Notes on Kraisri's and BernatzikI's word lists. พิมพ์ อีเมล
เขียนโดย William A. Smalley   






                                                  William A. Smalley


The Director of the Siam Society Research Center has very kindly
invited me to study the pre-publication manuscripts of the word lists of
Mrabri, Yumbri, Khmuˀ, and other languages listed for comparison in
this special issue of the Journal. I offer the following notes and observa-
tions in appreciation of the effort which has gone into the publication of
this very valuable linguistic material, and in the hope that they will enhance
the usefulness of these lists through the interpretations they offer.

The lists present several fundamental challenges to study which I
have not had opportunity to follow up. For one thing, it would be useful
to see what kinds of phonemic systems these languages present, in the light
of the amount of data included. I have not had access to the tape record-
ings which Mr. Kraisri collected for Mrabri, however, and an analysis of
the sound system should wait for that.

A detailed comparative study of the systematic correspondences
between these languages would also be in order, since so much of the data
is new. This would entail a word-by-word, or rather phoneme-by-phoneme
analysis of the regularities of correspondence, and the postulation of the
forms from which present speech is derived. It would also involve placing
this information in relation to the wider linguistic groupings of this language
family. Such an analysis will have to wait more time.

Two kinds of problems were amenable to study in the short time
available, however, and the results of this study follow in this paper. I
made a brief statistical comparison of basic vocabulary between Mrabri
and the other languages, and an analysis of the transcription systems used
in the one. case by Kraisri for Khmuˀ, and in the other case by Bernatzik
for Yumbri. In so doing some additional light is shed on the lists, making
them more useful to others.

Comparisons of Mrabri Vocabulary

Mr. Kraisri has done a great service in collecting and compiling a
comparison of Mrabri vocabulary with Bernatzik's Yumbri and with the







190                                         William A. Smalley

three dialects of Tin, with Khmuˀ of Muong Sai, and with dialects of
Khamed, Lawa, and Mon. The lists make fascinating study, and will be
a contribution to the data needed for comparative analysis of linguistic

Ultimately questions of linguistic relationship must be solved by
painstaking comparative analysis, word by word, between two languages or
within groups of languages. In such work linguists look for more than
"similarities". They look for regularities of correspondence, where the same
phoneme in one of the languages regularly corresponds to a phoneme in
the other language under the rules which can be described. In light of our
limitations here we will have to do something more provisional, less fully
diagnostic, based on similarities of "basic vocabulary" only.

Linguists find that for purposes of gaining a quick impression of
linguistic relationships not all words are equally useful. Some linguists
speak of "basic vocabulary" in referring to a small number of words not
very likely to be borrowed from one language to another because they are
all so basic to human experience. Such words include the words for 'man,'
'women,' 'person,' 'fish,' 'dog,' 'come,' 'eat,' 'sun,' etc.

One such list, often called the "Swadesh lish" after its compiler,
consists of two hundred words in its longer form. I have made a quick
analysis of Kraisri's and Bernatzik's material as found on this list, and a
comparison of Mrabri with each of the other languages in the lights of
this selection of words.

Of the two hundred items on the Swadesh list, Kraisri lists sixty-
six and Bernatzik fifty-eight. I compared these "basic vocabulary" items
in Mrabri with each of the following languages, and arrived at the number
of obviously related words (first column ), percentage of relationship (out
of the 66 or 58, as the case may be, second column ). I also included
Kraisri's own figures for his whole lists (not just basic vocabulary, third
column ).








I feel that this scanty evidence does not support Mr. Kraisri's
contention that the Mrabri and the Yumbri are the same people. Accor-
ding to these figures the Mrabri are linguistically just as closely related to
the Khmuˀ and the Tin as to the Yumbri. In the light of the evidence
I doubt very much that Mrabri and Yumbri would be mutually intelligible.
That Bernatzik's data are faulty is very possible, but the evidence as it
stands would make Mrabri just as different from Yumbri as it is from
Khmuˀ and Tin. These latter are certainly not "the same people". That
these languages are in some way related to Mon-Khmer I think is incon-
testible, but it might be wisest for now simply to say that they are part of
Austroasian (of which Mon and Khmer are also a part) without tying
them more closely to Mon-Khmer than to other branches.


