal-initial syllables, with the consequent fluctuation in spelling between o and u.l4 However, the possibility of comparing Lepcha o with Tibetan a in the nasal-initial syllable, too,gives the o spelling an advantage over the u spelling, e.g. ii. nasal-initial: Lep. ngo 'fish', 16:-mo 'mother', fa-ngo 'five' Tib. nya a-ma lnga The spelling with o, then, would be the Sino-Tibetan comparatist's preference; but the Lepchas are not Sino-Tibetan comparatists and seem to be moving towards the spelling with u in these nasalinitial-piece lexical items, e.g. ngU 1fish1 'stew'; 1a-mU 'mother' (Tamsang 1981) .15 229

PAGE 242

l. NOTES This was cited in Siiger (1967:27) but not available to me; similarly, I rely on Shafer (1955) since Marrison (1967) was also not available. 2. My romanization follows Mainwaring (1876) except for the following: Lepcha: ;:( g Mainwaring: eh-ehh-Sprigg: e-eh'q ay-crng 'y -am --ar-J -(i(-) s-u-J -U-J My -am is for the Lepcha symbol called literally 'sun-moon' (Tib.nyi-zla), resembling the eandra bindu of the Devanagari script (Lambert 1953:70). 3. For the expansion of Nepali as a lingua franea in Sikkim see Nakane 1966:261-2. 4. For a corresponding stylistic variation to that of Lepcha,compare the use of /u/ in English in the loanwords Jungfrau and Sung, in imitation of the German and the Chinese pronunciation, as opposed to the /A/ of 'original' English in velar-nasal-final syllables, e.g. and, indeed,an alternative pronunciation of Sung as (Jones 1977:280, 479, 558). My phonetic and phonological analysis is based on data in the Tamsangmo dialect from K.P. Tamsang, Research Assistant in Lepcha at the School of Oriental and African Studies, University of London, in 1952; K.P. Tamsang was, at that time, Mandal of Bong Bustee, Kalimpong, and Secretary of the Darjeeling Lepcha Association. I compared these data with the pronunciation of the late J. Rongong, of Kalimpong, and of the late Pastor P.S. Targain, a speaker of the Ilammo dialect, at Kalimpong in 1965. To all three, but especially to K.P. Tamsang, I am grateful for the patience and care that they showed in helping me towards this analysis. 5. I have symbolized the vowels in open syllables as long here; but they vary in length in accordance with differences in junction. 6. The Mainwaring (1898) spelling differs from that of Tamsang (1981), I have given both, with the Mainwaring (strictly speaking, the Grfulwedel) spelling following the Tamsang spelling after a conuna, e.g. and, so that examples may be grammatically comparable, I have used verbs where possible, but this list contains one noun: nye 'afterwards'. 230

PAGE 243

7. The qualification 'almost' is necessary here because the open/closed-syllable type of verb includes lexical items that have aspiration as a syllable-initial feature; and this feature suggests loanword status, e.g. thi/th"it., thi/th"it 'reach 1 ( ?Tibet.an thebs), khu/khut., khu/khum 1 able 1 ( ?Tib. 'khyud; Das 1902/1960:196; but GrUnwedel, in Mainwaring 1898:46, suggests khugs). It is significant that the aspirated initials, UI'llike some of the non-aspirated initials, do not combine with land r to form initial clusters; cf. kh k kl kr b bl br ml mr f fl fr. I find further support for my view of aspiration as a loan feature in Lepcha in Bodman (Jc968). In his lists of Lepcha-Adi cognates his occlusive-occlusive correspondences show only three, out of a total of 61 examples, in which the Lepcha word has aspiration (and, incidentally, there are no examples of aspiration in the Adi words). 8. The nasalization feature is prominent in syllables in which the (nasal) initial consonant is lingual, but less so where it is labial, indeed, I have not symbolized it in examples in which I have perceived it as weaker than in the nasalized vowels of French. The reason for this relative weakness is, presumably, that a labial closure is at the far end of .the oral cavity from the naso-pharynx, with the result that, the instant that the lips part, the whole of the oral cavity functions as a resonator in competition with the nasal cavity; and the nasal resonance is correspondingly less prominent. I believe that the same (aerodynamic) reason is responsible for the lesser prominence of nasalization in association with front vowels in labial-initial syllables, e.g. mi: mi., m"i 'fire', me: me 'that', as compared with syllables in which the obstruction caused by the raising of the tongue is further back in the mouth' e.g. ?my: m a: ?m;:;:, ?mo: p.206; 2.a-d; the rearward raising of thetongue, when combined with the lowered soft palate, impedes the flow of air into the oral cavity, and, as it were, directs it into the nasal resonance chamber. 9. The role of the glottal-stop type of cluster, e.g. ?m-, 2-, 2i-, as a criterion of borrowing from Tibetan is discussed in Sprigg 1966a. 231

PAGE 244

10. I should have preferred to give to the two members (i.e. 'terms') and to the four members of these two vowel systems a different set of phonological symbols from those used for the phonological vowel units of the two systems appropriate to the oral syllable-initial piece (4.B.; 4.C.), but this would mean going beyond the resources of the Roman and Greek scripts combined; so I find it necessary to use some of the same symbols as have already been used in those earlier section-s. Y, A, U, and 0, for example, can be re-used for the four units of the vowel system stated above for the back syllable-initial piece (5.A.2). Duplicating symbols in this way need not cause confusion provided that it is always made clear which of the systems a given symbol belongs to in any given instance, as, for example, whether the symbol U is being used for the appropriate member of the six-term vowel system that applies to the oral syllable-initial piece (4.c) or to the four-term vowel system appropriate to the nasal syllable-initial piece (5.A.2). Thus, the four sets of examples of the back nasal syllable-initial piece at (5.A.2) can also be treated as examples of each of that type of piece's four vowel units: a. Y: ?mV:, etc.; b. A: mO:: etc. ; c. u: ?mu:, etc. ; d. 0: ?m3:, etc. 11. My arguments in favour of classifying Lepcha as a stress language rather than as a tone language are in Sprigg 1966b: esp. 199-201. 12. cf. also Rischel, Siiger (1967:25): We do not want, however, to insist upon our transcription of /u/ and /o/ after /m/. It would be tempting to suggest that they do not commute at all in open syllable after nasal consonant. Our distinction is made mainly on the basis of the Lepcha orthography. 13. Lepcha shares the lip-rounding feature with certain related languages further east, especially Adi, e.g. abu-abbo 1 father 1 ; do 1 eat 1 ; -O!JO 1 fish 1 (Ao a!JO r; pil!JO -a!JO 'five' (Mikir phO!JO) ; cf. also Kachin u 'bird' (Bodman 1968). 14. cf. James (1929:120-1): The ear is less able to distinguish a :nasalized vowel from its near neighbour than it is to distinguish an oral vowel from its neighbour. It is harder to hear the difference between and E than between e and e. Hence the acoustic confu;ion from the 232

PAGE 245

existence of a number of nasalized vowel phonemes in French was considerable, and after a period of hesitation there emerged the four nasalized phonemes of the present language. Even now the process of reduction seems to be proceeding. 15. A recent publication by the Government of Sikkim (Anonymous 1972) is exceptional in this respect, e.g. 1a-mo3 ngo3 fa-ngo3 mlo (1, 4, 19, 24), and thop (21). Lepcha has recently been recognized, together with Sikkimese Tibetan and Nepali, as an official language of the State of Sikkim, so this use of the older spellings may be quite significant for Lepcha orthography. Anonymous Benedict, P.K. Bodman, N.C. Das, S.C. Forrest, R.A.D. Grierson, G.A. Henderson, E.J.A. REFERENCES 1972. rang la-zong kat-bo (Directorate Educ., Govt. Sikkim). Kalimpong: Mani Printing Works. 1972. Sino-Tibetan3 a conspectus (Prince-Cambridge studies in Chinese ling. 2). Cambridge: Univ. Press. 1968. [Handwritten Sino-Tibetan course material in which Lepcha is compared with cognates in 13 other languages, including Adi, Ao, and Mikir.J 1902/1960. A Tibetan-English dictionary (Bengal Secretariat Book Depot). [1960 reprint, Alipore: Superintendent, Govt. Printing, West Bengal Govt. Press.] 1962. The linguistic position of Rong (Lepcha). J. Amer. Orient. Soc. 82, 331-5. 1909/1967. Linguistic survey of India3 Vol.33 Tibeto-Burman family Pt.l. [1967 reprint, Delhi: Motilal BanarsidassJ. 1948. Notes on the syllable structure of Lushai. Bull. Sch. Orient. Afr. Stud. 12, 713-25. 233