Kraisri's Transcription of Khmu

       Because I have some knowledge of Khmu and have prepared a

descriptive grammar1 1 of it I have studied Mr. Kraisri's previous transcrip-
tion of Khmu in the Journal of the Siam Society2 and in his present
paper in this issue, where in both cases he lists Mrabri words with Khmuˀ


          1. William A. Smalley, Outline of Khmu Structure, Essay No. 2 of the Ameri-

 can Oriental Series. New Haven, Conn. American Oriental Society, 1961.

           2. Kraisri Nimmanahaeminda and Julian Hartland-Swann, "Expedition to the

 " Khon Pa," JSS Vol. L Part 2 (July 1962), pp. 180-183.






192                                       William A. Smalley


equivalents. Judging by Khmuˀ, Kraisri's remarks on his Thai transcrip-
tion as against his romanized one are well taken. For someone untrained
in phonetics and the adaptation of roman symbols to phonetic use, the
Thai script lends itself to transcribing Khmuˀ better than do English
writing habits.³ From a linguistic standpoint Mr. Kraisri's Thai tran-
scription of Khmuˀ is fully workable with some exceptions.

The Muong Sai dialect of Khmuˀ which Mr. Kraisri recorded is
slightly different from the Luang Prabang dialect with which I am most
familiar. I will not comment on differences which are obviously due to
dialect. There are some purely transcriptional problems, however.

In the writing of any language certain conventions have to be
established. It is almost always impossible to take over the writing system
of one language for use in another and have it represent the phonemes
( significant sound distinctions ) adequately. A phonetic or phonemic
transcription in romanized script uses the symbols in ways accepted by
convention among schools of linguists, explaining any deviation from that
convention. However, taking over English writing habits as such for the
transcription of another language is extremely difficult because of the
inconsistency in English spelling.

Thai transcription is potentially more consistent, and its inventory

of phonemes happens to be quite similar to those of Khmuˀ. But the

Thai writing system likewise has to be adapted for Khmuˀ. Some Thai
symbols have to be used with new values if all Khmuˀ phonemes are to be
recognized. Some new combinations of consonants are going to have to
occur with new values, and the tonal system of Thai will have to be dis-
carded. In making such an adaptation the new conventions need to be
made explicit for the benefit of the Thai-reading person, and of the
linguist who might try to figure out the phonemes by means of the Thai.
It is for a clarification of some of these points that the following discussion
is presented


            3. An extensive paper which I have written on the problems of writing Meo,
Yao, Kuy, Lawa, and Karen in Thai script will appear in William A. Smalley, and
others, Orthography Studies, to be published ( 1963 ? ) by the United Bible Societies,








Mr. Kraisri uses some Thai tone indications in transcribing
Khmu? and the other languages he lists. Khmu? ( that is any dialect I
have studied ) is not a tonal language. If the others are, it is to be proven. 4
When Thai consonants are used they must be used without any sense of the
tone classes they represent in Thai, ข and ค would mean the same thing
in Khmu'. There is no point in using both.

On Khmu? final consonant which are pronounced why does Mr.
Kraisri write--- or--- ? For example in his ฮิเอียร์ /hi?iar/5 why this mark
on the final ร์ /r/? It is enough to state that all consonant are pronounced
as written, whether initial or final.

Khmu? has a contrast ( at least in the Luang Prabang dialect )
between /c/ จ , /ch/ ช , and /j/ for which there is no symbol in Thai. (Mr.
Kraisri romanizes จ as j, which is satisfactory if he also has a symbol for
what we call /j/.) I cannot be positive that this contrast between /c/ and
/j/ exists in the Muong Sai dialect, but think it likely. It is not represented
in Mr. Kraisri's transcription. In Luang Prabang, for example, (cɉan) is
the name of a legendary culture hero, /jᵼan/ the word for 'foot'. I think
these would both be written จ in Mr. Kraisri's script.

            Khmu? has /c/ จ in final position, pronounced as such. People not
familiar with it, often miss it however. Mr. Kraisri transcribes it in
various ways. In the previous JSS article he transcribes /hmuuc/ 'ant'
as มุ้ยซ์. In this new material he makes it มุ้ยซ์ which is phonetically closer
in the final consonant. Actually, by the structure of the Khmu? language,
however, it should be หมูจ (on ห see the next paragraph). The ย /y/
sound is caused by the presence of the /c/ and should not be written
bceause it is automatic. Kraisri transcribes เม็ดช์ ' hear' for เม็จ /mec/.


             4. It is not true to say, as some have done, that Mon-Khmer languages are not
tonal. Some are. In fact, David Filbeck informs me that there is strong evidence for
a rudimentary tone system in Tin. However, it is unlikely that most of the languages
listed by Kraisri are tonal.

            5. Transcriptions in /. . ./ are my phonemic transcriptions. / ? / is the symbot
for glottal stop, or catch in the throat which serves as a consonant in Khmu? and many
other languages.






194                                            William A. Smalley


The previous paragraph contains another set of problems. The
Khmu? I know has /hw hy hm hn hñ hŋ/. The /h/ in these cases repre-
sents that these sounds are made without any vibration in the vocal cords
(voiceless, or sourd). The mouth is set the same as for /w y m n - ŋ/, but
the sound is very different. In Luang Prabang Khmu?, for example,
/ŋ?/ เงาะ is 'fear' but /------/ หเงาะ is 'paddy'. Mr. Kraisri transcribes
'tooth' as ราง (i.e., /raaŋ/) whereas in Luang Prabang it would be หราง
/hraarj/. Again Thai readers should be warned that the ห is used in this
combination by a convention which does not occur in Thai. It is used here
for a consonant sound, not as an indication of tone. Quite likely at an earlier
stage of Thai development หม, หน, etc., in Thai represented the sound we
are now using it for in Khmu?, and developed its tonal association when
the sound became voiced in the historical development of Thai.

I welcome Mr. Kraisri's very valuable Khmu? lists. These are the
best lists of Khmu vocabulary to date.6 They will aid comparative
studies. Perhaps these notes will help the linguists who use them to in-
terpret them somewhat for linguistic uses.


Bernatzik's Transcription of Yumbri


             The republication of Bernatzik's little-known word list of Yumbri

would be a great deal more valuable if it could be "matched" linguistically
with Kraisri's transcription of Mrabri. That is, if the same system of
transcription could be used for both, and if we could be sure that the
premises and conventions followed in the two transcriptions were some-
what alike it would make the parallel listings much more valuable. Since
I have some understanding of Kraisri's transcription of Mrabri, assuming
he followed the same conventions he used in recording Khmu?, I thought
it might be worthwhile to make an analysis of Bernatzik's transcription of
Yumbri to see where it might be clarified for purposes of comparison.


6. The only other list is in Henri Roux and Tran-van-Chu, " Les Tsa Khmu,"
in Quelques Minorités Ethniques du Nord-Indochine, Saigon: France Asie Nos. 92-93,
pp. 297-357.

My grammar contains nearly seven hundred different Khmu5 word used in the
various illustrations, but these are not in any list form. Rather, they are sprinkled
through the pages of the book.






                 NOTES ON KRAISRI'S AND BERNATZIK'S WORD LISTS               195


The method used was that of a brief structural analysis of some of
the most obvious features of Bernatzik's transcription in the light of ex-
perience in the analysis of several related languages. The analysis is not
exchaustive, and I do not pretend that it is " phonemic," but I think it
gives us a basis for a retranscription of much of Bernatzik's meterial in a
more useful form.

The consonant are easier to handle than the vowels, so we will
take them first. To simplify the problem even more, we will take the
consonant which occur at the end of Yumbri words as listed by Bernatzik
before we go on to others. Rather than listing these alphabetically (which
has no linguistic meaning) we will list them according to their presumed
phonetic qualities. Here we assume that Bernatzik used them in a way
conventional to European languages he knew or to phonetic usage. In
the following chart the first column contains sounds made by closing the
lips ( bilabial ), the second by closing the tongue behind the teeth, the
third a little farther back in the mouth, and the fourth in the back of the



Linguistically, the placement of each of the above symbols is fairly
easy. In the case of h we are aided by Bernatzik's own explanation.7


           Once we are statisfied that the symbols are correctly placed, as

nearly as we can tell according to their likely phonetic properties, we
look at the chart to see what linguistic problems it presents. In the first
place we are struck by two pairs of symbols b/p and g/k. In this part of
the world I know of no language where the sound [b] contrasts with the
sound [ p ], or the sound [g] with the sound [ k ] at the end of a syllable.
All of our symbols were taken from the ends of words, so there is probably


           7. Hugo Adolf Bernatzik, Die Geister der Gelben Blätter, München: Verlag
F. Bruckmann, 1938, p. 237,







196                                         William A. Smalley


no significant difference within these pairs. Here we are referring to
pronunciation. In Thai, for example, the characters , may both be
used as the final letter in a word, but they both carry the same sound.
There is no spoken contrast in that position, [b] and [p] make a difference
in many languages of this area when they occur at the beginning of a
syllable but not at the end. Our first reaction, on the linguistic level, is to
consider the writing of b and g at the end of the syllable superfluous.
We write p and k instead because in most of these languages the sounds
in question have more of the phonetic qualities of [p] and [k] than of
[b] and [d].