PAGE 246

Henderson, E.J.A. James, A.L. Jones, D. Lambert, H.M. Mainwaring, G .B Marrison, G.E. Nakane, C Shafer, R. Siiger, H. & J. Rischel 1957. Colloquial Chin as a pronominalized language. Bull. Soh. Orient. Afr. Stud. 20, 323-7. 1963. Notes on Teizang, a northern Chin dialect. Bull. Soh. O rient. Afr. Stud 26, 551-8. 1965. Tiddim Chin: a descriptive analysis of two texts. (London Orient. Ser. 15) London: Oxford Univ. Press. 1967. Khasi. 564-88. Vowel length and vowel quality in Bull. Soh. Orient. Afr. Stud. 30, 1929. Historical introduction to French phonetics. London: Univ. London Press. 1917/1977. English pronouncing dictionary (Everyman's Ref. Lib.). London: Dent, 14th rev. ed. 1953. Introduction to the Devanagari script .. London: Oxford Univ. Press. 1876 A grammar of the Rong (Lepeha) language .. Calcutta: Baptist Miss. Press. 1898. (comp.) Dictionary of the Lepeha language (rev. by A. Grlinwedel). Berlin: Unger Bros. 1967. The classification of the Naga languag e s of north-ea s t India. Ph.D. thesis, Univ. London, 2 vols. 1966. A plural society in Sikkim--a study of the interrelations of Lepchas, Bhotias and Nepalis. In Caste and kin in Nepal3 India3 and Ceylon ... (ed.) C. von FUrer Haimendorf. New York/London: Asia Publ. House, 213-63 1955. Classification of the Sino-Tibetan languages. Word 11, 94-lll. 1967. The Lepehas3 culture and religion of a Himala yan people (Publ. Nat. Mus., Ethnog. ser. 12 (2)). Copenhagen: Nat. Mus. Denmark. 234

PAGE 247

Sitling, Rev. G.T. Sprigg, R.K. Tamsang, K.P. Waddell, L.A. 1929. rang m:ng-hZO:p3 eho kcd-bU. Calcutta: Baptist Miss. Press. 1970. rang m:ng-hZap3 cho kat-bu. Kalimpong: Mutanchi Rong Shez;um. \'The Lepcha Assoc.), 3rd ed. 1966a. The glottal stop and glottal constriction in Lepcha, and borrowing from Tibetan. BuZZ. TibetoZogy 3, 5-14. 1966b. Lepcha and Balti Tibetan: tonal or non-tonal languages? Asia Major (NS) 12, 185-201. 1981. Lepcha-English encyc Zopaedic dictionary. Kalimpong: Mani Press [in press]. 1899. The 1Lepchas1 or 1Rongs1 and their songs. InternationaZes Archiv fllr Ethnographie 12, 41-57. 235

PAGE 250

Also published by the School of Oriental and African Studies Tiddim Chin: a Descriptive Analysis of Two Texts by Eugenie J.A. Henderson Linguistic Comparison in South East Asia and the Pacific Edited by H.L. SHORTO The Short Story in South ,. Asia: Aspects of a Genre Edited by .. >O H. CORDELL Lai Thai: Es. "S .. n aonour of E.H.S. Sinnnonds Edited by J. H. C.::. DAviDSON