The second column we notice has to do with column three in the
chart. Languages of this family usually have some more sounds in this
position. When we take another look at Bernatzik's material, to include
his clusters of two consonants at the end of his words we find the following:
dy, ty, yn, gn. We insert them in the chart as follows,



We immediately eliminate dy in favor of ty on the same basis as
b/p above. We assume that the ty is the same sound /c/ as is transcribed
จ in Thai ( and in Khmu?, as described above ), but that unlike Thai ( but
like Khmu5) it occurs at the end of a syllable in Yumbri. We are of
course referring to pronunciation, not just to writing. It is very common
at the end of the syllable in Cambodian and other Mon-Khmer languages.

We assign gn and yn to a sound [ñ] as in Spanish. There is
probably no distinction between gn and yn. The sound occurs at the
beginning of a syllable in Northern Thai and in Lao. It occurred at an
earlier stage of Bangkok Thai and was written ญ We assume that
Bernatzik meant this sound by gn because that is the way it is spelled in








French. As for yn, we are assuming that Bernatzik made a very common
mistake in transcription. In many of the Mon-Khmer languages, in-
cluding Khmer itself and Khmuˀ, when the [n] occurs finally after
certain vowels there is a little automatic [y] sound before it, and the
unwary transcribe it as yn ( cf. the discussion of Khmuˀ /c/ which is a
parallel problem ). We therefore retranscribe Bernatzik's gn and yn the
same way because they probably represent the same phoneme.

Another problem remains in the chart. There is no w, which we
would expect in this area. We find that Bernatzik has used three vowel
symbols, u, ȗ, and ǒ in final position after vowels For example, in the

word for ' chicken ' ặtḗŭ we assume that the ŭ represents /w/ because that
is what /w/ would sound like after this vowel. Our further reasoning for
this analysis here is a little technical in nature, and we will not go into it.
The chart now stands as follows.






Not all Mon-Khmer languages are identical, of course, in their
charts of final consonants, and we should not expect them to be, but
many are similar. Two differences emerge in the above comparison.
They are /ˀ/ and /h/, which are listed for Khmuˀ, but not for Yumbri.
It seems incredible to me that Yumbri lacks /ˀ / (glottal stop, or catch in





198                                         William A. Smalley


the throat), and unlikely that it lacks /h/, but we do not find a hint of
either in Bernatzik, so far as I can interpret him. It is very likely that
the word Yumbri has a glottal stop on the end /yᵼmbriˀ/ or /yᵼmbriˀ/
Kraisri's Thai transcription implies it for Mrabri, and the cognate form of
the last part of this word is /briˀ/ in Khmuˀ. It means ' wild ' or ' forest '.

Using a more consistent transcription, we can now retranscribe
all of Bernatzik's final consonants in phonetic or Thai characters, represen-
ting a normalized or possibly semi-phonemic analysis. The phonetic
symbols are the same as I have used for transcribing Khmuˀ






ny we are assuming to be another way of writing [ ñ ] ( Bernatzik's
third way ). The fact that /?/ and /h/ are missing is very suspicious. That
both are missing is highly improbable, but we seem to have no evidence
for them in Bernatzik's data. The fact that there are no clearly indicated
differences between aspirated and unaspirated stops (i.e., between /p/ ป
and /ph/ พ, /t/ ต and /th/ ท, etc.) is very suspicious. It is very probable,
that Bernatzik heard [th] as [ts], and [ch] as [Iš]. At any rate, ds,
ts, tš, and s in Bernatzik's transcription are a problem to me as I have not
found them in related languages in any phonetic form which I would






           NOTES ON KRAISRI'S AND BERNATZIK'S WORD LISTS               199


transcribe with these symbols. However, they are perfectly possible and
cannot be rejected a priori. I lean to the guess that ds does not differ from
ts, and that this is Bernatzik's transcription of /th/, and that tš is his
transcription of /ch/. We find no evidence for /ph/ or /kh/.

Bernatzik uses three symbols for /w/ and two for /y/, as indicated
in the chart. In the word sōŭḗ ' hungry ' the ŭ probably represents /w/,
coming as it does between two vowels. The reason for considering this a
consonant rather than a vowel is technical, and we will not go into it here.
In ' hunt ' the ŭ of gŭā́ is probably a /w/, coming as it does after a /g/. As
a principle we could state that Bernatzik's ŭ or ŏ ( note that they are short )
coming between two long vowels or between a consonant and a long vowel
should be transcribed /w/ ( or ว in Thai script ). Likewise Bernatzik's i in
the same position should be retranscribed /y/ or ย.

Note that in syllable initial position there is no problem about
accepting both /p/ and /b/, etc. They typically contrast in this position
in related languages. It was only in final position that their occurence was
suspect. Here is our retranscription.




                                           Retranscription of Bernatzik's Syllable Initial Consonants



Note that ฏ /j/, ฆ /g/, ญ /ñ/, and ฉ /š/ are not used for their usual
Thai values. They are here given the values of sounds which do not occur
in Thai

More guess work and interpolation from other languages is involved
n the analysis of the vowels, and more inconsistency in Bernatzik's tran-
scription is likely. Here is a chart of my interpretation of Bernatzik. The
columns headed B contain Bernatzik's symbols. The remaining symbols
are my retranscriptions in Thai and phonetic scripts.







200                                           William A. Smalley





The Thai symbols in the chart are intended only to indicate the

material in Thai symbols the usual Thai conventions would be followed,
the different symbols for the same sound being used in the appropriate
places. In other words, the chart does not try to be exhaustive of the Thai
symbols that would have to be used, but only indicates the presumed
pronunciations involved in the vowel system.

Certain problems become obvious immediately. By our interpreta-
tion, for one thing, many of the short vowels are written by two symbols,
one with a short mark over it, and the other without. It is possible, but
unlikely, that there are really three lengths which are distinctive and








contrastive, or that there are more vowels than we are postulating, and
that these additional ones do not occur long and short.

The assignment of Bernatzik's ā to /əə/ and his ā to /aa/ is pretty
arbitrary. It was done on the basis of frequency. Bernatzik apparently
makes a difference between ā and ā most languages of this family have a
contrast between /ee/ /aa/. /aa/ usually comes more frequently in words of
this family which I have analyzed, so the assignment was made.

/ᵼᵼ/ and /i/ (Thai – ือ and – ึ) are very difficult to reconstruct

When Bernatzik writes I suspect he means /ᵼr / — ึรSometimes when
he writes no vowel at all in a consonant cluster it should be read with /ᵼ/.
This is a common problem in languages of the family.

Bernatzik's transcription includes a set of vowels ȱ, o. which could
be a misinterpretation on his part, or may well represent the phonetic value
of an extra low back vowel, as indicated on our chart. This has no Thai
counterpart. Kuy (Souey ), a Mon-Khmer language of Northeast Thailand,
Cambodia, South Laos, has such a vowel.

Bernatzik's ' accent mark probably indicates stress on the final
syllable, as is common to all these languages. When it does not occur
there I suspect he simply neglected to record it. Where it occurs more
than once in a word, as in kruṅ ' camp ' I suspect he is trying to indicate
/ᵼ/: /kᵼruuŋ/ กูรูง

à is hard to interpret because it has a conflict of symbols ( short and
long marks) on the same letter.

With these points in mind, here are some sample retranscriptions
of Bernatzik's words.

              Bernatzik                       Phonetic Symbols                   Thai Symbols

tongue    ắ trlât                                  /atᵼrlaat/                                   อะตึรลาด

eye          mắt                                     /maat/                                          มาด

dog         laš5                                    /lašɔɔ/                                         ละฉอ

wind        rmôt                                  /rᵼmoot/                                         รึโมด




FileคำอธิบายFile sizeDownloadsLast modified
Download this file (vol 51 pt 2 page 189-201.pdf)vol 51 pt 2 page 189-201.pdf 1142 Kb16609/23/10 10:34




We have: 13 guests online
IP ของคุณ:
วันนี้: ๒๒ ก.ย. ๒๕๖๓